Campbell CR23X, Руководство оператора

Страница: 210 / 374

SECTION 13. CR23X MEASUREMENTS
13-10
TABLE 13.3-6. Maximum Lead Length vs. Error for Campbell Scientific Resistive Sensors
Sensor Maximum
Model # Error Range V
e
(µV) Length(ft.)
107 0.05
°
C 0
°
C to 40
°
C 5 1000
1
034A 3° @ 360
°
2083 380
2
227 - - - 2000
3
237 10 kohm 20k to 300k 1000 2000
3
1
based on transient settling
2
based on signal rise time
3
limit of excitation drive
The comparatively small transient yet large
source resistance of the 034A sensor indicates
that signal rise time may be the most important
limitation. The analysis in Section 13.3.2
confirms this.
The Model 227 Soil Moisture Block has a
relatively short time constant and essentially no
transient. Lead lengths in excess of 2000 feet
produce less than a 0.1 bar (0-10 bar range)
input settling error. With this sensor, the drive
capability of the excitation channel limits the
lead length. If the capacitive load 0.1 µfd and
the resistive load is negligible, V
x
will oscillate
about its control point. If the capacitive load is
0.1 or less, V
x
will settle to within 0.1% of its
correct value 150 µs. A lead length of 2000 feet
is permitted for the Model 227 before
approaching the drive limitation.
Table 13.3-6 summarizes maximum lead lengths
for corresponding error limits in six Campbell
Scientific sensors. Since the first three sensors
are nonlinear, the voltage error, V
e
, is the most
conservative value corresponding to the error
over the range shown.
MINIMIZING SETTLING ERRORS IN NON-
CAMPBELL SCIENTIFIC SENSORS
When long lead lengths are mandatory in
sensors configured by the user, the following
general practices can be used to minimize or
measure settling errors:
1. When measurement speed is not a prime
consideration, Instruction 4, Excite, Delay,
and Measure, can be used to insure ample
settling time for half bridge, single-ended
sensors.
2. An additional low value bridge resistor can be
added to decrease the source resistance, R
o
.
For example, assume a YSI nonlinear
thermistor such as the model 44032 is used
with a 30 kohm bridge resistor, R’
f
. A typical
configuration is shown in Figure 13.3-7A. The
disadvantage with this configuration is the high
source resistance shown in column 3 of Table
13.3-7. Adding another 1 K resistor, R
f
, as
shown in Figure 13.3-7B, lowers the source
resistance of the CR23X input. This offers no
improvement over configuration A because R’
f
still combines with the lead capacitance to
slow the signal response at point P. The
source resistance at point P (column 5) is
essentially the same as the input source
resistance of configuration A. Moving R
f'
out
to the thermistor as shown in Figure 13.3-7C
optimizes the signal settling time because it
becomes a function of R
f
and C
w
only.
Columns 4 and 7 list the signal voltages as a
function of temperature using a 5000 mV
excitation for configurations A and C,
respectively. Although configuration A has a
higher output signal (5000 mV input range), it
does not yield any higher resolution than
configuration C which uses the
±
1000 mV
input range.
NOTE: Since R
f
' attenuates the signal in
configuration B and C, one might consider
eliminating it altogether. However, its
inclusion "flattens" the non-linearity of the
thermistor, allowing more accurate curve
fitting over a broader temperature range.
3. Where possible, run excitation leads and
signal leads in separate shields to minimize
transients.
4. Avoid PVC-insulated conductors to
minimize the effect of dielectric absorption
on input settling time.