handbook of modern sensors

5.9 Noise in Sensors and Circuits 213
capacitance can be modeled by circuit shown in Fig. 5.48. Here, e n is a noise source.
It may be some kind of a part or component whose electric potential varies. C s is the
stray capacitance (having impedance Z s at a particular frequency) between the noise
source and the circuit impedance Z, which acts as a receiver of the noise. The voltage
V n is a result of the capacitive coupling. A noise current is defined as
and actually produces noise voltage
i n =
V n =
V n
Z + Z s
(5.83)
e n
1 + Z c /Z . (5.84)
For example, if C s = 2.5 pF,Z = 10 k (resistor), and e n = 100 mV, at 1.3MHz the
output noise will be 20 mV.
One might think that 1.3-MHz noise is relatively easy to filter out from lowfrequency
signals produced by a sensor. In reality, it cannot be done, because many
sensors and, especially the front stages of the amplifiers, contain nonlinear components
(p-n-semiconductor junctions) which act as rectifiers. As a result, the spectrum
of high-frequency noise shifts into a low-frequency region, making the noise signal
similar to the voltage produced by a sensor.
When a shield is added, the change to the situation is shown in Fig. 5.48B. With
the assumption that the shield has zero impedance, the noise current at the left side will
be i n1 = e n /Z C1 . On the other side of the shield, noise current will be essentially zero
because there is no driving source on the right side of the circuit. Subsequently, the
noise voltage over the receiving impedance will also be zero and the sensitive circuit
becomes effectively shielded from the noise source. One must be careful, however,
that there is no currents is flowing over the shield. Coupled with the shield resistance,
these may generate additional noise. There are several practical rules that must be
observed when applying electrostatic shields:
• An electrostatic shield, to be effective, should be connected to the reference potential
of any circuitry contained within the shield. If the signal is connected to
a ground (chassis of the frame or to earth), the shield must be connected to that
ground. Grounding of shield is useless if the signal is not returned to the ground.
• If a shielding cable is used, its shield must be connected to the signal referenced
node at the signal source side (Fig. 5.49A).
• If the shield is split into sections, as might occur if connectors are used, the shield
for each segment must be tied to those for the adjoining segments and ultimately
connected only to the signal referenced node (Fig. 5.49B).
• The number of separate shields required in a data acquisition system is equal to the
number of independent signals that are being measured. Each signal should have
its own shield, with no connection to other shields in the system, unless they share
a common reference potential (signal “ground”). In that case, all connections must
be made by a separate jumping wire connected to each shield at a single point.
• A shield must be grounded only at one point—preferably next to the sensor. A
shielded cable must never be grounded at both ends (Fig. 5.50). The potential