Campbell Scientific CR7 Measurement and Control System - Manual (Page 125)

13-1
SECTION 13. CR7 MEASUREMENTS
13.1 FAST AND SLOW MEASUREMENT
SEQUENCE
The CR7 makes voltage measurements by
integrating the input signal for a fixed time and
then holding the integrated value for the analog
to digital (A/D) conversion. The A/D conversion
is made with a 16 bit successive approximation
technique which resolves the signal voltage to
approximately one part in 30,000 of either the +
or - side of the full scale range (e.g., 1/30,000 x
5V = 166µV).
Integrating the signal removes noise that could
create an error if the signal were
instantaneously sampled and held for the A/D
conversion. The slow integration time provides
a more noise-free reading than the fast
integration time. One of the most common
sources of noise is 60 Hz from AC power lines.
The slow integration time of 16.67 milliseconds
is equal to one 60 Hz cycle so that during the
integration time the AC noise would integrate to
0.
There are several situations where the fast
integration time of 250 microseconds is
preferred. The fast integration time minimizes
time skew between measurements and
increases the throughput rate. The current drain
on the CR7 batteries is lower when fast
integration time is used because the I/O CPU is
switched on for shorter time periods. The fast
integration time should ALWAYS be used with
the AC half bridge (Instruction 5) when
measuring AC resistance or the output of an
LVDT. An AC resistive sensor will polarize if a
DC voltage is applied, causing erroneous
readings and sensor decay. The induced
voltage in an LVDT decays with time as current
in the primary coil shifts from the inductor to the
series resistance, a long integration time would
result in most of the integration taking place
after the signal had disappeared.
FIGURE 13.1-1. Timing of Single Ended
Measurement
Before making a series of measurements
prescribed by an Input Instruction, the CR7
makes a calibration measurement. The
calibration is accomplished by measuring two
known voltages which are sent through the
same amplifier circuit that will be used for the
measurements. The calibration for a single
ended measurement consists of measuring a
voltage which is 4/5ths of full scale and then
making a measurement with the input
grounded. A differential measurement is made
once with the inputs as connected and a second
time with the inputs reversed (Section 13.2):
calibration for differential measurements uses
voltages at ±4/5ths of full scale.
An offset error of up to 1 least significant bit can
occur in a slow, single ended measurement as
a result of dielectric absorption in the integrating
capacitor. This error is a function of the
previous measurement. If the CR7 is
programmed to make a single ended
measurement on the 5 volt range with the inputs
shorted, an error of -166 µV can be observed.
13.2 SINGLE ENDED AND DIFFERENTIAL
VOLTAGE MEASUREMENTS
NOTE: The channel numbering on the
Analog Input cards refers to differential
channels. Either the high or low side of a
differential channel can be used for single
ended measurements so each side must be
counted when numbering single ended
channels, e.g., the high and low sides of
differential channel 14 are single ended
channels 27 and 28, respectively.
The timing and sequence of a single ended
measurement is shown in Figure 13.1-1. A
single ended measurement is made on a single
input which is referenced to ground. A single
integration is performed for each measurement.
A differential measurement measures the
difference in voltage between two inputs. The
measurement sequence on a differential
measurement involves two integrations: first
with the high input referenced to the low, then
with the inputs reversed. (Figure 13.2-1). The
CR7 computes the differential voltage by
averaging the magnitude of the results from the
two integrations and using the polarity from the
first.