Campbell CR9000, Instruction Manual

Page: 61 / 218

Section 3. CR9000 Measurement Details
3-17
( ) V V
R
R R R R R
x
s
s f s f G
1
=
+ +
/
[3.6-1]
R
s
R
f
/R
G
is the source of error due to the ground loop. When R
G
is large the
equation reduces to the ideal. The geometry of the electrodes has a great effect
on the magnitude of this error. The Delmhorst gypsum block used in the 227
probe has two concentric cylindrical electrodes. The center electrode is used
for excitation; because it is encircled by the ground electrode, the path for a
ground loop through the soil is greatly reduced. Moisture blocks which consist
of two parallel plate electrodes are particularly susceptible to ground loop
problems. Similar considerations apply to the geometry of the electrodes in
water conductivity sensors.
The ground electrode of the conductivity or soil moisture probe and the
CR9000 earth ground form a galvanic cell, with the water/soil solution acting
as the electrolyte. If current was allowed to flow, the resulting oxidation or
reduction would soon damage the electrode, just as if DC excitation was used
to make the measurement. Campbell Scientific probes are built with series
capacitors in the leads to block this DC current. In addition to preventing
sensor deterioration, the capacitors block any DC component from affecting
the measurement.
3.2 CR9058E Isolation Module Measurements
Each CR9058E input channel has its own 24 bit sigma delta analog to digital
converter taking 10,000 measurements per second, or one measurement sample
per 100 microseconds. The effective resolution at this sample rate is 18.7 bits,
or +/- 9.4 microvolts when using the +/- 2 Volt range, because of the inherent
noise of the A/D converter and noise from other sources. The effective
resolution can be dramatically improved through filtering, and/or integrating,
multiple measurements. Thus, noise reduction and measurement speed can be
traded off using the Integration parameter. Noise is reduced by approximately
the square root of the number of samples within the integration time. Thus, if
the integration time is set to 10000 versus 100 microseconds, noise should be
reduced approximately by a factor of ten . This approximation assumes that
the noise is white noise, which is not
entirely true because some of the noise is
due to interference from sources at fixed frequencies. Noise reduction by
filtering can go just so far, and the best the CR9058E can achieve is
approximately 21 bits of resolution (+/- 1.9 microvolts on the 2 Volt range).
The CR9058E isolated input module is similar in operation to the CR9050
analog input module except for:
•
The CR9058E has ten differential input channels instead of 16 differential
/ 32 single-ended inputs.
•
The CR9058E has different voltage ranges of +/- 60 Volts DC, +/- 20
Volts DC, and +/- 2 Volts DC.
•
The CR9058E has a slower maximum scan rate than the CR9050, but this
is somewhat balanced by the fact that the CR9058E measures all of its
channels simultaneously, as each channel has its own 24 bit sigma delta
analog to digital converter. Conversely, the measurements from the