Blockbusters - Allt om Elektronik

5 

Elektor 

Blockbusters 

Top-15 articles from Elektor 

Elektor blockbusters Artikeltitel • 1

Contents 

Elektor RFID Reader ........................................................................................................3 

Stand Alone OBD-2 Analyser.........................................................................................11 

ElekTrack .......................................................................................................................18 

Making Waves at C ........................................................................................................24 

Digital Inspector ............................................................................................................27 

Profiler ..........................................................................................................................31 

ECIO PLC ........................................................................................................................36 

Universal JTAG Adaptor .................................................................................................41 

GBECG............................................................................................................................47 

Tilt Gamepad .......................................................................................................57 

Micro Webserver ..................................................................................................60 

Software Defined Radio ................................................................................................68 

Wireless USB in miniature .............................................................................................74 

A 16-bit Tom Thumb ......................................................................................................82 

USB Flash Board ............................................................................................................89 

Elektor blockbusters Contents • 2

HANDS-ON MICROCONTROLLERS 

ELEKTOR RFID Reader 

For MIFARE ® 

Gerhard H. Schalk 

RFID cards are becoming 

increasingly popular in 

many fields where 

previously barcodes and chip 

cards were used. They open 

up many new possibilities, such 

as applications in travel cards or 

even banknotes. As befits a 

premier electronics magazine, 

Elektor Electronics is offering its 

readers with this issue not only a 

free RFID card but also a professional 

RFID reader for your own applications. 

The design described here can both read 

from and write to all types of RFID card 

that are compatible with the MIFARE and 

ISO 14443-A international standards. 

In developing the Elektor Electronics 

RFID reader we have aimed to make 

the device as universal as possible. So, 

for example, the reader can be used in 

26 

and ISO 14443-A cards 

conjunction with a PC over a USB connection, 

or in stand-alone mode using 

its liquid crystal display. It is very simple 

to use the free PC-based program 

‘MIFARE Magic’ to read and write all 

kinds of MIFARE cards without 

installing special software in the 

reader. 

Elektor blockbusters Elektor RFID Reader • 3 

elektor electronics - 9/2006

Specifications 

Elektor Electronics RFID reader: 

• Near-field reader for 13.56 MHz RFID cards 

• Compatible with MIFARE and ISO 14443-A cards 

• Allows both reading and writing 

• USB interface for connection to PC 

• Ready for immediate use without programming 

• Free PC-based software available 

• Stand-alone (including portable) operation using LCD 

module 

• Dedicated MF RC522 reader IC 

• Dedicated microcontroller on reader board 

• SPI and I2C interfaces 

• Spare 8-bit microcontroller port 

• Buffered switching output 

MIFARE Magic directly supports a 

range of contactless 13.56 MHz 

MIFARE cards, including the Philips 

MIFARE UltraLight, MIFARE 1K and 

MIFARE 4K. The MIFARE Magic window 

(Figure 1) also offers the facility 

to send individual commands to the 

card with a click of the mouse. This 

allows you to determine the characteristics 

of different cards very easily. 

Examples of compatible cards include 

the MIFARE UltraLight RFID card supplied 

with this issue, and described in 

more detail in a separate article, and 

smart cards used on many public 

transport systems all over the world — 

for example, the London Underground 

Oyster card 

In stand-alone operation, for example 

in an access control application, the 

reader can be used directly with the 

firmware we have developed. On 

switch-on the reader immediately 

looks for cards within the range of the 

antenna (a few centimetres) and reads 

any cards it finds in that area. The LCD 

(if connected) then shows the card 

type along with its serial number, and 

the switching output of the reader is 

activated. 

The reader is constructed around the 

newest Philips reader IC type 

MF RC522 and a type LPC936 microcontroller. 

Since the reader IC is only 

available in an HVQFN32 package, we 

have decided to solve the problems of 

mounting and soldering by making 

available ready populated and tested 

reader boards fitted with pre-programmed 

microcontrollers. 

The Elektor Electronics RFID reader is 

naturally ideal for experimenting with 

the free MIFARE UltraLight card. The 

system includes a powerful microcontroller 

and I 2 C, SPI, UART and USB 

interfaces, and free development tools 

• Available as ready populated and tested SMD circuit 

board 

• Can be modified for user applications 

• Programming tools available 

MF RC522 reader IC: 

• Highly-integrated single-chip reader for ISO 14443-A and 

MIFARE cards 

• Supports contactless data transmission at 106 kbit/s, 

212 kbit/s and 424 kbit/s 

• 50 mm approx. read/write range (depending on antenna) 

• Integrated MIFARE Classic cryptography 

• Programmable over UART, I 2 C or SPI 

• 64 byte transmit and receive FIFO buffer 

• Programmable reset and power-down modes 

• Programmable timer 

• Internal oscillator allows direct connection of 27.12 MHz 

crystal 

are available. This makes it suitable for 

developing dedicated applications 

such as door and gate openers, membership 

card systems, storing passwords 

and configuration data, payment 

systems, security for domestic 

appliances such as televisions, video 

recorders and PCs, monitoring battery 

Figure 1. The MIFARE Magic program developed for the Elektor Electronics RFID reader allows MIFARE and 

ISO 14443-A RFID cards to be read, written and programmed. 


9/2006 - elektor electronics 27


packs and much more besides. The 

combination of secure identity, data 

storage and contactless interface 

opens up many opportunities for novel 

applications. 

Reader hardware 

Figure 2 shows the block diagram of 

the reader. The basic reader functions, 

including the creation of the HF magnetic 

field, modulation and demodulation, 

and the generation of the 

ISO 14443 data stream, are carried out 

in the MF RC522. It is simplest to think 

of the MF RC522 as a contactless 

UART driven directly by the microcontroller. 

In the Elektor Electronics reader 

we have used an 8051-compatible 

LPC936 microcontroller from Philips. 

The CPU takes only two cycles per 

instruction and is clocked at 16 MHz. 

This speed and the 16 kbyte Flash 

Elektor blockbusters 

28 

RFID Reader 

Analog 

Interface 

Antenna 

Matching 

MFRC5222 

MF RC522 

Voltage 

Regulator 

I 2 C 

Optional Power Supply 

5V 

LCD 

I /O 

�C / Philips 

LPC 935 

RS232 

FT232R 

USB / UART 

I 2 C 

060132 - 13 

Figure 2. Block diagram of the Elektor Electronics RFID reader. 

Contactless 

UART 

USB 

memory are sufficient for an enormous 

range of possible applications. Programs 

for the microcontroller can be 

simply written using any 8051 compiler. 

Communications with the PC are 

handled by an FT232R USB/RS232 

interface chip from our friends at 

Future Technology Devices (FTDI). 

The full circuit diagram is shown in 

Figure 3. When connected to a PC, 

power is taken from the USB via miniconnector 

K1. The FT232R USB interface 

chip is configured to report the 

reader as a high-power device when 

the bus is initialised (during ‘enumeration’). 

As a bus-powered device the 

reader can then draw a current of up to 

500 mA. When enumeration is complete 

the /PWRNEN signal on pin 11 of 

IC1 changes state, making P-channel 

MOSFET T2 conduct. The 5 V supply is 

then passed through to voltage regulator 

IC5. The output of the LM2937 pro- 

Register Bank 

FIFO Serial UART 

SPI 

I 

Host 

2C Figure 4. Block diagram of the Philips MF RC522 reader IC. 

060132 - 14 

vides the 3.3 V supply for the LPC 

microcontroller (IC3) and the 

MF RC522 (IC4). Red LED D6 shows 

when the 3.3 V supply is present. If 5 V 

power is not provided via the USB connector 

Schottky diode D4 allows an 

external power supply to take over 

automatically. Either four AA-size cells 

(the enclosure suggested in the parts 

list will accept these) or a 5 V mains 

supply capable of delivering at least 

300 mA can be used. 

Figure 4 shows an overview of the 

internal functions of the MF RC522 

reader IC in the form of a (greatly simplified) 

block diagram. The output drivers 

of the device allow direct connection 

of transmit and receive antennas 

without external active amplification 

circuitry. A few passive components 

provide the essential matching to the 

antenna characteristics. The analogue 

interface handles demodulation and 

decoding of the reply data sent by the 

card. The digital block is responsible 

for constructing the ISO 14443A or 

MIFARE protocol frames and accompanying 

error detection (parity and CRC). 

The FIFO buffer allows 64-byte blocks 

to be sent and received in ISO 14443 

mode (‘T=CL’ protocol). In MIFARE 

mode the largest data blocks 

exchanged are at most 16 bytes long, 

and so there is no need for the microcontroller 

to split up the command 

packets. The registers of the MF RC522 

can be programmed over the SPI, asynchronous 

serial or I 2 C interfaces. Since 

the LPC936 microcontroller only has 

one asynchronous serial interface, and 

this is required for communications 

with the PC, the I 2 C interface is used 

to talk to the MF RC522. 

If desired an LCD module can be connected 

to port P0 of the LPC936 via 

connector K2. P0.0 is buffered by a 

transistor and provides a switched output, 

and the SPI and I 2 C interfaces of 

the microcontroller afford plenty of 

opportunities to expand the reader by 

adding extra hardware. For example, a 

real-time clock could easily be added 

to allow for time monitoring, and the 

switched output could control a door 

opener; see also the pages about the 

RFID reader on the Elektor Electronics 

website. 

Get started 

The double-sided printed circuit board 

for the Elektor Electronics RFID reader 

is shown in Figure 5. It is only possible 

to reflow solder the reader IC, and 

so we are making the board available 

Elektor RFID Reader • 5 



060132 - 11 

LC DISPLAY 

LCD1 

VSS 

VDD 

VO 

RS 

R/W 

D7 

E 

D0 

D1 

D2 

D3 

D4 

D5 

D6 

D7 

K 

A 

1k 

R17 

BC517 

K2 

10k 

1 2 3 4 5 6 

7 8 9 

10 11 12 13 14 15 16 

R14 

T3 

P1 

10k 

10� 

16MHz 

27.12MHz 

2�2 

R10 

10� 

12p 

12p 

12p 

12p 

100n 

see text * 

voir texte * 

siehe Text * 

zie tekst * 

C29 

R11 

C7 

C8 

C13 

C14 

C16 

+5V 

100n 47p 

47p 

4 

GND 

MINI USB-B 

C3 

C1 

C2 

TEST 

GND GND GND 

17 20 24 

OSC 

IN OUT 

21 22 

X2 

* 

1n 

C24 

C15 

560nH 

27 

28 

OSCI 

OSCO 

26 

27p 

4�7 

L3 

5 

FT232RQFN 

220p 

* 

C23 

68p 

* 

R9 

Ant. 

3 

4 

15 

14 

USBDM 

USBDP 

C19 

9/2006 - elektor electronics 29 

1 

2 

DCD 

CTS 

8 

220p 

* 

C20 

C27 

68p 

* 

C28 

Ant. 

L5 

K1 

RTS 

DSR 

6 

7 

2 

3 

�1 

C17 

1 

P89LPC935 

P0.0 

P2.0 

P0.1 

P2.1 

P0.2 P2.2/MOSI 

P0.3 P2.3/MISO 

P0.4 

P2.4/SS 

P0.5 P2.5/SPICLK 

P0.6 

P2.6 

P0.7/T1 

P2.7 

VSS P3.1/X1 P3.0/X2 

7 8 9 

X1 

4 

18 

DVSS 

AVSS 

D6 

D7 

20 

19 

3 

26 

25 

24 

1 

2 

13 

14 

15 

16 

27 

28 

5 

D4 

D5 

23 

22 

EA 

MFRC522 

PVSS 

TX2 

13 

24 

32 

ALE SDA 

TXD 

RS 

R/W 

E 

30 

31 

D7 SCL 

TVSS 

D6 A0 

14 

8 

29 

D5 A1 

TVSS 

10 

IC2 = 74HC02 

9 

560nH 

IC2.A 

L2 

100n 

BAT54 

C31 

D5 

1k 

1k 

9 

8 

�1 

10 

R1 

R2 

19 1 

VCC VCCIO 

RESET 3V3OUT 

IC1 

CBUS0 TXLED TXD 

CBUS1 RXLED DTR 

CBUS2 

RXD 

CBUS3 PWRNEN RI 

CBUS4 SLEEP 

32 

27 

28 

D3 A3 

D4 A2 

27p 

C18 

C25 

C26 

RXD 

4�7 

L4 

10 

11 

2 

3 

C21 

R8 

22 

21 

31 

* 

30 

1k 

2k2 

C22 

R6 

R7 

18 

16 

5 

6 

�1 

4 

IC2.C 

IC2.B 

12 

11 

�1 

JP2 

21 

VDD 

4 

P1.7 

6 5 

P1.5/RST P1.6 

17 

10 

P1.1/RXD P1.4/INT1 

11 

P1.3/SDA/INT0 

18 

12 

P1.0/TXD P1.2/SCL 

IC3 

D1 A5 

D2 A4 

TX1 

25 

26 

VMID 

16 

11 

IC4 

1 

6 

23 

NRSTPD 

IRQ 

SVDD 

AUX1 

AUX2 

19 

20 

I2C 

PVDD 

DVDD 

AVDD 

TVDD 

RX 17 

15 12 

13 

GR 

YE 

100n 

100n 

IC2.D 

D1 

D2 

C4 

C5 

L1 

10k 

100n 

1k 

100n 100n 

100n 100n 

JP1 

4k7 

R16 

C6 

CPU PWR 

R15 

4k7 

4k7 

C9 

C10 

C11 

C12 

R5 

R3 

R4 

100k 

2�2 

1k 

R13 

R12 

D6 

C30 

14 

C32 

IC2 

2�2 

7 

Figure 3. Complete circuit diagram of the reader, which can operate either in stand-alone mode, using the LCD module, or in conjunction with a PC using the USB interface. 

7 

9 

2 3 

RE 

T1 

IRLM6402 

BAS19 

T2 

IRLM6402 

D3 

+4V8...+6V 

D4 

+3V3 

+5V 

IC5 

LM2937 

Elektor RFID Reader • 6


30 

T 

R17 

D7 

K2 

(C) ELEKTOR 

C6 

T3 

C7 

060132-1 

C8 

060132-1 

C9 

C10 

C29 

14 1 

R14 

X1 

IC3 

D6 

R15 

R4 

R3 

R5 

IC4 

R11 R10 

- 

X2 

C12 

C17 

L2 

C11 

L3 

C19 

C16 

R7 

R6 

C15 

C18 C20 

C21C25 

C27C23 

C22C26 

C28C24 

R8 R9 

C13C14 

JP2 

D3 

P1 

JP1 

IC2 

C32 

R12 

IC5 

T1 

R16 

D2 

R2 

R1 

C30 T2 

D1 

C3 

C5 

IC1 

C1 

C2 

C4 

R13 L1 

C31 D4 

+ + 

Figure 5. The double-sided printed circuit board incorporates the antenna. The reader IC is not suitable for 

hand soldering and so the board is available ready populated and tested. 


K1 

D5 

- 

ready populated and tested. Instructions 

are also provided for building the 

unit into the suggested enclosure, 

which we can also supply. 

The two jumpers on the reader board 

(JP1 and JP2) are not fitted for normal 

operation. Assuming the LCD module 

is connected to the reader board, the 

unit is ready for operation as soon as 

power is applied, and the serial number 

of any RFID card within range of 

the reader’s antenna will appear on 

the display. If the display appears 

blank, the contrast should be adjusted 

using P1. 

To use the reader with a USB connection 

to a PC, the free CMD-FDTI-USB 

driver must be downloaded from the 

Elektor Electronics website. This particular 

driver is required because the 

FT232R contains the Elektor Electronics 

Vendor ID and Product ID. 

When the RFID reader is connected to 

the PC using the supplied USB cable 

Windows will automatically detect the 

new USB device. The freshly-downloaded 

driver should be selected for 

the unit. If problems arise, the ‘Installation 

Guide’ on the FTDI website 

(www.ftdichip.com) can be consulted 

for assistance: this guide is also applicable 

to the modified driver. 

Installing the CMD-FTDI driver installs 

both the ‘D2XX’ (direct) and ‘VCP’ (virtual 

COM port) drivers. The VCP driver 

allows the USB link to be treated from 

the point of view both of the PC and of 

the microcontroller as an ordinary 

RS232 connection. 

The D2XX driver is required if it is 

desired to modify the unit in a way 

that requires changes to the internal 

configuration data stored in EEPROM 

in the FT232R. This can be done using 

the PC-based program MPROG, available 

as a free download from the FDTI 

website: MPROG will work only with 

the D2XX driver. 

MIFARE Magic 

Once the driver has been installed, 

MIFARE Magic, a specially-written PCbased 

program for the Elektor Electronics 

RFID reader, can be run. This is 

also available as a free download, from 

www.elektor-electronics.co.uk. After 

downloading the program the contents 

of the ZIP file must be copied into a 

subdirectory of your choice. Start the 

program with a double-click on 

MifareMagic.exe, with the reader 

already connected to the USB port. 

This allows MIFARE Magic to find the 

reader automatically. There is no need 




COMPONENTS 

LIST 

Resistors 

(all SMD case 0805, 5%) 

R1,R2,R6,R12,R15,R17 = 1k� 

R3,R4,R5 = 4k�7 

R7 = 2k�7 

R8,R9 = 4�7 

R10 = 270� 

R11 = 10� 

R13 = 100k� 

R14,R16 = 10k� 

P1 = 10k�-preset, SMD, 4 mm SQ 

Capacitors 

(all SMD case 0805, 16 V, ceramic) 

C1,C2 = 47pF NP0 

C3,C4,C5,C6,C9,C10,C11,C12,C16, 

C31 = 100nF 

C7,C8,C13,C14 = 12pF NP0 

C15 = 1nF NP0 

C17,C19 = 220 p NP0 

C18,C20 = not fitted 

C21,C23 = 27pF NP0 


C25,C27 = 68pF NP0 


C29,C30, C32 = 2μF2 

Semiconductors 

D1 = SMD LED (0805) green, 

low-current 

D2 = SMD LED (0805) yellow, 

low-current 

D3,D6,D7 = SMD LED (0805), red, 

low-current 

D4 = BAS19 (200 mA; SOT23) 

D5 = BAT54S (30V / 300 mA; SOT23) 

T1,T2 = 6402 (p-channel MOSFET, 20V / 

3.7A; SOT23) 

T3 = BC517 (npn Darlington; TO92 case) 

IC1 = FT232RQFN (QFN32 case, FTDI) 

IC2 = 74HC02 (TSSOP14 case; NOR 

gate) 

IC3 = P89LPC936FDH-S (SSOP28 case; 

Philips) 

IC4 = MFRC52201HN1 (HVQFN32case; 

Philips) 

IC5 = LM2937 (low-drop, 3V3, SOT223 

case) 

Miscellaneous 

X1 = 16MHz quartz crystal (18pF parallel 

capacitance; 5·3.2mm) 

X2 = 27.12MHz quartz crystal (18pF 

parallel capacitance; 5·3.2mm) 

K1 = miniature USB-B socket, SMD, 

5-way 

L1 = SMD ferrite (1.5 A; 0805 case) 

L2,L3 = 560nH SMD inductor (0805 case) 

JP1,JP2 = 0.1-in. jumper (see text) 

LCD1 = LCD module with 2x16 characters 

and backlight 

Enclosure, dim. 146x91x33 mm with LCD 

window and battery compartment for 4 

AA bateries 

PCB, order code 060132-91 (populated 

and tested, including USB cable; see 

Elektor SHOP pages and 

www.elektor.com) 

Compatible LC display (see Elektor SHOP 

pages and www.elektor.com) 

89LPC936 source & hex code files; free 

download from www.elektor.com 

Mifare Magic PC software incl. source 

code; free download from 

www.elektor.com 




32 

Figure 6. The ‘Terminal’ view of MIFARE Magic shows all the characters sent by the reader 

over the USB interface. 

Figure 7. The ‘MIFARE UltraLight’ and ‘Mifare Standard‘ windows allow 

simple programming of the RFID card. 

Figure 8. The free PC-based Flash Magic program can program the LPC microcontroller over the USB interface 

of the Elektor Electronics RFID reader. 


to select a COM port, as MIFARE 

Magic uses the D2XX driver internally. 

Figure 6 shows the ‘Terminal’ view of 

MIFARE Magic. This mode emulates a 

VT100 terminal and displays all the 

characters sent by the LPC microcontroller 

over the FTDI interface. 

The firmware in the LPC microcontroller 

defaults to ‘terminal’ mode on 

power-up. As soon as the reader 

detects a new card within its field it 

activates the card. The reader determines 

whether the card is a MIFARE 

UltraLight, MIFARE 1K or MIFARE 4K. 

The entire memory contents of the 

card are read out and displayed on the 

MIFARE Magic terminal. For MIFARE 

1K and 4K cards the standard MIFARE 

key is used. If the card uses a different 

key the data stored in certain sectors 

will not be readable. To use a different 

terminal program instead of MIFARE 

Magic (such as HyperTerminal or the 

built-in terminal in the LPC Flash 

Magic programming tool), the VCP 

driver must be used and the terminal 

program must be told the number of 

the relevant COM port. The parameters 

for the port are as follows: 115200 

baud, no parity, 8 data bits and one 

stop bit. 

The ‘Window’ menu allows MIFARE 

Magic to be switched between the 

‘Terminal’ view, the ‘MIFARE Ultra- 

Light’ and the ‘Show All Cards’ views. 

The ‘MIFARE UltraLight’ window (see 

Figure 7) allows various card commands 

to be executed with a click of 

the mouse. This makes it easy to program 

a MIFARE UltraLight card, such 

as the sample supplied free with this 

issue. When this window is opened 

the firmware in the LPC microcontroller 

on the reader board switches 

from terminal mode into PC reader 

mode. Here the microcontroller waits 

for a card command from the PC and 

calls the corresponding function in its 

software. This mode is useful when 

developing applications on the PC. 

The ‘Show All Cards’ window displays 

the serial numbers of all cards currently 

detected by the reader. This is 

useful for testing reader range and the 

capacity of the reader to deal with multiple 

cards simultaneously. 

Program-it-yourself 

For dedicated applications it is possible 

to modify or completely rewrite 

both the firmware in the LPC936 and 

the software running on the PC. Any 

updates to the reader firmware will 

also require reprogramming the 



LPC936. The most up-to-date software 

will always be available on the Elektor 

Electronics website for free download. 

Updates will be reported on the news 

pages of the website and in the magazine 

under ‘Corrections and Updates’. 

The LPC on the reader board can be 

programmed directly over the USB port 

using the free PC program ‘Flash 

Magic’ (see Figure 8). This program, 

from Embedded Systems Academy 

(www.esacademy.com) and sponsored 

by Philips (www.semiconductors.com) 

supports a range of Philips microcontrollers. 

Both jumpers JP1 and JP2 must be fitted 

on the reader board before the LPC 

microcontroller can be programmed. 

Interested readers will find a detailed 

discussion of how to program the 

device on the Elektor Electronics website 

along with a list of all the MIFARE 

UltraLight reader and card commands. 

The reader firmware was developed 

using the Keil mVision3 C compiler for 

the LPC microcontroller. All the commands 

necessary for developing dedicated 

applications are made available 

as functions and so it is not necessary 

to deal directly with the individual registers 

of the MF RC522. 

The listing shows the code necessary 

to activate a MIFARE UltraLight card 

and read a data block. The data will be 

transmitted using the serial interface 

of the microcontroller. 

As mentioned above, the PC 

reader mode of the LPC 

firmware allows a 

PC application 

to 

invoke 

card functions. 

Using 

this mode 

function invocation 

is done 

using a very simple 

serial protocol 

to communicate with 

the program running 

in the microcontroller. 

When the function has 

been executed the 

response is returned to the 

PC. The naming and parameters 

of the functions are identical 

in the PC software and in 

the microcontroller firmware. 

The source code for the PC-based 

MIFARE Magic program and for the 

microcontroller software can be 

downloaded for free from the Elektor 

Electronics website. 

(060132-1) 

Listing 

while(1) 

{ 

status = ISO14443_Request(WUPA, &bATQ); 

if(status != STATUS_SUCCESS) 

continue; 

} 

status = ISO14443_Anticoll(Level1,0,&abSNR[0]); 


continue; 

status = ISO14443_Select(Level1, &abSNR[0], &bSAK); 


continue; 

// Check if UID is complete 

if((bSAK & 0x04) == 0x04) 

{ 

// UID not complete 

status = ISO14443_Anticoll(Level2,0,&abSNR[4]); 


continue; 

} 

status = ISO14443_Select(Level2, &abSNR[4], &bSAK); 


continue; 

// Read UltraLight Block 0..3 

status = Read(0,abDataBuffer); 



PROJECTS VEHICULAR TEST EQUIPMENT 

Stand Alone OBD- 

Interpret 

‘trouble 

codes’ 

without a PC 

Folker Stange and Erwin Reuss 

This handy analyser makes a simple job of rummaging through the information stored by the client– 

accessible part of your car’s computer. It works with all current OBD-2 protocols and can read and erase 

trouble codes stored in the vehicle and reset the MIL display. All this without the help of a PC or a visit 

to a service station. 

Since the turn of the millennium more 

and more new car models have been 

fi tted with the latest version of the on 

board diagnostic interface OBD-2. With 

the increasing sophistication of modern 

engine management many new 

owners have seen the benefi ts of an 

OBD analyser such that it is fast becoming 

an essential part of the garage 

tool kit along with the torque wrench 

and spark plug spanner. 

It has been reported that sometimes 

when owners fi t a new car radio or satnav 

system to their car the vehicle 

management system unnecessarily 

registers a fault, similarly some owners 

who have modifi ed the engine to accept 

an alternative fuel have noticed 

that the engine management can incorrectly 

interpret the engine condition 


and trigger an error. In some cases the 

engine management can even be 

switched into an emergency condition. 

Whatever the cause the outcome is the 

same; a dashboard mounted MIL (malfunction 

indicator light) comes on, a 

fault condition is stored and it is necessary 

to make an (expensive) visit to the 

nearest garage to have the ‘trouble’ 

put right and the MIL reset. With the 

OBD analyser described here in your 

glove box it is a simple job to connect 

to the OBD socket, fi nd out what the 

trouble is, reset the error and continue 

on your journey. On cost grounds alone 

the price of the analyser will be more 

than repaid by avoiding just a single 

unnecessary visit to a dealership 

garage. 

A number of OBD analysers have been 

featured in the electronics press (including 

Elektor Electronics) describing 

an interface between the OBD connector 

and a laptop. The approach we 

have adopted here is however far less 

cumbersome, this stand-alone unit 

does not require a notebook or battery, 

recognises all the usual OBD-2 or EOBD 

protocols and is small enough to stow 

in your car’s glove box. Operation is 

quite straightforward using just two 

buttons, 580 of the commonest trouble 

codes can be recognised and described 

on its running text display. 

The Circuit 

The OBD-2 analyser employs an AT- 

90CAN128 microcontroller from the ATmega128 

family from Atmel. This par- 

Stand Alone OBD-2 Analyser • 11 

46 elektor electronics - 6/2007

2 Analyser 

socket 

ticular model has an on-board CAN 

bus interface as shown in Figure 1. 

The controller is supplied pre-programmed 

with the AGV4900 fi rmware 

which handles the user-interface including 

push buttons, buzzer, LEDs 

and LCD. 

Pin assignments the OBD-2 connector 

are given in Figure 2. In order to support 

all the current OBD-2 protocols 

the analyser needs to be able to interface 

to several bi-directional 

interfaces: 

6/2007 - elektor electronics 

ISO/KWP 

PWM 

VPWM 

CAN 

16 x 3 LCD DISPLAY 

A 

Key 

AGV4900 

B 

handheld scanner 

070038 - 12 

Figure 1. Block diagram of the OBD-2 analyser. 

K-Line 

CAN-H 

S-GND 

C-GND 

PWM+ 

VPWM 

8 7 6 5 4 3 2 1 

16 15 

14 13 12 11 10 9 

+12 V 

L-Line 

CAN-L 

PWM- 

070038 - 14 

Figure 2. Pin defi nitions of the OBD-2/EOBD connector. 


Specification 

• Automatic or manual selection of OBD-2 protocol 

• Very fast automatic protocol scan (0.1 to 2.6 s per protocol) 

• Fast software boot sequence (ready to go around a second after switch on) 

• Read and display important vehicle information (depending on the vehicle) 

• Real-time sensor reading (selectable) 

• Vehicle chassis number display (if supported by the vehicle manufacturer) 

• Readout and display of the trouble code memory 

• Read out and display of Freeze-Frame data 

• Erasure of trouble code memory 

• Language selection (English or French) 

• 580 trouble codes with description in running text 

• All existing OBD protocols for private vehicles are supported: 

ISO9141-2 

ISO14230-4 (KWP2000) 

J1850 PWM 

J1850 VPWM 

ISO15765-4 (CAN, 11/29 Bit, 250/500 kbits/s) 

• Power for the analyser is supplied from the vehicle’s OBD-2 connector (12 V) 

• Backlit 3-line LC display with adjustable contrast 

• Acoustic signal gives audible feedback and beeps when trouble codes recognised 

• LED Indicators for connection status and data traffi c fl ow 

• Simple operation using just two push buttons 

• Connection for a standard OBD-2 cable 

• Handheld format: 80x135x30 mm (wxhxd), weight 150 g (approx.) 

• Supplied as a kit through Elektor SHOP 

��K/L interface 

��PWM interface 

��VPWM interface 

��CAN interface 

Points to note 

The fi rst three of these in the circuit 

diagram (Figure 3) have been implemented 

using transistors and comparators 

confi gured to meet the interface 

standards. The specifi ed pull up resistors 

for the K and L signals have a rela- 

The OBD analyser is only suitable for vehicles fi tted with an OBD-2/EOBD connector. 

EOBD is fi tted to vehicles sold in the EU: 

- after 01.01.2001, for petrol engine vehicles. 

- after 01.01.2004, for diesel engine vehicles. 

Before the analyser is plugged into any vehicle manufactured before these dates, it is 

important to check compatibility with the OBD-2 standard. The website of Florian Schäffer 

[3] contains a databank of vehicles where you check to see if yours is OBD-2 compatible. 


47


K-Line 

L-Line 

Q4 

2x 

BS170 

+12V 

6 

7 

8 

9 

Q8 

BS250 D5 

1N4148 

PWM- 

D1 

+12V 

+12V+5V 

560 � 

560 � 

6k8 

6k8 

+5V 

8 RN1 

10k 

7 

VPWM 

PWM+ 

R6 

1N4004 

K2 

Q3 

1 

2 

3 

4 

5 

RN1 

10k 

9 

RN1 

10k 

5 

Q5 

7 

10 

6 

6 

IC2.B 

BS170 

8 RN2 

tively low impedance so MOSFETs 

have been used here as drivers. The 

CAN bus driver IC type PCA82C250 

takes care of the CAN interface. 

The user-interface software is logically 

designed such that just two push buttons 

are required to operate the analyser 

and select all possible menu options. 

Connections for the buzzer, the 

‘connect’ and ‘Data Traffic’ LEDs 

should not require any further explanation. 

Control of the three-line LCD is a 

little more complex, a fi ve wire SPI interface 

connects the display to the controller. 

LED backlights ensure that the 

display is night time readable. The relatively 

low controller clock speed 

(8 MHz) is a good compromise, producing 

a low level of EMI emissions while 

still giving ample operating speed for 

this application. 


3 RN1 

10k 

4 

R11 

R4 

100k 

R3 

100k 

RN3 

6k8 

6 

K/L-Interface 

5 

8 RN3 

7 

5 

4 

2 RN1 

10k 

1 

IC2.A 

PWM-Interface 

IC2 = LM339N 

CAN-H 

CAN-L 

1 

7 

2 

10k 

100 � 

2 RN3 

1 

R7 

CAN-Interface 

3 

100 � 

Q7 

BS250 

D4 

R8 

C6 C7 

470p 470p 

10 RN2 

10k 

9 

13 

7 

CANH 

6 

CANL 

8 

RS 

+5V 

S1 S3 

KL IN 

K OUT 

L OUT 

IC4 

Q6 

BS170 

PWM+ OUT 

PWM- OUT 

PWM IN 

+8V +5V 

VPWM-Interface 

RN2 

11 5 

IC2.D 

10 

1N4148 

8 

RN2 IC2.C 

2 

10k 

9 

1 

14 

4 RN2 

10k 

TXD 

1 

RXD 

4 

5 

VREF 

2 PCA82C250 

The Firmware 

3 

10k 6 

3 RN3 

6k8 

4 

VPWM 

OUT 

VPWM 

IN 

51 

(AD0)PA0 

50 

(AD1)PA1 

49 

(AD2)PA2 

48 

(AD3)PA3 

47 

(AD4)PA4 

46 

(AD5)PA5 

45 

(AD6)PA6 

44 

(AD7)PA7 

35 

(A8)PC0 

36 

(A9)PC1 

37 

(A10)PC2 

38 

(A11)PC3 

39 

(A12)PC4 

40 

(A13)PC5 

41 

(A14)PC6 

42 

(A15)PC7 

9 

8 

7 

6 

5 

4 

3 

2 

62 

The heart of the OBD-2 analyser is the 

pre-programmed microcontroller with 

the designation AGV4900 [1] available 

solely from Stange Distribution [2]. The 

software was developed by co-author 

Erwin Reuss. Like similar OBD projects 

the fi rmware for this analyser is only 

available pre-programmed into the microcontroller 

where it is copy protected. 

The source fi les are not available 

for download. Without this software 

copy protection is would not be possible 

to offer the analyser in kit form. 

There is no possibility for the home 

constructor to assemble a low-cost version 

of this design unless of course all 

the necessary software is written from 

scratch. 

A menu option switches all display information 

between either English or 

French (for the convenience of our Ca- 

nadian readers). Stange Distribution 

are specialists in OBD related equipment 

and produce several OBD-2 controllers 

for applications in the fi eld of 

OBD development. 

All the OBD analyser functions can be 

selected from the menu using just two 

keys. One feature of the software is the 

very quick boot procedure which ensures 

that the device is ready for use in 

little more than a second after switchon. 

The most important 580 trouble 

codes have a plain text description of 

the fault which is displayed in running 

text (in the language chosen). This feature 

helps promote a quick and effective 

diagnosis of the problem. In the 

vast majority of cases there will be no 

need to look up the code in an OBD 

trouble code book. 

48 elektor electronics - 6/2007 

C10 

100n 

(IC3/INT7)PE7 

(T3/INT6)PE6 

(OC3C/INT5)PE5 

(OC3B/INT4)PE4 

(OC3A/AIN1)PE3 

(XCK0/AIN0)PE2 

(TXD/PDO)PE1 

(RXD/PDI)PE0 

AREF 

29 

(IC1)PD4 

30 

(XCK1)PD5 

31 

(T1)PD6 

32 

(T2)PD7 

AGND 

63 

C5 

22p 

AVCC 

+5V 

64 52 21 

IC3 

PB7(OC2/OC1C) 

17 

PB6(OC1B) 

16 

PB5(OC1A) 

15 

PB4(OC0) 

14 

PB3(MISO) 

13 

PB2(MOSI) 

12 

PB1(SCK) 

11 

PB0(SS) 

10 

AT90CAN128 

AGV4900 

XTAL1 XTAL2 

Q1 

8MHz 

VCC 

VCC 

24 23 

PG4(TOSC1) 

19 

PG3(TOSC2) 

18 

PG2(ALE) 

43 

PG1(RD) 

34 

PG0(WR) 

33 

1 

C4 

22p 

20 

RESET 

PEN 

GND 

GND 

PF7(ADC7/TDI) 

54 

PF6(ADC6/TDO) 

55 

PF5(ADC5/TMS) 

56 

PF4(ADC4/TCK) 

57 

PF3(ADC3) 

58 

PF2(ADC2) 

59 

PF1(ADC1) 

60 

PF0(ADC0) 

61 

PD0(SCL/INT0) 

25 

PD1(SDA/INT1) 

26 

PD2(RXD1/INT2) 

27 

PD3(TXD1/INT3) 

28 

53 22 

1k 

R1 

1k5 

R12 

LED1 LED2 

+12V 

red green 

+5V 

VCC 

RST 

38 

CSB 

39 

RS 

Figure 3. The AVR microcontroller with on-board CAN interface is the main part of the circuit diagram. 

3 

IC2 

12 

GND 

R5 

33� 

1N4148 

R/W 

E 

VIN 

D2 

BZ1 

F/CM12P 

26 40 37 36 25 24 35 34 33 32 31 30 29 28 

VOUT 

D0 

D1 

D2 

D3 

D4 

D5 

D6 

D7 

LCD 

EA DOG-M163E 

PSB 

27 23 

C1 

100n 

IC5 

7805 

CAP1N CAP1B 

21 

+5V 

C2 

100n 

3 

IC1 

78L08 

+5V 

33� 

A1 

1 

A2 

20 

C1 

2 

C2 

19 

R9 

+8V 

C3 

100n 

070038 - 11 

Stand Alone OBD-2 Analyser • 13

Putting all the bits together 

Attention has been paid to the PCB 

layout (Figure 4); SMD components 

have not been used for this design to 

COMPONENTS 

LIST 

Resistors 

RR1 = 1k� 

R3,R4 = 100k� 

R5,R9 = 33� 

R6,R11 = 560� 

R7,R8 = 100� 

R12 = 1k�5 

RN1,RN2 = 10k� SIL-10 array 

RN3 = 6k�8 SIL-8 array 

Capacitors 

C1,C2,C3,C10 = 100nF 

C4,C5 = 22pF 



Figure 4. The PCB is an SMD-free zone. 

simplify component mounting. The 

PCB is produced to industry standard 

using FR4 type board with gold plating. 

