
Page 10
Programming Details
Chapter 3
Page 10
01270193.doc
Blue Chip Technology Ltd.
The input gain for the chosen channel when selecting manually, or for all
channels, when selecting automatically, may be set to 1, 10, 100 or 1000. At
higher gains or when extreme accuracy is required, it is recommended that
auto-calibration be performed (see “Auto-Calibration” below) to remove
inherent offset or gain errors.
Analogue Input Select Register
(Base 2 + 0D
h
) Read/write
Function
Bit no.
b7..b4
Channel number (Manual selection) -OR-
Number of channels to be scanned
(Automatic select & increment)
0 - 15 for single ended inputs
0 - 7 for differential inputs)
00 = 1
01 = 10
10 = 100
11 = 1000
00 = Single ended inputs
01 = Differential inputs
10 = Calibration zero volts input
11 = Calibration + 80% full scale input
b3..b2
Input gain:-
b1..b0
Input mode:-
The accuracy that may be achieved with and without auto-calibration, is listed
in the Specifications section.
Auto-Calibration
If auto-calibration is required, select the gain that will be used and zero input,
and then take ten samples over a period of between 1 millisecond and 1 second.
Take the mean of these samples. Then select a gain of 1 and positive
calibration input (this is actually 80% of full scale to avoid limiting) and take
ten more samples over a similar period. Take the mean of these samples. The
calibrated outputs may then be calculated with the following equation:-
Calibrated sample = (raw sample - mean zero sample) * 0.8
mean full scale sample
The reason for taking ten samples and using the mean of these in the
calculation is to minimise the effects of noise on the measurements, as any noise
on a single calibration sample will otherwise effect all future measurements.