Chapter 3
Hardware Overview
3-12
ni.com
When scanning among channels at various gains, the settling times may
increase. When the PGIA switches to a higher gain, the signal on the
previous channel may be well outside the new, smaller range. For instance,
suppose a 4 V signal is connected to channel 0 and a 1 mV signal is
connected to channel 1, and suppose the PGIA is programmed to apply
a gain of one to channel 0 and a gain of 100 to channel 1. When the
multiplexer switches to channel 1 and the PGIA switches to a gain of 100,
the new full-scale range is 100 mV (if the ADC is in unipolar mode).
The approximately 4 V step from 4 V to 1 mV is 4,000% of the new
full-scale range. For a 12-bit device to settle within 0.012% (120 ppm or
1/2 LSB) of the 100 mV full-scale range on channel 1, the input circuitry
has to settle to within 0.0003% (3 ppm or 1/80 LSB) of the 4 V step. It may
take as long as 100 µs for the circuitry to settle this much. For a 16-bit
device to settle within 0.0015% (15 ppm or 1 LSB) of the 100 mV full-scale
range on channel 1, the input circuitry has to settle within 0.00004%
(0.4 ppm or 1/400 LSB) of the 4 V step. It may take as long as 200 µs for
the circuitry to settle this much. In general, this extra settling time is not
needed when the PGIA is switching to a lower gain.
Settling times can also increase when scanning high-impedance signals due
to a phenomenon called
charge injection
, where the AI multiplexer injects
a small amount of charge into each signal source when that source is
selected. If the impedance of the source is not low enough, the effect of the
charge—a voltage error—does not have decayed by the time the ADC
samples the signal. For this reason, you should keep source impedances
under 1 k
Ω
to perform high-speed scanning.
Due to the previously described limitations of settling times resulting from
these conditions, multiple-channel scanning is not recommended unless
sampling rates are low enough or it is necessary to sample several signals
as nearly simultaneously as possible. The data is much more accurate and
channel-to-channel independent if you acquire data from each channel
independently (for example, 100 points from channel 0, then 100 points
from channel 1, then 100 points from channel 2, and so on).