Gold is chemically inert and gives 

C6,C7 = 470pF 


D1 = 1N4004 

D2,D4,D5 = 1N4148 

IC1 = 78L08 

IC2 = LM339N 

IC3 = AT90CAN128 (Atmel; QIL case), 

programmed as “AGV4900” (Stange 

Distribution) 

IC4 = PCA82C250 (Philips) 

IC5 = 7805 

LED1 = 3mm, red 

LED2 = 3mm, green 

Q3-Q6 = BS170 (TO92) 

Q7,Q8 = BS250 (TO92) 

Miscellaneous 

Q1 = 8MHz quartz crystal (HC49/S) 

the board almost unlimited shelf life. 

The plating also ensures that there 

should be no problems of corrosion 

which have been reported when lead- 

LC-Display 3x16 lines, type EA DOGM163E; 

with background light: EA LED55X31-A 

S1,S3 = PCB mount pushbutton type 40-XX 

B3F (Omron) with matching aluminium cap 

DC buzzer 

X2 = 9-way sub-D plug (male), PCB mount 

IC socket 14-way 

IC socket 8-way 

QIL socket (4 segments of 16 pins) 

PCB 

Case with front panel foil 

Mounting materials 

Note: Kit of parts no. 070038-71 contains 

all components, the case (with front panel 

foil fi tted), mounting materials and OBD- 

2 cable, see Elektor SHOP advert or www. 

elektor-electronics.co.uk 


49


free solder is used on unplated boards. 

Gold has excellent tinning properties 

and allows the use of either lead-free 

or lead/tin solder. Apart from the need 

to take care with component placement 

and soldering, no special electronic 

skills or programming competence 

is needed to complete this 

project. 

Except for the two LEDs ‘connect’ and 

‘data-traffi c’, the two pushbuttons and 

the LC display all other components 

are mounted on the PCB side printed 

with the component outlines and identification 

(Figure 5). Mounting the 

component starts with soldering each 

of the individual resistors into place 

followed by the diodes, capacitors, 

crystal, IC sockets, resistor networks 

(make sure they are the right way 

round), voltage regulator and then the 

transistors. The 7805 should fi rst be 

mechanically secured before the leads 

are soldered in place. Once the buzzer 


Figure 5. The PCB component side. Figure 6. The two push buttons, LEDs and LCD are mounted on the other side of the PCB. 

and the sub-D connector are fi tted the 

board can be flipped over and the 

pushbuttons, display and LEDs soldered 

into place. 

The AT90CAN128 chip from Atmel 

Figure 7. The controller board connector 

is made up of four sections. 

used in this project is unfortunately 

only available in either the TQFP or 

MLF/QFN outline and neither of these 

are really suitable for a self-build 

project. The controller is therefore supplied 

(in MLF outline) already mounted 

on a small carrier board. It is only necessary 

to fi t an intermediate pin/socket 

arrangement to connect the carrier 

board to the main PCB. The pin layout 

of this connector is the same as a 

QIL64 package (Quad in line, 64 Pins). 

All the components for this connector 

are included in the kit, to ensure success 

it will be necessary to follow the 

instructions carefully, a mistake here 

will be diffi cult to correct. 

The complete socket is made up of four 

strips (Figure 7) fi tted to the main PCB, 

each strip is fi xed in place initially by 

soldering just one pin of each of the 

strips, this allows the fi nal layout of the 

complete socket to be easily adjusted 

until it exactly matches the layout. 



Once you are sure that the four strips 

are accurately aligned (check that they 

are also all at the same height on the 

board) all the remaining pins can be 

carefully soldered to complete the 

socket. 

The carrier board can now be fitted 

with the pins. The supplied pin strips 

must be carefully separated 

into 8-pin lengths. Any rough 

corners can be smoothed 

down with a fine file. The 

strips are pushed fully into 

the socket Figure 8a (They 

only fit one way round: the 

thinner tapered pins go into 

the socket). 

The controller board can now 

be positioned onto the pins 

(Figure 8b) ensuring that 

pin 1 is correctly aligned (to 

the left by C10). The 64 pins 

can now be carefully soldered 

onto the controller board 

(Figure 8c). 

Once all the components have 

been fi tted a short test can be 

carried out by connecting a 

12-V supply to the sub-D connector 

(pin 9 = +12 V, pin 1 or 

2 = 0 V). The current drawn 

by the analyser should not exceed 

about 150 mA. The display 

backlight will be lit and 

the boot loader version 

number will appear on the 

display followed by the greeting 

message. The short selftest 

is now complete. 

The fi nished PCB can now be 

mounted in the case: Fit the 

pushbutton caps and the sub- 

D cover, remove the protective 

fi lm from the display and with 

the display facing downwards 



a 

b 

Figure 8. The three steps to mount the controller board. 

position the PCB into the front cover of 

the case. The small countersunk 

screws can now be carefully screwed 

in and tightened. Lastly, fi t the remaining 

half of the case and the OBD-2-Analyser 

is now ready for action! 

Figure 9. Displays: (a) Contrast setting, (b) Start menu, (c) Status display, 

(d) MIL/DTC PID Menu, (e) DTC trouble code number. 

Analyser operation 

The fi rst requirement before the OBD-2 

analyser can be used is that the car is 

fi tted with the corresponding OBD-2 

connector (see the advice given in 

‘points to note’ elsewhere in this article). 

If it is, the supplied OBD-2 cable 

is inserted into the OBD-2 connector in 

the car. The connector 

shouldn’t be too diffi cult to locate, 

regulations insist that it 

must be mounted in the vehi- 

a 

b 

c 

d 

e 

c 

cle within 1 metre of the driver’s 

seat. A concise operating 

manual for the analyser is 

available to download from 

www.elektor-electronics.co.uk. A 

shortform manual is also supplied 

in the kit of parts so we 

will not delve too deeply into 

the fi ner points here. An online 

simulator is also available 

on our website so you can familiarise 

yourself ‘virtually’ 

with the analyser operation. 

At switch-on it is possible to 

alter the display contrast (Figure 

9a). This is achieved by 

holding down keys A and B 

and plugging the analyser into 

the OBD-2 socket. The contrast 

changes each time key A 

is pressed. Once you are happy 

with the setting release 

key A and press B to store it. 

This basic method is used to 

control the analyser: Key A cycles 

you through the menu options 

while key B confi rms a 

selection or provides a response 

from the equipment. 

The display now shows the 

greeting ‘ELEKTOR OBD2 1.4’ 

with the start menu (Figure 

9b) following shortly after- 


51


wards with the options: Start Diag, 

Protocol and Language. When the analyser 

is regularly used on the same vehicle 

and you are sure of the correct 

protocol then it can be selected otherwise 

option code 0 tells the analyser to 

automatically fi nd the correct protocol. 

A press on key B begins the scan (if the 

vehicle interface is not compatible with 

OBD protocol the test ends with a failure 

message). When the scan has run 

the display will show the state of the 

MIL/DTC indicating if any troubles 

were present (Figure 9c). Selecting the 

option Live Data with button B will 

show the actual value of a parameter. 

The chassis number or vehicle ID can 

be read and the communication Protocol 

displayed as well as the option to 

rescan. 

The current reading of a sensor (live 

data) is given in the PID (parameter 

identifi er) menu. The example shown 

in Figure 9d is a reading of the intake 

Mass Air Flow (MAF in g/s). Pressing 

key B takes you back to the previous 

menu. When a failure has been detected 

by the engine management system 

the analyser will indicate that the MIL 

is on (MIL:ON), and the number of 

stored DTCs (Diagnostic Trouble Code) 

is given (Figure 9e) .There is now a 

choice between displaying the trouble 

codes or freeze frame data. For trouble 

codes the code number is displayed 


errors found: 

2 of 4 

error code 2 

error 2 

description 

select error 3 

with button B 

070038 - 13 

Figure 10. Trouble code menu showing a description of the trouble in running text. 

Figure 11. The Freeze frame menu (PID select). 

along with (in most cases) a detailed 

description of the fault (see Figure 

10). 

When the trouble codes are displayed, 

pressing key A brings up an option to 

clear the codes from the vehicle’s 

memory. 

More information about the failure can 

be gleaned by selecting Freeze Frame. 

When an error is detected by the engine 

management system the on 

board computer will take a snapshot 

or freeze frame of all sensor 

readings and store them 

in the vehicle’s memory. A 

check of this data can provide 

a valuable insight 

into the cause of the 

failure. In selecting 

freeze frame trouble 

codes it is possible 

to select successive 

sensor values 

stored 

around the 

time that 

the failureoccurred. 

The example in Figure 11 indicates a 

sensor reading when trouble was 

logged; trouble F000 a PID 0D (speed) 

of VSS = 33 km/h measured. Button A 

takes you through successive parameters 

while button B returns to the previous 

menu. 

The downloadable user’s manual contains 

overviews of all the menu options, 

selections and display messages. 

When you want to get more familiar 

with its operation, try the online simulator 

mentioned earlier, or better still 

put your order in and build your own 

OBD-2 Analyser! An extra fi le containing 

soldering and assembly hints can 

also be downloaded from the Elektor 



(070038-I) 

Literature 

[1] Datasheet for the AGV4900-Controller: 

www.obd-diag.de 

[2] Source of the AGV microcontroller: www. 

stange-distribution.de 

[3] http://www.blafusel.de/misc/OBD-2_ 

scanned.php 

Stand Alone OBD-2 Analyser • 17

PROJECTS GPS 

ElekTrack 

Tracking & tracing 

with GPS 

Chris Vossen 

Position determination is all the rage. The manufacturer of the well-known TomTom 

navigation system has become a publicly traded company, and the alarm systems of 

expensive cars and other vehicles often comprise positioning systems so they can report 

where the vehicle is located. However, such systems are rather expensive, so we decided to 

take the DIY approach and develop our own version, dubbed ElekTrack. 

Nowadays we want to know the current 

position of everything. Where’s 

that package I ordered? Is the book I 

want already back in the library? Has 

my nephew’s train arrived already? 

Technology has advanced so much in 

recent years that there’s almost no situation 

we can imagine that doesn’t already 

have a solution. And otherwise 

we provide a solution! 

Big Brother 

People are often a bit nervous about 

organisations that keep track of everything, 

but in some cases this is exactly 

what we want. For example, consider 

car security and alarm systems. Those 

of us who can afford the latest Maybach 

or Mercedes SLR will doubtless 

encounter stringent security requirements 

when they take out insurance 

for their vehicles. Quite often these 

cars must be fitted with security systems 

that include vehicle tracking and 

tracing capability in addition to standard 

anti-start and alarm functions. This 

means that they have a built-in GPSbased 

positioning system that reports 


the vehicle location to a message centre. 

Stolen vehicles are indicated by red 

spots on a map in the control room. 

Passenger cars are not the only vehicles 

being fitted with tracking systems. 

Lorries and boats can also benefit 

from such systems. More and more 

excavating machines are also being 

fitted with security systems, because 

they are stolen by the truckload. With 

a built-in track & trace module, these 

pricy machines can be tracked down 

and recovered. 

Of course, not everybody has an excavating 

machine or a Bentley in the garage, 

but it’s still possible to find other 

interesting uses for a GPS tracker. For 

example, on a scooter or motorcycle. 

And if you’re a private detective, such 

a system is bound to make your eyes 

light up. 

We developed the ElekTrack to give 

our readers an opportunity to experiment 

with GPS tracking. Due to the 

large number of SMD components 

and the difficulty of soldering such 

components, we decided to supply 

this module fully assembled only. 

Objective of the design 

What must a tracking aid be able to 

do in its simplest form? Naturally, you 

want to be able to track the unit’s location. 

We chose the most obvious solution 

for the position determination 

part: GPS. With this, the system can 

identify its position nearly everywhere 

in the world. In addition, GPS is presently 

very accurate, and as long as the 

European Galileo system [1] is not yet 

operational, it is the best ‘alternative’. 

We decided on SMS for data transmission. 

Although data transmission is 

not actually live with this approach, 

it is possible almost everywhere and 

at all times. The GPRS network would 

have been another good option for data 

transmission. GPRS works with a direct 

link via the Internet, so data can 

Figure 1. As you can readily see from the schematic, 

everything revolves around the GSM modem. 

ElekTrack • 18 

36 elektor - 10/2007

R7 

3V3 

56R2 

C15 

be displayed live on a computer 

with Internet access. The location 

could thus be queried 

without any time delay, 

and the data could be 

stored ‘live’ in a database. 

However, 

we gave the preference 

to the 

simpler SMS 

system. 

The design 

T3 MMBT3906 2x 

T4 

4k7 

T1 

T2 

GND 

R10 

Vbat 

100k 

10R 

R5 

10/2007 - elektor 

D6 L4 

BLM21P300S 

RS3K 

D7 

R3 

IRLML6402 

BC848 

1N4148 

K1 

GND 

4k7 

D4 

R11 

100n SM8S33A 

250V 

VCC 1 

GND 2 

RST 3 

VPP 4 

CLK 5 

IO 6 

10 

9 

Simcard Holder 

K4 

5 

4 

3 

2 


1 

1k 

L3 

100nH 

C7 C8 

100n 18p 

R4 

R6 

1k 

T5 

1V8 

GND 

C16 C17 C18 

4u7 

50V 

22n 220p 

BOOST 6 

1 

GND 

8 

GND 

9 

GND 

16 

GND 

17 

GND 

C1 

100n 

SW 2 

3 

VIN 

4 

VIN 

SW 5 

BIAS 10 

11 

VC 

FB 12 

15 

SHDN 

IC5 

680n 

L5 

47uH 

R12 

LT3430 

R13 

3k3 

10k 

R8 

MMBT3904 

10k 

R9 

1 

3 

4 

6 

2 

5 

ESDA6V1SC6 

D1 

1 

GND 

2 

GND 

4 

GND 

13 

GND 

14 

GND 

15 

GND 

27 

GND 

28 

GND 

GND 

3V3 

VCC 12 

3 

RF-IN 

5 

LNA 

7 

OPEN 

SHORT 

10 

BOOT 

11 

XRESET 

TXD-B 

8 

24 

TXD-A 23 

RXD-A 21 

RXD-B 20 

PPS 19 

XSTANDBY 16 

9 

RESERVED 

IC4 

Copernicus GPS Receiver 

14 

SYNC 

C11 

D8 

1 

3 

4 

6 

D5 

1N4148 

30BQ060 

thus consists of two main modules: a 

GPS module and a GSM module. 

The GSM module is a type Q2686 from 

Wavecom. This module can be controlled 

using Wavecom OpenAT commands, 

and it has a built-in microcontroller. 

The main advantage of this is 

shorter development time compared 

with using a separate modem, since 

it is not necessary to implement a microcontroller 

or any peripheral logic. 

Everything is located in a single module. 

This also keeps the overall design 

nicely compact. 

For the GPS module, we decided on the 

Copernicus from Trimble. This is a successor 

to the Lassen iQ, which already 

showed what it could do in the USB 

12k7 

4k99 

2 

5 

C10 

100n 

GND 

R14 

D2 

DALC208 

4V3 

C12 

100u 

10V 

8 

IVCC1 

1 

OVCC2 

C19 

C2 

470p 

1V8 

VL 3 

2 

100k 

GND 

R1 

4V3 

6 

THREE-STATE 

1 

VIN 

3 

ON/OFF NC 4 

VOUT 5 

LP2985 

2 

GND 

4V3 1V8 2V8 

9 

SIM_VCC 

11 

SIM_IO 

12 

SIMPRES 

13 

SIM_RST 

14 

SIM_CLK 

VBATT 1 

VBATT 2 

VBATT 3 

VBATT 4 

19 

ON_OFF 

17 

FLASH LED 

15 

BUZZ_OUT 

16 

BOOT 

18 

/RESET 

20 

BAT_TEMP 

21 

AUX_ADC 

22 

SPI1-CS 

23 

SPI1-CLK 

24 

D12 SPI1-I 

25 

SPI1-IO 

28 

SPI2-CS 

27 

SPI2-IO 

SPI2-I 

29 SPI2-CLK 

26 

34 

MIC2N 

35 

SPK1P 

36 

MIC2P 

37 

SPK1N 

38 

MIC1N 

39 

SPK2P 

40 

MIC1P 

41 

SPK2N 

30 

RXD2 

TXD2 

31 

32 

CTS2 

33 

RTS2 

42 

GPIO0 

43 

GPIO44 

4V3 IC6 

3V3 

C13 

VCC 7 

100n 1u 

16V 

IC3 

K5 

C9 

100n 

GND 

IVL2 4 

OVL1 5 

MAX3375E 

GND 

470R 

R25 

470R 

R2 

D3 

1V8 

C14 

10u 

10V 

101 

GND 

102 

GND 

IC2 

VCC_1V8 5 

CHG_IN 6 

BAT_RTC 7 

CHG_IN 8 

VCC_2V8 10 

Wavecom Q2686 

K2 

103 

GND 

Vbat 

GND 

GND 

PCM_SYNC 77 

PCM_IN 78 

PCM_CLK 79 

PCM_OUT 80 

RI1/GPIO42 69 

DCD1/GPIO43 70 

TXD1/GPIO36 71 

RTS1/GPIO38 72 

RXD1/GPIO37 73 

DSR1/GPIO40 74 

CTS1/GPIO39 75 

DTR1/GPIO41 76 

COL0/GPIO4 59 

COL1/GPIO5 60 

COL2/GPIO6 61 

COL3/GPIO7 62 

COL4/GPIO8 63 

ROW4/GPIO13 64 




ROW0/GPIO9 68 

104 

GND 

EXT_INPUT1 

EXT_INPUT2 

EXT_INPUT3 

TXD_OUT 

RXD_IN 

GPIO22 57 

GPIO24 58 

VPAD-USB 52 

GPIO2 53 

USB-DP 54 

GPIO23 55 

USB-DM 56 

SCL 44 

GPIO19 45 

SDA 46 

GPIO21 47 

GPIO20 48 

INT1 49 

INT0 50 

GPIO1 51 

GPS receiver design published in the 

May 2005 issue of Elektor Electronics. 

Schematic diagram 

A glance at the schematic diagram in 

Figure 1 quickly shows that everything 

is built around the GSM modem. The 

GPS module is connected to UART2 of 

the modem via a logic-level converter 

(IC3). The circuitry around T3, T4 and 

T5 detects whether the antenna circuit 

is shorted, open, or connected normally. 

The GPS module supplies power to 

the active antenna via T1 and T2. 

Two supply voltages are necessary for 

proper operation of the circuit: 4.5 V 

and 3.3 V. They are provided by IC5 and 

IC6. IC5 (a LT3430) is a buck convert- 

GND 

R21 

82k 

R20 

82k 

K3 

4k7 

R16 

82k 

T6 

MMBT3904 

R24 

C22 D9 

10n 1N4148 

10k 

DI0 

DI1 

DI2 

R17 

3V3 

2 

C1+ 

4 

C1- 

5 

C2+ 

6 

C2- 

11 

T1IN 

9 

R1OUT 

1 

EN 

FORCEON 12 

FORCEOFF 16 

R1IN 

10 

INVALID 

8 

T1OUT 13 

V- 7 

V+ 3 

C4 

100n 

C5 

C3 

100n 

100n 

L1 

BLM21B102 

MAX3221E 

R15 

L2 

BLM21B102 

C6 

100n 

10k 

2V8 

T7 

MMBT3904 

R23 

C21 D10 

10n 1N4148 

10k 

VCC 15 

IC1 

14 

GND 

2V8 

4k7 

GND 

GND 

R18 

4k7 

T8 

MMBT3904 

R22 

C20 D11 

10n 1N4148 

10k 

TXD_OUT 

RXD_IN 

R19 

040161 - 11 


37

PROJECTS GPS 

er operating a frequency of 200 kHz. 

Resistors R12 and R14 determine the 

output voltage, which is set to approximately 

4.5 V. 

IC6 provides the supply voltage for the 

GPS module. It generates a fixed 3.3-V 

output voltage from a 4.5-V input voltage. 

Three identical level converters 

are formed using T6, T7 and T8. They 

convert the relatively high input voltages 

to logic levels that comply with 

the specifications of the GSM modem. 

IC1 is a standard RS232 converter, 

which among other things can be used 

for connecting the unit to a PC in order 

to program the modem. You can also 

use IC1 as a port for your own applications. 

D1 and D2 protect the input 

of the GSM module against static discharges 

that can occur when the SIM 

card is inserted in the socket. 

LED D3 shows the status of the GSM 

link. If it is continuously on, the modem 

is not logged in to the network. It 

starts blinking as soon as the modem 

logs in successfully. LED D12 shows 

the status of the GPS module. 

External connections 

The power supply can work with an 

external 12-V or 24-V system. Based on 

tests, the ElekTrack requires an external 

supply voltage of at least 8 V for 

proper operation. The external power 

source must also be able to supply 

sufficient current. Relatively high peak 

currents occur when the GSM modem 

is transmitting data. The average power 

consumption of the unit is around 

500 mW, so a 9-V battery would only 

last for approximately 2 hours. A scooter 

or motorcycle battery will hold out 

for a lot longer. Naturally, it would always 

be possible to modify the soft- 


ware in order to reduce the current 

consumption. For instance, you could 

implement a function to place the GPS 

module in sleep mode with an SMS 

command and awaken it with another 

command. 

Naturally, the unit also has antenna 

connectors: one for the GPS module 

and one for the GSM module. The modem 

also has several logic inputs. They 

can be used for purposes such as connecting 

an alarm to the ElekTrack. A 

voltage above approximately 8 V will 

cause the digital input of the GSM modem 

to be triggered. When this happens, 

a text message can be sent to 

one of two previously programmed telephone 

numbers. You will have to implement 

this in the software yourself, 

since this function is not yet included 

as standard. Perhaps it will be implemented 

in a firmware update. 

The GPS module can be configured 

using various commands. For this purpose, 

the module must be linked to a 

PC by a cable. A program such as HyperTerminal 

for Windows must be used 

to configure the ElekTrack unit. The serial 

port of the PC must be configured 

as shown in Figure 2. 

Startup 

The first thing you have to do is to 

check whether the SIM card is protected 

by a PIN code. Use the following 

command for this: 

at+cpin? 

If the modem responds with ‘+CPIN: 

READY’, the SIM card does not need a 

PIN code. If you receive the response 

‘+CPIN:SIM PIN’ instead, the SIM 

card needs a PIN code, and this will 

first have to be eliminated. The current 

firmware does not yet support 

automatic PIN code use. This may be 

implemented in a future version of the 

firmware. Check our website for the 

latest version. 

If you enter the command 

at+cpin=xxxx 

(where ‘xxxx’ is the pin code of the 

SIM card), you will receive ‘OK’ in 

response. 

Next you have to request disabling of 

PIN checking by entering the following 

command: 

at+clck=”SC”,0,xxxx (where 

‘xxxx’ is again the PIN code of the SIM 

card). The modem will again respond 

with ‘OK’. 

In order to check whether the modem 

can now log into the network automatically, 

you have to reset it with 


38 elektor - 10/2007

at+cfun=1 

If everything goes well, the upper LED 

(D12) will start blinking after a bit less 

than a minute. This means that the mo- 

Better than GPS 

The Global Positioning System (GPS) was originally 

developed for military use. In response to 

steadily increasing demand for accurate position 

determination, the GPS system was released for 

civilian use a bit at time. At first, the accuracy of 

the satellite signals was intentionally degraded 

by the US authorities who originally established 

the system. This restriction was removed a few 

years ago, so consumers can now use the full 

resolution of the system. This allows GPS to be 

used to determine positions with an potential 

error of up to three metres. The positioning accuracy 

depends on the number of ‘visible’ satellites 

and whether a WAAS or EGNOS signal is 

received. 

The terms WAAS and EGNOS relate to systems 

that can be used in combination with GPS to 

increase positioning accuracy. WAAS stands for ‘ 

Wide Area Augmentation System’, while EGNOS 

stands for ‘European Geostationary Navigation 

Overlay Service’. The EGNOS and WAAS 

systems do the same thing, but the former is for 

Europe and the latter for North America. Each 

system consists of a network of satellites and 

ground stations that generate a GPS correction 

signal that can be used to increase positioning 

accuracy by a factor of up 5 on average. A receiver 

that supports WAAS or EGNOS generates 

positions that differ from true positions by less 


10/2007 - elektor 

dem has successfully logged in to the 

network. If you wish, you can check 

this by entering the command 

at+cops? 

The modem will respond with 

+COPS: 0,2,20408 

OK 

(here ‘20408’ is the operator number, 

than 3 metres in more than 95% of all cases. 

EGNOS currently consists of three geostationary 

satellites and several ground stations distributed 

over Europe. The ground stations collect information 

from each other and generate a correction 

signal. The ground stations know their own 

locations with high accuracy, and they compare 

their positions with the signals received from 

the satellites, since the paths of the satellite signals 

can be distorted by atmospheric conditions 

and other causes. The correction signal is then 

transmitted to the geostationary satellites. This 

data has the same format as the standard GPS 

signal, so it can be read by any GPS receiver 

that supports WAAS. 

EGNOS is as joint project of ESA, Eurocontrol 

and the European Commission, and it works 

with the American GPS navigation system and 

the Russian Glonass navigation system. The EG- 

NOS satellite number for Europe is 33. Many 

GPS systems indicate whether they are receiving 

a correction signal. The term ‘differential’ is 

also used for this. 

The MSAS system, which operates the same 

way, is used in Asia. 


39

PROJECTS GPS 

Figure 2. The settings shown here must be used for serial 

communication. 

which depends on the provider.) 

Now the modem will log into the network 

automatically as soon as power 

is applied to the module. 

Security 

The first thing you have to do is to 

change the password. The default 

password is ‘elektor’. To change the 

password, first enter the following 

command to request the password: 

at+password? 

The modem will respond with ‘+PASS- 

WORD:: elektor’. To change the password, 

use the command: 

at+password=”gpsmodule” 

Note: you must enter the quotation 

marks as shown; otherwise you will receive 

an error message. The maximum 

password length is 20 characters. 

GPS status 

The GPS module is connected to the 

serial port of the modem. Two commands 

for requesting data from the 

GPS module are implemented in the 

first version of the firmware. The first 

command is 

at+gpshealth? 

The following is an example of a possible 

response from the modem: 

Rcvr status code = 0x01 

(Don’t have GPS time yet) 

Receiver health byte = 0x11 

Battery backup: BBRAM not 

available at start-up 

Antenna feedline fault: An- 


tenna line open/short 

Type of fault: Open detected 

OK 

From this, you can deduce that the GPS 

module does not have any valid coordinates 

at present. The antenna status 

indicates a fault situation. The fault 

type indicates that the antenna connection 

is open-circuited. The antenna 

is probably not connected. 

If the command is issued again after 

the antenna has been connected, you 

might receive the following response: 

Rcvr status code = 0x00 (Doing 

position fixes) 

Receiver health byte = 0x01 

Battery backup: BBRAM not 

available at start-up 

Antenna feedline fault: OK 

OK 

Here you can see that the antenna 

status is ‘OK’. If the antenna input is 

short-circuited, a ‘Short circuit’ status 

message will appear. The receiver 

status code reports that the receiver 

Figure 3. ElekTrack knows exactly where to find Elektor’s new head office! 

was able to determine a valid position. 

When the module is first enabled, 

the receiver status code indicates the 

number of satellites being received. 

When the receiver has managed to determine 

a valid position, the lower LED 

(D3) starts blinking. If this LED is on 

continuously, something is wrong. It 

could mean that the module is not receiving 

enough satellites, but it could 

also mean that the antenna is connected 

incorrectly. 

Normal use 

Now that you have determined that a 

link has been established, you would 

like to know the latitude and longitude 

coordinates. Enter the following command 

for this: 

at+gpsposition? 

The modem will return a set of coordinates 

in response, such as 

Long: 5.803043 E; 

Lat: 50.941492 N; 

OK 


40 elektor - 10/2007

Practical problems 

Anyone who writes software knows that little bugs always 

find a way to creep into the code. When were first developing 

the software, we had problems with sending SMS 

messages. It seemed like the SMS service of the modem 

somehow didn’t want to work. After spending several 

hours looking for the source of the problem and reading 

through documentation, we came up with the simple but 

brilliant idea of trying a different SIM card. That meant 

a quick trip to the shop to pick up a new SIM card. And 

just imagine our surprise when we discovered that there 

were no problems at all with the new card! 

Of course, we found it rather remarkable that everything 

worked OK with the new SIM card, so we contacted the 

supplier. It turned out that there was indeed a bug in the 

modem firmware. Logging in to the SMS service evidently 

did not work properly with some types of SIM cards. 

This problem has been corrected in the latest version of 

the firmware. 

Another problem that is probably familiar to most software 

developers is the difference between little-endian 

and big-endian memory organisation. The GPS module 

outputs latitude and longitude coordinates as doubles in 

radians. The GPS module operates in little-endian mode, 

while the GSM module operates in big-endian mode. 

When you import the data (as ASCII values), the order 

of the bytes must therefore be reversed. If you convert 

little-endian radian data in a big-endian processor, the 

results are naturally all over the map. We overlooked this 

detail at first. Your first impulse is to think that you made 

a mistake in the code that performs the conversion. In 

retrospect it’s all pretty obvious, but in the midst of the 

fray it’s a hard nut to crack. 

Of course, you want to be able to request 

the coordinates via SMS so you 

can request the position of the module 

remotely. To prevent unauthorised 

persons from receiving a reply 

from the module if they send a text 

message to it, the module must have 

a password. Send the following message 

to the module via SMS to assign 

it a password: 

info:: 

or 

INFO:: 

Note: ‘info’ must be written either 

entirely in upper case or entirely in 

lower case. The password you configured 

using the at+password command 

must be entered here in place of 

. 

The parameter is option- 


10/2007 - elektor 

al. The reply is returned to the sender 

by default. If you want to have the reply 

be sent to a different number, you 

can use this parameter to specify the 

desired number. 

After a few seconds, the ElekTrack will 

send a text message via SMS with the 

longitude, latitude and altitude data, 

which can be used to determine the location 

of the ElekTrack unit. For example, 

you can do this online at [2] (Figure 

3) or [3]. 

The ElekTrack is supplied as fully assembled 

unit, and you can order it via 

our webshop at www.elektor.com. 

The latest version of the software is 

also available on our website. And of 

course, we always appreciate hearing 

from our readers about interesting ideas 

and applications. 

(040161-1) 

Web Links 

[1] en.wikipedia. 

org/wiki/Galileo_positioning_system 

[2] www.gpscoordinates.eu 

[3] http://boulter.com/gps 


41

TECHNOLOGY E-BLOCKS 

Making Waves 

John Dobson 

In last month’s article on making a sound waveform we used the E-blocks SPI bus D/A and 

memory board to generate a sound waveform of around 400 Hz. In using Flowcode I read that 

the flowchart is first converted into C and then into assembly code, and that you can embed C 

into your Flowcode programs to speed the operation up. In this article I describe my attempts to 

use C to speed up the sound waveform generator. At the same time I 

discovered how you can use Flowcode to learn C programming. 

To generate code for the PIC microcontroller, 

Flowcode processes a flow 

chart in a number of steps. Let’s have a 

look at how this works: 

Step 1 

Flowcode initially takes your flow 

chart and generates C from it. As an 

example of this let’s look at a simple 

counter program (COUNTER1.FCF) in 

Flowcode and its C equivalent. 

In Figure 1 you can see the counter in 

Flowcode: we declare the value of variable 

COUNT as 0; then we have a loop 

icon. The Loop WHILE 1 declares an 

endless loop as ‘1’ is always true. 


56 

Inside the loop we have icons to 

increase the value of COUNT by 1, wait 

for one second and then output the 

value of COUNT to port C. I have a 

16F877 PIC micro with LEDs counting 

up in a binary sequence on port C. 

Step 2 

If you open Windows Explorer and look 

in the directory where the 

COUNTER1.FCF Flowcode file is saved 

you will find that Flowcode has generated 

a number of files one of which is 

COUNTER1.C. This is the C equivalent 

of the COUNTER1 flow chart program. 

If you open this file in Notepad you will 

see the program in Figure 2. 

Here we have shown the program in 

two parts, side by side – to save space 

on this page. Firstly, Flowcode has 

defined some of the constants the C 

compiler needs with statements like 

‘char PORTC@0x07;’. This defines the 

C variable ‘PORTC’ as being hex 

(that’s what the ‘0x’ stands for) 

address 07. Similarly TRISC is defined 

as being hex 87. TRISC is the data 

direction register on the PIC micro and 

the TRISC register on the 16F877 I am 

using is at hex 87. Similarly, many of 

the other PIC micro and circuit specific 

functions – such as the clock speed, 

Making Waves at C • 24 


at C 


the pins for the internal USART etc. are 

defined. (Note that in C when a line 

begins ‘//’ this indicates that the line 

is a comment and not a line of code.) 

There are no macros or subroutines in 

our program so these sections are 

blank, but there is a variable defined 

as type CHAR called FCV_COUNT. 

This is our COUNT variable from the 

Flowcode program. This is the first 

hint in how you can use the C icon in 

Flowcode to embed code into a Flowcode 

program: all variables in Flowcode 

are prefixed by ‘FCV_’ when 

transferred to the C compiler. This 

means that when referring to a Flowcode 

variable in a C icon you must also 

prefix the variable with ‘FCV_’. (incidentally, 

FCV stands for FlowCode 

Variable.) 

After the declaration of variables you 

see the first line of C: 

Void main() 

{ 

This is a function declaration within C 

to indicate that this is our main program. 

This is the equivalent to the 

START icon in Flowcode. The pair of 

braces indicate that there are no variables 

passed to this function, and the 

‘open’ curly bracket ’{‘ indicates the 

start of the main function, and you can 

see a ‘close’ curly bracket ‘}’ at the end 

of the program. 

After this we have two more declarations 

– for the A/D converter – to 

declare these pins as analogue inputs – 

and to turn the timer interrupt on. 

These are declared inside the main 

program loop as they can be altered by 

the user within the Flowcode program. 

After this we have the main program: 

FCV_COUNT is declared as being 0, 

corresponding to our first flow chart 

icon, then there is a while (1) statement 

followed by further code inside 

curly braces: this means execute the 

routine inside the curly braces forever. 

Inside the braces we have the 

FCV_COUNT increment, and a delay of 

1 second. Note that these lines of C 

code end with a semicolon. Lines of C 

E-Blocks and Flowcode 

do your C programming 

Figure 1. A simple Flowcode counter. 

Figure 2. C equivalent of COUNTER1.FCF. 



TECHNOLOGY E-BLOCKS 

code are always terminated with a ‘;’. 

Then we have the line ‘TRISC=0x00’. 

We saw earlier that TRISC is the data 

direction register. This line of code 

writes value 0 to TRISC which declares 

all of the port C pins as outputs. A 

value for hex FF would declare all the 

pins as inputs. Our last line of C code 

in the loop writes the value of our 

COUNT variable to port C. 

The last line here is a kind of safety 

net: in the case where we do not have 

an endless loop in our Flowcode program, 

then the C program will get 

stuck at this point and execute this 

line forever. This is the equivalent to 

the END icon in Flowcode. If you have 

a C program without an END loop like 

this then you get a curious effect: the 

program counter in the PIC micro 

device keeps incrementing until it rolls 

over back to address 0000, at which 

point your program will start to run 

again. 

Well, that was not as hard as I had 

expected: C is a mysterious language 

but it seems that if we start from a flowchart 

then, with a little explanation, the 

C becomes readable. Admittedly all the 

declarations are difficult to understand 

but Flowcode seems to take care of 

them, so I won’t worry too much what 

they all mean. Besides, a good introduction 

to C was supplied free of charge 

with last month’s Elektor. 

Back to the plot 

So back to our original objectives of 

increasing the speed of the waveform 

generator program I wrote. Well, I duly 

compiled my program ‘SINE WAVE 

GEN.FCF’ from last month’s instalment, 

and looked at the C produced. 

What I noticed was that every time 

Flowcode used a DAC_SEND_CHAR 

icon it went through a lot of lines of C 

to declare further variables and 

address references that the C compiler 

needs. The reason Flowcode does this 

is that it has to assume that you don’t 

know what you are doing (thank goodness!) 

and it has to set all the parameters 

of Port C for serial communication, 

each time you use the SPI DAC macro 

icon, just in case you are using port C 

I/O pins for other functions. Well 

maybe we can take out some of these 

lines of C to speed things up? I 

decided to tackle this in two stages: 

firstly replace a DAC_SEND_CHAR 

icon with the C equivalent to make 

sure I had replicated the function of the 

program, and secondly to then start 

altering the C code to see if I could 


58 

make it more efficient. 

Going back to the main program I simply 

replaced one of the 

DAC_SEND_CHAR icons with a C icon. 

Then I looked up the equivalent C of 

the DAC_SEND_CHAR icon and 

pasted it into the new C icon. Then I 

adjusted the variable declarations to 

make sure the C icon picked up my 

OUTVAL Flowcode variable, and 

recompiled the program to make sure 

that it still worked. Eventually it did. 

Having got to first base I then commented 

out all the unnecessary lines of 

C – of the original 34 lines of C code, 

which a DAC_SEND_CHAR icon produces 

only 9 were actually needed! 

There was a certain amount of trial and 

error here – I just kept commenting out 

lines until the program stopped working! 

The program produced is called 

SINC3.fcf and you can download it from 

the Elektor Electronics website. The file 

number is 065032-11.zip and you can 

find it under Magazine � April 2006 

Unfortunately however I was unsuccessful 

in making the DAC go any 

faster – the speed limitation turned out 

to be the speed of the SPI but itself – 

let me explain! 

