D20/D200 Technical Overview
GE Grid Solutions
50
PRPI-019-3.00-10
GE Information
Analog Inputs
The D20A analog input module is capable of handling 32 differential analog inputs. The
conversion process is performed using a +/- 5 V full scale voltage only, with a software
controlled programmable gain amplifier providing full-scale input voltage ranges of +/- 1.0 V,
+/- 5.0 V and +/- 10 V on any input. Any required current input is accommodated using precision
shunt-resistor analog-scaling networks on each input. Adaptors at the input of each analog point
convert current and unusual voltage inputs to the required 1, 5 or 10 V full-scale input. Refer to
the Index for a list of available analog input adapters.
Conversion
After the input has been appropriately scaled, it is fed into a voltage-to-frequency converter. To
reject line noise (NMR), the output frequency is counted over a 50 or 60 Hz period (user
configurable). This value is compared to the microprocessor count over the same period to obtain
a raw analog count. It is this raw analog value that is converted to a 14-bit (plus sign) number
(the digitized raw analog value) representing the input voltage. The microprocessor can convert
all of the 32 inputs within a scan period of 550 ms (60 Hz) to 656 ms (50 Hz.)
Voltage References
The D20A uses a high-precision voltage reference chip to provide seven on-board references for
auto-calibration. After a specified number of scans (user configurable), the microprocessor
converts each of the references. The new calculated reference values are then stored over the
previously converted values in the reference data table. If the new values are within a certain
range (user configurable) of the old values, they are kept. If they are out of this range, an error is
reported.
Auto Zero
Eight internal references, which include a temperature reference, provide auto zero and auto
calibration functions. This results in an accuracy and temperature stability of the module equal to
the internal precision voltage reference.
The calculation used to obtain the digitized raw analog value (during the conversion process)
would normally be based on a linear curve like Line A shown in Figure 33. Due to temperature,
gain and offset inaccuracies, the curve may be slightly non-linear (Line B, exaggerated for
demonstration purposes). The D20 minimizes this error by a technique called three-point
linearity correction. By separately calculating the linearity difference (slope
m
) of the 0 to +5 V
and 0 to -5 V ranges, the software can use a different equation for either positive or negative
values.
