I'm using an AD7124-8 in a temperature data logger. I have an issue of large offsets being caused by switching between reference sources.
My channel sequence is as follows:
Enable Iout0, 100uA on ain6, and Iout1, 100uA on ain7
Channel 1 : gain 32, internal reference, ain0+, ain1- (TC)
Channel 2 : gain 32, internal reference, ain2+, ain1- (TC)
Channel 3 : gain 2, refin1, ain12+, ain11- (RTD using excitation from ain6 and ain7)
Enable Iout0, 100uA on ain8, and Iout1, 100uA on ain10
Channel 4 : gain 32, internal reference, ain3+, ain4- (TC)
Channel 5 : gain 32, internal reference, ain5+, ain4- (TC)
Channel 6 : gain 2, refin1, ain14+, ain15- (RTD using excitation from ain8 and ain10)
Loop to start
All inputs are buffered, and sampling rate is 16.7Hz. After each sample, I disable the current channel and enable the next channel in the sequence (the auto-sequencer is of no use when switching excitation sources between channels). I've followed the guidelines in CN-0383 for the RTD inputs, and CN-0384 for the thermocouple inputs.
If RTDs are connected to both channel 3 and channel 6, everything works ok. But if one of the RTDs is disconnected, the subsequent thermocouple channels have large offsets. I'm guessing it must be the switching between an out of range reference on refin1 and the internal reference that is causing this. It takes about 15 samples before the offsets have reduced to zero. If I enable reference buffering, the problem is even worse.
Please suggest how I can eliminate this issue. Using the internal reference for the RTD channels leads to large inaccuracies so this is not an option, and waiting 15 samples is also not an option.
Thanks in advance!