Within the code of the 

DAC_SEND_CHAR Flowcode command, 

the 8-bit SPI data is sent in two 

bytes. The last four bits of the first byte 

contain the most significant data nibble, 

and the four most significant bits 

of the second byte contain the least 

significant nibble. Other bits in each 

SPI data byte are reserved for chip configuration 

etc. The DAC_SEND_CHAR 

Flowcode icon takes OUTVAL and 

processes it like this: 

dac_val = (FCV_OUTVAL & 0xF0) > 4; 

sspbuf = dac_val; 

delay_us(3); 

dac_val = (FCV_OUTVAL & 0x0F) 

4’), then we set the 

SSPBUF register in the chip with the 

result and wait 3 �s. Whatever gets 

placed in the SSPBUF register will be 

sent by SPI. The second routine is similar: 

take OUTVAL, shift left by four 

bits, set SSPBUF and wait 3 �s. The 

last statement is the key: SSPBUF is 

the serial buffer in the PIC device and it 

will take 3μs for the buffer to clear as 

the information is sent one bit at a 

time. If you write to the buffer again 

before 3 �s has elapsed you will overwrite 

the data mid way through the 

serial send operation which produces 

some very strange results, as the SPI 

bus ceases to function correctly. 

So, it turns out that the frequency of 

our oscillator is determined by the 

speed of the SPI bus. For those of you 

that want the maths: each sample 

takes 6 �s; 256 samples per waveform 

gives a theoretical minimum period of 

1.5 ms or around 650 Hz. 

Oh well…time to reduce the number of 

samples to 64… 

Programs available for 

download 

(065032-1) 

• COUNTER1.fcf 

• SINC3.fcf 

File number: 065032-11.zip 

Location: MAGAZINE � April 2006 � 

‘E-Blocks making Waves at C’ 

Earlier in this series 

Electronic Building Blocks, 

November 2005. 

E-blocks and Flowcode, 

December 2005. 

E-blocks in Cyberspace, 

January 2006. 

E-blocks – now you CAN, 

February 2006. 

E-blocks Making Waves, 

March 2006. 

Articles may be downloaded individually 

from our website. 

For a complete overview of available 

E-blocks visit SHOP at 

www.elektor-electronics.co.uk 



PROJECTS LOGIC ANALYSER 

Digital 

Inspector 

Four-channel 

logic analyser 

Ronald de Bruijn 

When checking digital signals a logic analyser is indispensable, especially since many circuits use 

microcontrollers these days. In this article we describe an easy to build circuit that can cope with most 

digital signals and also has a memory function. 

Specification 

Sample frequency: 200 Hz-2 MHz 

Channels: 4 

Range: 0 to 5 V 

Memory: 1024 samples per channel 

Trigger levels: +Ve and –Ve 

Trigger pattern: can be set for each input 

Dot matrix LCD: 64 x 128 pixels 

Supply: 9 V PP3 battery 

The best way to inspect digital signals 

is with a logic analyser. Sometimes it’s 

useful to be able to do this on-site, or 

you may have to take a ‘floating’ measurement. 

The four-channel logic analyser 

described here is suitable in both 

situations due to its compactness and 

because it can be battery powered. 

The maximum sampling rate is 2 MHz 

and the circuit has sufficient memory 

to store 1024 samples of the signal. The 

dot-matrix display with a resolution of 

64 by 128 pixels shows a clear representation 

of the digital signals. 

Schematic 

At the heart of the circuit is IC2 (a 

PIC18F4850, see Figure 1). This PIC 

controller samples the signals and 


drives the display. It is 

controlled via five push 

buttons (S1 to S5). The 

crystal (X1, 10 MHz) determines 

what the maximum 

sampling rate is. 

The internal PLL of the 

microcontroller is used 

to give the controller an internal clock 

frequency of 40 MHz, which is the maximum 

frequency recommended by Microchip 

for this type of chip. 

Diodes D1 to D8 protect the inputs 

against too high or negative voltages. 

The input signals are fed to IC1, a 

74HC04N, which is used as a buffer. 

The fact that the signals are inverted 

doesn’t matter in this case, since we 

can easily convert the signals back via 

the software. The signals go directly 

from the buffers into the controller via 

RA1 to RA4, where the software takes 

over (see Control). 

Preset P1 is used to set the contrast 

of the display and T1 turns on the 

background light of the display. Bz1 

gives an audible warning when a new 

sampling cycle starts and when you 

change between run and hold mode. 

The five switches used to control the 

circuit don’t require a hardware debounce 

circuit, since this is taken care 

of by the software. 

The power supply for the circuit consists 

of two parts: a stabilised 5 V supply 

and a 9 V supply for the display 

light. The source for these voltages can 

either be a 9 to 12 V mains adapter or 

a 9 V rechargeable battery. 

A simple charging circuit for the battery 

is also included (T2, R1, R17, D12), 

which comes into action whenever a 

mains adapter is connected. Assuming 

a standard LED with a forward voltage 

drop V f of 1.5 V is used, the charging 

current for the battery will be: 

(1.5 – 0.6) / 56 = 16 mA. 

A 9 V NiMH battery with a capacity C 

of 170 mAh is then charged at about 0.1 C, 

so no damage will occur if it is charged continuously. 

The battery will be fully charged 

in about 10 hours with this circuit. During 

the charging LED D12 will be on. If an 

ordinary (non-rechargeable) battery is 

Digital Inspector • 27 


CH4 

CH3 

CH2 

CH1 


R5 

330 � 

R6 

330 � 

R4 

330 � 

R3 

330 � 


D8 

D5 

4x 

4x 

+5V 

+5V 

D7 

D4 

14 

IC1 

7 

1N4148 

D2 

D1 

D6 

D3 

1N4148 

IC1 = 74HC04 

C6 

100n 

R9 

100k 

IC1.D 

9 8 

1 

R10 

100k 

IC1.C 

5 6 

1 

R8 

100k 

IC1.B 

3 4 

1 

R7 

100k 

IC1.A 

1 2 

1 

D10 

D9 

S5 

S1 

S2 

S3 

S4 

used, the circuit around T2 can be left 

out. 

Control 

Switch S1 is used to select the sampling 

frequency. The rates that can be 

selected are 5/10/20/50/100/200/500 � 

s/div and 1/2/5 ms/div. S2 selects the 

channel that is used to trigger the circuit. 

S3 is used to tell the PIC if it is to 

trigger on a rising or falling edge and 

S4 can arm and stop the circuit, or clear 

the display. One short press of S4 arms 

2x 

1N4148 

10k 

R15 

10k 

R14 

+9V 

R11 

10k 10k 

10k 

12V 

R12 

R13 

10k 

K3 

R16 

3 

2 

1 

C2 

22p 

D11 

1N4004 

680 � 

R1 

D12 

+5V 

11 

20MHz 

C1 

22p 

BC337 

R17 

56� T2 

C5 

100n 

1 

MCLR/VPP/RE3 RC0/T1OSO/T13CKI 

15 

RC1/T1OSI 

16 

2 

RA0/AN0/CVREF 

RC2/CCP1 

17 

3 

4 

RA1/AN1 

RA2/AN2/VREF- 

IC2 

RC3/SCK/SCL 

RC4/SDI/SDA 

18 

23 

5 

RA3/AN3/VREF+ 

RC5/SDO 

24 

6 

RA4/T0CKI 

RC6/TX/CK 

25 

7 

RA5/AN4/SS/HLVDIN RC7/RX/DT 

26 

PIC18F4580-I/P 

RD0/PSP0/C1IN+ 

19 

RD1/PSP1/C1IN- 

20 

RD2/PSP2/C2IN+ 

21 

RD3/PSP3/C2IN- 

22 

RD4/PSP4/ECCP1/P1A 

27 

33 

RB0/INT0/FLT0/AN1 RD5/PSP5/P1B 

28 

34 

RB1/INT1/AN8 

RD6/PSP6/P1C 

29 

35 

RB2/INT2/CANTX RD7/PSP7/P1D 

30 

36 

RB3/CANRX 

37 

RB4/KBI0/AN9 

38 

RB5/KBI1/PGM 

RE0/RD/AN5 

8 

39 

RB6/KBI2/PGC RE1/WR/AN6/C1OUT 

9 

40 

RB7/KBI3/PGD RE2/CS/AN7/C2OUT 

10 

OSC1 OSC2 

12 13 

X1 

14 31 

BT1 

9V 

BZ1 

the circuit. After the trigger signal occurs 

it will take 1024 samples per channel 

and store them. Pressing S4 briefly 

again will make the circuit read in a 

new set of 1024 samples after the next 

trigger signal. When S4 is held down 

for longer the display is cleared. The 

last settings for the sampling frequency, 

the trigger channel and the trigger 

condition are stored inside the EEP- 

ROM of the microcontroller. These settings 

are then used as the initial state 

when the circuit is next turned on. 

S5 turns the backlight on or off. After 

Figure 1. From the circuit diagram it is clear that the microcontroller takes care of just about everything. 

32 

220 � 

R21 

1k 

S6 

R19 

T1 

BC337 

+9V 

+5V 

P1 

20k 

C3 

100n 

11 10 

1 

+5V 

IC1.E 

IC1.F 

13 12 

1 

10k 

R2 

47� 

+5V +9V 

IC3 

7805 

R20 

C4 

K2 

+5V 

100n 

060092 - 11 

LC DISPLAY 


39

PROJECTS LOGIC ANALYSER 

I4 

I3 

I2 

I1 

I5 

about one and a half minutes, or when 

in a ‘Lo_Batt’ condition, the microcontroller 

automatically turns off the 

backlight. 

Operation 

D3 

D5 D4 

D1 

D2 

D7 

D8 

D6 

R8 

P1 

R10 

K2 

C5 

IC1 

S1 

R3 

R4 

R6 

R5 

R7 

R9 

R12 

C2 

C1 

In order to obtain the highest possible 

sampling rate we initially let the microcontroller 

store the samples in its RAM 

when the trigger event occurs. For this 

we use the following software instruction: 

movff port a, postinc0. This instruction 

copies the contents of port a 

to the RAM and increments the RAM 

address by one. This cycle is then re- 

X1 

R11 IC2 

S2 

R16 

R19 

R20 

R13 

R14 

C6 R1 

D10 

D9 R21 

R2 

peated 1024 times. At the end of this, 

128 samples are read from the RAM 

and shown on the display. This process 

is repeated once a second. 

If no new trigger event occurs for about 

three seconds (depending on the sampling 

rate), the circuit reads in 128 

samples and shows them on the LCD. 

In this way we can tell what condition 

(high or low) the inputs are. 

A quick press of S4 turns on the memory 

function. This is indicated by an ‘R’ 

on the right of the screen. The circuit 

then waits for the trigger event. Once 

this has occurred and the 1024 samples 

have been stored the ‘R’ changes into 

an ‘S’ and the display shows the first 

128 samples of each channel. Switches 

S1 and S2 can now be used to scroll 

through the memory. A short press of 

S1 or S2 causes small jumps through 

the memory; a longer press of S1 or S2 

creates larger jumps. The cursor at the 

bottom of the display shows which 

area of memory is currently shown. 

Another quick press of S4 makes the 

circuit read in a new set of samples 

and store them in memory. The display 

keeps showing the same area of memory 

as for the previous samples. This 

is of course very useful when you’re 

studying the signals that follow a short 


R15 

S6 C3 

1 

3 

D11 2 

C4 

Figure 2. As can be seen from the component layout, the construction of the circuit isn’t difficult. Connector K2 is placed such that the display can be mounted directly above the double-sided PCB. 


3 

1 

3 

1 

T2 

T1 

R17 

BZ1 

S3 

K3 

IC3 

D12 

S5 

BT1 

S4 

Components 

list 

Resistors 

R1 = 680� 

R2,R11-R16 = 10 k� 

R3-R6 = 330� 

R7-R10 = 100k� 

R17 = 56� 

R19 = 1k� 

R20 = 47� 

R21 = 220� 

P1 = 20k� preset, multiturn, vertical 

mounting 

Capacitors 

C1,C2 = 22pF 

C3-C6 = 100nF 


D1-D10 = 1N4148 

D11 = 1N4001 

D12 = LED, 5mm diam. 

T1,T2 = BC337 

IC1 = 74HC04 

IC2 = PIC18F4580-I/P, programmed, 

Elektor SHOP # 060092-41 

IC3 = 7805 

Miscellaneous 

Bz1 = AC buzzer 

X1 = 10MHz quartz crystal 

Graphic LC display, 128 x 64 pixels, 

e.g. DEM128064A or NLC128x64 

(Conrad Electronics # 187429) 

Case 186 x 123 x 41mm with compartment 

for 9V battery, e.g. Strapubox 

(Conrad Electronics # 522775) 

S1-S5 = pushbutton Multimec 

RA3FTL6 w. knob AQC09-24.2 

S6 = on/off switch 

9-V battery clip 

5 wander sockets, screw mount (for 

connection to I1-I5) 

Kit of parts incl. case: Elektor SHOP # 

060092-71 

PCB layout: free download from 

www.elektor.com, file # 060092-1 

Digital Inspector • 29

time after the trigger event. 

If you hold down S4 a bit longer, until 

the buzzer gives a beep, the circuit 

reads in a new set of samples and 

stores them in memory. But this time 

the display won’t show the same area 

of memory; instead it jumps right back 

to the beginning. 

If you hold down S4 longer still (until 

you’ve heard two beeps), the logic analyser 

comes out of memory mode and 

returns to the standard mode where 

128 ‘live’ samples are always shown 

on the display. 

Construction 

In this design we haven’t used any 

SMDs. The layout is fairly sparse, with 

all components easily accessible. The 

soldering should therefore not cause 

any problems. 

We would like to come back to the connection 

between the display and the 

board. There is enough room above the 

board for the display. The easiest way 

to connect the display to the board is 

to first solder a single pin-header strip 



to the display board. Next, plug a wirewrap 

socket into this pin-header and 

plug the other end into the main board. 

Check that the display is at the right 

height and then solder the wirewrap 

socket to the main board. 

When you use the recommended enclosure 

for this circuit you should first 

file off the corners of the board at the 

side of the input signals. The board 

will then fit perfectly. 

Comments 

When you’re not using all of the channels 

it is advisable to connect the unused 

channels to ground. You’ll often 

find that open inputs can pick up interference, 

which results in a garbled 

display. 

It should be clear that this analyser is 

not suitable for use with very high frequencies. 

Applications for this device 

are found with ‘slower’ microcontrollers, 

serial communications, etc. Even 

so, this simple circuit can make your 

life a lot easier during the development 

of a digital (microcontroller) circuit. 

PIC Microcontrollers 

Silent alarm, RGB fader, poetry box, 

night buzzer and more! 

For this project we’re offering a complete 

kit of parts (order code 060092- 

71), which consists of the display, the 

main board, a programmed microcontroller, 

the components and the enclosure. 

All that’s left for you to do is solder 

the components to the board and 

mount it in the enclosure. After connecting 

a battery or mains adapter you 

can start analysing straight away. 

For those of you who want to etch 

the board yourselves, the layout can 

be downloaded from our website at 

www.elektor.com, under file number 

060092-1.zip. And if you have the facility 

to program the PIC microcontroller, 

you can also download the source code 

from our site (file # 060092-11.zip). 

This hands-on book covers a series of exciting and fun 

projects with PIC microcontrollers. You can built more 

than 50 projects for your own use. The clear explanations, 

schematics, and pictures of each project on a breadboard 

make this a fun activity. You can also use it as a study 

guide. The technical background information in each 

project explains why the project is set up the way it is, 

including the use of datasheets. All software used in this 

book can be downloaded for free, including all of the 

source code, a program editor, and the JAL open source 

programming language. 

446 pages • ISBN 978-0-905705-70-5 • £27.00 • US$ 54.00 

Elektor 

Regus Brentford 

1000 Great West Road 

Brentford TW8 9HH 

United Kingdom 

Tel. +44 20 8261 4509 

Order quickly and safe through www.elektor.com/shop 

(060092-I) 

Publicité 


41

HANDS-ON MILLING 

Profiler 

Construction kit 

for a general-purpose 

milling machine 

Harry Baggen 

Have you always dreamed of having your own milling machine 

but found them just too expensive? If so, we have 

the perfect solution for you. Working in close 

collaboration with the Belgian manufacturer 

Colinbus, we have put together a construction 

kit for our readers that enables you to build a 

professional milling machine for a fraction of 

the cost of a ready-made model. This machine 

is suitable for a wide variety of jobs, ranging from 

making parts for models to milling circuit boards. 

As an electronic hobbyist or professional, 

you have to work with more 

than just pure electronics, and you often 

have to deal with a lot of mechanical 

tasks. This includes jobs such as 

making printed circuit boards, routing 

wiring, and fashioning a suitable 

enclosure with a corresponding front 

panel. Some of these jobs require 

suitable tools if you want to achieve 

good results – at least if you want to 

do it all yourself. 

Many hobbyists, as well as designers 

of prototypes and small development 

labs, would certainly be able to 

make good use of a small milling ma- 


chine for this sort of work. Investing 

in such a machine may be affordable 

for commercial use, but the situation 

is a bit different for home use. A 

good, accurate milling machine can 

easily cost several thousand euros, 

and even then you only have a basic 

model without all the bells and whistles 

(such as vacuum swarf removal 

and a high-speed spindle motor). 

We launched this project especially 

for all electronics hobbyists and professionals 

who regularly pursue their 

job and/or hobby at home. The idea 

for this construction project arose 

spontaneously during a conversation 

with Frank Jacops of the Belgian 

company Colinbus, which specialises 

in milling machines. When you 

talk with someone who not only sells 

milling machines but also designs 

them from the ground up, the conversation 

quickly turns to the fact that 

most electronics types have splendid 

ideas about all the nice things 

they could do with such a machine, 

but the price is an obstacle in most 

cases. Frank Jacops understood this 

immediately (he has been an avid 

reader of Elektor Electronics for many 

years), and he suggested offering a 

construction kit at a special price for 

a limited time, exclusively for readers 

of Elektor Electronics. 

�� 


Naturally, we must 

admit that the kit is not exactly 

inexpensive at £ 1099 or � 1599 plus 

shipping costs, but this still represents 

a savings of nearly £ 1400 relative 

to a comparable ready-assembled 

model. That’s an attractive reward for 

a day or so of assembly work. 

Of course, several small milling machines 

are also commercially available 

at a lower cost, but they are all 

quite small and have limited features. 

If you look at the photos in this article, 

you can see that we’re talking 

about a completely different category 

here. Robust construction, high accu- 



‘Profiler’ technical 

specifications 

Dimensions 453 (w) x 583 (d) x 468 (h) mm 

Max. working area 300 (X) x 400 (Y) x 100 (Z) mm 

Interface Serial 

Power 240 V, 50–60 Hz 

X/Y/Z linear transport MultiStab guideways 

X/Y/Z drive Stepper motors 

Positioning speed 60 mm/s 

Mechanical resolution 0.0075 mm 

Software resolution 0.025 mm 

Software Colinbus User Interface 

Conversion program for Gerber and Excellon files 

racy, and a generous working area of 

30 � 40 cm — now that’s something 

you can sink your teeth into! This machine 

is suitable for all sorts of jobs, 

such as dispensing, potting, camera 

inspection, boring holes in boxes, 

milling front panels, 

and even 

3D 

modelling (using 

separate software). You can also 

mill circuit boards with this machine, 

although the manufacturer explicitly 

wishes to state that this model is not 

designed primarily for this purpose, 

since it requires even higher accuracy. 

However, based on our experience 

it yields excellent results for average 

PCBs, and the necessary software 

is included. 

Assembly of the machine is quite 

straightforward, and the clearly 

written instructions practically exclude 

any problems. The accompanying 

circuit boards with the drive 

electronics are fully assembled and 

tested, so all you have to do is install 

and connect them. 

Design 

The ‘Elektor Profiler’, as we have 

christened the machine, is the smallest 

milling machine produced by Colinbus. 

Its construction is largely the 

same as the commercial CBR-40 model 

(see www.colinbus.com). 

The machine is made from steel and 

aluminium parts. This combination of 

materials provides sufficient weight 

and stability to withstand the motion 

of the machine, while the accuracy 

of the guideways and the speed of 

travel are very high thanks to the use 

of aluminium extrusions. 

The MultiStab system, which uses 

three rollers per edge travelling along 

precision steel rods pressed into aluminium 

extrusions, provides good 

mechanical guidance with minimum 

play. This design is used for all three 

axes. One roller of each set of three 

for each guideway assembly can be 

manually positioned to adjust the 

amount of play as necessary. Each of 

the three axes is driven by a spindle 

with a special zero-backlash nut. 

The milling machine is controlled 

by a built-in processor board fitted 

with a Renesas H8/3003 and a driver 

board fitted with three ST L6208 

ICs, which look after driving the 

three stepper motors. This approach 

makes the timing independent of the 

connected PC. The computer simply 

sends commands and coordinates 

to the processor board, which processes 

and executes the commands 

independently. 

The processing power of the builtin 

microcontroller is not sufficient to 

drive all three stepper motors at the 

�� 

15


same time (this trick is reserved for 

the larger and more expensive Colinbus 

models), so it can only drive 

two at the same time. The machine 

can thus execute smooth motions 

in a two-dimensional plane. For 3D 

motions, it switches rapidly back 

and forth between two axes, which 


Figure 1. The mechanical parts of the construction kit. 

makes it appear that all three motors 

move simultaneously. This has little 

noticeable effect in practice, since 

the steps resulting from the interpolation 

are very small. 

The control board in the milling machine 

has a serial interface, but a 

standard USB to RS232 adapter can 

Figure 2. The assembled circuit boards and the spindle motor and bracket. 

You can order the construction kit 

for the Profiler milling machine 

by filling in the Order Form on 

the Elektor Electronics website at 

www.elektor.com (click ‘Milling 

Machine’ under ‘Quick Service’). 

The price is 1099 pounds (1599 

euros) including VAT, plus 

shipping charges. The shipping 

charges depend on the country 

and are stated on the website 

Order Form. 

The kit is supplied and invoiced 

directly and exclusively by the 

manufacturer, i.e. not by Elektor. 

There are also several optional 

extras available for the machine. 

Consult the Colinbus website for 

information about these options. 

be used without any problems to operate 

the machine via the USB port 

of the PC. 

The spindle motor supplied with the 

milling machine is a fairly basic model 

from Ferm, which allows you to 

start using the machine right away. 

Naturally, you can also fit other types 

of motors, but we should remark here 

that a true professional-quality spindle 

motor can easily cost more than 

the entire construction kit. 

Construction kit 

What do you get for the price of £1099 

or 1599? The photos in Figure 1 and 

Figure 2 show almost all the parts 

included in the kit. They include all 

the mechanical parts, screws and 

bolts, guideway rollers, spindles, 

bearings, stepper motors, cables 

and so on, plus the drive electronics 

on the two pre-assembled circuit 

boards. The previously mentioned 

Ferm spindle motor (and associated 

bracket) is also included in the kit. A 

MDF panel (also included in the kit) 

serves as the base plate. 

The companion software for the 

milling machine consists of two programs: 

a user-friendly interface for 

configuring and operating the machine 

and a conversion program for 

PCB layouts (see below). 

The kit includes assembly instruc- 

�� 


tions that provide a step-by-step description 

of how to put the machine 

together. 

There’s not much that can go wrong 

during assembly. The various parts 

fit together very accurately thanks to 

the combination of precision holes in 

the steel panels and the steel rods 

in the aluminium extrusions, so no 

further alignment is necessary. You 

have to provide the necessary wiring 

for the two circuit boards for the 

control and drive electronics, and the 

connectors and transformer must be 

soldered in place. 

All the assembly steps are clearly 

described in the instructions. However, 

we would like to clarify a few 

details here. 

Once the machine is partially assembled, 

the guideways must be adjusted. 

The accuracy of the entire milling 

machine depends entirely on this 

adjustment. The guideways for the 

bridge (on the sides) have of three 

rollers, of which one can be adjusted 

using an adjustment plate. It must 

be adjusted so there is very little 

play, but it should not be made too 

tight, since otherwise the rollers will 

jam. A similar construction is also located 

at the top of the bridge and in 

the motor column, but the latter part 

must be partially disassembled to 

access the rollers. 

The drives use spindles with trap- 


Figure 3. Cross-section drawing of the boring head mount on the bridge. Figure 4. Detail of the plastic nut with the transverse pins, which provide a zero-play coupling. 


ezoidal threads that run in special 

plastic nuts (see Figure 4). These 

nuts are fitted with transverse pins 

that provide the coupling to the 

bridge or the boring column. The X 

and Y spindles must also be carefully 

aligned to achieve the specified 

accuracy and linearity. Note that the 

Figure 5. The boring column and the guideway in the bridge. 

special plastic nuts on these spindles 

must be lubricated using only 

the special grease included with the 

kit. Do not use normal grease! 

The rest of the assembly process is 

adequately described by the instructions 

included with the kit. 

�� 

17


Software 

The user interface specially written 

for the Colinbus milling machine provides 

the operator interface for the 

machine (Figure 6). This program is 

designed such that even inexperienced 

users can work with it easily. 


Figure 6. The user interface. The working surface of the machine is shown at the upper right. 

The screen is composed of several 

windows. The effective working surface 

of the machine is shown in the 

right-hand window. After loading a 

file, you can use the mouse to place 

the object to be machined in the desired 

location on this surface. The ma- 

Figure 7. Gerber and Excellon files can be read in with the PCB contouring program 

and converted into contour data for the milling machine. 

chine will then start working at exactly 

this position. You can also specify 

the values of all the settings and 

preferences for the machine. In addition, 

you can operate all machine axes 

manually using this program. Everything 

you can think of can be configured 

here, such as reading memory 

points, relative zero points and so on 

with micrometre accuracy. 

The user interface includes a PCB 

contouring program (Figure 7), which 

you can use to edit and merge the 

Gerber and Excellon files generated 

by commonly used PCB programs 

and then convert the coordinates in 

these files into the contours used by 

the milling machine. You can manually 

specify the reference points to be 

used on the circuit board and then 

use them to ensure that the data in 

the Gerber and Excellon files are registered 

exactly with each other (the 

Gerber file contains the data for the 

PCB tracks, while the Excellon file 

contains the hole data). The file generated 

by the program can then be 

imported into the user interface program 

so you can see the PCB layout, 

which you can drag to a specific location 

on the working surface. 

If you want to process 3D files (such 

as DXF files), you will need the separate 

RAMS3D program, which is not 

included in the construction kit due 

to the special price. 

Finally, there’s something for diehard 

DIYers: the entire command set for 

controlling the milling machine is 

freely available, so you can also write 

your own code. 

1001 jobs 

This robust, versatile milling machine 

opens up a world of opportunities 

for handy hobbyists and professionals 

who aren’t afraid to roll up their 

sleeves. With it, making rectangular 

openings in a box is a piece of cake 

and a machining a slot in a front panel 

for a slider is no sooner said than 

done. What’s more, this milling machine 

is a handy tool for things that 

have nothing to do with electronics. 

For instance, you can use it to make 

your own parts for a model airplane 

or toys for your kids (and donate your 

jigsaw to local charity!). 

In short, there are more than enough 

things you can do with this machine. 

Once you’ve used it, you won’t want 

to do without it. 


(060232-1) 

��

PROJECTS PROGRAMMABLE LOGIC CONTROL 


ECIO PLC 

Cheap, DIY and with CAN! 

Ben Rowland (Matrix Multimedia) & Luc Lemmens (Elektor Labs) 

Here’s the first real-life application of ECIO modules introduced in the October 2007 issue 

of Elektor. An ECIO acts as the brains of a PLC board that has relays, opto-isolators, CAN (!) 

connectivity and an LCD. All this I/O capacity together with Flowcode allows the board to 

act as a versatile, powerful PLC for quite complex control and automation projects. The LCD 

module is used to display ASCII characters to the user as a means of troubleshooting during 

the software development stage or for monitoring the system. 

ECIO PLC Features 

• CAN bus connection 

• 4 optically isolated inputs 

• 4 relay driven outputs 

• 2x16 character alphanumeric LCD 

• Flowcode-programmed but will also take PIC18 hex files 

• Free Flowcode for ECIO 

• USB connectivity for programming 

A full PLC application board comprising isolated outputs, inputs, 

LC display and CAN bus connectivity…sounds good! 

But you’d also like to hear that it’s low cost, based on an 

ECIO module and easy to program using Flowcode, the all 

graphics method of mastering PICs (and other micros, too). 

PLC 

A PLC (programmable logic control) is a device typically 

used as the central, intelligent, control element in a flowchart-designed 

industrial process, usually for mass manufacturing 

or quality checking and goods sorting. Conveyor 

belt controls will typically be PLCs. 

PLCs exist commercially and their (high level) programming 

languages have been standardized to a wide extent. Unfortunately 

the price tag of almost any commercial PLC puts it 

well out of reach of enthusiasts. A pity, because many Elektor 

readers have affinity with industrial control systems. 

PLC command sequences are in a way like computer programs 

— they follow a predefined sequence of events with 

all the options for conditional loops, event timers, logic 

conditions, event counters, analogue values (temperature; 

liquid level; pressure) and simple result logs. 

If you have a slightly complex process you’d like to con- 

trol using electronics on a machine, then a PLC is a good 

choice because of its (electrical) robustness and flexibility 

when it comes to connecting the real world through relays 

and optocouplers. 

To be able to program a PLC (i.e. define the sequence of 

events to happen in a process) you need to write a program. 

Once debugged and simulated on a PC, you can 

download the program to the PLC and from then on it’s 

fingers crossed. 

PLC programs can be edited, debugged, extended, optimized 

and of course stored and retrieved. Right, just like 

any PC or microcontroller program! 

Circuit description 

The circuit diagram of the ECIO PLC is shown in Figure 1. 

The circuit comprises a number of distinct elements which 

we’ll discuss below. 

ECIO 

The ECIO module is the brain of the board. It is used to control 

all of the peripheral devices on the board. ECIO modules, 

introduced in the October 2007 issue of Elektor [1], 

represent an ultra low-cost way of entering the world of PIC 

microcontroller programming. ECIOs are available from the 

Elektor Shop with good discounts for volume orders. Here 

we use the 40-pin version called ECIO-40P. 

Relays 

The relays the PLC is using to switch electrical devices on 

and off are controlled via ECIO pins RB4–RB7. Outputting 

a logic Low to these pins will switch off the corresponding 

relay and outputting a logic High will switch on the corresponding 

relay coil. If a relay is enabled then the LED next 

to that particular relay will light to give a visual on/off representation. 

The relays are used to provide electrical isolation 

between the ECIO module and the external switched 

voltage (which could be the 230 V mains). This means that 

power devices running from high voltages like 48 VDC or 

ECIO PLC • 36 

64 elektor - 3/2008

K1 

K2 

K3 

K4 

OI0 

K6 

+9V - +20V 

3 

1 

2 

VDD USB 

230 VAC can be controlled safely via the ECIO module. 

Each set of relay contacts (NO and NC) is brought out to 

three pins of a PCB mount screw terminal block — the centre 

pin (C) is the pole. 

Opto-isolators 

The opto-isolated inputs are special in not having a plus (+) 

and a minus (–) input terminal. Inside IC1, the two diodes 

on each input are in fact LEDs so the polarity of the control 


3/2008 - elektor 

T1 

D1 

+12V 

IC2 

VDD USB 

μA78M05CKC 

R7 

3 

B1 

1 

1 3 

VDD_EXT 

C1 C3 

C5 C6 

D5 

B40C1500 

10u 

25V 

100n 

100n 10u 

25V 

VDD USB 

OI1 

2 

4 

VDD USB 

T2 

9 

VSS 

8 

OSC1 

VDD 18 

7 

OSC2 

22p 20MHz 22p 

BC547 

BC547 

BC547 

BC547 

R2A R3A R2B R3B R2C R3C R2D R3D 

D2 

OI2 

2 

TXCAN 1 

RXCAN 2 

CLKOUT/SOF 3 

TX0RTS 4 

TX1RTS 5 

TX2RTS 6 

RX0BF 10 

RX1BF 11 

17 

RESET 

13 

SCK 

14 

SI IC3 

15 

SO 

16 

CS MCP2515 

12 

INT 

1 

TXD 

4 

RXD CANL 6 

VDD USB 

C7 

10u 

25V 

VDD USB 

K9 

IC1 

R1A 

R1B 

R1C 

R1D 

1 

2 

3 

4 

5 

6 

7 

8 

16 

15 

14 

13 

12 

11 

10 

9 

OI0 

OI1 

OI2 

OI3 

R4A 

R4B 

R4C 

R4D 

RA0 

RA1 

RA2 

RA3 

RESET 

R5A 

RB1 

R5B 

RC7 

R5C 

RB0 

R5D 

RB3 

R6 

RB2 

220R 

IC4 

MCP2551 

VDD USB 

R1 = 4x 4k7 

TLP620-4 

R4 = 4x 220R R5 = 4x 220R 

X1 

ECIO 

C2 

C4 

R2 = 4x 10k 

R3 = 4x 330R 

VDD USB 

T3 

D3 

OI3 

VDD USB 

T4 

D4 

390R 

2 

VDD 3 

VSS 

GND 

RESET 

RA0 

RA1 

RA2 

RA3 

RA4 

RA5 

RD0 

RD1 

RD2 

RD3 

RD4 

RD5 

RD6 

RD7 

VREF 5 

7 

CANH 

8 

RS 

1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

15 

16 

17 

18 

19 

20 

CAN H 

CAN L 

R8 

120R 

40 

39 

38 

37 

36 

35 

34 

33 

32 

31 

30 

29 

28 

27 

26 

25 

24 

23 

22 

21 

JP1 

END 

NODE 

VDD_EXT 

RB7 

RB6 

RB5 

RB4 

RB3 

RB2 

RB1 

RB0 

RC7 

RC6 

RC2 

RC1 

RC0 

GND 

RE2 

RE1 

RE0 

RC6 

RC2 

RC1 

RC0 

RD7 

RD6 

1N4001 

R9A 

RB4 

1N4001 

R9B 

RB6 

D6 

D7 

2 

4 

6 

8 

10 

12 

14 

+12V 

D8 

T5 

+12V 

R10A 

K5 

BC547 

D9 

R10B 

T6 

BC547 

1 

3 

5 

7 

9 

11 

13 

Re1 

Re2 

VDD USB 

VDD USB 

P1 

10k 

CONTRAST 

RESET 

RA4 

RA5 

RE0 

RE1 

RE2 

K14 

NO 

C 

NC 

voltage you wish to apply to ECIO PLC does not matter! The 

opto-isolator outputs are connected to ECIO pins RA0–RA3. 

These input pins will be at logic zero for no input voltage 

and at logic one for a voltage of 3.5 V or more. The optoisolators 

are used to provide an isolation layer from input 

voltages. This means that relatively high voltages can be 

used to safely control the ECIO module. The signals you 

want to process in the ECIO PLC are applied to the circuit 

through PCB mounted 2-way screw terminal blocks. LEDs 

D1-D4 show the logic status of the opto-isolator inputs. 

K15 

NO 

C 

NC 

RD4 

RD5 

RD0 

RD1 

RD2 

RD3 

1N4001 

R9C 

RB5 

1N4001 

R9D 

RB7 

VDD USB 

R11 

33R 

D10 

D11 

R9 = 4x 2k2 

R10 = 4x 1k 

+12V 

D12 

T7 

+12V 

R10C 

BC547 

D13 

R10D 

T8 

LCD1 

1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

15 

16 

BC547 

Re3 

Re4 

LC Display 2 x 16 

K18 

K19 

070786 - 11 

NO 

C 

NC 

NO 

C 

NC 

Figure 1. 

Circuit diagram of the 

ECIO PLC board. Minimal 

hardware — great 

potential in terms of I/O. 


65


Figure 2. 

The unstuffed ECIO PLC 

board is available 

from Elektor. 


K6 

K1 

K2 

K3 

K4 

B1 

R11 

CAN 

The CAN (controller area network) interface is used for adding 

the ECIO PLC onto a CAN network. CAN is ‘hip’ and 

few PLCs we have seen offer this connectivity. Here we use an 

MCP2515 CAN controller chip and an MCP2551 line driver. 

The CAN controller chip is connected ECIO peripheral pins 

set up to do SPI comms. There is also an interrupt pin, which 

is connected to ECIO pin RB2 and a chip select pin, which is 

connected to ECIO pin RB3. The CAN controller has its own 

20 MHz clock derived from a quartz crystal, X1. 

The actual connection of the CAN bus to the ECIO is by 

way of K9, a 2-way PCB mount screw terminal block. Jumper 

JP1 has to be installed only if the MCP2551 is the end 

node of the bus. 

C1 

R1 R3 

IC1 

COMPONENTS LIST 

Resistors 

R1 = 8-pin SIL array 4 x 4k�7* 

R2 = 8-pin SIL array 4 x 10k�* 

R3 = 8-pin SIL array 4 x 330�* 

R4,R5 = 8-pin SIL array 4 x 220�* 

R6 = 220� 

R7 = 390� 

R8 = 120� 

R9 = 8-pin SIL array 4 x 2k�2* 

R10 = 8-pin SIL array 4 x 1k�* 

R11 = 33� 

P1 = 10k� preset 

* see text 

Capacitors 

C1,C6,C7 = 10μF 25V radial 

IC2 

C3 C5 C6 

D5 R7 

D1 

D2 

D3 

D4 

T1 

T2 

R2 

T3 

T4 

R4 

C2,C4 = 22pF 

C3,C5 = 100nF 

LCD 

The LCD module is used to display ASCII characters to the 

user as a means of troubleshooting during the software development 

stage, or for monitoring the system as it’s busy 

checking and controlling the events in the automated process. 

