So I thought the AD9106 was the perfect solution to my problem but now I realized that it has a fatal flaw, the 267 PPM gain drift, specified with the rather drifty internal reference, which has a 160PPM / degree C drift.
If a very stable external reference is used, how is gain drift re-calculated? I hope that maybe the datasheet writers tested the drift with a poor tempco Rset resistor as well.... (I could use a vishay foil here if it helps!)
I was hoping to knock it down an order of magnitude to make it usable..
Thank you.