handbook of modern sensors

5.4 Analog-to-Digital Converters 183
discharging its capacitor through S 2 . The charge at the capacitor gained during the
input signal integrate phase is precisely equal to the charge lost during the reference
deintegration phase. Therefore, the following holds:
V in
T t
= V ref , (5.33)
R in C in R in C in
which leads to
V in
= t
V ref T . (5.34)
Therefore, the ratio of the average input voltage and the reference voltage is replaced
by the ratio of two time intervals. Then, the counter does the next step—it converts
the time interval t into a digital form by counting the clock pulses during t. The
total count is the measure of V in (remember that V ref and T are constants).
The dual-slope converter has the same advantage as the charge-balance converter:
They both reject frequencies 1/T corresponding to the integrate timing. It should be
noted that selecting time T = 200 ms will reject noise produced by both 50 and 60 Hz,
thus making the converter immune to the power line noise originated at either standard
frequency. Further, the conversion accuracy is independent of the clock frequency
stability, because the same clock sets timing T and the counter. The resolution of the
conversion is limited only by the analog resolution; hence, the excellent fine structure
of the signal may be represented by more bits than would be needed to maintain a given
level of scale-factor accuracy. The integration provides rejection of high-frequency
noise and averages all signal instabilities during the interval T . The throughput of a
dual-slope conversion is limited to somewhat less than 1/2T conversions per second.
To minimize errors produced in the analog portion of the circuit (the integrator and
the comparator), a third timing phase is usually introduced. It is called an auto-zero
phase because during that phase, the capacitor C in is charged with zero-drift errors,
which are then introduced in the opposite sense during the integration, in order to
nullify them. An alternative way to reduce the static error is to store the auto-zero
counts and then digitally subtract them.
Dual-slope converters are often implemented as a combination of analog components
(OPAMs, switches, resistors, and capacitors) and a microcontroller, which
handles the functions of timing, control logic, and counting. Sometimes, the analog
portion is packaged in a separate integrated circuit. An example is the TS500 module
from Texas Instruments.
5.4.4 Successive-Approximation Converter
These converters are widely used in a monolithic form thanks to their high speed (to 1
MHz throughput rates) and high resolution (to 16 bits). The conversion time is fixed
and independent of the input signal. Each conversion is unique, as the internal logic
and registers are cleared after each conversion, thus making these A/D converters
suitable for the multichannel multiplexing. The converter (Fig. 5.29A) consists of
a precision voltage comparator, a module comprising shifter registers and a control