The LCD is connected to ECIO pins RD0–RD5 with the 

four data bits taking up bits 0–3, the RS bit taking up bit 4 

and the Enable bit taking up bit 5. P1 is the LCD contrast 

adjustment. 

Power supply 

Nothing special in this department — the usual 7805 voltage 

regulator (IC2) and the traditional set of decoupling capacitors. 

A bridge rectifier is used ahead of the regulator to 

66 elektor - 3/2008 

14 

13 

R9 

T6 T7 

T5 

T8 

P1 

K5 

R10 D8 

Re1 

Re3 

D12 

Re2 

D9 

Re4 

D13 

R5 


D1-D5,D8,D9,D12,D13 = 3mm LED 

D6,D7,D10,D11 = 1N4001 

B1 = B80C1500 (round case; 80V piv @ 

1.5A p ) 

T1-T8 = BC547 

IC1 = TLP620-4 

IC2 = 7805 

IC3 = MCP2515-I/P 

IC4 = MCP2551-I/P 

Miscellaneous 

Re1-Re4 = 12V relay, SPDT, e.g. Omron 

G5LE-1 


2 

1 

C7 

IC3 

R6 

X1 

C2 

R8 

IC4 

D10 

C4 

D6 

D7 

D11 

K11 

K8 

K10 

K7 

JP1 

K9 

K1,K4,K9 = 2-way PCB screw terminal 

block, lead pitch 5mm 

K5 = 14-way boxheader 

K6 = AC/DC low-V adapter socket, 

PCB mount, e.g. CUI Inc. # PJ-002B 

(Digikey # CP-002B-ND) or Cliff Electronic 

Components # DC10B (Farnell 

# 224960) 

K7,K8,K10,K11 = 3-way PCB screw terminal 

block, lead pitch 5mm 

ECIO = ECIO-40P processor module 

(Elektor Shop) 

LCD1 = LCD, alphanumerical, 2x16 

characters, e.g. Displaytech 162 

JP1 = 3-way SIL pinheader with jumper 

PCB, order code 070786-1 from 

Elektor Shop 

ECIO PLC • 38

The ECIO family of USB programmable microcontrollers provides 

a simple way of adopting microcontroller technology 

into your projects. The device behaves just like a normal microcontroller 

— but when you plug the USB lead in and press 

the reset switch you can send a new program to the device. 

This makes the ECIO 

one of the lowest cost 

USB-compatible PIC 

programmers in the 

world. 

Currently there are two 

products in the range: 

ECIO-28P and ECIO- 

40P. These devices 

are based on PICmicro 

18 series devices; 

the 18F2455 and the 

18F4455 respectively. 

allow the ECIO PLC board to be powered by AC as well as 

DC wall warts with an output voltage between 9 and about 

20 volts. For the DC connection, polarity is irrelevant. 

Phew, that ECIO PLC board has a lot of connections to the 

outside world and to prevent losing track of all those I/O 

lines and associated devices we’ve made a summary in 

Table 1. 

Construction 

The component mounting plan of the PCB designed for 

ECIO PLC is shown in Figure 2. As usual the copper track 

layout may be downloaded from our website for those with 

the wherewithal to etch & drill their own circuit board at 

home. All others will like to hear that the bare board is 

available ready-made from the Elektor Shop. 

No SMD components are used in this project so stuffing the 

board should be mostly plain sailing if you use care and 

precision in handling and soldering the parts. The 8-pin resistor 

arrays containing four individually wired resistors 

(i.e. not top commoned!) may prove difficult to get. Instead 

of arrays, you can also use four individual resistors mounted 

vertically side by side. 

Give your completed board a good visual inspection before 

inserting ICs and powering up for the first time. 

ECIO hardware Testing — I/O and CAN 

To help you find your bearings we have written a simple 

and instructive ECIO test program, of which the Flowcode 

listing is shown in Figure 3. Download it, load it into Flowcode, 

run the simulation, compile and blow it into the ECIO 

module. 


3/2008 - elektor 

ECIO – the cheapest USB PIC progger out there 

VDD USB 

GND 

/RESET 

RA0/AN0 

RA1/AN1 

RA2/AN2 

RA3/AN3 

RA4/AN4 

RA5/AN5 

RE0/AN5 

RE1/AN6 

RE2/AN7 

RD0 

RD1 

RD2 

RD3 

RD4 

RD5 

RD6 

RD7 

ECIO40 CONNECTIONS 

1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

15 

16 

17 

18 

19 

20 

ECIO40P 

40 

39 

38 

37 

36 

35 

34 

33 

32 

31 

30 

29 

28 

27 

26 

25 

24 

23 

22 

21 

VDD EXT 

RB7 

RB6 

RB5 

RB4 

RB3/AN9 

RB2/AN8/INT2 

RB1/AN10/INT1/SCK 

RB0/AN12 

RC7/RX/SDO 

RC6/TX/CK 

RC2 

RC1 

RC0 

GND 

VDD USB 

GND 

/RESET 

RA0/AN0 

RA1/AN1 

RA2/AN2 

RA3/AN3 

RA4/AN4 

RA5/AN5 

The ECIO microcontrollers are pre-programmed with a bootloader 

program which allows you to send a new program to 

the microcontroller via USB, in principle as many times as you 

like. ECIO is compatible with hex code from any appropriate 

compiler. ECIO is directly compatible with Flowcode — a 

graphical programming language which greatly simplifies the 

code generation process 

ECIO28 CONNECTIONS 

1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

ECIO28P 

28 

27 

26 

25 

24 

23 

22 

21 

20 

19 

18 

17 

16 

15 

VDD EXT 

RB7 

RB6 

RB5 

RB4 

RB3/AN9 

RB2/AN8/INT2 

RB1/AN10/INT1/SCK 

RB0/AN12 

RC7/RX/SDO 

RC6/TX/CK 

RC2 

RC1 

RC0 

— but can also be used 

with any C compiler or 

Microchip’s own development 

suite MPLAB. 

ECIO is well supported 

with a wide range of 

learning and development 

tools including (free) 

Flowcode and inexpensive 

E-blocks. 

The program includes a general PIC setup routine that does 

all the port and I/O pin configuring on the ECIO. For example, 

ECIO lines RB4-RB7 are set up as output lines (for 

the relay drivers). The test program is simple to use: if you 

drive the opto-isolator on RA0, the relay on RB4 will be energised. 

Similarly with the combination RA1/RB5, and so 

Table 1. ECIO connections overview 

Relay Output ECIO I/O Pin 

Relay Re1 (K14) RB4 




Opto-Isolator Inputs ECIO I/O Pin / LED 

K1 RA0 / D1 

K2 RA1 / D2 

K3 RA2 / D3 

K4 RA3 / D4 

CAN (from CPU) ECIO I/O Pin 

Serial Data Out RC7 

Serial Data In RB0 

Serial Clock RB1 

/Interrupt RB2 

/Chip Select RB3 

LCD ECIO I/O Pin 

D0 RD0 

D1 RD1 

D2 RD2 

D3 RD3 

RS RD4 

Enable RD5 


67


Figure 3. 

This Flowcode program 

runs a hardware test 

on the ECIO PLC. 


on. The display will show: ‘Elektor ECIO PLC board’ (no? 

adjust that LCD contrast!). Neither family members nor PIC 

geeks may be too impressed but if this works then a whole 

lot of PIC code is being executed and you can safely assume 

that all hardware is functioning as it should. 

The CAN test utility discussed in the inset is also included 

in the free software download for this project. The two test 

programs provided are educational and certainly worth 

looking at even if you do not actually build the project (hint: 

use the free Flowcode version to begin with). The archive 

file number is 070786-11.zip. 

PLC programming 

All ECIO PLC programming bears great resemblance to Eblocks 

and Flowcode so if you have any experience with 

these you’re in luck. If not, there’s a mass of information 

out there on the Internet [1], in previous issues of Elektor 

and in the Flowcode package itself. Best of all, Flowcode 

for ECIO is free [1]. 

Writing a process control program for the ECIO PLC underscores 

what Flowcode is all about: rather than worrying 

about syntax and PIC assembly code you are working at 

a fairly high level, setting up a full-blown flowchart of the 

program and let Flowcode arrange all the compiling, initializing, 

error tracking and code downloading to the ECIO 

module. Of course you can simulate your PLC program so 

there’s a good chance of instant success when the ECIO 

PLC board is connected up to the real world. 

More advanced users may want to rely on their own methods 

of producing PIC18 code using C++ compilers or similar. 

The ECIO even accepts straight hex code from all you 

assembly code diehards out there. Simply use the free USB 

I/O drivers to connect ECIO to your PC. 

(070786-I) 

CAN interface testing 

CAN is essentially good for sending complex message structures 

between a number of ECUs (microcontrollers). The test 

program written for Flowcode simply looks for an echo of 

the outgoing CAN message. This could be used to see how 

many nodes are on the network or used to calculate the distance 

between nodes based on echo time, etc. 

Basically the ECIO test program is sending out a specific 

CAN message with an standard ID of 12. When the Multiprogrammer 

receives a CAN message it checks the ID and 

if it is equal to 12 then it resends the message ID 12. The 

ECIO continues to send out this message ID until it receives 

a CAN message back. Once it has received a message back 

it checks the ID and then confirms or denies if the echo was 

successful. 

For this test you will need another ECIO PLC board or an 

E-blocks Multiprogrammer connected to an E-blocks CAN 

module. One ECIO PLC board will send several messages 

over the CAN bus. The other one will ‘listen’ until a predefined 

message appears and then replies to the ECIO. When 

the CAN transmitter is started the display reads ‘Startup’, 

followed by ‘Done’ a bit later once the CAN controller is initialized. 

If the CAN connection is okay the display shows ‘Message 

returned’, if not, you’ll see ‘Message failed’. 

Web Link and reference 

[1] http://www.matrixmultimedia.com/ECIO-X.php 

[2] EasyControl I/O, Elektor October 2007. 


68 elektor - 3/2008

PROJECTS JTAG ADAPTOR 

Universal 

JTAG Adaptor 

Marcel Cremmel 

For 

programming and emulation 

This adaptor was originally intended to allow programming of the memory and CPLD of the PSD813 

used in the GBECG Gameboy cartridge, which converts this games console into an electrocardioscope 

(see October 2006 issue). But it’s much more universal than that (see box entitled ‘In-Circuit JTAG’) Our 

adaptor connects to a PC parallel port and uses the JTAG IEEE 1149.1 protocol. 

Informed microelectronics amateurs 

will of course be aware that other ‘In- 

Circuit’ programmable devices use this 

same port (parallel) and an identical 

protocol. Unfortunately, the programmer/emulators 

intended for these devices 

are not compatible — far from it 

in fact: so there’s no point hoping for a 

mixed marriage! 

However, closer examination of the circuit 

diagrams of certain programmers 

suggested by the IC manufacturers 


shows that the differences are relatively 

minor and in fact concern the interconnections 

between the LPT port signals 

and the JTAG connectors. So a few multiplexing 

functions is all it takes to produce 

a ‘universal’ adaptor. 

Had it been achieved using conventional 

logic components, the circuit of our 

adaptor would have been quite complex, 

with different electronics for each 

of the sections for the different types 

of processor. Using an EP900 program- 

mable logic circuit (Altera, on free offer 

from Elektor) makes it possible to offer a 

very cheap and simple programmer. 

Many manufacturers have adopted the 

JTAG (Join Test Action Group) protocol 

for programming, debugging, and testing 

their ICs in situ on the board (IC for 

In Circuit). Fortunately, you don’t need 

to know all the details of this protocol 

to be able to use it: the PC software 

(usually free) and the target components 

each include a JTAG core that al- 

Universal JTAG Adaptor • 41 


lows them to communicate completely 

transparently. 

The devices involved have special 

‘JTAG’ pins that you merely need to 

connect to the pins of the same name 

on the programmer connector. The size 

(number of contacts) and pinning of this 

connector differ from one manufacturer 

to another. This information is given in 

the various diagrams shown in the boxes 

of Figures 1–4, concerning respectively 

Altera CPLDs and EPLDs (Byteblaster 

II) (Figure 1), Xilinx CPLDs and 

EPLDs (Parallel Download Cable) (Figure 

2), MSP430 microcontrollers from 

Texas Instruments (LPT IF 4 wire JTAG 

Communication) (Figure 3) and the 

PSD, uPSD and DSM families (Flashlink 

FL101) from ST Microelectronics (Figure 

4). It should also be noted that there is 

a certain discrepancy in the naming of 

the signals between the different JTAG 

connectors. 

ADAPTOR CIRCUIT 

The heart of the circuit (Figure 5), 

which with its 44 pins could hardly go 

unnoticed, is an EP900 PLD. This PLD 

forms the link between the PC’s parallel 

port, K1, and the four DIL pin headers 

for the JTAG connections to the four targets, 

named respectively MSP430 (K2), 

FLASHLINK (K3), XILINX (K4) and AL- 

TERA (K5). SW, a dual-gang DIP switch 

comprising contacts JP1 and JP2, allows 

selection of one of the 4 types of 

programmer recognized by the JTAG 

adaptor (see truth table in the circuit 

diagram, also given on the component 

overlay on the board). These four options 

appear in the form of the same 

number of HE-10 headers in the bottom 

right-hand part of the circuit. Each option 

has its own logic structure within 

the EP900; all these various sub-assemblies 

using logic gates are shown 

in Figure 6. 

Each of these structures is drawn from 

the manufacturers’ programmer circuits. 

For reasons of effi ciency, the EP900’s 

logic structure is described in Altera’s 

AHDL language. The circuit diagram is 

easier for an electronics technician to 

read, but the ‘AHDL’ form is more efficient 

here. Just for information, the 

‘source’ fi le (.tdf) for the contents of the 

EP900 is given in the inset. 

At the bottom left we fi nd the… 

POWER SUPPLY 

The EP900 PLD is quite an old IC already! 

It requires a 5 V supply, but as 

its consumption is quite high, the pro- 



JTAG ‘In-Circuit’ – 

some applications 

– PSDs, uPSDs and DSMs from ST Microlectronics 

– MSP430 microcontrollers from Texas Instruments 

– EPLDs and CPLDs from ALTERA 

– EPLDs and CPLDs from XILINX 

TCK 1 2 GND 

TDO 3 4 VCC 

TMS 5 6 

7 8 

TDI 9 

10 GND 

VCC 

2 

4 

6 

8 

10 

1 

3 

5 

7 

9 

TCK 

TDO 

TMS 

Figure 1. CPLD and EPLD (Byteblaster II) from Altera: 10-pin DIL connector. 

Software: Quartus II Web Edition, Quartus II Programmer [1] 

1 2 VCC 

GND 3 4 TMS 

GND 5 6 TCK 

GND 7 8 TDO 

GND 9 10 TDI 

GND 11 12 

GND 13 

14 

1 2 

3 4 

5 6 

7 8 

9 10 

11 12 

13 14 

V CC 

TDI 

1k 

TMS 

TCK 

TDO 

TDI 

V CC 

1k 

1k 

XILINX 

Figure 2. CPLD and EPLD (Parallel Download Cable) from Xilinx: 14-pin DIL connector. 

Software: ISE WebPACK [2] 

TDO 1 2 VCC out 

TDI 3 4 VCC in 

TMS 5 6 TCLK 

TCK 7 8 TEST 

GND 9 10 

RST 11 12 

13 

14 

VCC TOOL 2 

VCC TARGET 4 

6 

TEST/VPP 8 

10 

12 

14 

J1 

J2 

1 

3 

5 

7 

9 

11 

13 

TDO/TDI 

TDI/VPP 

TMS 

TCK 

RST 

47k R1 

C1 

10n/2n2 

1k 

TMS 

TCK 

TDO 

TMS 

TDI 

060287 - 12 

V CC 

VCC 

GND 

VCC 

VCC 

Target 

Altera 

Device 

GND 

TDI TDO TDI TDO TDI TDO 

C2 

TCK 

C3 

100n 

TMS 

TCK 

060287 - 14 

TMS 

TCK 

060287 - 13 

V CC 

VCC /AVCC /DVCC 

TDO/TDI 

TDI/VPP 

TMS 

TCK 

RST/NMI 

MSP430Fxxx 

TEST/VPP 

VSS /AVSS /DVSS 

Figure 3. MSP430 microcontrollers (LPT-IF 4-wire JTAG Communication) from Texas Instruments: 14-pin DIL connector. Software: 

IAR-Kickstart [3] 


57


About the author 

Marcel Cremmel, the author, has been a qualifi ed lecturer in Electrical Engineering, electronics 

option, since 1979 (state certifi ed by the French National Education system). 

After completing his fi rst years of teaching in the School of Engineering in Rabat in Morocco, 

under the Co-operation scheme, in 1982 he was assigned to the Louis Couffi gnal College in 

Strasbourg, in the BTS SE section (Higher Technician’s Certifi cate, ‘electronics systems’). 

His job requires him to cover all fi elds of electronics, though his preference is for telecommunications, 

video, microcontrollers (MSP430 and PIC) and programmable logic devices 

(Altera). 

Alongside electronics, his other passion is motorbikes in all their forms: touring, competitions, 

etc. His personal website is at http://electronique.marcel.free.fr/ 

1 2 

GND 3 4 

TDI 5 6 

VCC 7 8 RST 

TMS 9 10 GND 

TCK 11 12 GND 

TDO 13 

14 

RST 

1 

3 

5 TDI 

7 

9 TMS 

11 TCK 

13 TDO 

gramming adaptor can’t be powered directly 

from the outputs of the PC’s LPT 

port. To simplify implementation and allow 

us to dispense with a special dedicated 

power supply, we have decided 

to power the adaptor from the power 

rails in the target systems. But these 

are usually content – especially nowadays! 

– with 3 V or 3.6 V, which is not 

enough to power the EP900. 

So we’ve fi tted the adaptor with a very 

fl exible switched capacitor voltage converter 

that supplies a regulated 5 V output 

from an input voltage anywhere between 

2.7 and 5.5 V! Yes, that’s right: 

the converter works just as well with 

an input voltage either lower or higher 

than the output voltage, with an effi - 

ciency of around 90%! Bravo to the Burr 

Brown engineers (that company since 

taken over by Texas Instruments, which 

explains why the spec. sheet has to be 

obtained from the TI website). However, 

the current is limited to 30 mA. 

10k 

2 

4 

6 

8 

10 

12 

14 

System Reset Circuity 

(connect directly to RST 

input on PSD) 

10n 

100k 

100k 

100k 

100k 

User I/O Signals 

Figure 4. PSD, uPSD and DSM families (Flashlink FL-101) from ST Microelectronics: 14-pin DIL connector. 

Software: among others, PSDsoft Express [3] for programming the PSD813 in the ECG cartridge for Game Boy. 


The only awkward point for amateurs 

is the size of the regulator IC (it’s only 

available in an SM version), making it 

tricky to solder. But luckily it only has 

six pins. So its now or never, to try your 

hand with an SM device. Position IC2 

accurately on its pads. Apply a little solder 

to one of the pad + legs. Once the 

solder has set, solder the leg diametrically 

opposite the previous one. If everything 

is OK, now solder the remaining 

legs. If you create a solder bridge 

between two legs, remove it using desoldering 

wick. 

CONSTRUCTION 

USER 

PC BOARD 

PSD or PSD Port C 

TDI - PC5 

VSTBY or PC2 

TMS - PC0 

TCK - PC1 

TDO - PC6 

General I/O - PC3 



060287 - 15 

As shown in Figure 7, the board designed 

for this project is double sided; 

it uses only a very few SM components, 

mainly around the EP900. Naturally, 

these are to be fi tted on the track side 

of the board. So let’s get stuck in! For 

reasons of practicality, we recommend 

starting with the SM components. 

Watch out – certain of them, in particular 

capacitor C1, are tucked away at the 

centre of the board, right between the 

legs of the PLCC44 socket (into which 

the EP900 is going to be plugged, on 

the other side). Take care to solder the 

regulator IC2 carefully, as without this, 

nothing else will work. It’s surrounded 

by capacitors that are bigger than 

it is. Take care to identify the values of 

the SM components correctly (resistors 

often have coded value information: 

103 means 10 k�, 1203 means 120 k�; 

things are trickier with the capacitors, 

which are often not identifi ed or identifi 

able. Once the SM components are 

fi tted, you can fi t the row of resistors, 

the rest of the conventional components, 

the selector SW, the headers K2 

(MSP430) to K5 (ALTERA), the PLCC44 

socket, fi nishing off with the 25-pin sub- 

D connector K1. Make sure you pick the 

male version of the printer connector 

(LPT); the female version won’t make 

for a very good connection! One little 

note about the dual selector SW: it’s not 

always easy to get hold of a dual DIP 

switch, so we’ve left enough room to fi t 

a quad one, but you’ll need to cut off the 

spare legs before you fi t it. 

If you’re making your own board, it’s 

equally possible to make it single-sided 

– the second side of the double-sided 

board is in fact only used to avoid the 

need for the wire links that a single-sided 

version will require. Construction is 

the same, but in this case, it’s preferable, 

for reasons of practicality, to start 

off by fi tting various wire links, using 

tinned copper wire. 

Take care to avoid shorts with the wire 

links positioned between the ‘FLASH- 

LINK’ and ‘XILINX’ connectors, which 

are relatively close together. 

All that remains is to plug the EP900 

into its socket. Check the quality of your 

construction one last time (soldering, 

component values – luckily there’s only 

one value for the conventional resistors), 

as there is no way of testing the 

proper operation of this circuit except 

by trying it out for real! 

Note about the EP900 PLD (order 

code 060287-41): this is available programmed, 

free of charge (apart from 

standard postage and packing charges) 

from the Elektor SHOP. If you order PCB 

# 060287-1, the programmed IC will be 

automatically supplied with it. 

TARGET CONNECTIONS 

Watch out – you must only use one 

connector at a time! In most cases, a 



K1 

SUB D25 

VCC IN 


1 

14 

2 

15 

3 

16 

4 

17 

5 

18 

6 

19 

7 

20 

8 

21 

9 

22 

10 

23 

11 

24 

12 

25 

13 

STROBE 

AUTOFDX 

D0 

ERROR 

D1 

INIT 

D2 

SLCTIN 

D3 

GND1 

D4 

GND2 

D5 

GND3 

D6 

GND4 

D7 

GND5 

ACK 

GND6 

BUSY 

GND7 

PE 

GND8 

READY 

100R 

R1 

100R 

R10 

100R 

R9 

100R 

R17 

100R 

100R 

R8 

R11 

100R 

100R 

R7 

R12 

100R 

R6 

100R 

100R 

100R 

100R 

100R 

100R 

100R 

C3 

R5 

R4 

R3 

R13 

R14 

R15 

R16 

220n 

6 4 

PUMP+ PUMP- 

100k 

5 

IN 

IC2 

OUT 

1 

REG710NA-5 

3 

EN 

R28 R29 R30 R31 R32 

100k 

+5V 

100k 

C5 C2 2 

C4 C7 

100k 

100k 

STRB 

AFDX 

D0 

ERR 

D1 

INIT 

D2 

SLCT 

D3 

D4 

D5 

D6 

D7 

ACK 

BUSY 

PE 

RDY 

2 

24 

C1 

CLK1 

CLK2 

20 

IN 

19 

IN 

21 

IN 

30 

IN 

25 

IN 

26 

IN 

27 

IN 

41 

IN 

32 

I/O 

31 

34 

33 

35 

37 

38 

40 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

100n 

+5V 

1 

IC1 

44 

EP900LC 

22 

23 

VCC IN 

K4 

XILINX 

1 2 

3 4 TMS TDI 

5 6 TCK TMS 

7 8 Nstat TDO 

9 10 TCK RST 

11 12 

13 14 

I/O 

10 

SEL0 

43 

SEL1 

42 

17 

NC 39 

NC 

IN 

IN 

IN 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

I/O 

3 

4 

5 

11 

12 

13 

14 

15 

16 

18 

28 

29 

36 

6 

7 

8 

9 

10k 

R35 

+5V 

10k 

R37 

JP1 

10k 

R20 

100R 

R24 

100R 

R25 

100R 

100R 

R19 

R21 

100R 

R22 

100R 

R23 

100R 

100R 

100R 

100R 

R26 

R18 

R27 

R36 

JP2 

VCC IN 

10k 

R33 

100k 

R34 

TDO 

TDO F 

Nstat TDO 

TMS TDI 

TCLK 

TCK TMS 

TDI TMS 

TCK RST 

TDO TCK 

TCK A 

VCC IN VCC IN K3 

K5 

FLASHLINK 

ALTERA 1 2 

TCK A 1 2 

3 4 

TDO 3 4 

TMS TDI 5 6 

TMS TDI 5 6 TCK TMS 

7 8 TDO TCK 

Nstat TDO 7 8 TDO TCK TDI TMS 9 10 

TDI TMS 9 10 

TCK RST 11 12 

TDO F 13 14 

+5V 

TDO 

TMS TDI 

TCK TMS 

TDO TCK 

TCK RST 

JP1 JP2 

ON ON ALTERA 

OFF ON XILINX 

ON OFF PSD 

OFF OFF MSP430 

VCC IN 

VCC IN 

K2 

MSP430 

1 2 

3 4 

5 6 TCLK 

7 8 TDI TMS 

9 10 

11 12 

13 14 

060287 - 11 

Figure 5. The EP900 takes pride of place in the centre of the circuit for the universal JTAG programmer. It’s available ready-programmed, free of charge, when you order the PCB 060287-I. 



59


ALTERA 

TRI 

D0 TCK 

D1 

D2 

D3 

D6 

AFDX 

WIRE 

D4 ACK 

TDO 

PIN7 

TRI 

TRI 

TRI 

TRI 

WIRE 

WIRE 

WIRE 

TMS 

PIN8 

PIN6 

TDI 

BUSY 

READY 

ERROR 

D0 

D1 

D2 

D3 

D6 

D4 

TDO 

VCC 

XILINX 

TRI 

TRI 

TRI 

NOT 

WIRE 

WIRE 

WIRE 

WIRE 

AND2 

TDI 

TCK 

TMS 

BUSY 

PE 

ERROR 

ACK 

READY 

D0 

D1 

D2 

D3 

FLASHLINK 

TRI 

TRI 

TRI 

NOT 

TCK 

TMS 

TDI 

NOT 

D5 RSTN 

D6 

TDO 

WIRE 

NOT 

WIRE 

WIRE 

WIRE 

READY 

PE 

ACK 

BUSY 

ERROR 

simple 10- or 14-way ribbon cable, with 

a crimped connector at each end (watch 

out for the orientation!) is used to establish 

the electrical links between the 

target and the adaptor (see the circuit 

diagrams of Figures 1 through 4 and the 

pinning of the relevant connectors). 

If you have direct access to the rear of 

the PC, the adaptor can be inserted directly 

into the LPT port, without using 

an extender cable interconnecting the 

PC and the JTAG adaptor. 

USB adaptors 

The latest offi ce and notebook PCs no 

longer have parallel ports (LPT) – a 

highly regrettable decision, especially 

for this project! To make up for this, 

you can fi nd USB/LPT adaptors, but do 

make sure you check their compatibility 

with our JTAG programmer; many of 

them will only accept printers. We can’t 

go into details of the programming procedures 

for all the possible targets, so 

we’re going to confi ne ourselves to one 

example, the… 

GameBoy ECG cartridge 

The cartridge uses an SMD connector 

with a pitch of 1.25 mm (K3). To make 

the cable, we recommend you follow 

the following procedure. 

1. Press a piece of 14-way ribbon cable 

to a female DIL14 connector; 

2. Use the Molex connector and the 

wires already prepared in the components 

list (see Elektor Electronics October 

2006) to make up the appropriate 

6-way connector for K3; 

3. Solder the four wires TCK, TDI, TDO 

and TMS and the two power supply 

wires to both connectors; 

4. Check the connections with a continuity 

tester and then insulate the soldering 

with heat-shrink sleeving. 

And there you go, all ready to program 

the PSD813s in the GameBoy ECG 

cartridge. 

One last remark: the adaptor is compatible 

with Byteblaster II (Altera); it does 

not work with the fi rst version of the 

driver (Byteblaster on its own, without 

the II). This old driver was used by 

the MaxPlus II software, and has been 

replaced by Quartus for two or three 

years now). 


D0 

D1 

D2 

MSP430 

TRI 

TRI 

TRI 

NOT 

TDI 

TMS 

TCK 

TRI 

TDO PE 

SLCT 

WIRE 

INIT TEST 

TRI 

STRB RST 

Figure 6. Nothing like it to illustrate the fl exibility of a PLD like the EP900! A single device can fulfi l several complex logic functions. 

060287 - 16 

Figure 7. Component overlay for the board designed for this project. The track layout is available for free download. 


AFDX 

NOT 

WIRE 

WIRE 

WIRE 

WIRE 

ACK 

BUSY 

READY 

ERROR 

(060287-I) 

Universal JTAG Adaptor • 45

Bibliography and Internet links 

[1] https://www.altera.com/support/software/ 

download/sof-download_center.html 

[2] http://www.xilinx.com/ise/logic_design_ 

prod/webpack.htm 

[3] http://focus.ti.com/docs/toolsw/folders/ 

print/iar-kickstart.html 

[4] http://mcu.st.com/mcu/modules.php? 

name=Content&pa=showpage&pid=57 

REF710-5 data sheet: 

http://focus. 

ti.com/lit/ds/symlink/reg710-5.pdf 

Supplementary information, fi le # 060287- 

11.zip, free download from: www.elektorelectronics.co.uk 


list 

Resistors 

R1,R3-R27 = 100� 

R28-R32,R34 = 100k� (SMD) 

R33,R35,R26,R37 = 10k� (SMD) 

(R2 not fi tted) 

Capacitors 

C1 = 100nF (SMD 1206) 

C2,C4 = 2μF2 (SMD 1206) 

C3 = 220nF (SMD 1206) 

C5,C7 = 47μF 10V radial 

(C6 not fi tted) 


IC1 = EP900LC (programmed, order 

code 060287-41) * 

IC2 = REG710-NA5 

Miscellaneous 

K1 (K_LPT) = 25-way sub-D plug, (male), 

right-angled pins, PCB mount 

K2 (FLASHLINK), K3 (MSP430), K4 (XI- 

LINX) = 14-way 2-row pinheader 

K5 (ALTERA) = 10-way 2-row pinheader 

J1,J2 (SW) = 2-way DIP switch 

PLCC-44 socket 

Project software, fi le # 060287-11.zip, 

free download from Elektor website 

PCB, order code 060287-1 

* Ready-programmed PLD supplied free 

when ordering PCB # 060287-1 from 

the Elektor SHOP 

Optional 

Parts for the cable connection to K3 on 

the GBECG: 

- 14-way (2x7) press-on IDC socket 

- Molex socket, 6-way, 1.25mm lead 

pitch (RS Components # 279-9178) 

- 6 wires with crimped contacts for 

Molex connector (RS Components # 

279-9544) 



‘AHDL’ source fi le for the EP900 

Contrary to fi rst impressions, an AHDL fi le can tell you a lot. Looking at this one a little more 

closely, it’s easy to spot the various options (->). 

subdesign prog_jtag_univers 

( 

TDO,Nstat_TDO,TDO_F : input; 

STRB,AFDX,INIT,SLCT : input; 

D[6..0] : input; 

SEL[1..0] : input; -- 0->ALTERA,1->XILINX, 

-- 2->FLASHLINK,3->MSP430 

ACK,BUSY,READY,ERROR: output; 

TCK_A,TMS_TDI,TCK_TMS,TDO_TCK,TDI_TMS,TCK_RST,PE : bidir; 

) 

variable 

TCK_A,TMS_TDI,TCK_TMS,TDO_TCK,TDI_TMS,TCK_RST,PE : tri; 

begin 

TCK_A.in=D0; TCK_A.oe=AFDX; 

case SEL[] is 

when 0 -- ALTERA 

=> TMS_TDI.in=D1 ; TMS_TDI.oe=AFDX; 

TCK_TMS.in=D3 ; TCK_TMS.oe=AFDX; 

TDO_TCK.in=D2 ; TDO_TCK.oe=AFDX; 

TDI_TMS.in=D6 ; TDI_TMS.oe=AFDX; 

TCK_RST.in=GND; TCK_RST.oe=GND; 

ACK =D4; 

BUSY =TDO; 

PE.in=GND; PE.oe=GND; 

READY=Nstat_TDO; 

ERROR=GND; 

when 1 -- XILINX 

=> TMS_TDI.in=D2 ; TMS_TDI.oe=!D3; 

TCK_TMS.in=D1 ; TCK_TMS.oe=!D3; 

TDO_TCK.in=GND; TDO_TCK.oe=GND; 

TDI_TMS.in=GND; TDI_TMS.oe=GND; 

TCK_RST.in=D0 ; TCK_RST.oe=!D3; 

ACK =GND; 

BUSY =D6; 

PE.in=D6; PE.oe=VCC; 

READY=Nstat_TDO & D4; 

ERROR=VCC; 

when 2 -- FLASHLINK 

=> TMS_TDI.in=D2 ; TMS_TDI.oe=!D3; 

TCK_TMS.in=GND; TCK_TMS.oe=GND; 

TDO_TCK.in=!D5; TDO_TCK.oe=VCC; 

TDI_TMS.in=D1 ; TDI_TMS.oe=!D3; 

TCK_RST.in=D0 ; TCK_RST.oe=!D3; 

ACK =GND; 

BUSY =GND; 

PE.in=!TDO_F; PE.oe=VCC; 

READY=D6; 

ERROR=GND; 

when 3 -- MSP430 

=> TMS_TDI.in=D0 ; TMS_TDI.oe=!SLCT; 

TCK_TMS.in=D1 ; TCK_TMS.oe=!SLCT; 

TDO_TCK.in=D2 ; TDO_TCK.oe=!SLCT; 

TDI_TMS.in=INIT; TDI_TMS.oe=VCC; 

TCK_RST.in=STRB; TCK_RST.oe=!AFDX; 

ACK =GND; 

BUSY =GND; 

PE.in=TDO; PE.oe=!SLCT; 

READY=GND; 

ERROR=GND; 

end case; 

end; 

For info: the ‘Jedec’ programming fi le (prog_jtag_univers.jed) is available from the Elektor 

website (www.elektor-electronics.co.uk). 


61

HANDS-ON ELECTROCARDIOGRAPH 

GBECG 

Marcel Cremmel 

Lots of electronics hobbyist dream 

of recording an electrocardiogram 

(ECG) using a circuit built at home. Usually 

out of technical curiosity, as numerous problems have to be 

resolved in order to properly sample the heart’s electrical activity. Alternatively, some people 

require personal medical monitoring while under a cardiologist’s care. And then it’s great to be 

able to make your own ECG and show it to your GP or clinical staff. 

The idea of using a Nintendo Gameboy 

games console equipped with a special 

cartridge was inspired by the 

world-famous Elektor GBDSO [1] (a big 

thank you goes out to Steve Willis for 

his help with this project). 

Our electrocardiograph utilizes three 

electrodes: one on each wrist, the 

third on the left leg. The electronic 

device, built on a cartridge that slips 

into any Gameboy model, processes 

the sampled signals and produces a 

very high quality ECG scrolling across 

the LCD (see the various illustrations). 

The electrocardiogram implements 

the method of M. Einthoven (see the 

inset on the next page). It only uses 

two active electrodes, a third being 

used to set the no-signal level of the 

first two. All leads are single-ended. 

Despite this simplicity, the results are 

noticeable and even recognized as 

usable by a cardiologist. The electrocardiograph 

easily meets the initial 

specifications for which it was 


32 

designed: to monitor tolerance to the 

anti-malaria medication. 

To do that, we measure the QT interval 

(see Figure 1) which should remain 

‘normal’. Figure 1 [2] matches up the 

electrical activity sampled and the cardiac 

cycle phases, as follows: 

P-Wave: Auricular contraction; the 

blood coming from the veins is 

pushed into the ventricles. 

QRS Complex: Ventricular contraction; 

the blood contained inside is 

pushed into the arteries. 

Both of these waves produce the ‘lubdub’ 

heartbeat sounds. 

T-Wave: Repolarisation of the ventricles; 

the ventricular muscle returns 

to rest. 

The electronics! 

After this little ‘dose’ of general knowledge, 

let’s deal with our favourite subject: 

describing the GBECG electronic 

structures and making the board. 

Just as with the GBDSO [1] (Elektor 

Electronics October and November 

2000), the specific electronics and software 

(in Flash memory) are grouped 

onto a cartridge that slips into the console’s 

connector. In this way, the 

Gameboy is transformed into a powerful 

electrocardiograph! 

The electronic device processes the 

very low voltages sampled between 

the two active electrodes. The singleended 

leads are designated DI, DII and 

DIII according to their localisation (see 

drawing in Figure 2). 

The most common lead is DI. 

Due to its low peak-to-peak amplitude 

(of the order of one mV), the EMF (electromotive 

force) measured is considerably 

amplified (about 1000�) before it 

can be converted to 8-bit digital. The 

sampling frequency selected is 477.84 

Hz, compatible with the spectrum of 

an ECG signal. 

The digital signal is then taken care of 

by the console processor. It is then 

placed in an 8 kBytes cyclic buffer 

GBECG • 47 


GameBoy 

ElectroCardioGraph 


It is practically impossible to grasp the operation of this 

home-made electrocardiograph instrument without a minimum 

of medical knowledge. We deal with the heart of the 

subject. First of all... 

A bit of history… 

In passing, let us pay 

homage to Willem 

Einthoven who discovered 

the relationship 

between electrical phenomena 

and muscular 

contraction of the human 

heart about 100 years 

ago. He received the 

Nobel Prize in 1924 for 

that discovery. 

