How to calibrate offset and gain in the DDC devices?

Hi,

This is actually a common question and the answer actually a pretty straightforward one. One would do an offset and gain calibration like one would do any other calibration. Not apply a signal (disconnect the output or at least have zero signal) and measure the output; or apply a known signal almost full-scale and measure the output. Then correct all data respect to a line going through those two points. 

The usual question comes from the fact that every channel in the DDC has two integrators that alternate each other taking the samples. Then basically the calibration procedure is the same, just applied to each stream of data separately (A or B side) and then, after correction, both streams can be combined if needed.

Another approach is to sample at 2x the final speed and get every individual sample at the final rate by adding two of those samples. Now we can calibrate that as a single stream, as every sample contains an error from A and one from B.

Interestingly enough, this last method can actually result in lower noise. Please see the attached Word document where I explained these points at length.

Regards,
Eduardo

A_B Mismatch correction 2.2 E2E.docx