Hi,
For a design of mine where I need to sense a DC voltage supply level, I'm planning to use the lower INL variant ADS7828EB. I'm curious about the estimated worst case absolute DC error expected from this ADC at room temperature. I'm planning to use it to sense pseudo differential inputs where the +ve input can vary between 700mV till 2V, and the -ve input will be close to GND, say ~3mV or so. The plan is to use it with 2.7V VDD and the internal 2.5V VREF.
I understand that the absolute DC error will comprise of the INL (+/-0.5LSB typical), gain error (+/-0.75LSB typical), offset error (+/-0.75LSB typical) and the quantization error (rms value of 1LSB/sqrt(12) and peak value of +/-0.5LSB). But obviously since these are mostly uncorrelated simply adding up the max values would lead to pessimistic result. Will the rms sum of these 4 items give a better representation, which comes out to be ~1.2LSB rms? Or can you please give me a more accurate computation of the expected worst case absolute DC error if available?
Also, the maximum values mentioned for INL, gain error and offset error are +/-1LSB, +/-3LSB and +/-2LSB respectively. This degradation from the typical values quoted is primarily due to the temperature? If I maintain my ambient temperature at say 25degC these parameters are bound to remain within the typical specifications listed? Are the typical specifications listed the worst case results from tests at typical conditions or are they derived based on an average of the results from tests at typical conditions which means there may be cases where these results will be exceeded?
Finally, is external VREF recommended over internal VREF for maximal accuracy? Because from the performance plots the external VREF results seem to look a tiny bit better?
Thanks,
Anoop