Willem Einthoven, inventor of 

the electrocardiograph. 

and a bit of biology… 

The heart is an autonomous muscle: it is the only one not 

controlled by the brain. The ‘sinus node’, located in the right 

auricle, triggers nerve flows that control the heart muscles. 

These contract (‘depolarisation’ in medical lingo) and relax 

(‘polarisation’) in order to make up the blood pump that 

gives us life. The contraction is caused by a change in electrical 

polarity on each side of the cellular membranes. 

During the relaxation phases, the electrical charges find their 

state of equilibrium before being stimulated again. 

The resulting potentials are transmitted to the skin surface. 

They can then be sampled by cutaneous electrodes, as the 

human skin is sufficiently conductive. 

A wise placement of the electrodes allows a cardiologist to 

deduce the heart’s mechanical behaviour (and its defects!) 

by analysing the electrical activity. 

Characteristics: 

• Cartridge compatible with Nintendo Gameboy consoles 

type Classic, Pocket, Colour or Advance 

• Single-ended connections using 3 electrodes 

• Sensitivity: 1.6 mV full scale 

• Common mode rejection: 100 dB 

• Trace memory : 68 s 

• Scrolling display 

• Temporal window: 2.6 s in acquisition mode (1.3 s or 

2.6 s in consultation mode) 

• Heartbeat indicator 

• Battery power supply required 

• Approx. 2 hours use from battery power 

The Electrocardiogram (ECG) 

Technology has greatly evolved since the 1920s. The first patients dipped their 

hands and feet in basins full of very salty water! 

String galvanometer, The U-shaped magnet ends are enveloped in water cooling 

tubes (well before PCs!) Photos : Stichting Einthoven Foundation 

GBECG • 48 



34 

ECG 


P 

R 

Q S 

auricular contraction diastole 

(repolarisation) 

T 

ventricular contraction 050280 - 27 

Figure 1. Relationship between the measured electrical activity and the cardiac cycle phases. 

I 

EA 

EB 

EC 

F1 

differential 

amplification 

AD1 

F5 

K 

IC1.D 

IC1.A + IC1.B 

� 

F4 

1 /2 

summing 

IC1.C 

II 

Figure 2. These single-pole leads are used to implement the electrocardiograph. 

F2 

differential 

amplification 

AD2 

IC2.D 

F6 

auto-zero 

control 

IC2.A 

Figure 3. Block diagram of the analogue part of the circuit. 

III 

050280 - 12 

F3 

lowpass 

filter 

A3 

IC2.C 

reference: 

2V5 

050280- 13 

ECG 

memory and reread to show the ECG 

in real time on the screen, in ‘scrolling’ 

mode. 

The analogue part 

Presenting an adequate signal to the 

input of the digital analogue converter 

presents a challenge to the electronics 

engineer because there are a number 

of technical problems to analyse and 

resolve. 

Differential amplifier 

The peak-to-peak amplitude of the signals 

sampled between the electrodes 

is very low at just 2 mV max. 

Also, both the human body and the 

connecting wires to the electrodes are 

strongly influenced by high noise levels 

radiated by mains wires and other 

power carrying leads inside buildings. 

Capacitive coupling, although very 

low, produces a relatively high voltage 

(often over 1 V) to appear with respect 

to ground, despite the relatively low 

frequency of just 50 Hz or 60 Hz. 

To begin with, it would seem difficult 

to isolate the useful signal because its 

amplitude is 1,000 times lower than all 

the interference around! Moreover, the 

mains frequency is included in the useful 

spectrum; so the filtering solution 

does not work here. 

However, considering the wavelength 

of the mains voltage (6,000 km!), it is 

safe to assume that that each point on 

the skin receives the same induced 

potential thanks to its conductivity. 

Therefore, a common-mode voltage is 

developed with respect to the electrodes. 

In this case, the solution becomes 

obvious: we’re going to use a differential 

instrumentation amplifier with an 

adequate common-mode rejection ratio 

(CMRR): 

Sp 

S 

CMRR � 

SECG 

N 

dB dB 

� � 

� � � 

� � 

� � 

� � 

�� 

�� 

In this formula: 

S P = amplitude of the interference: 1 V 

S ECG = ECG amplitude: 1 mV 

S/N= signal to noise ratio: 40 dB 

Or : CMRR � 100 dB. 

In addition, the amplifier must be 

characterised by a very high input 

impedance (> 10 M�) and a low offset 

voltage. 

Numerous integrated instrumentation 

amplifiers exist (the AD624, for exam- 

GBECG • 49 



100n 

TP 

GND 

050280 - 11 

C14 

BAV99 

1 

BAV99 

AUDIOIN 

GND 

12 

3 

2 

RESET 

14 

IC1.D 

13 

11 

D3 

R21 

47k 

R22 

100k 

9 

19 46 

2n2 

RESET 

TP 

GND 

ADD 

TP 

2V5 

RESET 

RST 

C12 

TP 

41 

100n 

PA6/D6 

PA7/D7 

15 

14 

4 

5 

6 

7 

8 

9 

10 

D1 

D0 

1n 

R20 

390k 

100n 

D5 

100k 

D1 

D0 

8 

+5V 

1M 

C16 

D3 

D2 

R25 

1k 

R24 

3 

BC848B 

C13 

PSD813F2A 

-90M 

PA4/D4 

PA5/D5 

17 

16 

D3 

D2 

C10 

IC1.C 

D5 

D4 

R19 

1 

IC2.A 

5 

+5V 

2 

IC2.B 

D6 

7 

1� 

4 

IC2 

11 

JTAG 

IC4 

PA2/D2 

PA3/D3 

20 

18 

D5 

D4 

10 

9 

6 

100n 

TCK 

GND 

PC0/TMS 

PA0/D0 

PA1/D1 

D7 

BAV99 

1% 

1M 

T1 

C8 

C11 

100k 

100k 

210k 

12 

2M2 

R23 

BAV99 

12 

13 

22 

21 

PC1/TCK 

D7 

D6 

+5V 

TMS 

PC2/VSTBY 

A14 

A15 

470p 

10 

11 

PC3/TSTAT 

ADIO14 

ADIO15 

38 

39 

A14 

A15 

R17 

R18 

R10 

R16 

14 

13 

D2 

+5V 

+5V 

TDO 

TDI 

PC4/TERR 

K3 

A12 

A13 

C9 

6 

7 

PC5/TDI 

ADIO12 

ADIO13 

36 

37 

A12 

A13 

GND 

GND 

16 

15 

+5V 

P1 

22k 

D4 

A10 

A11 

K1 

ELECTRODES 

EA 

EB 

EC 

10/2006 - elektor electronics 35 

47k 

4 

5 

PC7/DBE 

PC6/TDO 

ADIO10 

ADIO11 

34 

35 

A10 

A11 

5 

1% 

+5V 

18 

17 

A9 

R15 

IC1.B 

47k 

220k 

1% 

ADIO9 

+5V 

ADIO8 

19 

A8 

7 

1% 

32 

33 

A8 

A9 

R13 

R14 

6 

20 

A7 

100k 

100n 

100n 

100n 

524.25kHz 

131.0625kHz 

45 

44 

PB6 

PB7 

ADIO6 

ADIO7 

29 

30 

A6 

A7 

21 

A6 

Gameboy Connector 

12k 

22 

A5 

1% 

R11 

C7 

C5 

C4 

4 

23 

A4 

R12 

48 

47 

PB4 

PB5 

ADIO4 

ADIO5 

27 

28 

A4 

A5 

24 

A3 

10k 

2 

6 

VIN+ D0 

IC3 

3 

7 

VIN- CLK 

ADC08831IM 

5 1 

VREF CS 

TP 

50 

49 

PB2 

PB3 

ADIO2 

ADIO3 

25 

26 

A2 

A3 

25 

A2 

33n 

5% 

26 

A1 

1% 

R9 

PB0 

PB1 

PB0 

PB1 

ADIO0 

ADIO1 

23 

24 

27 

A0 

C2 

TP 

52 

51 

A0 

A1 

100k 

10 

IC2.C 

4k7 

13 

1% 

PD2/CSI 

CNTL2/PSEN 

RAMCS 

8 

R7 

2 

1% 

IC2.D 

220k 

1% 

42 

22k 

PD1/CLKIN 

CNTL1/RD 

R8 

14 

9 

2 

1 

29 

28 

IC1.A 

47k 

R5 

R6 

8 

PD0/ALE/AS 

CNTL0/WR 

1 

12 

3 

40 

43 

WR 

RD 

30 

47k 

R4 

3 

100n 

ECLK 

WR 

RD 

2M2 

1% 

R2 

C1 

5% 

560p 

22k 

8 

31 

32 

31 

+5V 

C6 

262.125kHz 

Figure 4. The bulk of the work is handled by the ISP flash memory, IC4 and IC3, an A/D converter, Serial I/O. The Gameboy processor handles the processing. 

DO 

CLK CAN 

CS 

CLK 

R1 

R3 

100n 

K2 

C15 

+5V 

+5V 

+5V 

100n 10� 

BAV99 

4 C3 

IC1 

11 100n 

IC1 = TLV2254AID 

IC2 = TLV2254AID 

TP TP TP 

CS CLK DO 

TP 

ECG 

TP 

+5V 

C17 

C19 

D1 

1,0485MHz 

+5V 

+5V 

+5V 

GBECG • 50



36 

255 

N ECG 

0 

1V75 4V25 

VECG 050280 - 14 

Figure 5. The transfer function is determined by divider bridge R3/R12. 

RAM_CS 

RD 

WR 

ECLK 

+5V 

1 

top view 

A0 - A15 D0 - D7 

GAME BOY 

32 

RESET 

AUDIO IN 

GND 

050280 - 15 

Figure 6. Connector pinout for the Gameboy cartridge (view from above). 

ple). These are very high-performance 

devices and need no adjustment. But 

quality comes at a cost. 

We decided to make the differential 

amplifier using more economical operational 

amplifiers. This also allows significant 

savings in cost and power consumption. 

Moreover, these opamps 

function perfectly with a single 5 V 

power supply (this is not the case for 

the AD624). The disadvantage is the 

presence of an adjustable potentiometer 

to optimise the CMRR. 

Block diagram and wiring diagram 

Figures 3 and 4 respectively give the 

block diagrams for the analogue part 

and the complete circuit diagram. The 

references associated with each function 

(ICx.y) identify the operational 

amplifiers for the structural diagrams 

that show the functionality. 

The instrumentation amplifier is made 

up of functions F1 and F2. Function F3 

is a 2nd order low pass filter with a 

roll-off of 170 Hz and a damping factor 

m of 0.73 (i.e., near Butterworth). It will 

faithfully attenuate all unwanted components 

outside of the useful frequency 

spectrum and replaces the 

anti-aliasing filter for the DAC 

(digital/analogue converter) that follows 

it. 

The gain distribution in the circuit is as 

follows: A1 = 21�, A2 = 4.7� and A3 = 

10�. The total amplification is 987, in 

compliance with our objectives. The 

other functions (F4, F5 and F6) assist 

the instrumentation amplifier in order 

to ensure its proper operation. In fact, 

the operational amplifiers have a supply 

voltage of between 0 V and 5 V. 

The ideal no-signal voltage on each of 

the terminals is 2.5 V. Setting this level 

is not a problem in most cases: a 

divider bridge with two resistors is 

appropriate (R23 and R24). It is harder 

for the two input amps because we 

must take care not to compromise their 

input impedances. 

The problem is solved using the third, 

common, ECG electrode (see Figure 3) 

and the functions of F4 and F5. 

It can be shown that the voltage S 

equals half the sum of the voltage (EA 

+ EB). It is compared to the ‘2.5-V’ setting, 

and the error voltage is amplified 

to produce an ECG signal that can be 

processed. As there is no current flow 

in the electrodes, voltages EA and EB 

are equal to EC (give or take a few 

mV). In this way, the human skin actually 

helps to keep EA and EB equal to 

the target level of 2.5 V. That is the goal 

GBECG • 51 



we are seeking: the no-signal voltage 

of the opamps is as desired without 

reducing the input impedance. 

Moreover, a natural, but very annoying 

phenomenon occurs when we place 

the electrodes: an electromotive force 

(EMF) contact potential is produced 

between the skin and the electrode 

metal. This ‘micro-cell’ is very weak (a 

few mV) but it is not eliminated by the 

instrumentation amplifier. On the contrary, 

it is amplified! 

Functions F4 and F5 partially reduce 

this effect, but the offset in S1 and S2 

may still reach 1 V in differential mode. 

This value is unacceptable and is 

therefore compensated by function F6. 

F6 compares the average S3 signal 

value with the 2.5 V setting. The error 

voltage is integrated (time constant 

R16C8 = 2.2 s) in order to produce the 

ZERO signal. This continuous voltage 

offsets the S3 signal until its average 

value is stabilised at 2.5 V. 

To increase the amplitude of this compensation, 

two diode pumps C9-D4- 

C11 and C12-D5-C13 produce –3 V and 

+8 V supply voltages for IC2. 

The digital part 

The digital-analogue conversion is performed 

by IC3. It integrates a real differential 

amplifier, but requires an 

external reference voltage. This is simply 

derived from the 5 V power supply 

using a potential divider (R23/R24) 

buffered by T1. The precision and the 

stability are average at best, but still 

sufficient for this application. The 

divider R3/R12 determines the transfer 

function (see Figure 4). 

The asymmetry with respect to 2.5 V is 

justified by the asymmetrical shape of 

an ECG in relation to its average value. 

The DAC provides its NECG results in 

‘serial’ format. It is controlled by the CS 

and CLK signals — the first triggers 

the conversion (its frequency is 477.84 

Hz) and the second clocks the data 

output (on DO). 

The PSD813F2 

A Gameboy cartridge plugs into a connector 

which accommodates the console’s 

microprocessor buses: 

• Addresses: A15 to A0 

• Data: D7 to D0 

• Check: ECLK, WR, RD and RESET 

The first Gameboy games consoles on 

the market from about 1989 had a 

microprocessor similar to the old Z80, 

which explains the size of the buses. 

The most recent consoles are equipped 

with much more powerful CPUs, but 

for commercial reasons the old cartridges 

(and even older than those) still 

operate on the current Gameboys. That 

is also the case for our electrocardiograph! 

The PSD813F2 is an integrated circuit 

that’s perfectly adapted for making a 

cartridge for a Gameboy console. 

Unfortunately, we do not have enough 

space in this article to describe its 

complete functionality (see [3]). In 

summary, the PSD813F2 includes: 

• a configurable microprocessor interface 

that can be adapted to all 8-bit 

microprocessors on the market, 

The 

author 

including the vintage Z80; 

• 128 kBytes of Flash memory (only 

32 kBytes are utilised by the electrocardiograph 

object code, which 

leaves space for extensions or other 

things...); 

• a complex programmable logic 

device (PLD) that takes care of the 

address decoding; 

• a 16-cell sequential Complex Programmable 

Logic Device (CPLD). It 

is responsible for the serial-to-parallel 

conversion of the DAC frames to 

relieve the CPU and produce the 

squarewave signals required for the 

diode pumps; 

• 27 I/O configurable ports; 

• 2 kBbytes RAM (not used). 

Moreover, a JTAG interface is used to 

completely configure the ‘on circuit’ 

component via connector K3, which 

should not fail to interest all of you 

electronics hobbyists. The PSDSoftExpress 

development environment can 

be downloaded for free on the manufacturer’s 

website. 

As you can see from the diagram (Figure 

4), the connection between the 

console bus and the PSD813 is simple: 

the signals of the same name are connected. 

The only particularity that could be 

The author, Marcel 

Cremmel, has been a qualified 

teacher in Electrical 

Engineering (with 

Electronics option) since 

1979 (French National 

Education state diploma). 

After having initially taught 

at the Mohammedia de 

Rabat Engineering School 

in Morocco as a participant 

in the Cooperation program, 

he was assigned to 

the Louis Couffignal High 

School in Strasbourg in 

1982, in the BTS EL section 

(Certificate, Senior 

Electronics Technician). 

Although his profession forces him to deal with all domains of electronics 

Marcel has a preference for telecommunications, video, microcontrollers 

(MSP430 and PIC) and programmable logic circuits (Altera). In addition to electronics, 

his other passion is motorcycles in all forms: touring, races, etc. 

His personal website is at http://electronique.marcel.free.fr/ 

interpreted as an error: the data bus 

connections are crossed! That enabled 

us to simplify the board layout and the 

binary file of the GBECG control program 

was changed to suit. 

The software 

The software is entirely written in 

assembly language. The author used 

the ‘Gameboy Assembler Studio’ environment 

by Nicklas Larsson (freeware 

available on the web [4]). 

The assembly language was necessary 

because the specifications require a 

‘scrolling’ display in real time. That 

GBECG • 52 



occupies the CPU of the first consoles 

at a level of 80% due to the ‘old’ manner 

in which the screen memory is 

organised (separate screen memory 

and character memory). 

The software can be thought of as handling 

four tasks: 

1. Initialisations 

This task is executed at switch-on or 

after a reset, 

• Initial assignment of variables. 

• Configuration of I/O ports. 

• Initialisation of the LCD. The screen 

has 160 x 144 pixels, but for technical 

reasons, the useful part is reduced 

38 

EPS 

050280 - 1 

COMPONENT 

LIST 

(all SMD, except K1) 

Resistors 

(all 0805 case) 

R1 = 2M�2 1% 

R2,R15,R21 = 47k� 

R3 = 22k� 

R4,R13 = 47k� 1% 

R5,R14 = 220k� 1% 

R6 = 22 k� 1% 

R7 = 4k�7 

R8,R11 = 100k� 1% 

R9 = 10k� 

R10 = 210k� 1% 

R12 = 12k� 

R16 = 2M�2 

R17,R18,R22,R25 = 100k� 

R19 = 1k� 

R20 = 390k� 

R23,R24 = 1M� 

P1 = 22 k� preset (Bourns 3314G) 


EPS 

050280 - 1 

Figure 7. Copper track layout and component mounting plan of the double-sided board designed for the GBECG. 

Soldering of IC3 is particularly difficult, so we’re supplying the board with all components pre-soldered. 

to 160 x 96 pixels. The lower part 

(160 x 48) is utilised for fixed messages. 

• Internal timer: this is programmed 

to produce interrupts at a rate of 

477.84 Hz (sampling frequency). 

• Sound generator: this is pre-programmed 

to produce a cardiac 

‘beep’ when required. 

2. Main loop 

The main loop simply detects the 

action of certain keyboard keys and 

modifies the operating mode: 

• Start: acquisition mode 

• Select : stop mode 

Capacitors 

(all 0805 case except C8 and C19) 

C1 = 560pF 5% 

C2 = 33nF 5% 

C3-C7,C11,C13-C17= 100nF 

C8 = 1μF (1208) 

C9 = 470pF 

C10 = 1nF 

C12 = 2nF2 

C18 = not fitted 

C19 = 10μF (1208P) 


IC1,IC2 = TLV2254AID 

IC3 = ADC08831IM (Analog Devices) 

or TLC0831CD (Texas Instruments) 

IC4 = PSD813F2A-90M 

(STMicroelectronics), programmed, 

order code 050280-41 

D1-D5 = BAV99 

T1 = BC848B 

Miscellaneous 

K1 = Molex connector, 5-way, Dubox 

EPS 

050280 - 1 

• �: zoom _1 in stop mode 

• �: zoom _2 in stop mode 

3. Timer interrupt program 

This task is executed a rate of 477.84 

times per second. It carries out the following 

functions: 

• Debounced readout of the keyboard 

state. 

In run mode: 

• triggers a new conversion; 

• acquisition of the last sample (result 

of the previous conversion); 

• all 4 samples (or 119.46 times per 

second); 

• calculation of the ‘average sample’ 

89882-405, Digikey # 90148-1102- 

ND 

Programming connections: 

K3 = Molex connector, 6-way, 1.25mm 

lead pitch, type 53261-0671 (Digikey 

# WM7624CT-ND) 

Optionally: FlashLink programming 

connection: 

Molex 6-way connector, 1.25mm lead 

pitch, female (Digikey # WM1724- 

ND) 

6 wires with pin terminations for Molex 

connector (Digikey # WM1775-ND) 

Electrodes: 

Cutaneous probes or clips are available 

from medical supplies outlets 

5-way SIL pinheader 

4 mm plate (3x) 

6 m screened audio cable 

PCB, with all components ready fitted, 

tested, order code 050280-91 

GBECG • 53 



= average of the last 4 samples; 

• detection of the R-wave in order to 

trigger the cardiac ‘beep’; 

• loading the 8 kBytes cyclic buffer 

with the average sample. 

4. Interrupt program, V-Blank 

This interrupt is produced at the end 

of each LCD vertical sweep. The rate 

is: f v = 59.73 Hz; or every two averaged 

samples. 

The program also takes care of refreshing 

the display. 

• Run mode: the LCD shows the last 

samples (or 320 averaged sampled 

values on the cyclic buffer), which 

traces the last 2.68 s on the screen 

width. 

• Stop mode: depending on the zoom 

value: �1 or �2, the LCD shows the 

last 320 or 160 samples, which 

traces the last 2.68 s or 1.34 s on the 

screen width. 

In stop mode, the program also detects 

action on the � or � keys allowing 

the use to move around within the 

screen memory. 

During the display of the ECG trace, 

the program draws the vertical and 

horizontal scales. The latter moves 

with the trace in order to improve readability. 

The source code file of the 

The electrodes 

A good ECG can only be obtained with good, well-placed 

and properly-wired electrodes. 

To limit the effect of undesirable signals, it is recommended 

that you use shielded cables. Electronically, ‘audio’ cables 

are perfectly suited to this function. In practice however they 

are much too fragile, We therefore propose making small 

adaptors on which the cable clasps eliminate practically all 

risk of breaking (see Figure F), 

You will see that the shielding is only connected at the cartridge 

side and that it is isolated on the electrode side in 

order to avoid any contact with the skin. 

A 

GBECG control program is available as 

a free download from the Elektor Electronics 

website, the file number is 

050580-11.zip, see under month of 

publication. Improvements and additions 

are welcome. 

Building it 

The use of SMD components is 

unavoidable. In fact, the underside of 

the board must be perfectly smooth 

(therefore, no through wires or pins) so 

that it can slip into the cartridge case. 

To save you the trouble of struggling 

with (and losing) those tiny SMD parts, 

we are supplying the GBECG printed 

circuit board with all components 

already soldered in place, and the 

PSD813 programmed, all at an affordable 

price (see the inset). The order 

code is 050280-91. 

All that is left to do is find an old 

Gameboy cartridge case and clear out 

the two half-shells a little. Photo A 

illustrates the cutouts to make. 

The wider cutout in the upper halfshell 

is due to potentiometer P1 which 

otherwise prevents closing the case. 

Adjustment 

The only adjustment consists in optimising 

the CMRR of the differential 

‘4 mm’ type plugs make it possible to use commercially available 

electrodes (Figure G). 

The clip is very practical and adapted for children. But the 

price of these two electrodes may discourage many readers 

(more than £6 each and you will need three off!) 

Homebrew electrodes can be made from coins as illustrated 

in the opposite photo. The author used French Florin (FF) 

coins which are made from nickel. Solder a 4-mm socket and 

the electrode (Figure H is ready to use. 

Three rubber bracelets keep them in place on the wrists and 

the lower calf. These bracelets can be made by cutting ribbon 

for straps to the proper length and gluing them to the 

self-adhesive ‘Velcro’ tape at both ends. We can also use 

sections from air chambers from a motorcycle or a scooter. 

E F G H 

GBECG • 54 



B 

C 

amplifier. For that, you need a function 

generator and an oscilloscope or an AC 

voltmeter. Begin by making the measurement 

circuit in Figure B (the capacitor 

has a capacitance of 10 μF). Top to 

bottom: EA, EB, EC, GND. 

Plug the device into the fixed connec- 

D 


40 

tor K1, observing the orientation, and 

only then connect the generator. In this 

way, we inject a common-mode test 

signal. Adjust the generator: 50 Hz 

sine wave with amplitude 1 V. Insert 

the cartridge in the Gameboy with the 

upper half-shell removed in order to 

access the ECG test point. Turn on the 

power and measure the AC component 

of the ECG signal. Then adjust P1 to 

minimise the peak-to-peak amplitude. 

It should be less than 25 mV in order to 

obtain an S/N ratio in excess of 40 dB. 

Final check 

This step is not strictly required. Its 

purpose is to ensure proper operation 

of the electrocardiograph by injecting 

a signal and by verifying the result on 

the screen. 

Most benchtop signal generators are 

incapable of reliably producing the 

very low levels required for this check. 

Therefore, we must greatly attenuate 

the GBF signal. That is what the 

device shown in Figure C does. 

The signal is effectively divided by 

100. In this way, we inject a 1 Hz sinusoidal 

signal and a 140-mV DC amplitude 

signal. These should result in an 

image on the LCD screen similar to 

that in Figure D. 

The sinewave is aligned with the first 

dotted line and has a DC amplitude of 7 

divisions, or 7�200 μV = 1.4 mV. 

The fixed connector K1 on the cartridge 

is not very sturdy. To limit the 

risk of deformation or of it being torn 

out, restrain the three shielded cables 

on the cartridge cover with two cable 

ties, as illustrated in photo E. 

The cable ties we used on our prototype 

of the GBECG do not touch the 

internal components. They require four 

2 mm holes to be drilled which do not 

significantly weaken the shell. 

(050280-1) 

We would like to extend our thanks to Professors 

Schalij and Maan, Leiden University Hospital, the 

Netherlands, for their valuable assistance. 

Bibliography 

[1] GBDSO — Gameboy Digital 

Storage Oscilloscope, 

Elektor Electronics 

October and November 2000. 

Internet 

references 

and links 

[1] http://chem.ch.huji.ac.il/ 

~eugeniik/history/einthoven.html 

GBECG • 55 


Operating 

instructions 

Placement of the electrodes 

It is absolutely necessary to clean the skin and the electrodes 

well with a cotton and ether or alcohol, In this way, the electromotive 

force EMF on the contacts, which may saturate the 

amplifiers, are limited considerably. 

The standard lead is ‘DI’: 

• EA electrode: right wrist 

• EB electrode: left wrist 

• EC electrode: left foot (lower calf) 

Utilisation of ‘contact’ products based on potassium chloride 

considerably improve the quality of the measurements. 

The best ECGs are obtained when the patient is calm and 

lying down so that the only muscle in action is the heart. 

Operation 

• Live start-up: welcome screen is displayed. 

• Go to the acquisition screen: press Start, A, B or Select. 

[2] http://www.e-cardiologie.com/ 

examens/ex-electro2.shtml 

[3] http://www.st.com/stonline/products/ 

[4] http://www.devrs.com/gb/ 

Additional documents and 

program sources on 

http://www.elektor.com 

http://www.infoscience.fr/histoire/ 

biograph/biograph.php3?Ref=128 

PSD813 datasheet 

http://www.st.com/stonline/ 

products/literature/ds/7833.pdf 

ADC08831IM datasheet 

http://www.ortodoxism.ro/ 

datasheets2/6/ 

0rcoik1yuwhx1dj2ogg8wid7sfcy.pdf 

Attention 

The GBECG electrocardiograph described 

in this article has not received medical 

approval and is therefore not intended 

• ECG acquisition: if 

the electrodes are for professional use. The instrument 

properly placed must always be powered by batteries in 

and the patient is 

order to respect protection category III. 

calm, the reading 

should be stabilised 

in a few seconds and look like the one illustrated in 

Figure I. 

Do not use this ECG as a reference; the shapes may vary 

appreciably from one individual to another. 

If no trace appears after thirty seconds, clean the skin below 

the electrodes using a pad and ether or alcohol. 

Irregularity of the trace may be reduced by using a ‘contact’ 

product. 

• Stop mode: pressing the Select button stops the acquisition. 

You can then analyse the memory which contains 

68.6 s worth of ECG. 

�: zoom x 1 

�: zoom x 2 (see below) 

�: Move ahead 

�: Move backwards 

I Scales (zoom 1x) 

A beep sound will be heard each time an R wave is detected. 

The volume can be adjusted with the volume button on 

the console. 

Attention: The screen memory is erased when power is 

removed. 

Elektor blockbusters GBECG • 56 


PROJECTS GAMEPAD CONVERSION 

Tilt Gamepad 

Upgrade your Gamepad 

with acceleration sensors 

Xin Wang and Marko Westphal 

Users of the Nintendo Wii and Play station 3 ‘tilt’ controllers have raved about the more intuitive 

control these devices offer. Up until now there hasn’t been a comparable gamepad available for the 

dedicated PC gamer but why should they be left out of all the fun? Join in by adding this two-axis 

tilt sensor to a standard gamepad, it is particularly good for vehicle and flight simulation as well as 

adventure games. Give those thumbs a rest and start waving your arms around! 

A tilt gamepad senses the angle 

at which the handheld controller is 

moved and converts that measurement 

into equivalent digital outputs 

which would be produced by pressing 

the up/down left/right buttons 

on the gamepad. It is not necessary 

to press any of the buttons 

to control direction; the 

on-screen object is controlled 

simply by tilting the gamepad. 

In this design the movement 

is detected by an acceleration 

sensor manufactured by Analog 

Devices and sensor values 

are processed by an Atmel 

ATmega8 microcontroller. The 

2 

entire circuit fits onto a small 

PCB which converts a standard 

gamepad into a tilt gamepad. 

The sensor 

The novel component in this 

design is the analogue acceleration 

sensor type ADXL322 

from Analog Devices. This 2axis 

device produces two independent 

output voltages 

proportional to the inclination 

of the sensor in the x and 

y planes. The supply voltage 

can be in the range of 2.4 V to 


Vs 14 

6 V. The two analogue output signals 

have a sensitivity of 420 mV/90�. The 

sensor range is ±2 g and it is supplied 

in an SMD CP-16 package which cannot 

be soldered into place using a con- 

+5V 

C2 

GND 

100n 

ST 

Yout 10 

Xout 12 

IC2 

ADXL322 

GND 

GND 

GND 

GND 

3 

5 

6 

7 

GND 

Vs 15 

GND 

8 

20 

XTAL1 

9 

ventional soldering iron so the PCB is 

supplied with this component already 

mounted. 

Signal processing 

The output signals from the acceleration 

sensor are analogue 

so it is necessary to process 

them using a microcontroller 

with an on-board A/D converter. 

The Atmel ATmega8 is an 

8-bit microcontroller with six 

multiplexed analogue inputs 

which can be selected internally 

as an input to the 10-bit 

resolution A/D converter. Up to 

23 of its pins can be configured 

as general-purpose digital I/O 

pins. 

The two analogue signals representing 

the X/Y tilt from 

the acceleration sensor are 

connected directly to the A/ 

D converter inputs of the microcontroller. 

The signals are 

digitised, filtered and then 

converted into digital output 

signals which emulate the up/ 

down, left/right function of the 

direction buttons on the original 

gamepad. 

The X and Y values are sam- 


7 

1 

PC6 (RESET) 

PD0 (RXD) 2 

PD1 (TXD) 3 

PD2 (INT0) 4 

PD3 (INT1) 5 

PD4 (XCK/T0) 6 

VCC 

PD5 (T1) 11 

PD6 (AIN0) 12 

PD7 (AIN1) 13 

PB0 (ICP) 14 

PB1 (OC1A) 15 

PB2 (SS/OC1B) 16 

PB3 (MOSI/OC2) 17 

PB4 (MISO) 18 

PB5 (SCK) 19 

AVCC 

IC1 

28 

PC5 (ADC5/SCL) 

27 

PC4 (ADC4/SDA) 

26 

PC3 (ADC3) 

25 

PC2 (ADC2) 

24 

PC1 (ADC1) 

23 

PC0 (ADC0) 

ATmega8-16PI 

21 

AREF 

C3 

100n 

C4 

22p 

+5V 

X1 

4MHz 

GND 

XTAL2 GND 

10 

C5 

22p 

22 

C1 

GND 

100n 

+5V 

GND 

K1 

070233 - 11 

Figure 1. Besides the acceleration sensor and microcontroller there are very 

few other components required. 

0 

Tilt Gamepad • 57

pled alternately, the 2.56 V reference 

for the A/D converter is produced onchip 

and decoupled by capacitor C3 on 

Pin 21 (AREF). The I/O pins have good 

sink/source current capability which 

together with selectable internal pullup 

resistors means that there is no requirement 

for additional drivers for the 

output signals. 

Simple circuitry 

It can be seen in the circuit diagram in 

Figure 1 that apart from the microcontroller 

and sensor there are very few 

additional components required. The 

layout of the double-sided PCB shown 

in Figure 2 is therefore quite simple. 

Figure 3 gives the flow chart describing 

the main software functions. The 

microcontroller ADC port is sampled 

every 10 ms, raw values of acceleration 

are converted into tilt values which are 

then filtered. The signals output by the 

gamepad depend on the direction of 

tilt and tilt angle. 

The 6-way pin header (K1) is fitted 

to the PCB for all the connections to 

the gamepad. The circuit is powered 

directly from the USB interface (+5 V 

and ground). 



Putting it together 

The finished PCB can be fitted into the 

casing of a standard PC gamepad if 

sufficient space is available. In principle 

any gamepad can be used providing 

the direction buttons are ‘active 

Low’ i.e., when you press a button the 

output signal goes from a high to a low. 

The author used a ‘Firestorm Digital 3’ 

while in the Elektor Electronics lab a 

‘MAXFIRE G-08X4’ from Genius happened 

to be available for conversion 

(it must have been used earlier by one 

of our team for some serious research 

work…). 

In addition to the gamepad and finished 

PCB a short length of 6-core cable 

is required and possibly a small 

plastic enclosure for the finished PCB 

if it will not fit in the gamepad case. 

Do not insert the programmed microcontroller 

in its socket yet. The microcontroller 

can be ordered ready-programmed 

from the Elektor Electronics 

website. Alternatively, the hex file 

(object code) can be downloaded from 

the same website at no cost if you prefer 

to program the device yourself. The 

original source files are protected by 

licences and copyrights and are not 

freely available. 

Assembly begins by first dismantling 

the gamepad; undo the screws at the 

K1 

C5 

C1 

IC1 

C3 

C2 

IC2 

Figure 2. The double-sided PCB is supplied with the 

SMD-outline tilt sensor already mounted 

(near the bottom of the board). 

X1 

C4 


list 

Capacitors 

C1,C2,C3 = 100nF 

C4,C5 = 22pF 


IC1 = Atmega8-16PI, programmed, 

Elektor SHOP # 070233-41* 

IC2 = ADXL322 

Miscellaneous 

K1 = 6-way SIL pinheader 


PCB with ADXL322 sensor fitted, Elektor 

SHOP # 070233-91 

* hex code file: free download # 

070233-11.zip from www.elektor.com 

Tilt Gamepad • 58 

43

PROJECTS GAMEPAD CONVERSION 

back of the unit which hold the two 

parts of the shell together. Once inside 

it is necessary to find out which parts 

of the circuit are connected to +5 V 

and which are connected to ground. 

The simplest method is to trace wires 

from the USB connection, pin 1 (usually 

black) is ground and pin 4 (usually 

red) is +5 V. Similarly check out the 

wiring to the gamepad buttons; a close 

inspection reveals that each of the direction 

buttons have two contacts, 

one of which is usually connected to 

ground (as in the Thrustmaster gamepad 

but some use +5 V for this connection) 

the other contact goes to the microcontroller. 

This contact will be used 

later to solder wires to the new PCB 

connector K1 pins 2 to 5. If it is necessary 

to mount the PCB externally in a 

small plastic enclosure (as is the case 

with the “Firestorm Digital 3”), a hole 

will need to be drilled in the rear of 

the gamepad housing (5 mm diameter 


System Init 

Reset Timer 

Start Timer 

Select 

ADC Channel 

Start ADC 

ADC 

Ready 

? 

yes 

Filter and 

Calculate 

Output 

Timer = 10 ms 

? 

yes 

no 

no 

070233 - 12 

Figure 3. The software flow diagram. The sensor 

is sampled every 10 ms. 

Gamepad/sensor board connections 

Wiring between the gamepad and K1 on the sensor PCB using 6-core colour 

coded cable. 

K1 GND Up Left Down Right +5 V 

Cable Black Orange Yellow Green Blue Red 

Gamepad Earth/Ground Up Left Down Right +5 V 

should be sufficient) to run the multicore 

cable through. 

Wiring between the tilt PCB connector 

K1 and gamepad can now begin by soldering 

the wires +5 V, ground and the 

four direction button contacts. 

The pin assignment for K1 is detailed 

on the circuit diagram in Figure 1. 

Pins 1 and 6 carry the power supply 

while 2 to 5 are the digital output 

signals wired to the direction button 

contacts (active low, the idle state is 

high). 

The Table below shows the wiring 

connections in detail and the cable colours. 

Printed arrows on the PCB next 

to connector K1 indicate the direction 

in which the PCB should be moved to 

produce an output at that pin. 

Once the wiring is complete the preprogrammed 

microcontroller can now 

be fitted in its socket on the PCB. The 

PCB can be secured in the gamepad 

using hot glue (roughen the internal 

surface of the housing for good adhesion) 

or if an external enclosure is used 

it can be attached externally to the rear 

of the gamepad again with hot glue. 

Screw the two halves of the gamepad 

together. 

The tilt gamepad is now finished! The 

PC has no way of knowing that the 

gamepad internals have changed so 

it’s not necessary to load any new software 

drivers. Revisit all your favourite 

games but this time experience a 

whole new level of intuitive control. 


(070233-I) 

Tilt Gamepad • 59

Jürgen Wickenhäuser 

Remote measurement and 

control is possible via 

the Internet. 

Unfortunately, 

webservers 

usually sit in 

large, 

humming 

grey 

cabinets. 

