Specifications

51Sensors
Complete Signal-Chain Solutions
The sensor signal chain must handle
extremely small signals in the presence
of noise. Accurately measuring
changes in the output voltage from a
resistive transducer requires circuitry
that provides the following electrical
functions: excitation, amplication,
linearization, oset nulling, ltering,
and acquisition. Some solutions may
also require the use of digital signal
processing (DSP) techniques for signal
manipulation, error compensation, gain,
ltering, and user programmability.
Discrete vs. Integrated Solutions
In this chapter we discuss functional
blocks, keeping in mind that Maxim
oers more highly integrated solutions
when the application warrants their
use. Some examples are given.
Excitation
Accurate and stable voltage or current
sources with low-temperature drift are
required for sensor excitation. To easily
eliminate eects of reference voltage
tolerance, it is common practice to
use the same reference for both the
sensor excitation and the analog-to-
digital converter (ADC). This makes
the signals ratiometric, eliminating
rst-order tolerances allowing the
use of less accurate references,
or alternately providing higher
performance from a given reference.
Amplification and Level
Translation—the Analog
Front-End (AFE)
In some designs the transducer’s output
voltage range will be very small, with
the required resolution reaching the
nanovolt range. In such cases, the
transducer’s output signal must be
amplied before it is applied to the ADC’s
inputs. To prevent this amplication
step from introducing errors, low-noise
ampliers (LNAs) with extremely low
oset voltage (V
OS
) and low temperature
and oset drifts must be selected. A
drawback of Wheatstone bridges is that
the common-mode voltage is much
larger than the signal of interest. This
means that the LNAs must also have
excellent common-mode rejection ratios
(CMRR), generally greater than 100dB.
When single-ended ADCs are used,
additional circuitry is required to remove
large common-mode voltages before
acquisition. Additionally, since the signal
bandwidth is low, the 1/f noise of the
ampliers can introduce errors. Chopper-
stabilized ampliers are, therefore,
often used. Some of these stringent
amplier requirements can be avoided
by using a small portion of the full-scale
range of a very high-resolution ADC.
Acquisition—the ADC
When choosing the ADC, look at
specications like noise-free range or
eective resolution that indicate how
well an ADC can distinguish a xed
input level. Alternate phrasing for
these applications might be noise-free
counts or codes inside the range. Most
high-accuracy ADC data sheets show
these specications as a table of peak-
to-peak noise or RMS noise vs. speed;
sometimes the specications are shown
graphically as noise histogram plots.
Other ADC considerations include low
oset error, low temperature drift, and good
linearity. For certain low-power applications,
speed vs. power is an important criterion.
Filtering
The bandwidth of the transducer signal is
generally small and the sensitivity to noise
is high. It is, therefore, useful to limit the
signal bandwidth by ltering to reduce the
total noise. Using a sigma-delta ADC can
simplify the noise-ltering requirement
because of the inherent shaping of the
noise spectrum out of the band of interest
by the oversampling in that architecture.
Digital Signal Processing
(DSP)—the Digital Domain
Besides the analog signal processing,
the captured signals are further
processed in the digital domain for
signal extraction and noise reduction. It
is common to nd focused algorithms
that cater to particular applications and
their nuances. There are also generic
techniques, such as oset and gain
correction, linearization, digital ltering,
and temperature (and other dependent
factors)-based compensation that are
usually applied in the digital domain.
The DSP function necessitates sucient
processing capability in the signal path.
Integrated Solution
In more highly integrated solutions, all
required functional blocks are integrated
into a single IC commonly called a
sensor signal conditioner. A sensor signal
conditioner is an application-specic
IC (ASIC) that performs compensation,
amplication, and calibration of the
input signal, normally over a range
of temperatures. Depending on the
sophistication of the signal conditioner,
the ASIC integrates some or all of
the following blocks: sensor, sensor
excitation circuitry, digital-to-analog
converter (DAC), programmable
gain amplier (PGA), ADC, memory,
multiplexer (mux), CPU, temperature
sensor, and digital interface.
There are two types of sensor signal
conditioners: analog signal-path
conditioners and digital signal-path
conditioners. Analog signal conditioners
have a faster response time and
provide a continuous-output signal,
immediately reecting changes at the
input. They generally have a hardwired
(inexible) compensation scheme.
Digital conditioners, which are usually
microcontroller-based, have slower
response times because of latencies
introduced by the ADC and DSP routines,
and they introduce quantization error
in the output signal. The magnitude of
the quantization error depends on the
resolution of the ADC used and on the
resolution of data processed within the
microprocessor. The main benets of
digital signal conditioners are the exibility
of the compensation algorithms that can
be adapted to the user’s application, and
the ease with which the output can be
interfaced to an external microcontroller.
Maxim oers both fully analog path
and digital signal-path conditioners.
www.maxim-ic.com/psi