That’s not 

the ideal 

solution for 

keeping an 

eye on your 

refrigerator, 

coffee machine 

or central heating 

system. The Elektor 

Electronics Micro 

Webserver provides an 

alternative. 


Web 

�� Micro 

Micro Webserver • 60

server 


The Elektor Electronics Micro Webserver 

is a full-fledged node for Internet 

traffic, despite its quite modest 

dimensions and complexity. It consists 

of a microcontroller board 

with a network interface. 

Thanks to its compact construction 

and the versatility 

of the microcontroller 

board, the Micro Webserver 

is an ideal 

choice for measurement 

and control 

applications. Naturally, 

the fact 

that it can be 

read and operated 

from anywhere 

in the 

world via the 

Internet is a 

major bonus. 

Despite these 

unprecedented 

features, the 

necessary 

hardware is 

actually minimal. 

In principle, 

two ICs are all 

you need for a 

complete webserver. 

To avoid any 

misunderstanding, 

this is not some kind of 

demo or prototype, but a 

fully functional device 

suitable for industrial applications, 

and its potential uses 

extend far beyond what we can 

describe here. 

Basic design 

The underlying technology is rather 

complex. Consequently, in this article 

we must omit a large number of inter- 

esting details that are not essential for 

a ‘simple’ webserver. However, readers 

who want to know all the details 

will find what they’re looking for in the 

accompanying software. The interface 

is without question unusually userfriendly. 

For example, the program 

variables can be used directly in websites. 

It’s hardly possible to make 

things any easier. 

The Micro Webserver is programmed 

using the C language. But don’t let 

yourself be discouraged if you aren’t 

familiar with C, since this project is certainly 

suitable for beginners as well. 

Connection 

control and regulation 

via the Internet 

Internet and Ethernet are closely 

related. Ethernet is a standard that 

defines the connection. The transmission 

speed is normally 10 or 100 

Mbit/s, and it is automatically configured 

when the connection is established. 

We use the 10-Mbit/s variant in 

this project, since it is more than adequate 

for an embedded webserver. 

We assume you already have an Ethernet 

network. The webserver can thus 

be connected directly to a hub or 

switch, so the Internet can be 

accessed via Ethernet. There are also 

agreed conventions regarding how 

Internet communication takes place 

(via Ethernet, for instance). All of this 

is specified in the TCP/IP protocol. 

Here we assume that the network to 

which the Micro Webserver is connected 

can also ‘speak’ this protocol. 

From a technical perspective, there’s 

no reason why the Micro Webserver 

cannot also be directly connected to a 

PC using a crossover cable. However, 

describing this in more detail is 

beyond the scope of this article, since in 

some cases the PC settings must be 

changed for such a connection. 

Hardware 

After all these introductory diversions, 

it’s time to get down to brass tacks. 

The hardware platform is the by now 

well-proven MSC1210 board (originally 

described in the 2003 Summer Circuits 

issue). If you do not already own a 

copy of this outstanding board, you 

can obtain one from Elektor Electronics 

together with the extension 

described here (Figure 1). 

The extension is thus new. In principle, 

it’s simply a ‘custom’ network card for 

the MSC board. This card is built 

around the CS8900A Ethernet driver IC 

(refer to the schematic diagram in Figure 

2). As usual with network cards, 

there are two LEDs (D1 and D2) to indicate 

the status of the network connection. 

D1 flashes for 6 ms each time a 

data packet is received or transmitted, 

or if there is a collision between two 

packets. The second LED indicates 

whether the CS8900A is receiving 

proper link pulses. These pulses are 

used in Ethernet networks to synchronise 

transmitters and receivers, and D2 

will be on if this synchronisation is 

successful. 

The network IC also has a complete 

10Base-T transceiver. 10Base-T is the 

standard for 10-Mbit/s Ethernet over 

twisted-pair cable. The circuit requires 

only a few external components. The 

transformer just ahead of the RJ54 connector 

provides electrical isolation from 

the rest of the world. 

The printed circuit board (Figure 3) 

has a ‘prototyping’ area to provide 

extra space for user applications, in 

addition to the space on the MSC1210 

board. Several spare signal lines are 

available in the leftmost row of the prototyping 

area (see Figure 2). Two extra 

LEDs and a pushbutton switch are 

also placed on the LAN board. The 

Micro Webserver • 61 

13

Figure 1. The MSC1210 board with the network extension: a powerful pair! 

Figure 2. The network card is built around the CS8900 network IC. 

14 

VCC1 

DGND2 

P37 

P36 

INT1 

INT0 

P17 

P16 

P15 

P11 

P10 

DGND3 

VIN_D2 

VCC2 

P00 

P01 

P02 

P03 

P04 

P05 

PO6 

P07 

ALE 

DGND4 

P27 

P26 

P25 

P24 

P23 

P22 

P21 

P20 

VCC3 

DGND5 

K1 

Peripheral Bus 


RD 

WR 

INT1 

INT0 

NET_RES 

P16 

P15 

P11 

P10 

AD0 

AD1 

AD2 

AD3 

AD4 

AD5 

AD6 

AD7 

ADR15 

ADR03 

ADR02 

ADR01 

ADR00 

USER 

LED1 

D4 D3 

1k 

INT1 

INT0 

P16 

P15 

R9 

USER PAD 

P11 

P10 

VCC4 

DGND6 

INT1 

INT0 

PORT1_6 

PORT1_5 

PORT1_1 

PORT1_0 

USER 

LED0 

1k 

R8 

1k 

R10 

LAN RESET 

(ACTIVE HIGH) 

DATA- 

BUS 

AD0 

AD1 

AD2 

AD3 

AD4 

AD5 

AD6 

AD7 

CS 

ADR00 

ADR01 

ADR02 

ADR03 

ADDRESS 

BUS 

S1 

USER 

C10 C4 C5 C6 

100n 100n 100n 100n 

X1 

DVDD1 

61 

IOR 

62 

IOW 

33 

IOCS16 

75 

RESET 

97 

XTAL1 

98 

XTAL1 

20MHz 

65 

SD0 

66 

SD1 

67 

SD2 

68 

SD3 

71 

SD4 

72 

SD5 

73 

SD6 

74 

SD7 

27 

SD8 

26 

SD9 

25 

SD10 

24 

SD11 

21 

SD12 

20 

SD13 

19 

SD14 

18 

SD15 

63 

AEN 

37 

SA0 

38 

SA1 

39 

SA2 

40 

SA3 

41 

SA4 

42 

SA5 

43 

SA6 

44 

SA7 

45 

SA8 

46 

SA9 

47 

SA10 

48 

SA11 

50 

SA12 

51 

SA13 

52 

SA14 

53 

SA15 

54 

SA16 

58 

SA17 

59 

SA18 

60 

SA19 

9 

DVDD2 

DVDD3 

DVDD4 

DVSS1 

DVSS1A 

DVSS2 

DVSS3 

DVSS3A 

U+ 

22 56 69 90 85 95 

IC1 

CS8900 

DVSS4 

AVDD1 

AVDD2 

AVDD3 

DMA 

LED 

AUI 

10BT 

INT 

E2PROM 

C7 C8 C9 

100n 100n 100n 

16 

DMACK0 

14 

DMACK1 

12 

DMACK2 

15 

DMARQ0 

13 

DMARQ1 

11 

DMARQ2 

TEST 

36 

SBHE 

64 

IOCHRDY 

49 

REFRESH 

29 

MEMR 

28 

MEMW 

34 

MEMCS16 

77 

SLEEP 

CHIPSEL 

AVSS0 

AVSS1 

AVSS2 

AVSS3 

AVSS4 

LAN 

LINK 

100 

99 

HC1 

78 

IRQ0 

32 

IRQ1 

31 

IRQ2 

30 

IRQ3 

35 

RES 

EESK 

4 

EECS 

3 

EEDI 

6 

EEDO 

5 

ELCS 

2 

CSOUT 

17 

8 10 23 55 57 70 1 89 86 94 96 

76 

DO+ 

83 

DO– 

84 

DI+ 

79 

D1– 

80 

CI+ 

81 

CI– 

82 

TXD+ 

87 

TXD- 

88 

RXD+ 

91 

RXD+ 

92 

93 

7 

R2 

24�9 

R3 

24�9 

4k99 

LAN 

ACTIVITY 

R5 R6 

4k7 

C3 

68p 

100� 

R1 

C1 

100n 

R4 

D1 

1k 

8 

7 

6 

3 

2 

1 

R7 

1k D2 

T1 

HALO 

TG43-1406N 

C2 

100n 

LINK 

PULSES 

9 

10 

11 

14 

15 

16 

044026 - 11 

1 

2 

3 

4 

5 

6 

7 

8 

K3 

TX+ 

TX- 

RX+ 

RX- 

RJ45 LAN 


elektor electronics - 7-8/2004


Applications The Micro Webserver is ideal for the following applications: 

Automatic online weather station: 

Access control and registration in combination 

– temperature 

with: 

– precipitation 

– badge readers 

– lightning detection 

– light barriers 

– wind strength and direction 

– door openers 

– relative humidity 

– RFID tags 

– rain barrel level 

– light intensity 

Web interface for home appliances and fixtures: 

– refrigerator or freezer temperature monitoring 

– remote control for coffee machine, central heating or lighting 

– controlling sun awnings or roller shutters 

– outside lighting 

– intruder detection 

– greenhouse climate control 

placement of the connector for the link 

to the ‘motherboard’ allows the extension 

card to be located next to the 

motherboard or underneath it. In the 

latter case, the two boards can sandwiched 

together using standoff 

bushes. 

Although the design of this project is 

especially simple, there is one thing 

that must be mentioned. The current 

consumption of the LAN IC is 

100–120 mA, which is relatively high 

compared with the current drawn by 

the microcontroller. The 5-V supply 

voltage is taken from the MSC1210 

board. To prevent the voltage regulator 

on that board from becoming overheated, 

we strongly recommend that 

the entire circuit be powered from a 

voltage of 7.5 to 9 V, but definitely no 

higher than this. 

Online 

There’s actually not much more to say 

about the hardware. Configuring the 

board is fully described in the text box. 

Once you’ve gotten the server ‘up’, you 

can start testing. 

This is where things start to get interesting. 

To start off, simply connect the 

board to the network. LED D2 will be 

continuously on if an Ethernet signal is 

detected. This is a promising start, but 

the real test comes next. It consists of 

trying to ‘ping’ the server using the 

Windows Command Prompt window 

(DOS command window). On a PC connected 

to the network, type the following 

command in the command line: 

ping 192.168.1.156 

(of course, the IP address here must be 

the address previously assigned to the 

Webserver). LED D1 should start blinking 

as an indication that data is being 

transferred via the Ethernet, and a 

reply from the server should appear in 

the command win- 

dow. 

Ping is a simple protocol 

that allows a 

few bytes to be 

transmitted and 

waits for an ‘echo’. 

It’s a really handy 

way to quickly check 

a network connection. 

If the ping test is OK, 

you can then access 

the webserver using 

a web browser. In the 

browser window, 

enter the following 

address: 

http://192.168.1.156 

(use the address that has previously 

been assigned to the webserver). And 

that’s it: what you see next comes from 

that little board (see Figure 5). 

In the terminal download window, you 

can also see which page was 

requested. 

How it works 

What actually happened when you 

requested the web page? First, you 

made a connection to an IP address. 

Actually, it’s a bit more complicated 

than that: you made a connection to a 

‘socket’ at a particular address. A 

socket is a sort of ‘connector’, in this 

case one that only fits web links. Each 

socket is also assigned a specific port 

Monitoring and controlling machinery: 

– rpm 

– voltage and current 

– temperature 

– liquid level 

– flow rate / discharge rate 

– pressure 

– valve control 

– relay control or PWM (servo) control 

Terminal for a central database (in combination 

with an LC display and barcode reader) 

Internet references 

[1] www.wickenhaeuser.com 

μC/51 compiler with source code 

[2] www.mikrocontroller.info/kabelsalat/ 

Wiring diagram for a null-modem cable 

[3] www.ti.com/msc MSC121x home page 

[4] groups.yahoo.com/group/TI-MSC 

MSC121x users group. Definitely worth the effort. 

Free, but registration is required. 

[5] groups.yahoo.com/group/TI-MSC/files 

You can find tools for the MSC121x here, such as 

the original TI downloader. 

[6] www.cirrus.com/en/pubs/proDatasheet/ 

cs8900a-4.pdf 

Data sheet for the CS8900A network driver 

number. Port 80 is frequently used for 

webservers. You can see this in the 

program line SOCKET_SETUP(i, 

SOCKET_TCP,80,FLAG_PASSIVE_OPEN). 

The final parameter 

here indicates that 

the socket is passive, 

which means it waits 

for requests from 

clients. The sockets 

are created in a FOR 

loop. The number of 

sockets created 

determines how 

many clients can be 

connected to the 

server at the same 

time. As each socket 

costs memory, the 

total number is limited. 

The CS8900A IC 

used here also has a 

buffer (approximately 

4 kB) for incoming Ethernet packets. 

That’s not especially large if several 

users want to connect to the server at 

the same time, or if large items such as 

images are requested. Actually, this 

doesn’t matter all that much, since 

TCP allows the occasional packet to 

remain unanswered. If necessary, the 

client resends unanswered packets on 

its own initiative. 

After the sockets have been created, 

ELM_FLEX.C initiates the A/D converter 

of the microcontroller a few lines 

later in the code. For more information 

about the A/D converter, see the companion 

Micro Webserver article ‘Measurement 

and Control via the Internet’ 

in this issue. 

After this, the program enters a endless 

FOR loop. In this loop, poll_web- 


7-8/2004 - elektor electronics 15


16 

COMPONENTS 

LIST 

Resistors (SMD): 

R1 = 100�, shape 0603 

R2,R3 = 24�9, shape 0603 

R4 = 4k�99, 1 %, shape 0603 

R5,R10 = 4k�7, shape 0603 

R6-R9 = 1k�, shape 0603 

Capacitors (SMD): 

C1,C2,C4-C10 = 100nF, shape 0603, 

ceramic 

C3 = 68pF, shape 0603, ceramic, NP0 

Semiconductors (SMD): 

IC1 = CS8900A-CQ (5 V), shape 

TQFP100 

D1-D4 = chip-LED, shape 0805 

Recommended colours: D1 green; D2 

yellow; D3,D4 red 

Miscellaneous: 

T1 = Ethernet transformer type TG43 

(Halo) or ST7010T (Valor), see also ref. 

[6] 

X1 = 20MHz quartz crystal, HC49_SMD 

case 

K1 = 34-way DIL pinheader 

K2 = 8-way pinheader 

K3 = RJ45 connector (screened) 

S1 = mini pushbutton 

For software, bare PCBs and fully 

assembled boards, 

see the ‘What you need’ box. 

Figure 3. The network card for the 

MSC1210 board. 

Figure 4. The web page sent by 

the micro webserver. 




Configuring the board 

The Micro Webserver only works in a TCP/IP network. Just like all other computers in a 

TCP/IP network, the microcontroller is assigned a unique address, which is its IP 

address. Before you start programming the microcontroller, you must manually specify 

this address, since the Micro Webserver does not work with automatic address assignment. 

The default IP address is set to 192.168.1.156. It belongs to a range of addresses 

that are specifically reserved for networks that are not directly connected to the 

Internet. Subscribers to ADSL or cable Internet use addresses in this range for their local 

networks. Addresses having the form 10.0.0.x also belong to this category. It may also 

be possible to request a ‘real’ Internet address for your Micro Webserver, but that 

depends on your provider. In any case, you must personally check which address range 

is used in your network and which addresses are available to be assigned to the server. 

After choosing an address, you can turn your attention to the necessary programming 

software and C files. Part of the required source code (the part that implements the 

actual webserver) is included with the uC/51-compiler (from version 1.20 onwards). A 

fully functional demo version of this compiler can be downloaded free of charge from 

the author’s home page (see reference [1]). The only difference between the demo version 

and the registered version is that code size for the Micro Webserver is limited to 

16 kB, but that’s more than enough for this application. Sample source code for initialising 

the webserver and implementing web pages (including several sample pages) is 

included in the package. 

After installing the uC compiler, you must first use MakeWiz to create a workspace. In 

MakeWiz, open the file ...\SRC\MSC1210\ELM_FLEX\ELM_FLEX.MAK. Then change 

something in the text (for example, add your own version number) so that the Save 

button will be enabled. Tick the ‘Write JFE-Workspace File’ check box and save the file 

(Figure 5). 

Now you can start the JFE editor (with thanks to Jens Altmann). In JFE, use ‘Open 

Workspace’ to open the file ...\SRC\MSC1210\ELM_FLEX\ELM_FLEX.WSP. All of the files 

belonging to the project will appear in the editor window. Now you have to specify the 

previously determined IP address in the ELM_FLEX.C file. You can do so in the line 

COMPOSE_IP(my_ip, 192.168.1.156). 

A workspace that has been created using MakeWiz causes three special buttons to 

appear in JFE: ‘MAKE’, ‘RE-MAKE’ and ‘DL.BAT’. The MAKE button causes the project to 

be compiled, but it limits processing to the files that have actually been modified. The is 

the usual (and fastest) way to generate 

the hex file you need for programming 

the microcontroller. RE-MAKE must be 

used if something not present in the 

workspace has been modified, such as a 

header file (.h). This command causes 

everything to be recompiled. Finally, 

DL.BAT causes the result to be sent to the 

MSC board. This actually amounts to simply 

executing a batch file, to which JFE 

passes a parameter. This parameter is 

always the name of the target file, which 

in this case is ELM_FLEX (with no extension). 

The specific command line that initiates 

downloading to the MSC board is stated 

in the batch file (which is also located in 

the project folder). In this case, the command 

line is download 

/F%1.hex /X11 /P1 /T 

/B9600. 

Parameter P1 indicates that COM1 of the 

PC must be used for programming. This 

can be changed if necessary. 

So far, so good: you’ve modified the IP 

address in ELM_FLEX.C, you’ve compiled 

the project, and your finger is just itching 

to press DL.BAT — but hang on a 

Figure 6. With JFE, all files are easily accessible. 

moment! Before you can download anything to the board, you have to acquire a copy 

of the original Texas Instruments downloader (Downloader.exe). You can obtain this 

from the MSC group site at Yahoo (reference [4]), among other places, and it can be 

placed in the project folder. If you wish, you can also place it in a more general location, 

but in that case you naturally have to specify its new location in DL.BAT. 

Be sure to fit jumpers J1 and J2 on the MSC1210 board (J3 must remain open). If J1 

and J2 are not fitted, the board is protected against resetting and modifying the 

firmware via the PC. Finally, you need a null modem cable to connect the board to the 

PC, but that should be obvious. After you’ve found a place for the downloader, modified 

DL.BAT if necessary (to specify a different COM port or change the path to the 

downloader), connected the board to the proper PC port, and powered up the board, 

you’re finally ready to click on DL.BAT in JFE. 

If everything goes as it should, the MSC1210 board will return a short greeting message, 

and if ‘’ is not included in this message, the Ethernet board has 

been successfully recognised. In addition, one of the red LEDs on the MSC board should 

blink slowly. 

After downloading the code, don’t forget to remove jumpers J1 and J2. 

Figure 5. Use MakeWiz to store the project. 


7-8/2004 - elektor electronics 17

What you need 

The Micro Webserver consists of: 

– the MSC1210 ‘Precision Measurement Central’ board (see 

the July/August 2003 issue of Elektor Electronics) 

– the 10-Mbit Ethernet network card (RJ45, twisted pair) 

– the μC compiler with the necessary software (Readers 

Services order code 044026-11) 

– the TI downloader program (Downloader.exe) 

The MSC1210 microcontroller card and the associated network 

extension are available from Elektor Electronics. 

The μC compiler, including all the necessary source code, 

can be downloaded at no charge from www.wickenhaeuser.com 

or from the Elektor Electronics website. 

The programming software for the MSC board 

(Downloader.exe) is available from reference [4]. Updates 

will be available from the author’s website. 

server() is called periodically. As long 

as the result returned by this call is ‘0’, 

other (user-written) routines can also 

be executed in this loop. However, it’s 

important to ensure that user-written 

extensions do not take up too much 

processor time, since the webserver 

will otherwise become inaccessible. 

The FlexGate TCP/IP stack works with 

events. The Micro Webserver only 

responds to EVENT_HTTP_REQUEST 

(page request) and 

EVENT_SOCKET_IDLETIMER (which 

has a period of approximately 0.5 s). If 

a client wants to access a page, the 

name is first requested using webpage_name(). 

Following this, webpage_bind() 

is used to prepare the cor- 


18 

Prices: 

responding page for the reply. Pages 

that are to be externally available must 

be declared in advance as array extern 

code uchar (see ELM_FLEX.C). 

This completes the process if the 

requested page does not contain any 

dynamic data. However, dynamic data 

is exactly where the power of this 

handy little device lies. An example of 

dynamic data is measurement data 

coming from the microcontroller board. 

Such data can easily be incorporated 

into a web page. And in the other 

direction, you can remotely control the 

microcontroller outputs via a web page. 

To find out more about this, see the 

companion article ‘Measurement and 

Control via the Internet’ in this issue. 

– ready-made MSC1210 board: £69.00 (US$112.50) 

(assembled and tested; Readers Services order code 

030060-91) 

– ready-made network extension for the MSC1210 board: 

£41.95 (US$73.95) (assembled and tested; Readers 

Services order code 044026-91) 

– combined package: assembled MSC1210 board, network 

extension and all related Elektor Electronics articles on 

diskette: only £103.50 (US$184.95) (Readers Services 

order code 044026-92) 

For die-hard DIYers, bare PCBs are also available for the 

MSC1210 board (Readers Services order code. 030060- 

11) and the network extension (Readers Services order code 

044026-11). Note that most of the components are SMD 

types, and some of them are very difficult to obtain as oneoffs 

as well as solder by hand. 

Naturally, there’s a lot more we could 

say about the Internet portion of the 

software (the TCI/IP stack), but that 

goes beyond what we had in mind for 

this article. If you want explore this 

question in more detail, have a look at 

the manual for the stack. You’ll find it 

in the folder ...SRC\FLEXGATE\ that 

comes with the microcontroller compiler. 

In addition, Texas Instruments is 

presently preparing an application 

note for this project. The details will 

appear in due time on the TI website. 

(044026-1) 

About the introductory illustration: 

This jumble of lines may appear chaotic, but it 

actually represents a reasonably well-organised 

entity: the Internet. This ‘map’ was automatically 

generated by a program that literally combs the 

Internet. In its travels, the program also came close 

to the server where the Elektor Electronics website is 

hosted. See www.opte.org. 

Indicators: 

Cyan: Asia Pacific 

Pink: Europe, Middle East, Central Asia, Afrika 

Yellow: North America 

Blue: Latin America and Caribbean 

Red: RFC1918 IP Addresses 

Black: Unknown 




non reflected 

reflected 

Micro Webserver • 67

PROJECTS SDR 

Software Defi ned 

With USB Interface 

Burkhard Kainka 

SD (software-defi ned) radio receivers use a bare minimum of hardware, relying instead on their 

software capabilities. This SDR project demonstrates what’s achievable, in this case a multi-purpose 

receiver covering all bands from 150 kHz to 30 MHz. It’s been optimised for receiving DRM and AM 

broadcasts but is also suitable for listening in to the world of amateur transmissions. 

The designer’s aim for this project was 

to create a receiver displaying high linearity 

and phase accuracy. Development 

was focussed on the characteristics 

that were most important for a 

top-notch DRM receiver and the end 

result is a receiver with remarkable interference 

rejection characteristics. Reception 

of DRM stations using DREAM 

software produced signal-to-noise ratio 

(S/NR) values of well over 30 dB. The 

design principle of the receiver guarantees 

an extremely fl at fi lter-curve 

response. This works out rather well 

not only for DRM but also for the audio 

quality of AM broadcasts, which sound 

almost as good as VHF FM. It’s worth 

noting too that some transmitters that 

do not conform to the normal bandwidths 

laid down for medium wave (9 

kHz) and shortwave (10 kHz) as rigidly 

as perhaps they should. Whilst these 

stations produce no observable sound 

improvement for listeners using normal 

receivers (since their IF fi lters limit the 

bandwidth and in the process the frequency 

response too), this is not the 

case with SDR, where it’s no problem 

to select a wider bandwidth at will. It 

gets even better: in software receivers 

the fi ne-tuning capabilities of PC decoder 

programs give you the capability 

of determining the desired bandwidth 


with notch fi lters to the automatic level 

control (ALC) settings along with 

selecting all the usual receive modes 

from AM by way of DRM and SSB to 

CW. 

Further refi nements can be added for 

SWL (shortwave listening) applications. 

If for instance you wish to increase 

the sensitivity on the upper amateur 

bands this is easily arranged by 

using two switchable antenna inputs 

and providing an optimised preselector 

circuit or preamplifi er in one of them. 

The receiver’s printed circuit board 

itself provides a pretty basic RF front 

end, which is nevertheless perfectly 

adequate for broadcast reception. A 

long wire antenna of adequate length 

will lift the strength of signals above 

atmospheric noise level to ensure you 

miss virtually nothing. 

Hardware requirements 

Most SDR programs [1] require the Windows 

XP platform to operate satisfactorily. 

The most important hardware necessary 

then is an SDR-capable sound 

card. We have developed a small circuit 

for testing sound cards, described 

elsewhere in this issue under the heading 

‘Developer Tips’. Without performing 

this test fi rst it’s utterly pointless 

starting to make the SDR receiver! 

All about USB 

The receiver is controlled over a USB 

connection and powered with +5 V 

in the same way (no additional mains 

power supply needed). For the USB interface 

in the receiver circuit (Figure 1) 

we selected the FT232R from our Scottish 

friends at FTDI. This modern USBto-serial 

converter works without the 

need for a quartz crystal, as it is equipped 

with an internal R-C oscillator of 

adequate stability. The module (IC4) is 

used here in its ‘bit-bang’ mode along 

the lines of a fast parallel port. Eight 

data lines are available for use and these 

can be driven in whichever way we 

choose. Two of the lines are used as an 

I²C Bus and control the frequency of the 

receiver. Three wires connect the input 

multiplexer to one of up to eight antenna 

inputs with and without fi ltering. 

Two additional inputs serve to control 

the IF amplifi cation of the receiver. In 

this way the receiver operates entirely 

under remote control. Kiss good-bye to 

all those knobs and controls of bygone 

radio days… 

Please pay particular attention to decoupling 

the power supply. One reason 

for this is because the USB chip 

�� 


Radio 

K1 

5 

6 

3V3 

K3 

C8 

10n 

GND 

GND 

1 

2 

3 

4 

L2 

USB-B connector 

ANT 

GND 

L6 

L3 

R12 

470 

2200uH 

GND 

TEST_CLK 

GND 

PC1 

C16 

100n 

L5 

R24 

1k 

47uH 

C29 

220p 

C31 

100p 


VCC 20 

CBUS0 23 

RI 6 

CBUS1 22 

RTS 3 

TXD 1 

DTR 2 

CTS 11 

4 

VCCIO 

RXD 

17 

3V3OUT 

5 

DSR 9 

DCD 10 

CBUS2 13 

CBUS3 14 

CBUS4 12 

R2 

330 

16 

R3 

USBDM 

330 

15 

USBDP 

19 

RESET 

27 IC4 

OSCI 

28 

OSCO 

FT232R 

26 

TEST 

C23 

100n 

C25 

100n 

GND 

C32 

470 

100n 

R15 

GND 

C36 

4.7 

GND 

100n 

C38 

100n 

R25 

AGND 

GND 

GND 

GND 

25 

7 

18 

21 

GND 

L1 

VCC 

10uH 

C1 C4 

C2 C3 

100n 4u7 

16V 

100n 4u7 

16V 

GND GND GND GND 

13 

A0 

14 

A1 

15 

A2 

12 

A3 

1 

A4 

5 

A5 

2 

A6 

4 

A7 

INH 6 

COM 3 

A 11 

B 10 

C 9 

IC6 

74HC4051 

7 

VCC_HF 

VEE 

VCC 16 

8 

GND 

GND 

C21 

100n 

GND 

C30 

100n 

R16 

1M 

C7 

100n 

GND 

C26 

100n 

GND 

100k 

R17 

VCC_HF 

GND 

470 

VDD 2 

VDD 19 

VDD 14 

VSSL 7 

VDDL 

10 

PDM/OE 

6 

AVSS 

11 

4 

AVDD 

13 

SCL 

17 

VCXO/WP 

SDA 

CLOCK1 

5 

8 

CLOCK2 9 

CLOCK3 12 

CLOCK4 15 

CLOCK5 18 

CLOCK6 3 

IC3 

CY27EE16ZE 

TEST_CLK 

L4 

10uH 

T1 

BF245 

R18 

1 

XIN 

X1 

3V3 

20 

XOUT 

16 

C12 

C13 

10p 10MHz 10p 

Figure 1. Diagram of the receiver circuit, which in fact comprises just a tuning oscillator and a mixer. 


R7 

100 

R19 

100 

VSS 

GND 

I2 

5 

13 

Q2 

6 

12 

4 

3 

1 2 

8 

9 

11 10 

Q3 

IC2C 

IC2D 

Q1 

VCC 

S 10 

12 

D 

9 

11 C 8 

13 

R 

VCC 

VCC 

2 

D 

IC1A 

5 

3 C 6 

C33 

100n 

C35 

2n2 

GND 

S 4 

1 

R 

VCC 

IC1B 

Q_SW_N 

Q_SW 

I_SW_N 

I_SW 

I3 

C19 R6 

10k 

R5 

100k 

IC2B 100n 

C22 

2n2 

6 

GND 

5 

IC5B 

IC2A 

I1 

C24 

2n2 

100n 

R9 

R14 

10k 

4k7 C27 

4k7 C39 

13 

R13 

100k 

IC5D 

12 

100n 

C37 R21 

2n2 

070039 - 11 

7 

14 

IC1 = 74AC74 

IC2 = 74HC4066 

IC5 = TL084CN 

IC7 = 74HC4066 

10k 

R10 

10 

9 

C28 

100n 

GND 

10k 

R22 

3 

2 

C40 

100n 

GND 

VCC 

IC5C 

R11 

27k 

IC5A 

R23 

27k 

C5 

100n 

R8 

100k 

R20 

100k 

C9 

4u7 

16V 

R4 

100 

8 

1 

VCC 

14 

IC1C 

7 

GND 

1 2 

IC7B 

4 3 

5 

13 IC7A 

8 9 

6 IC7C 

IC7D 

11 10 

12 

C6 

100n 

VCC 

100 

GND 

C20 

100n 

C34 

100n 

R1 

14 

IC2E 

7 

C10 

4 

IC5E 

14 

IC7E 

C11 

100n 

11 7 

100n 

C14 

470u 

16V 

C15 

4u7 

16V 

C17 

100n 

VCC_HF 

GND 

C18 

10n 

GND 

�� 

K2 

15

PROJECTS SDR 

FT232R operates internally at the same 

frequency range that we are receiving 

through the antenna downlead and we 

don’t want any of this RF to leak across 

from one stage to another. That said, 

the decoupling within the chip itself 

is remarkably good and the residual 

RF on the control port lines is barely 

detectible. Consequently we can control 

the HC4051 RF input multiplexer direct 

from the control port lines, without 

traces of the processor clock appearing 

in the wanted signal region. 

Using its built-in 3.3 V voltage regulator, 

the FT232R provides the operating 

supply for the programmable clock 

generator CY27EE16ZE, avoiding 

the need for an additional voltage regulator. 

The rest of the circuit (Figure 

1) operates exclusively on 5 V. Several 

different smoothed and fi ltered voltages 

are produced, to guarantee good 

RF decoupling on one hand and to ensure 

suppression of audio frequency 

interference on the other. This is particularly 

crucial at the RF input stage of 

the receiver, from which the signal is 

fed via the mixer to the IF circuitry. For 

this reason a large electrolytic is provided 

at this point (VCC_HF) to ensure 

proper ‘peace and quiet’. 

Programmable VFO 

The SDR calls for an oscillator frequency 

running four times higher than that 

of the signal received, in order that the 

necessary phase fi ltering can be divided 

by four. If we are aiming to receive 

signals up to 30 MHz, then the oscillator 

needs to run right up to 120 MHz. 

DDS oscillators are very popular in HF 

projects today but at 120 MHz a DDS 

is dearer, more power-thirsty and far 

less controllable. Accordingly we shall 

look away from DDS oscillators and use 

a programmable clock oscillator with 

internal PLL here. Many Elektor Electronics 

readers will remember the CY- 

27EE16ZE back from the February 2005 

issue. This clock oscillator, developed 

specially for digital applications, performs 

equally well in RF circuitry. The 

frequency resolution does not quite 

match that of a DDS but the phase accuracy 

of the output signal achieves 

comparable results. Restricting power 

consumption to a relatively modest 

amount is important with this project, 

since we must not draw too much current 

from the USB port. 

The chip is programmed over the I²C- 

Bus using lines SCL and SDA. The internal 

VCO operates in the frequency 

range 100 to 400 MHz, stabilised by 


means of the 10-MHz crystal and a 

PLL. Its output signal then goes via counters 

to the desired outputs. Here we 

select the clock output Clock5, where 

a VFO signal between 600 kHz and 120 

MHz is available for further processing 

in the 74AC74 counter. 

The principle of the I-Q mixer has been 

described already in Elektor Electronics 

12/2006. A two-stage mixer is created 

here from a total of four analogue 

switches inside an HC4066 IC. This is 

controlled by two phase-shifted oscillator 

signals, which themselves are 

produced with a 74HC74 counter. Supposing 

the programmable clock oscillator 

produces 24 MHz, then the mixer 

would need a drive of 6 MHz. The receiver 

would in this case operate in a 

region of around ±24 kHz either side of 

the centre frequency of 6 MHz. 

Vital here is a phase shift of exactly 90 

degrees between the two oscillator signals. 

Any deviation will lead to reduced 

suppression of the image frequencies. 

A 74HC4053 or 74HC4052 integrated 

changeover switch device would 

not make a good choice for the analogue 

switch because the signal transit 

delays in the internal decoders would 

then cause different phase errors to 

appear in every frequency range. Our 

chosen solution using the rather more 

basic switches of an HC4066 retains all 

four phases in sync. Since the 74AC74 

counter is confi gured as a synchronous 

counter we would not expect to fi nd 

any phase errors here either. In fact the 

receiver displays image frequency suppression 

of around 40 dB up to 15 MHz 

or so, although this value decreases 

beyond about 20 MHz (which we can 

tolerate given that these frequencies 

are not so heavily occupied). 

Signal processing 

The receiver is provided with several 

inputs, selected by the 74HC4051 input 

multiplexer (IC6). The antenna input 

ANT is fed by way of fi lters to the fi rst 

three inputs. The fi rst switch setting 

(wideband) uses only one input choke 

(L6), which shunts any audio frequency 

signals at the input to ground. In the 

second position (Medium Wave) there 

is a low-pass fi lter with a boundary frequency 

of 1.6 MHz, using resistor R12 

to attenuate excessive resonance. This 

fi lter suppresses interference to medium 

wave reception from overtone mixing 

with stations in the short wave 

range. The third position makes use of 

a simple R-C high-pass fi lter to attenuate 

strong medium wave signals. 

An additional input (PC1) can be selected 

if you wish to connect external 

tuned input circuits or preamplifi ers. Finally 

three more inputs are provided for 

future developments. The input fi lters 

on the printed circuit board are good 

to be getting on with and are certainly 

adequate for most applications. You 

can of course introduce steep low-pass 

fi ltering ahead of the fi lters provided 

if you want to be certain of blocking 

out overtone mixing in every possible 

situation. Or you might choose to fi t 

different resonant circuits, selected by 

input switching. 

The particular input that is active at 

any given time is connected to the 

common output COM (pin 3). Coupling 

capacitors are provided either side of 

the switch. A bias voltage of about 2.5 

V is provided to the switch from the 

source connection of the BF245 via a 

1-megohm resistor. This eliminates any 

distortion from large input signals that 

might arise when signals are limited 

by the protection diodes on the analogue 

inputs to the ICs. 

Input A7 delivers a calibration signal 

from Output 3 (Test-Clk) of the programmable 

crystal oscillator. The oscillator 

produces a square-wave signal of 

3.3 V peak-to-peak at 5 MHz. A signal 

voltage of around 5 mV at 5 MHz is 

produced at the voltage divider, corresponding 

to a signal strength of S9 + 

40 dB. This enables the fi eld strength 

meter created in software to be calibrated 

without any further expenditure. 

JFET BF245 on the output of the input 

multiplexer serves as an impedance 

transformer. This provides a relatively 

high impedance termination of 100 

k� for the RF signal, enabling for instance 

a high-Q resonant circuit to be 

connected even to input In2. At the 

low-impedance output of the source 

follower we arrange to have a voltage 

of circa 2.5 V, fed via the mixer and the 

following op-amp all the way to the 

output. It is important that no audio 

frequency signal remnants appear at 

the source connection and for this reason 

the ‘critical’ supply Vcc_HF is also 

fi ltered very thoroughly. The FET itself 

provides additional decoupling of the 

supply voltage, but we don’t want any 

signal escaping from the Gate either 

that might fall in the IF region below 

24 kHz. This is why an RF choke is connected 

directly to the antenna input, 

to shunt for instance any 50 Hz mains 

hum signal. 

Leading off from the Source connection 

are two 100-� resistors that go to the 

two mixers for the I and the Q signals. 

�� 


COMPONENTS 

LIST 

Resistors 

R1,R7,R19 = 100� 

R2,R3 = 330� 

R4 = 100� 

R5,R8,R13,R17,R20 = 100k� 

R6,R10,R14,R22 = 10k� 

R9,R21 = 4k�7 

R11,R23 = 27k� 

R12,R15,R18 = 470� 

R16 = 1M� 

R24 = 1k� 

R25 = 4�7 

Capacitors 

C1,C2,C5,C6,C7,C10,C11,C16,C17,C1 

9,C20,C21,C25-C28,C30,C32,C33,C 

34,C36,C38,C39,C40 = 100nF 

C3,C4,C9,C15 = 4μF7 16V radial 

C8,C18 = 10nF 

C12,C13 = 10pF 

C14 = 470μF 16V radial 

C22,C24,C35,C37 = 2nF2 

C29 = 220pF 

C31 = 100pF 


IC1 = 74AC74 

IC2,IC7 = 74HC4066 

IC3 = CY27EE16 (Cypress) 

IC4 = FT232R (FTDI) 

IC5 = TL084CN with socket (see text) 

IC6 = 74HC4051 

T1 = BF245 

Inductors 

L1-L4 = 10μH 

L5 = 47μH 

L6 = 2.2mH 

Miscellaneous 

K1 = USB-B socket, PCB mount 

K2 = stereo jack socket, 3.5mm, PCB 

mount 

K3 = 2-way PCB terminal block, lead 

pitch 5mm 

PC1 = solder pin 


Ready-populated and tested PCB, order 

code 070039-91 

Project software, free download 

070039-11 

Supplementary document, free download 

PCB, bare, ref. 070039-1 from www.thepcbshop.com 

They improve the symmetry of the mixers, 

the ‘on’ resistances of which let 

through a certain amount of leakage. 

The mixers themselves are HC4066 

analogue switch ICs ganged as changeover 

switches. The voltage of these 

too is set around 2.5 V, allowing them 

to be controlled without overdriving up 

to about 5 V peak-to-peak. 

The IF amplifi er consists of two exact- 



Figure 2. The SDR receiver board. 

�� 

17

PROJECTS SDR 

ly equal branches that together produce 

an attenuation of up to 40 dB at 

all times. When you are using 5 V supplies, 

the gain bandwidth (GBW) of 

the selected op-amp is important, in 

order to achieve tenfold amplifi cation 

without phase errors for signals around 

20 kHz. In the author’s test samples a 

TL084 turned out to be adequate. If 

you provide a socket for IC5 you will 

be able to try other, faster op-amps. 

The input stage works as a differential 

amplifi er. In dimensioning the resistors 

what we are looking for is not the best 

common-mode suppression but rather 

an input resistance that’s as equal as 

possible across the inverting and noninverting 

inputs. Tests show that good 

phase accuracy (and consequently high 

image-frequency suppression) depend 

on equal impedance existing on all four 

phases of the mixer. The input impedance 

amounts to around 5 k� at all 

of the inputs. Note the load resistance 

of 4.7 k� on the non-inverting input as 

opposed to 10 k� on the inverting one. 

This is correct, since signal transit on 

this input gets dispersed in exact antiphase 

by inverse feedback, halving 

the input resistance to 5 k�. In this 

way both inputs offer the same input 

resistance as close as matters. 

The 2.2 nF capacitors together with 

the mixer’s internal resistance and the 


Figure 3. This lab sample board is not quite equivalent to the production version supplied through the Elektor SHOP. 

100 � series resistors form simple lowpass 

fi lters with a limiting frequency of 

over 100 kHz, so as to keep remnants of 

RF well away from the audio frequency 

stages. The limiting frequency lies 

far above the transfer frequency range, 

meaning that capacitor tolerances do 

not produce any discernible phase errors. 

You can use even ceramic disc capacitors 

here. Tolerances between 10 

and 20 % are not a problem with any of 

the capacitors in the signal path acting 

as high-pass elements with a limiting 

frequency of around 300 Hz. 

The fi nal stage has a tenfold gain (20 

dB), which can, however, be reduced to 

unity gain by the analogue switches. 

A total of three attenuation steps are 

provided: 0 dB, –10 dB und –20 dB. 

To avoid it being driven too hard, the 

gain can be reduced in software. As 

the receiver’s input displays high resistance 

to being overdriven the attenuator 

is placed in the fi nal stage, so 

as to prevent overdriving of the output. 

This corresponds to gain control in an 

IF amplifi er. 

Construction 

The printed circuit board shown in Figure 

2 uses standard wire-ended components 

as far as possible, with the exception 

of the LSI (large scale integra- 

tion) chips FT232RL and CY27EE16, 

which unfortunately are available only 

in SSOP case format with a pin spacing 

of 0.65 mm. Figure 3 shows the 

laboratory prototype PCB with components 

fi tted. 

The best way to begin is by soldering 

the two surface-mount device (SMD) 

chips in place. It pays to start fi rst at 

the four corners, before soldering all 

the other pins generously. Superfl uous 

solder can be removed with desolder 

braid, followed by thorough checking 

with a magnifying glass to avoid unwelcome 

surprises later on. 

The components with wire leads will 

present no diffi culty. The circuit does 

not call for any special RF components 

or test points. Capacitors C12 and 

C13 should not be fi tted initially. The 

CY27EE16 has presettable internal 

capacitors that should enable you to 

achieve a frequency of exactly 10 MHz 

without diffi culty. C12 and C13 will be 

needed only if the crystal used requires 

greater loading capacity. 

Once all construction is complete you 

need to make a quick round-up with a 

multimeter checking for any short circuits 

around the USB connections, as 

you certainly don’t want to damage 

the PC. 

�� 


Hook-up and alignment 

Before connecting the receiver to the 

computer’s USB port for the fi rst time 

you will need to install the driver software 

for the FT232R. You can fi nd this 

on the manufacturer’s website (www. 

ftdichip.com/FTDrivers.htm) or alternatively 

in the software download for 

this article. Installation using CDM_ 

Setup.exe automatically removes any 

traces of older FTDI drivers on your 

computer. After this has been done 

Windows will fi nd the correct driver 

automatically as soon as the receiver 

is connected. The same process provides 

the PC automatically with an 

additional virtual COM-port interface. 

For this you do not even need to know 

which COM number is allocated to the 

device, as it effectively sets up its own 

direct connection to the FT232R. FT- 

D2XX.dll controls the eight data lines 

of the chip as for a parallel port, eliminating 

at the same time all timing 

problems. To save time the multiple level 

changes involved in controlling the 

I2C bus are loaded conveniently into 

a buffer and then fed out to the data 

lines in short order. The program ElektorSDR.exe 

enables you to control all 

functions of the receiver (Figure 4) and 

can be found in the download archive 

as an executable fi le together with the 

Delphi source code. Also available for 

download is a supplementary .pdf document 

that describes initialisation 

and commissioning. 

Decoder software 

Nearly all signifi cant characteristics of 

the receiver are determined by settings 

in the decoder software on your PC. As 

the survey in [1] indicates, there are a 

number of different programs to choose 

from. You could perform your fi rst test 

with SDRadio [2] for example. After this 

you will discover additional possibilities 

in DREAM [3] or G8JCFSDR [4]. 

Whichever program you select, it’s vital 

to set up the sound card correctly (this 

is described in the supplementary document). 

Information on the programs 

can be found on the relevant Web 

pages and in the Elektor Electronics 

articles listed below. Further advice 

may be found on the author’s homepage 

(www.b-kainka.de) and will appear 

also in due course on the project 

page at www.elektor-electronics.co.uk 

and, if necessary, in the Forum on the 

same website. 



(070039-1) 

Figure 4. Elektor Electronics SDR Tuning control program. 

Figure 5. Four AM stations in tuning range spectrum, as displayed by the SDRadio program. 

Web links: 

[1] www.nti-online.de/diraboxsdr.htm 

[2] www.sdradio.org/ 

[3] http://sourceforge.net/projects/drm 

[4] www.g8jcf.dyndns.org/ 

Literature: 

Burkhard Kainka: DREAM Team –Software for DRM reception, Elektor Electronics 4/2004, pp. 

20 ff. 

Wolfgang Hartmann and Burkhard Kainka: ‘Radio listening with Matlab—Diorama software 

DRM receiver’, Elektor Electronics 4/2006, pp. 76 ff. 

Burkhard Kainka: I-Q: a highly intelligent approach to quality radio, Elektor Electronics 

12/2006, pp. 38 ff. 

�� 

19

TECHNOLOGY WIRELESS 

Dr. Erik Lins and Christian Meinhardt 

Wireless 

iDwaRF: a network 

iDwaRF brings together a Cypress WirelessUSB transceiver and an Atmel AVR microcontroller to create a 

networkable 2.4 GHz radio module featuring a free protocol stack and development environment. 

Besides standard applications such 

as mobile radio, WLAN and Bluetooth, 

highly-integrated low power radio devices 

open up many new possibilities, 

Figure 1. The iDwaRF radio module with 2.4 GHz WirelessUSB 

transceiver and ATmega168 AVR microcontroller. 


including wireless sensor networks 

and even radio-controlled robotic football 

teams able to orchestrate an offside 

trap in the blink of an eye! And, 

with iDwaRF, we can do all this without 

complex protocols or licensing 

problems. 

An alternative to ZigBee 

For wireless sensor network applications 

ZigBee [1] is often the protocol of 

choice. The protocol is relatively complicated, 

and only members of the Zig- 

Bee Alliance are permitted to use it in 

commercial products. Cypress [2] offers 

a simpler alternative in its WirelessUSB 

technology [3]. The devices are cheap 

and the radio protocol makes only moderate 

demands in terms of hardware 

and memory in the microcontroller. 

WirelessUSB supports wireless manyto-one 

links and is thus ideal for use in 

wireless sensor networks. The protocol 

details are freely available and can be 

used without restriction in combination 

with Cypress radio chips. 

iDwaRF 

The iDwaRF-Net software that accompanies 

the iDwaRF-168 module (Figure 

1) is a port of the Cypress WirelessUSB 

protocol to the ATmega168 

AVR-family microcontroller [4]. The 

iDwaRF-168 module can be freely reprogrammed 

and can equally well play 

the role of hub or sensor in a manyto-one 

wireless sensor network. It is 

easy to add extra application-specifi c 

functions. 

WirelessUSB operates in the 2.4 GHz 

ISM band. Each WirelessUSB radio network 

uses a selection from a total of 79 

channels: even when multiple WirelessUSB 

devices are operating simultaneously 

the protocol will be able to fi nd a 

free channel to use. 

Transmission uses a robust DSSS (direct 

sequence spread spectrum) modulation 

scheme [5]. Even at a 10 % error 

rate the data can still be received correctly, 

and if there should be long-term 

interference the protocol provides for 

Wireless USB in miniature • 74 


USB in miniature 

able WirelessUSB radio module 

an automatic change of channel. Like 

Bluetooth, WirelessUSB comes in shortrange 

(up to 10 m) and long-range (up 

to 50 m) versions. The latter is used in 

the iDwaRF-168 module. 

Hub and sensors 

An iDwaRF module programmed to 

act as a hub forms the centre point 

of a star-topology many-to-one radio 

network, which can consist of many 

sensors (Figure 2). Normally the hub 

operates continuously and can be connected 

to a PC or another microcontroller, 

acting as a host, using a serial 

link. For simple applications the 

iDwaRF module itself can be programmed 

to carry out the required 

dedicated host functions. 

ANT2 

1 


1 

ANT_WIGGLE_L 

ANT1 

ANT_WIGGLE_R 

L1 

3nH9 


C1 

1p8 

C4 

10p 

C2 

NOLOAD 

VCC 

C11 

100n 

NC 48 

NC 47 

RFIN 46 

VCC 45 

VCC 44 

NC 43 

X13IN 

5 

RFOUT 

35 

X13OUT 26 

/PD 33 

PACTL 34 

SCK 25 

6 

VCC 

9 

VCC 

VCC 28 

VCC 29 

VCC 32 

1 

NC 

2 

NC 

3 

NC 

4 

NC 

7 

NC 

8 

NC 

10 

NC 

NC 

11 

NC 

12 

NC 

27 

NC 30 

NC 31 

NC 36 

U1 

CYWUSB6935-48LFX 

VCC 42 

VCC 41 

38 

X13 

NC 37 

NC 39 

NC 40 

14 

/RESET 

20 

DIO 

DIOVAL 

13 

GND 

15 

NC 

16 

VCC 

17 

NC 

18 

NC 

19 

21 

IRQ 

22 

/SS 

23 

MOSI 

24 

MISO 

VCC 

C5 

10u 

C6 

100n 

C10 

100n 

VCC 

C8 C7 

100n 100n 

E 

/INT0 

MISO 

MOSI 

/ SS 

X1 

13MHz 

VCC 

X13OUT 

SCK 

C9 

100n 

/WUSBRESET 

Figure 2. The star-topology radio network consists of a hub and several sensors. 

/PD 

/WUSBRESET 

/INT0 

14 

13 

12 

11 

10 

9 

8 

7 

6 

5 

4 

3 

2 

1 

CON1 

/RESET 

VCC 

PORT6 

PORT6 

VCC 

GND 

/RESET 

MOSI 

SCK 

VCC 

MISO 

PORT7 

PORT6 

PORT5 

PORT4 

PORT3 

PORT2 

PORT1 

PORT0 

Figure 3. Circuit diagram of the iDwaRF module. 

X2 

C12 

xxMHz 

C13 C14 

100n 22p 22p 

MISO 

MOSI 

/ SS 

SCK 

PORT4 

PORT5 

PD1 TXD 31 

32 

PD2 INT0 

PORT0 

PORT1 

PORT0 

PORT7 

PORT7 

PORT2 

PORT3 

PORT0 

PORT1 

1 

INT1 OC2B PD3 

2 

XCK T0 PD4 

3 

GND 

4 

VCC 

5 

GND 

6 

VCC 

7 

XTAL1 TOSC1 PB6 

8 

XTAL2 TOSC2 PB7 

PB5 SCK 17 

AVCC 18 

ADC6 19 

AREF 20 

GND 21 

ADC7 22 

PC0 ADC0 23 

PC1 ADC1 24 

U2 

ATmega168-20MI 

OC0B T1 PD5 

OC0A AIN0 PD6 

AIN1 PD7 

CLKO ICP PB0 

OC1A PB1 

/SS OC1B PB2 

MOSI OC2 PB3 

MISO PB4 

9 

10 

11 

PD0 RXD 30 

PC4 ADC4 SDA 

26 

PC3 ADC3 

25 

PC2 ADC2 

27 

PC5 ADC5 SCL 28 

PC6 /RESET 29 

12 

13 

14 

15 

16 

050402 - 11 

VCC 

C15 

100n 


45


Figure 4. The iDwaRF printed circuit board 

includes a printed antenna. 

Figure 6. The iDwaRF module is mounted 

on the hub printed circuit board. 

A sensor unit consists of a suitablyprogrammed 

iDwaRF module with 

sensors attached. To save power we 

use the AVR’s internal RC oscillator. 

Also, the module is only activated at 

intervals (as determined by the ‘beacon 

time’). Communications are initiated 

by the sensor and terminated by 

the hub; a reverse channel (transmitting 

information from hub to sensor) is 

also available. 

Module printed circuit board 

Figure 3 shows the circuit diagram 

of the iDwaRF-168 module. The Cypress 

CYWUSB6935-LR transceiver is 

connected to an ATmega168 microcontroller 

over SPI, using the MISO, 

MOSI, SCK and /SS (chip select) signals. 

An interrupt signal (/INT0) from 

the radio device indicates the reception 

of data. The ATmega168 can put 

the radio chip into power-down mode 

using an I/O pin connected to the / 

PD signal, and can reset it using the 

/WUSBRESET signal. The radio chip 

needs just an external 13 MHz crystal 

(X1) and decoupling capacitors (C6 to 

C11) for operation. The transmit and 

receive antennas are separate from 

one another and integrated directly 

into the circuit board layout as meander 

lines (Figure 4). Having separate 

antennas gives greater range and sim- 


CON1 

CON3 

CON1 

CON3 

1 PORT0 

2 PORT1 

3 PORT2 

4 PORT3 

5 PORT4 

6 PORT5 

7 PORT6 

8 PORT7 

9 MISO 

10 VCC 

11 SCK 

12 MOSI 

13 RST 

14 GND 

1 MISO 

2 VCC 

3 SCK 

4 MOSI 

5 RST 

6 GND 

1 PORT0 

2 PORT1 

3 PORT2 

4 PORT3 

5 PORT4 

6 PORT5 

7 PORT6 

8 PORT7 

9 MISO 

10 VCC 

11 SCK 

12 MOSI 

13 RST 

14 GND 

CON2 

1 PORT0 

2 PORT1 

3 PORT2 

4 PORT3 

5 PORT4 

6 PORT5 

7 PORT6 

8 PORT7 

9 MISO 

10 VCC 

11 SCK 

12 MOSI 

13 RST 

14 GND 

1 

2 

3 

4 

5 

6 

MISO 

VCC 

SCK 

MOSI 

RST 

GND 

12 SUSPEND 

9 

RST 

SUSPEND 


* 

1k5 

R1 

see text 

VCC 

VCC 

LED1 

green 

MOSI 

SCK 

RST 

SW1 

24 

RTS 

CTS 

23 

27 

28 

DTR 

DSR 

25 

26 

TXD 

RXD 

1 

11 RI 

2 DCD 

NC 16 

NC 17 

NC 18 

NC 19 

NC 20 

NC 21 

NC 22 

10 

NC 

NC 

13 

C1 

10u 

CP2102 

USB to UART 

Bridge 

29 

GND 

GND 

30 U2 

U1 

4 

IN 

2 

NC 

1 

BYPSS 

C2 

10n 

* 

14 

NC 

NC 

15 

3 

SHD 8 

GND 

Figure 5. Circuit of the hub board. 

BATT 

Battery Clip 3xAAA 2479 

PORT7 

1k5 

PORT3 

PORT2 

PORT6 

R1 

green 

PORT4 

PORT5 

VCC 

LED1 

J1 

SW1 

1k5 

R2 

1k5 

1 

2 

U3 

1 

SI 

2 

SCK 

3 

RST 

4 

CS 

R3 

* 

1k5 

Vbatt 

VCC 

R4 

7 

VCC 6 

GND 

VBUS 8 

REGIN 7 

D- 5 

D+ 4 

VDD 6 

GND 3 

Vbus 

OUT 5 

SNS 6 

ERR 7 

D- 

D+ 

C5 

100n 

Vbus 

VCC 

LP2989IM-3.3 

C3 

10u 

SO 8 

WP 5 

AT45DB081D-SSU 

U2 

1 

SDA 

2 

SCL 

3 

O.S. 

C1 

10u 

4 

C5 

100n 

VCC 

VCC 8 

GND 

4 

2 

1 

C2 

10n 

A0 7 

A2 5 

A1 6 

LM75 

U1 

IN 

NC 

MISO 

BYPSS 

3 

SHD 8 

GND 

PORT0 

PORT1 

1 

2 

3 

4 

5 

C4 

1u 

OUT 5 

SNS 6 

ERR 7 

9 

8 

6 

7 

050402 - 12 

CON2 

USBMINIB 

LP2989IM-3.3 

C3 

10u 

PORT7 

Figure 7. Circuit diagram of the node board with temperature and light sensors. 

* 

C4 

100n 

J2 

4k7 

* 

LDR 

R7 

see text 

VCC 

VCC 

Vbatt 

1M 

1M 

050402 - 13 

R5 

R6 

Wireless USB in miniature • 76

Table 1. CON1 connection groups 

iDwaRF-168 CON1 

(port pin) 

plifi es matching (L1, C1 and C4). The 

antennas are laid out in the manner 

recommended by Cypress. 

The microcontroller is equipped with 

a crystal in an HC49 package, and so 

it is straightforward to change it for a 

different frequency. As supplied the 

ATmega168 is confi gured to use its internal 

RC oscillator. 

A 14-way header (CON1) brings out 

the ISP (in-system programming) 

signals of the ATmega168 (MOSI, 

MISO, SCK and /RST), power, and 

eight spare I/O port pins. Some of the 

header pins are connected to more 

than one signal on the microcontroller 

to allow as many as possible of 

the peripheral functions of the device 

to be used. This means that you must 

ensure that any two microcontroller 

pins that are connected to the same 

header pin are never simultaneously 

confi gured as outputs, or damage to 

the microcontroller may result. Table 

1 shows the CON1 pinout and 

signals in detail. 

Application boards 

In the simplest wireless network scenario 

one iDwaRF-168 module is programmed 

as a hub and one or more 

modules are programmed as sensors. 

It is of course necessary to build the 

necessary interfaces to the sensors 

themselves and connect them to the 

modules. To simplify building such 

systems we have developed three application 

boards, to each of which can 

be attached an iDwaRF-168 module: a 

hub board, a node board and a prototyping 

board. 


First connection 

(ATmega168 pin number) 

Second connection 


The hub board supports the iDwaRF 

module with a USB interface (the 

CP2102), a 3.3 V LDO voltage regulator, 

button and LED (Figure 5). 

The node board is used to make a sensor 

unit using an iDwaRF module. The 

circuit (Figure 6) includes an LDR as 

a light sensor, an LM75 temperature 

sensor, an (optional) AT45DB801D serial 

fl ash memory, a button and an LED. 

Power is provided by three AAA cells 

and a 3.3 V LDO voltage regulator. 

The two printed circuit boards (Figure 

7 and Figure 8) are chiefl y populated 

using SMD components. Figure 9 

shows the boards with iDwaRF-168 

modules fi tted. 

Third connection 


PORT0 OC0B / T1 / PD5 (9) AIN1 / PD7 (11) ADC3 / PC3 (26) 

PORT1 OC0A / AIN0 / PD6 (10) ADC2 / PC2 (25) - 

PORT2 SCL / ADC5 / PC5 (28) - - 

PORT3 SDA / ADC4 / PC4 (27) - - 

PORT4 TXD / PD1 (31) - - 

PORT5 RXD / PD0 (30) - - 

PORT6 INT1 / OC2B / PD3 (1) XCK / T0 / PD4 (2) - 

PORT7 CLKO / ICP / PB0 (12) OC1A / PB1 (13) - 


Figure 9. Node board (rear) and hub board (front) 

with iDwaRF radio modules fi tted. 

Figure 8. The node board converts the iDwaRF module 

into a complete sensor unit. 


47


Table 2. Example programs 

empty Empty program: framework for new applications. 

chat Creates a wireless serial connection between two host PCs, allowing ‘chatting’ between two terminal programs. 

tutorial 

terminal 

Example program that switches the LEDs on the hub and sensor modules on and off remotely when a button is pressed. The implementation 

of this example is explained step-by-step in a separate ‘how to’ document (see text box). 

This basic sensor network application supports the components on the node board or iDwaRFSensorBox. Data packets (including 

battery voltage, potentiometer setting, button state and temperature) from several sensors are displayed in plain ASCII text. The terminal 

program can also send data to individual sensors. 

quad_adc This program reads four ADC channels and transmits the readings to the hub at regular intervals. 

K2 

1 PORT0 

2 PORT1 

3 PORT2 

4 PORT3 

5 PORT4 

6 PORT5 

7 PORT6 

8 PORT7 

9 MISO 

10 VCC 

11 SCK 

12 MOSI 

13 RST 

14 GND 

iDwaRF 

K4 

1 MISO 

2 VCC 

3 SCK 

4 MOSI 

5 RST 

6 GND 

AVRISP 

VCC 

K3 

1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

EXT 

K1 

3 

1 

2 

BATT 

D1 

1N4001 

VCC 

Battery Clip 3xAAA 2479 

SOT-23/6 

SOIC-24 

1 

2 

SOT-23/6 

1u 

25V 

1 3 

The prototyping board has a generous 

area for building your own circuits 

and is suitable for creating more 

specialised applications using the 

iDwaRF-168 module. There are SMD 

pads on the reverse of the board. The 

only active component in the circuit 

(Figure 10) is the LDO voltage regulator, 

in a TO-92 or TO-220 package according 

to the expected current draw. 

As well as the regulator and socket to 

accept the iDwaRF-168 module there 

is an AVR programming connector, a 

battery holder and, as an alternative, 

a socket for a mains adaptor with a 

reverse polarity protection diode. The 

printed circuit board (Figure 11) is 

half-Eurocard sized and can be used 

in the place of a node board. For reasons 

of space we have made the parts 

lists for the circuit boards and layouts 

available for download from the Elektor 


Software 

The iDwaRF-Net software package [6] 

includes a library of fi rmware for use 

in hub and sensor modules with corresponding 

header fi les (in the ‘iDwaRF’ 

directory), along with a few ancillary 

functions, for example to support serial 

communications (‘USART’ directory). 

There are also four example programs 

which can either be used as they stand 

or form the basis for dedicated applications. 

Table 2 gives more details of 

these example programs. 

Each example program consists of hub 

source code (userMain_hub.c) and 

sensor source code (userMain_sensor. 

c). The fi rmware provides the facility 

to register so-called ‘callback’ functions, 

which the fi rmware calls regularly 

in the course of normal operation. 

These functions can be used for application-specifi 

c code. The most important 

callback functions are explained 

in the text box. 

Ready, steady… 

Assembling the hardware is relatively 

straightforward. The SMD printed circuit 

boards (the iDwaRF module, the 

node board and the hub board) are 

available as ready-made units (see 

the ‘Elektor Shop’ pages at the back of 

this issue). A kit of parts is available 

for the prototyping board. Separately-ordered 

iDwaRF-168 modules are 

supplied unprogrammed and without 

header or crystal fi tted in order to give 

the user maximum fl exibility. 

The fi rmware for programming a hub 

or sensor module is freely available 

[6]. Modules ordered bundled with a 


C1 

IC1 

LP2950 

LF33CV 

Figure 10. Circuit diagram of the prototyping board for dedicated iDwaRF module applications. 

Figure 11. The prototyping board has an experimentation area 

with SMD footprints on the reverse. 


2 

C2 

VCC 

1u 

25V 

050402 - 14 

Wireless USB in miniature • 78

node or hub board come ready-programmed. 

A crystal is always required 

for baud rate generation on the hub 

board (a 7.3728 MHz crystal is supplied 

as standard), and the microcontroller 

must be suitably programmed 

for crystal operation. The correct values 

for the ATmega168 with a crystal 

oscillator are: extended byte, 0xF9; 

high byte, 0xDF; and low byte, 0xFC. 

It is recommended to use the same 

values and same crystal frequency for 

the sensor, and the fi rmware is currently 

set up to work on this assumption. 

If the frequency is to be changed 

(using a different crystal or the internal 

RC oscillator) the relevant #defi ne 

in the fi rmware must be changed and 

the code compiled afresh. 

The AVR ISP connectors have to be 

soldered on to the node board and 

hub board, and the node board also 

needs to be connected to the battery 

holder. Finally the iDwaRF module can 

be fi tted and (in the case of the node 

board) the batteries inserted. The 

USB connection on the hub board requires 

the corresponding CP2102 virtual 

COM port driver to be downloaded 

and installed [6]. Drivers are available 

for both Windows and Mac OS X. A 

CP210x module is provided as stand- 


ard in current Linux kernels, allowing 

the iDwaRF module to be used with 

non-Windows PCs. 

The pre-compiled hex fi les [6] for hub 

and sensor are downloaded using the 

six-pin AVR ISP connector. Note that 

the supply voltage for the iDwaRF 

module on the node and hub boards is 

only 3.3 V, and so care must be taken 

Table 3. Commands available in ‘terminal’ program 

Command Description 

to ensure that the programming adaptor 

also works at this voltage. Simple 

STK200-compatible programming 

adaptors that connect to the PC’s 

printer port, which draw power from 

the target device, do not always work 

reliably at 3.3 V. Modern USB programming 

adaptors (compatible with the 

STK500-V2) such as the CrispAVR-USB 

work without problems. 

rst Restarts the hub fi rmware. All sensors are unregistered and lose their ID codes, and must register again with the hub. 

gps Returns an internal Cypress fi rmware state variable. 

bon 

Activates automatic bind mode on the hub. New sensors are probed for using PN code 0 and channel 0. In normal use automatic 

bind mode is activated. 

bof Deactivates automatic bind mode. New sensors can no longer register with the hub. 

enu 

cln 

cnf 

snd 

del 

Displays a list of the currently registered sensors on the hub. The list includes the sensor ID assigned during registration and the 

unique manufacturing ID stored in the radio chip. 

Cleans up the hub’s list of registered sensors. Sensors which have registered more than once with the hub and which have therefore 

been assigned different ID codes are removed from the list and only their current ID code remains valid. 

Confi gures network parameters. There are 8 different PN codes and channel subsets, and this command allows the hub to be switched 

to a new PN code and channel. The facility to change channel number is for test purposes only, as the channel is changed 

automatically in normal operation. The PN code, however, remains fi xed. The format is 

cnf 

where 0 < pncode < 9 and 0 < channel < 80. The cnf command automatically deletes all registered sensors from the list. 

Sends beacon time and other data to a sensor. The data packet is buffered in the hub and when the sensor in question next makes 

a transmission the packet is sent back in the back channel to the sensor. All parameters are given in decimal. The format is 

snd ... 

A beacon time of -1 indicates that the beacon time is not to be changed. 

Removes a sensor from the list of registered sensors in the hub. The format is 

del 

hex Causes sensor data to be displayed in hex rather than as plain text. 


Figure 12. Prototyping board with iDwaRF radio module plugged on. 


49


The principal callback functions 

In the hub: 

cbSensorPacketReceived(): called when a packet is received from a sensor. Direct access to the current sensor data is possible, although if 

lengthy processing is to be carried out the data should be copied into a global buffer and the work done in the main program. 

cbSerialDataReceived(): called when a byte is received over the serial interface. Usually the byte is simply stored in a global buffer and its reception 

signalled using a fl ag, so that more time-consuming processing can be carried out in the main program. 

cbProcessRxData(): further processes the bytes received by cbSerialDataReceived(), for example to implement a complex communications protocol 

with the host PC. In the ‘terminal’ example the commands entered on the host PC are parsed and processed in this function. 

In the sensor: 

cbConfi gForSleep(): called shortly before the sensor switches into power-down mode. This allows for particular sensor devices to be switched 

off to reduce power consumption. 

cbExitFromSleep(): called when the sensor leaves power-down mode. At this point particular sensor devices can be powered up again. 

cbTxProcess(): assembles the data packet to be sent to the hub. At this point sensors can be read and the readings stored in the global transmit 

buffer. The current data packet is then automatically sent to the hub. 

cbBackchannelProcess(): called with data received from the hub in the reverse channel. This can be used to create an output signal on the 

sensor, or to generate an analogue voltage using PWM. The ‘terminal’ example switches on the LED when a data packet is received. 

... go! 

The ‘terminal’ example program is 

the best one to use to demonstrate 

all the important functions of a wireless 

sensor network. iDwaRF modules 

programmed as sensors automatically 

connect to the module programmed 

as the hub and transfer 

data. With one node board and a hub 

board connected to a PC it is possible 

to see immediately on the PC’s 

Data format 

screen when light falls on or is shaded 

from the light sensor, when the 

button is pressed, or when the temperature 

sensor is warmed or cooled. 

Setup proceeds as follows. 

Hub board 

The virtual COM port driver creates a 

virtual COM port when the USB cable 

is plugged in. The number of the port 

can be obtained from the Device Manager 

(reached from the Control Panel). 

Now we can run a terminal program, 

set to the relevant port, and talk to the 

hub in plain ASCII. If there are no sensors 

there will initially be no reports 

from the hub. The hub is reset by typing 

the command ‘rst’ and pressing 

‘enter’: the hub will then emit its startup 

message. It is best to confi gure the 

terminal program to expect CR+LF at 

the end of each line and to enable local 

echo, as the hub does not echo the 

characters it receives. 

S5: ID 0 ldr 212 temp 22.5°C batt 2.9V button OFF (5 6 7 8 9 10) : 11 

a) b) c) d) e) f) g) h) 

Legend: 

a) Packet type: S0 (BIND_REQUEST); S1 (BIND_RESPONSE); S2 (PING, hub only); S3 (ACK); S4 (DATA, hub only); S5 (DATA, sensor only) 

b) Sensor ID 

c) ADC value from the light intensity sensor 

d) Temperature 

e) Battery voltage 

f) Button state (ON or OFF) 

g) Six unused bytes displayed as decimal indices 

h) Data byte count 

The values from c) to g) above form the packet data payload. The standard packet size is set to 17 bytes, of which six are protocol overhead, 

leaving 11 bytes of payload. 




Sensor 

If a sensor is switched on the data 

packets received will be displayed 

line-by-line in the terminal window. 

The first packet is called a ‘bind request’ 

where the sensor registers with 

the hub and, in return, receives an ID 

code assignment and a value called the 

‘beacon time’. This period, which has 

a default value of fi ve seconds, is the 

interval between the transmission of 

successive data packets. The format of 

the data packets is described in detail 

in the text box ‘Data format’. A brief 

fl ash of the LED on the sensor board 

shows when it is active; the LED is extinguished 

when the sensor returns to 

power-down mode. 

If the node board is moved to a warmer 

location, or if the ambient light intensity 

changes, the data packets displayed 

will refl ect the new sensor values. 

If the button on the node board is 

pressed a data packet is transmitted 

immediately: this shows that it is possible 

to react immediately to external 

events, without waiting for the preset 

beacon time to elapse. 

Command line 

Commands can be typed into the terminal 

window at any time. Typing ‘enu’ 

(‘enumerate’) lists the sensors that are 

registered with the hub, showing their 

ID code and unique serial number. 

Our fi rst sensor will appear in this list 

with ID ‘0’. To send data to this sensor 

we use the ‘snd’ command. In its 

simplest form this has two parameters: 

‘snd 0 40’ sends the value 40 to 

the sensor with ID code ‘0’. The fi rst 

data byte is always interpreted by 

the sensor as a new beacon time, in 

units of 125 ms. In this example, the 

value of 40 corresponds to the default 

beacon time of fi ve seconds. So, if we 

type ‘snd 0 8’ we will set the beacon 

time for sensor 0 to one second, and 

data packets from this sensor will be 

displayed in the terminal window at 

this rate. The command ‘snd 0 40 1’ 

sets the beacon time back to fi ve seconds 

and also sends an extra data 

byte with value ‘1’, which will cause 

the LED on the node board to light. In 

the terminal example the code in the 

sensor simply checks whether there is 

an extra data byte beyond the beacon 

time or not, and sets the LED on or off 

accordingly. The actual value of the 

extra byte is not taken into account. 

The complete set of commands provided 

by the hub in the ‘terminal’ ex- 



ample is listed in Table 3. 

If now a further sensor is activated 

another ‘bind request’ packet will appear 

among the data packets being received 

from the fi rst sensor, followed 

by a series of data packets. The ‘enu’ 

command can be used to list the registered 

sensors, and should now display 

two entries with ID codes ‘0’ and ‘1’. 

We can test the new sensor by adjusting 

its beacon time: ‘snd 1 8’ will set 

it to one second, and we should now 

receive packets from sensor 1 at fi ve 

times the frequency of those from sensor 

0. If an ‘x’ is used in place of the ID 

code in the send command, the beacon 

times for all sensors are set simultaneously. 

Using ‘snd x 40’ we can therefore 

reset the beacon times for both sensors 

to fi ve seconds. 

If the data packets are to be processed 

on the host PC, the ‘hex’ command 

can be used to switch the display 

from readable form to pure hex 

values. These can easily be read by 

another application, for example using 

the scanf() function. 

Room for expansion 

The supplied programs can form the 

basis of dedicated applications using 

iDwaRF modules: the possibilities are 

endless. Often a couple of extra components 

are all that is needed, for example 

to make measurements using an 

ADC, generate PWM waveforms, scan 

keys or drive an LCD. 

The iDwaRF-Net software can in theory 

work with up to 255 sensors, although 

at present the limit is 32. For larger 

sensor networks the xHub is planned, 

using an ATmega128 with more fl ash 

memory and an external SRAM. An external 

antenna will increase the range 

of the hub. 

Users of the iDwaRF radio module [7] 

and the iDwaRF-Net fi rmware can use 

a forum [8] organised by the author. 

This is in addition to the forum on the 

Elektor Electronics website. Further example 

programs will of course appear 

for download at [6] as soon as they become 

available. 

(050402-I) 

References 

and links: 

[1] www.zigbee.org 

[2] www.cypress.com 

[3] Thomas Biel: ‘Wireless USB’, 

Elektor Electronics, September 2004, p. 8 

[4] www.atmel.com/avr 

[5] Stefan Tauschek: ‘Wireless Connectivity’, 

Elektor Electronics, February 2005, p. 14 

[6] www.elektor-electronics.co.uk or 

www.chip45.com/iDwaRF-168_Downloads 

[7] www.chip45.com/iDwaRF-168 

[8] www.chip45.com/iDwaRF-Net 


51


A 16-bit Tom Th 

Gunther Ewald and Burkhard Kainka 

Thanks to the efforts of Elektor Electronics and Glyn, for the first time now a European 

electronics magazine supplies a complete microcontroller starter board and accompanying 

software CD-ROM at a very attractive price. We already introduced the Renesas R8C in the last 

issue. Now it’s time to start using it. 

Now it’s for real – from this issue you 

can order a circuit board with an 

R8C/13 microcontroller and the necessary 

software, at price consisting of 

P&P and handling only. The service is 

strictly limited to stocks and provided 

exclusively for Elektor Electronics 

readers in co-operation with Glyn. 

So if you want to grab an R8C starter 

kit, order it straight away — the online 


38 

method using our website is by far the 

fastest. 

There are three good reasons for using 

the Renesas R8C/Tiny family: first, it 

provides 16-bit computing power at a 

low price; second, it comes with a free 

but nevertheless very capable C compiler; 

and third, no programming hardware 

is necessary, because code can 

be easily downloaded to flash mem- 

ory via the RS232 port. We already 

introduced the board and the software 

in the January 2006 issue. If you 

missed that article, you can download 

it free of charge from the Elektor Electronics 

website at www.elektor-electronics.co.uk 

(select the Magazine 

tab, then January 2006). Our objective 

in this article is to help you start 

using the board. 

A 16-bit Tom Thumb • 82 


umb 


When installing the necessary 

software, follow 

the instructions exactly as 

given in order to ensure 

that what you later see 

on your PC matches 

what we describe here. 

Install the KD30 monitor/debugger 

first, followed 

by the NC30 C 

compiler with the HEW 

development environment 

and an update. 

Installation in this order 

is important so the HEW 

will find the debugger 

already installed and link it in properly. Next, install the 

debugger package in order to integrate the debugger into 

the development environment. Later on, you only need to start 

the HEW in order to have everything together on your 

screen. Finally, you have to install the Renesas Flash 

Development Toolkit (FDT), which enables you to download 

finished programs to the microcontroller. 

After you insert the CD, a product summary in PDF format 

will be displayed first. You should see the main directory of 

the CD after you close the product summary. Alternatively, 

you can right-click on the drive icon and select ‘Open’ in 

order to skip the PDF intro. The most important directories on 

the CD are \Software and \Sample_NC30. They contain all 

the necessary programs (which should now be installed) and 

the initial sample projects. 

1. KD30 

The executable installer file KD30V410R1_E_20041203.exe. 

is located in the \Software\kd30400r1\ directory on the 

CD. Start the installation and confirm the default path 

C:\MTOOL\. 

2. NC30 

Run the setup file nc30wav530r02_2_ev.exe in the 

\Software\ nc30v530r0_hew\ directory on the CD. First you 

have to select whether you want to install the Japanese version 

or the English version. Most of our readers will probably 

prefer the English version. The program suggests C:\Program 

Files\Renesas as the default installation path. You should confirm 

this path. 

A second default path is shown for the tool chain: 

C:\Renesas\NC30WA\V530R02. You should also accept the 

default path name, as well as all subsequent default path names. 

At the end of the installation, an individual site code is displayed. 

You can ignore it, because it is only necessary for 

registering the software. If you purchase the full version of the 

compiler, you will automatically receive a CD with your own 

Our R8C starter kit is available 

– now you can get going! 

The software 

code for enabling the 

software. 

After you have installed 

the HEW, you have to 

run the installer for the 

AutoUpdate program. As 

shown in the figure, confirm 

that you wish to 

receive weekly updates 

from the Renesas Tools 

website. That enables 

you to keep your software 

‘fresh’ all the time. 

If the PC on which your 

software is installed does 

not have access to the Internet, you can cancel installation of 

AutoUpdate without causing any problems. The most recent 

update file is on the CD, and it must be installed next. 

After being installed, the updater immediately checks to see 

whether there is anything new and automatically downloads 

the latest changes. The first update will then be installed automatically. 

The PC must then be restarted before you continue. 

3. Update HEW 

This step is only necessary if you have not already downloaded 

the latest update from the Internet and installed it. Run 

the first HEW update from the CD by starting 

hewv40003u.exe in the 

\Software\HEW_V.4.00.03.001_Update directory on the 

CD. That will update your compiler to the latest version 

(Database Version 7.0). That is important, because the sample 

projects have been generated for this version. Although 

you can use the new version to load older projects, which 

will then be updated automatically, regressing to an older 

version of the compiler is not that easy. 

4. Debugger package 

Run the installer file m16cdebuggerv100r01.exe in the 

\Software\Debugger Package\ directory on the CD. Follow 

the instructions of the installer program and confirm your 

agreement to the licence conditions. Everything else is automatic. 

You must restart the computer after this installation. 

5. Flash Development Toolkit 

Install the FDT by running the installer file fdtv304r00.exe in 

the \Software\Flasher_FDT directory on the CD. Confirm all 

default settings. Everything else is automatic. 

When everything has been installed, you will find the 

Renesas program group under Start / All Programs. The two 

relevant programs in the group are High-performance 

Embedded Workshop and Flash Development Toolkit. 




40 

Figure 1. The hardware you can order from us at low cost forms a complete microcontroller system. 

SUB-D9 

+9V 

1 

2 RXD 

3 TXD 


6 

7 

8 

4 

9 

5 

27k 

1N4004 

Figure 2. The fully assembled board with the pin headers fitted. 

10k 

4k7 

BC 

548C 

78L05 

100n 100n 

R8C/13 

1 

RXD1 TXD1 

32 

2 

AN6 

31 

3 

RESET AN5 

30 

4 

AN4 

29 

5 

VSS MODE 

28 

6 

AN3 

27 

7 

VCC 

AN2 

26 

8 

P17 

AN1 

25 

9 

P16 

AN0 

24 

10 

P15 

23 

11 

P14 

P30 

22 

12 

P13 

21 

13 

P12 

P31 

20 

14 

P11 

19 

15 

P10 

P32 

18 

16 

P45 P33 

17 

Figure 3. The circuit diagram of a minimal system for initial tests. 

100k 

10k 

BC 

558C 

050179 - 2 - 11 

The hardware 

The low-profile PCB with pre-assembled 

SMD components is supplied with 

two pin headers that you must fit and 

solder yourself (Figure 1). That yields 

a complete processor module in the 

format of a 32-pin DIL IC (Figure 2). 

There is also space reserved on the 

board for a 14-way pin header, but it 

does not have to fitted right away 

because it is only needed for the E8 

debugger. 

The actual microcontroller (the R8C/13) 

is contained in the 32-pin LQPF SMD 

package, which measures 7 by 7 mm 

and has a lead spacing of 0.8 mm. The 

marking ‘R5F21134FP#U0’ reveals that 

it is an R8C/13 with 16 KB of flash 

ROM. We selected the R8C/13 because 

it has the same characteristics as its 

‘siblings’ (R8C/10, R8C/11 and 

R8C/12). The board also comes fitted 

with a 20-MHz crystal and the necessary 

capacitors, as well as several 

other capacitors and resistors. Altogether, 

this amounts to a complete 

microcontroller system. Once a program 

has been loaded, all you have to 

do is connect a 3.3-V or 5-V supply voltage 

and you’ve got a working system. 

Program code can be loaded using a 

serial interface; no special programming 

hardware is necessary. That’s 

because the microcontroller has a 

debug interface and a corresponding 

boot program that can be used to copy 

the software into the flash ROM. 

First contact 

In the next issue of Elektor Electronics, 

we will describe a complete development 

system with RS232 and USB 

interfaces. But you naturally don’t 

want to wait that long to try out the 

board. For that reason, we describe a 

solution here that only requires a few 

components from your parts bin: 

• a source of 5-V power, preferably stabilised 

by a voltage regulator 

• an inverting level converter for connection 

to the RS232 port 

• a pushbutton reset switch 

• a mode switch to select the programming 

mode 

The microcontroller board feeds out all 

the microcontroller leads one to one. 

As we already mentioned, a crystal, a 

few capacitors and some resistors are 

fitted on the board. The schematic diagram 

in Figure 3 shows only the con- 




nections that are essential for a ‘quick 

start’, so you can get your bearings as 

fast as possible. The connections not 

shown in the figure must always 

remain open. Figure 4 shows an experimental 

setup on a prototyping board. 

The serial port of the PC is connected 

here via transistor inverters. Although 

a MAX232 could be used just as well 

for this purpose, the transistors will 

probably be easier to find in your parts 

bin. A BC548C NPN transistor inverts 

the TXD signal from the PC and feeds 

it to the RXD1 input of the microcontroller. 

This input does not have an 

internal pull-up, so the collector resistor 

is essential. In the opposite direction, 

TXD1 drives a BC558C PNP transistor. 

The RXD input of the PC has its own 

pull-down, so the collector resistor can 

be omitted here. 

The MODE input of the microcontroller 

determines whether the internal 

boot program or a downloaded 

user program is run after a reset. The 

MODE input is pulled high by a 10-k� 

resistor when the switch is open, 

which causes the user program to be 

started. If you want to load a program 

into the flash memory, you first have 

to close the switch in order to pull 

MODE low. Then you must briefly 

press the Reset switch. That causes 

the microcontroller to start up in the 

debug mode, which allows new software 

to be loaded into the flash ROM. 

After the software has been transferred, 

open the Mode switch and 

press the Reset switch again. That 

causes the downloaded program to be 

started. However, duty comes before 

pleasure, and in this case the duty is 

installing the software on the PC. 

That’s described step-by-step in the 

inset. Follow the instructions exactly 

as given in order to ensure that what 

you later see on your PC matches 

what we describe here. 

Ready, set, flash! 

The first thing you should do is to try 

out the Flash Development Toolkit 

(FDT) by downloading a finished program 

to the microcontroller. For the 

time being, we’ll skip the process of 

developing you own programs, so you 

can enjoy some tangible results as 

soon as possible. 

After installing the Renesas software 

package, you will find the FDT program 

in the Windows ‘Start’ menu. The 

program is shown there in two versions: 

a full version and a compact 

Figure 4. A prototype of the test system built on a lab prototyping board. 

‘Basic’ version. Start ‘Flash Development 

Toolkit Basic’ (Figure 5). 

You must configure several settings 

the first time you start the program. 

If necessary, you can access them 

later on via the Options / New Settings 

menu. Select ‘R5F21134’ as the 

microcontroller type, and select the 

upper of the two core protocol options 

(Figure 6). 

In the next window, select the serial 

interface port from the range 

COM1–COM4. The third window asks 

you to specify a baud rate for the link 

to the microcontroller. Enter ‘9600 

baud’ here (Figure 7). 

Finally, you have to specify whether 

you want to enable readout protection 

for the microcontroller. As the risk of 

criminal industrial espionage is rather 

small with the initial familiarisation 

programs, you can dispense with any 

form of protection. Save the settings as 

shown in Figure 8. That completes the 

preparations. 

Hurrah, it blinks! 

Now connect the board to the specified 

COM port on your PC. Close the Mode 

switch and briefly press the Reset 

switch. The microcontroller will enter 

boot mode and wait for you to send it 

data. 

The next step is to download a fully 

compiled program to the microcontroller. 

On the CD, you will find the 

‘Sample NC30’ folder with several sample 

projects. That includes the project 

folder ‘Sample_NC30\ port_toggle’, 

which contains the folder ‘port_toggle\ 

Release’. The file port_toggle.mot is 

located in the latter folder. It is a program 

in Motorola hex format that can 

be downloaded directly to the microcontroller. 

Specify the path to this file, and then 

start the download process by selecting 

‘Program Flash’. The download 

takes around two seconds. The flash 

memory is first erased, and the new 

program is then copied over. If everything 

goes properly, the message 

‘Image successfully written to device’ 

will be displayed. Open the Mode 

switch and briefly press the Reset 

switch. That will start the program 

that you just downloaded. 

The program toggles the first four 

lines of Port 1 (P1_0 to P1_4) at a slow 

rate, so you can observe their states 

using a LED with a series resistor. The 

ports of the R8C and the other M16 

microcontrollers have a low impedance 

in the output direction, regardless 

of whether they are in the high or 

low state. That means you must 

always use a series resistor (1 k�, for 




example) when connecting a LED to 

them (Figure 9). 

All of the sample programs described 

below can be copied to the microcontroller 

in the manner just described. If 

the R8C cannot be flashed due to a 

communication error, it must be left 

without power for one minute to erase 

the internal RAM loader. This error can 

occur if you have been working with 

the debugger before attempting to 

download a program. 

The R8C as a musician 

Like to try another hardware test? 


42 

Figure 5. The Flash Development Toolkit, Basic version. 

Then download the ‘Jingle_Bells’ 

Motorola file from the R8C project to 

the microcontroller. Connect a small 8- 

� loudspeaker or a headphone with a 

1-k� series resistor to the board (Figure 

10), and then start the microcontroller. 

You will hear a simple melody. 

Incidentally, this program uses the 

internal high-speed ring oscillator 

(8 MHz) and does not require the crystal 

oscillator. If you touch an oscilloscope 

probe to the crystal lead Xin 

(pin 6) or Xout (pin 4), you will see that 

no clock signal is present (in contrast 

to the situation with the port_toggle 

project). As you can hear from the 

Figure 6. Selecting the microcontroller type and transmission protocol. 

absence of any sour notes coming from 

the speaker, the internal RC oscillator 

is fully adequate for this task. That 

means you could flash the program 

into an R8C/13, connect it to a piezoelectric 

speaker and a 3-V button cell, 

and hang it from your Christmas tree 

or insert it in a Christmas card. 

Now that you know the hardware is 

OK, your fingers are probably itching 

to start programming something, 

aren’t they? OK, it’s time to start up 

the integrated hardware development 

environment (HEW). In order to keep 

things fairly simple at first, let’s start 

by making a few changes to an existing 

project. We’ll also work without the 

debugger to start with. 

High C 

You can use HEW to generate assembly-language 

projects. However, programming 

the R8C in assembly language 

is significantly more difficult 

than programming it in C, because 

there are so many different data formats 

and addressing modes. The C 

compiler looks after all that for you. You 

don’t even have to know whether the 

microcontroller is processing a word, a 

byte or a bit. In this case, C is easier 

than assembly language even for people 

who are only used to working in 

assembly language. The unfamiliar 

notation will quickly become second 

nature after you work through a few 

examples, so there’s no need to be 

afraid of C! 

First copy the entire ‘port_toggle’ project 

from the CD to your PC. When you 

start a new project, use a directory 

such as C:\WorkSpace, which is also 

the default directory suggested by 

HEW. In the Renesas program group, 

start the ‘High-Performance Embedded 

Workshop’ program. A selection 

window is displayed when the program 

starts, and you can specify 

whether you want to generate a new 

project or open an existing project. 

Use File/ Open Workspace to open the 

port_toggle file. All the files belonging 

to the project will then be shown at the 

left. Click on port_toggle.c to open the 

source code file. Everything should 

then appear as shown in Figure 11. 

Next, you should try to compile the 

project, just for the exercise. First you 

have to decide whether you want to 

generate a debug version or a release 

version. You should work without the 

debugger for now, which means you 

should select the ‘Release’ option 




under Build / Build Configurations. 

Then start the compilation by selecting 

Build / Build All. The C source code 

will be compiled, linked, and written to 

the output directory \Release in the 

form of a Motorola hex file. The entire 

process is listed at the bottom of the 

Build window. At the end, you’ll see 

the longed-for message that signals 

success: 

Build Finished 

0 Errors, 1 Warning 

No errors – that’s very good! Warnings 

occur relatively often and aren’t all that 

dramatic. In this case, the warning 

reads: ‘Warning (ln30): License has 

expired, code limited to 64K (10000H) 

byte(s)’. You don’t have to worry about 

that, because you’re using the free version 

of the compiler, and 64 kB is anyhow 

more than R8C/13 can hold. If you 

wish, you can download the output file 

to the microcontroller again and run 

the program. It will work just as well 

as the Motorola hex file on the CD. 

Now let’s have a look at the source 

code: 

while (1) /* Loop */ 

{ 

p1_0 = 0; 

p1_1 = 0; 

p1_2 = 0; 

p1_3 = 0; 

for (t=0; t


cm05 = 0; /* Xin on */ 

cm16 = 0; /* Main clock = No 

division mode */ 

cm17 = 0; 

cm06 = 0; /* CM16 and CM17 

enable */ 

asm(“nop”); /* Waiting for 

stable oscillation */ 

asm(“nop”); 



ocd2 = 0; /* Main clock 

change */ 

prc0 = 0; /* Protect on */ 

The above listing shows the relevant 

instruction lines at the beginning of 

the source code. It’s necessary to 

change a few control bits in System 

Clock Control Registers 0 and 1, but 

they are initially in protected mode. 

This protection is disabled in the first 

line so the relevant control bits can be 

altered. After the bits have been 

44 

P1- 0 


1k 

050179 - 2 - 20 

LED 

Figure 9. Connecting a LED to P1_0. 

P1- 0 

switched, we have to wait a little 

while for the oscillator to stabilise after 

being started up. After this, the clock 

source is changed and write protection 

is re-enabled. From now on, all programs 

will run at 20 MHz. 

If you now remove this entire block of 

code from the listing, including prc0 = 

1 and prc0 = 0, the oscillator will not 

be changed and the microcontroller 

will continue operating at 125 kHz. To 

make it possible to observe the result 

within a finite length of time, you must 

also shorten the wait loops by a factor 

of 100: 

void main(void) 

{ 

pd1 = 0x0F; /* Set Port 1.0 - 

1.3 be used for output*/ 

1k 

050179 - 2 - 21 

Figure 10. Connecting a loudspeaker to the 

microcontroller. 

while (1) /* Loop */ 

{ 

p1_0 = 0; 

Figure 11. The port_toggle project in the development environment. 

} 

p1_1 = 0; 

p1_2 = 0; 

p1_3 = 0; 

for (t=0; t

PROJECTS MICROCONTROLLERS 

USB Flash Board 

An 8051-based system for rapid software 

Alexander Kniel 

Flash microcontrollers are easy to program, which makes them suitable for rapid software development 

environments and educational uses. In the past, program code was usually downloaded via a serial 

interface, but nowadays many PCs (especially laptops) only have USB ports. Our versatile Flash Board 

provides a solution to this problem. It is built around an AT89C5131A, which is an extended 8051-family 

microcontroller with an 80C52 core and a Full Speed USB port. As a sort of bonus, the IC has a complete 

update interface for downloading new firmware. Atmel also provides suitable software in the form of its 

FLIP program, which is available free of charge. 

The Flash Microcontroller Board originally 

published in December 2001 

is well known to Elektor readers, 

and it has helped many readers get 

started in the world of microcontrollers. 

That’s hardly surprising, since 

microcontrollers with flash memory, 

such as the AT89C8252 used in the 

original Flash Board, are easy to program. 

As with many other similar 

boards used for educational purposes, 

the code is downloaded from a 

development PC to the microcontroller 

via a serial interface. Unfortunately, 

the good old RS232 interface is 

becoming increasingly rare. Laptops 

in particular often have only USB 

ports and no longer come with printer 

ports or serial ports. If a teacher 

wants to give his students training 

boards that they can also program at 

home using a laptop, a different approach 

is necessary. 

The author, an electronics instructor 

at a vocational/technical school in 

Heilbronn (Germany), thus developed 

a version of the Flash Board based on 

a modern microcontroller with a USB 

interface. For this purpose he selected 

the Atmel AT89C5131AM, which has 

an 80C52 core and thus belongs to the 

8051 family, just like the AT89C8252. 

The IC incorporates an Full Speed USB 


port, and it is specifically designed for 

use in USB devices such as printers, 

cameras, and so on. As a sort of bonus, 

the microcontroller has a complete 

update interface for downloading new 

firmware. This in particular enabled 

the author, who has a weakness for 

hardware and all sorts of programming 

languages, to build an extremely simple 

USB Flash Board because Atmel also 

provides suitable software in the form 

of its FLIP program, which is available 

free of charge. All you have to do is provide 

the program code in a hex file and 

you’re ready to go. 

Generation-2 Flash Board 

Many copies of the first version of the 

new board developed by Alexander 

Kniel have already been built by students 

and used with laptop computers. 

The board design was modified 

slightly in the Elektor lab, and among 

other things Elektor designer Chris 

Vossen added an LCD interface. The 

board thus follows in the footsteps of 

the 2001 version of the Elektor Flash 

Microcontroller Board and is suitable 

for not only learning how to program 

microcontrollers, but also for mature 

applications in device controllers, robots, 

and many other areas. 

Everything revolves around the abo- 

ve-mentioned AT89C5131AM, which 

is an extended member of the 8051 family. 

Its core is an 80C52X2 with six 

clocks per instruction cycle. Besides 

32 KB of flash memory, the IC has 

1024 bytes of extended RAM, additional 

EEPROM, and many other useful 

peripherals. Another helpful feature is 

that the M version of the microcontroller 

can also operate at 5 V, and a version 

in the user-friendly PLCC52 package 

is available. However, probably the 

most important feature is the USB 1.1 

/ USB 2.0 Full Speed module (for the 

experts: with endpoint 0 for control 

transfers and six additional endpoints 

with up to 512 bytes of FIFO memory). 

If you want to develop USB software, 

this gives you everything you 

could wish for, although you do need 

a bit of technical expertise. Everyone 

else can regard the microcontroller as 

a normal 8051 device that can be programmed 

via USB. 

The schematic diagram (Figure 1) 

shows a dual power supply that can 

draw power either from the USB bus 

or (with jumper JP4 fitted) via voltage 

regulator IC2 from an AC adapter connector 

K9. The D+ and D– pins of the 

Figure 1. Schematic diagram of the USB Flash Board. 

USB Flash Board • 89 

22 elektor - 11/2007

development 


11/2007 - elektor 

100R 

K2 

5 

6 

R5 

C4 

10n 

K9 

S1 

1 

2 

3 

4 

+5V 

C3 

2n2 

+5V 

+5V 

4k7 

R6 

4k7 

R11 

R4 

1k5 

R2 

27R 

R3 

27R 

K4 

1 2 

3 4 

5 6 

1 2 P1.0 

P1.1 3 4 P1.2 

P1.3 5 6 P1.4 

P1.5 7 8 P1.6 

P1.7 9 10 

C8 

10u 

16V 

K6 

JP2 

2 

JP3 

4k7 

R10 

1k 

24 

VREF 

23 

D+ 

D- 

22 

21 

C1 

100n 

PLLF 

26 

EA 

34 

NC 

NC 

46 

7 

P4.0 

P4.1 

8 

27 

ALE 

47 

P1.0 

48 

P1.1 

49 

P1.2 

50 

P1.3 

51 

P1.4 

4 

P1.5 

5 

P1.6 

6 

P1.7 

43 

RESET 

28 

PSEN 

18 

UCAP 

C9 

1u 

16V 

D9 

3 

1 

D10 

IC2 

7805 

1 3 

BAT46 

JP4 

2 

1N4001 

C12 

C13 C14 

1000u 

16V 

100u 

16V 

100n 

R7 

P1.0 

P1.1 

P1.2 

P1.3 

P1.4 

P1.5 

P1.6 

P1.7 

25 

UCAP 

19 

AVSS 

AVDD 17 

12 

X1 

VDD 16 

IC1 

AT89C5131 

C10 

D11 

6V2 

22p 

+5V 

+5V 

X1 

24MHz 

13 

X2 

C11 

22p 

D12 

C15 C16 

100n 100n 

C2 

100n 

41 

52 P0.0 

P0.0 

45 P0.1 

P0.1 

44 P0.2 

P0.2 

42 P0.3 

P0.3 

40 P0.4 

P0.4 

38 P0.5 

P0.5 

37 P0.6 

P0.6 

36 P0.7 

P0.7 

1 P2.0 

P2.0 

2 P2.1 

P2.1 

3 P2.2 

P2.2 

9 P2.3 

P2.3 

10 P2.4 

P2.4 

11 P2.5 

P2.5 

14 P2.6 

P2.6 

15 P2.7 

P2.7 

20 P3.0 

P3.0 

29 P3.1 

P3.1 

30 P3.2 

P3.2 

31 P3.3 

P3.3 

32 P3.4 

P3.4 

33 P3.5 

P3.5 

35 P3.6 

P3.6 

39 P3.7 

P3.7 

VSS 

1k5 

R12 

+5V 

P0.6 

P0.4 

P0.2 

P0.0 

P2.3 14 

P2.1 12 

10 

8 

P2.4 6 

P2.7 4 

2 

K3 

1 2 

3 4 

5 6 

7 8 

9 10 

K7 

K8 

+5V 

P0.7 

P0.5 

P0.3 

P0.1 

13 P2.2 

11 P2.0 

9 

7 

5 

3 

1 

1 2 P3.0 

P3.1 3 4 P3.2 

P3.3 5 6 P3.4 

P3.5 7 8 P3.6 

P3.7 9 10 

P3.2 

P3.3 

P3.0 

P3.1 

P0.0 

D1 

P0.1 

C6 C7 

100n 100n 

1 

2 

3 

4 

5 

6 

7 

8 

9 

D2 

P0.2 

C5 

100n 

+5V 

D3 

P0.3 

+5V 

P1 

10k 

9 

8 

D4 

S3 S4 S5 S6 

+5V 

P0.4 

JP1 

7 

6 

D5 

1 

P0.5 

5 

4 

D6 

R1 

1k5 

P0.6 

3 

2 

1 2 3 

D7 

4 

P0.7 

D8 

R9 

10k 

P3.4 

P3.5 

P3.6 

P3.7 

S2 8 7 6 5 

070125 - 11 


23


K9 

14 

13 

K7 

K4 

2 

1 

R6 

R7 

070125-1 

(C) Elektor 

IC2 

C11 

C10 

X1 

C2 

C9 

C4 

1 2 3 

C14 

C3 

C12 

D10 

microcontroller are for the USB data 

lines. To activate the internal USB 

boot loader, a low signal level must 

be applied to PSEN via JP2 (jumper 

toward the edge of the board). 

When reset switch S1 is pressed, the 

boot loader starts up and receives 

data via the USB port. Jumper posi- 

6 

5 

+ 

C1 

R5 

C5 

JP3 

+ 

+ 

C13 

R12 

D9 

IC1 

D11 

R10 

R2 

D12 

JP4 

Figure 2. Assembling the circuit board should not present any problems. 


list 

Resistors 

R1 = 1k�5 8-way SIL array 

R2,R3 = 27� 

R4,R12 = 1k�5 

R5 = 100� 

R6,R7,R11 = 4k�7 

R9 = 8-way 10k� array 

R10 = 1k� 

P1 = 10k� potentiometer 

P1 

Capacitors 

C10,C11=22pF 

C3 = 2nF2 

C4 = 10nF 

C1,C2,C5,C6,C7,C14,C15,C16 = 100nF 

C8 = 10μF 16V 

C9 = 1μF 16V 

C12 = 1000μF 16V 

C13 = 100μF 25V 


2 

1 

10 

9 

JP1 

JP2 

1 

2 

10 

9 

K6 D1 D2 

K3 

D3 D4 D5 D6 D7 D8 

R11 

R9 

S3 S1 

C15 

R4 

R3 

S5 S4 

S6 

ON 

1 234 

K2 

tion JP3 must be closed (jumper toward 

IC2) to activate the USB port. 

This connects pullup resistor R4 to 

the D+ line, which indicates a Full 

Speed USB device to the PC. If you 

would like to have a more convenient 

way to switch between run mode 

and download mode, you can connect 

a changeover switch to JP2 and JP3. 

It can be fitted directly on the PCB or 

mounted on a front panel. 

Four full-fledged 8-bit ports 

The microcontroller has four fullfledged 

8-bit ports, each of which is 

accessible via a connector and/or assigned 

a specific peripheral function. 

Port P0 is available on K3, and it also 

drives eight LEDs that can be connected 

to VDD (+5 V) via series resistors. 

Port P1 is freely usable and 

accessible via K6. Port P2 is wired to 

LCD connector K7. An LCD module 

can be operated in 4-bit mode via this 

connector, and a contrast adjustment 

trimpot is provided on the board. Finally, 

port P3 is specifically intended 

to be used for inputs, and it can be 

accessed externally via K8. For testing 

user-developed programs, the board 

is fitted with pullup resistors, four 

pushbutton switches (P3.0…P3.3), 

and four DIP switches (P3.4…P3.7) on 

port 3. Switches normally require debouncing, 

which can usually be implemented 

in software. The P3.2 and P3.3 

lines have supplementary hardware 

debouncing in the form of capacitors 

C6 and C7, since these lines are connected 

to the interrupt inputs of the 

microcontroller. We also mustn’t forget 

port P4 with the P4.0 and P4.1 lines, 

which form the I2C bus interface and 

are accessible via K4. 

The bare PCB for the USB Flash Board 

(Figure 2) is available from the Elektor 

Shop under order number 070125- 

1. Alternatively, you can purchase a 

complete kit with all the components 

under order number 070125-71. Assembling 

the board is not difficult. 

Be sure to avoid creating any shorts 

between D+, D– or the 5 V supply 

line and ground in the area around 

the USB socket. As there is no special 

protection for the D+, D- and 5-V 

supply lines, it’s a good idea to check 

this with an ohmmeter – but be sure 

to remove the microcontroller from its 

socket first. There is room for an extra 

100-nF ceramic capacitor beneath 

the IC socket, which should be fitted 

first. It provides optimal supply voltage 

decoupling. 

Initial operation 

You should use an AC adapter (8– 

12 V DC) for initial testing. Fit jumper 

JP4 to select this power source. LED 

D1 should light up now. If you have 

already connected an LCD module, it 

24 elektor - 11/2007 

1 

2 

R1 

+ 

S2 

C7 C6 

C8 

2 

1 

K8 

10 

9 


D1-D8,D12 = LED, red, low-current 

D9 = BAT46 

D10 = 1N4001 

D11 = zener diode 6V2 

IC1 = AT89C5131AM 

IC2 = 7805 


Miscellaneous 

JP1,JP4 = 2-way SIL pinheader 

JP2,JP3 = 3-way SIL pinheader 

K2 = USB-A socket 

K3,K6,K8 = 10-way boxheader 

K4 = 6-way (2x3) pinheader 

K7 = 14-way boxheader 

K9 = mains appliance socket, PCB mount 

S1,S3-S6 = miniature pushbutton 

PLCC socket 

PCB, # 070125-1 from Elektor SHOP 

Kit of parts, # 070125-71 from Elektor 

SHOP 

USB Flash Board • 91

should display dark pixels in the top 

line. If necessary, adjust the contrast 

trimpot until both lines are clearly distinguishable. 

The upper line will remain 

dark until the board has been initialised 

with a program. If you have an 

oscilloscope, you can also check the 12- 

MHz clock signal on the crystal. This 

clearly shows that the microcontroller 

is running. 

You have to download a program for 

the first real software test. For this 

purpose, you can use the Flexible In- 

System Programmer (FLIP) software, 

which you can download free of charge 

from Atmel’s home page (www.atmel. 

com). Enter ‘Flip’ as the search term 

to find FLIP 2.4.6 for Windows (4 MB, 

Version 2.4.6, updated May 2006). First 

extract the contents of archive file flip- 

2_4_6.zip to a separate folder and then 

run the Setup.exe file in that folder. Follow 

the installation instructions and 

accept the licence conditions and suggested 

installation location. You will 

then see a short list of instructions for 

what you have to do next (Figure 3). 

The program is installed by default in 

C:\Program Files\Atmel\FLIP 2.4.6\. 

Now connect a cable to the USB connector 

and fit jumper JP2 in the ‘USB’ 

position (toward the edge of the 

board). To be on the safe side, press 

reset switch S1 and close JP3. This 

starts the USB download firmware, 

which waits for contact with the PC to 

be established. The program reports 

vendor ID 03EB and product ID 2FFD, 

which enable Windows to assign a suitable 

driver. Windows will recognise 

a new device and ask you to select a 

suitable driver. Select the driver located 

in folder C:\Program Files\Atmel\ 

FLIP 2.4.6\usb (see Figure 4). After it 

is installed, you will see the new device 

in the Device Manager window. It 

can be recognised by its name ‘Jungo 

AT89C5130/AT89C5131’. 

If something goes wrong during this 

process, you have to track down the 

problem. Possible problem sources include 

incorrectly fitted jumpers. For 

instance, if you activate the USB port 

with JP3 (pullup connected to D+) but 

do not start the internal firmware (JP2 

still in the ‘Run’ position or no reset 

executed after switching over), Windows 

will report a new device – but not 

the right one. By contrast, you might 

start the update firmware correctly but 

fit JP3 incorrectly. In this case, Windows 

will not recognise that a device 


11/2007 - elektor 

Figure 3. The free FLIP programming software displays a list of what you have to do to start using the board. 

Figure 4. The microcontroller is recognised by Windows as a new device. 

Figure 5. After you click Run, FLIP downloads the program to the flash memory of the microcontroller. 


25


Figure 6. Main menu of the BASCOM compiler. 

is connected, and thus no communication 

will be established. After a bit of 

practice, you won’t have any problems 

making the right settings, and you can 

establish a communication session 

with the PC whenever you need it. 

Program download 

Now launch FLIP. First you have to 

use F2, Device à Select, or the IC icon 

to select the correct IC (AT89C5131). 

Then use F3, Settings à Communication 

à USB, or the cable icon to select 

and open the USB interface. Finally, 

you have to use F4 or File à Load Hex 

Hardware test in Bascom-51 

‘Simple test for inputs, 

‘outputs and LCD 

‘********************** 

Dim X As Byte 

P1 = 0 

Cls 

Lcd “ 8051-Test “ 

Wait 1 

Lowerline 

Lcd “ Elektor “ 

Wait 3 

For X = 1 To 13 

Shiftlcd Right 

Waitms 200 

Next 

Cls 

Lcd “ Test Port 0 “ 

Lowerline 

Lcd “ Bit 2 exp 0 “ 


File to load a suitable hex file. Select 

program file: 5131_TEST_ELEKTOR. 

HEX, which you can obtain along 

with the BASCOM AVR source code 

from the Elektor home page. Click the 

Run button (see Figure 5) to download 

the program code to the flash memory. 

After this, you must change over 

JP2 and press the Reset button to run 

the program. Caution: the BLJB option 

is enabled automatically with a new 

microcontroller. You must deselect 

(uncheck) it the first dime you download 

a program, since otherwise it will 

not be possible to run the program after 

it has been downloaded. 

P0 = &B11111110 

Wait 1 

Lowerline 


P0 = &B11111101 

Wait 1 

Lowerline 


P0 = &B11111011 

Wait 1 

Lowerline 


P0 = &B11110111 

Wait 1 

Lowerline 


P0 = &B11101111 

Wait 1 

Lowerline 


P0 = &B11011111 

Wait 1 

Lowerline 


P0 = &B10111111 

Wait 1 

Lowerline 

If you want to download a new hex 

file after this test, you must first disconnect 

the USB cable and then reconnect 

it – and of course, with the right 

jumper settings and a Reset first. After 

this, you must establish the connection 

again in FLIP. 

Alternatively, you can leave the USB 

cable connected and simply open 

JP3, which will also isolate the device 

from the USB without interrupting the 

supply voltage. In order to download 

a new program, you must first change 

the setting of JP2 again. The press 

Reset, wait two seconds, and fit JP3 

again. This initialises the USB device. 

You will have to open the interface in 

FLIP again, after which you can start 

the download. 

Programming with BASCOM 

The BASCOM-51 Basic compiler is an 

ideal tool when you are just getting 

started with developing programs 

for the system, although you can also 

write programs for the microcontroller 

in C or assembly language. You can 

download a free demo version of BAS- 

COM-51 from the site of its producer, 

MCS Electronics (www.mcselec.com). 

The free version can generate up to 

4 KB of code, which is sufficient for 

many applications. 

Figure 6 shows the main menu of the 

compiler. In order to ensure correct 

operation of the board, you must as- 


P0 = &B01111111 

Wait 1 

Lowerline 

Lcd “ All Bits “ 

P0 = &B00000000 

Wait 1 

Cls 

Lcd “ Test Port 3 “ 

Lowerline 

Lcd “ Test Port 0 (LED) “ 

Wait 3 

Status: 

P0 = P3 

X = P0 

Cls 

Lcd “ Inputs “ 

Lowerline 

Lcd “Port 3 = “ ; X ; “ “ 

Waitms 60 

Goto Status 

26 elektor - 11/2007 

End 

USB Flash Board • 93

sign the LCD pins to port P2 under 

Options (Figure 7). 

BASCOM supports configuration of 

different register files for individual 

8051 derivatives. Although there are 

no specific settings for the AT89C5131, 

this microcontroller is largely compatible 

with the 8052, so you should use 

register file 8052.dat. 

The listing shows the source code of 

the test program. It is easy to read and 

largely self-explanatory. After an introductory 

message is displayed on the 

LCD, a running-light routine is executed 

to check all the LEDs on Port P0. 

Following this, the inputs on port P3 

are read in an endless loop and their 

states are copied to output port P0 


11/2007 - elektor 

Figure 7. The assignment of the LCD pins to port P2 must be configured under Options. 

and 

shown on 

the LCD. 

You can actuate 

the 

DIP switches 

(S2) and 

pushbuttons 

S3–S6 to check 

that they are properly 

assigned to the 

port pins. The associated 

output 

LED will light 

up for each 

switch. 

The test 

program 

thus 

exercises 

practically all 

of the hardware. 

A couple of ideas 

Finally, a couple of ideas for further 

projects. The microcontroller has an 

internal EEPROM, similar to what is 

found in the 89S8252 and the 89S8253. 

However, in this case it is governed by 

different control registers (SFRs). This 

means that you cannot escape a careful 

study of the data sheet if you want 

to use the supplementary hardware. 

Like the 8052, the AT89C5131 has another 

serial interface that can be used 

with BASCOM by instructions such 

as Print and Input. However, this requires 

connecting an additional line 

driver (such as a MAX232), since the 

USB Flash Board does not have a serial 

interface port. This opens the door to 

typical interface applications, which 

means that you can use the microcontroller 

as a PC-based measuring instrument, 

counter or motor controller. 

Of course, the AT89C5131 can also do 

a lot more, including implementing a 

complete USB device. This is described 

in several application notes and 

accompanying source code on the Atmel 

website. The archive file c5131usb-kbd-stand-alone-1_0_2.zipdemonstrates 

how to construct a USB 

keyboard. 

With this USB microcontroller and the 

extensive software archive, you have 

essentially everything you need to develop 

your own USB applications. 

(070999-1) 


27

5 

Elektor blockbusters Artikeltitel • 95

Blockbusters - Allt om Elektronik

Create successful ePaper yourself

Delete template?

Save as template?