Morning,
I was scanning Audio ADC's specs for example parts to drive from a 50kHz measurement mic, scanning for input spec - this looks like an error, should be Input probably, Are there CM input tolerance specs on this, or other more recent, Audio ADC's - can you point to an example?
Hey Steve, I guess you are not going to reply, not easy if the specs are not there - have to move on and put something in this next audio interface to PCM ADc's article, that looks like this right now -
"Table 1. Recent precision FDAs with <-135dBc audio frequency distortion.
The DC accuracy concerns at the ADC inputs include CM shifts from midscale and small differential offsets. The audio sigma delta ADCs will have an allowed input CM tolerance – rarely specified, the precision devices in Table 1 all offer <20mV CM offset even with the control input floated. That by itself should not impair ADC performance, but include supply voltage variation as well if the control input is floated as that output CM is just a divider off the FDA supply. Use a buffered ADC reference voltage source into the VOCM control pin if this becomes a concern."
moving on to CM noise to diff output noise conversion via feedback ratio mismatch - I originally assumed that would not be a big issue, but, if this floated input curve from the THS4551 PDS is correct, that is kind of scary below 100Hz - probably need to drive that Vocm input pin from a buffered version of the desired ADC Vcm input to bypass this issue completely.
Thanks Colin, I was off looking at other things, but this is useful,
Yes, I have decided you cannot just float the Vocm on the FDA due to low frequency noise (here,I am aiming at a 2Hz to 50kHz span for a measurement mic). I could drive the Vocm directly from the PCM output 2.5V - is that an ok idea or should there be some kind of unity gain buffer amp. I do worry about popcorn in that buffer stage - which a chopper would fix, but then most folks worry about the switching in a chopper, If we use a buffer amp, I am almost of the opinion no output C should be used since the Vocm input to the FDA is not a sampling event, just a reference voltage to a very high impedance. (in the background, the issue about output Vcm noise for the FDA is tied into the CM to DM conversion through feedback ratio mismatch).
Ideally,if we wanted to use a chopper as a Vocm buffer, that would be a low cost, higher frequency device to move that chopper up out of band - the filter I am doing is looking for 100dB atten at 6Mhz,
Point 2, the buildout R's for an outside the loop RC are driving the differential 3kohm input of the PCM4202, not a DC level shift but a gain loss that can be recovered in the filter gain target. And again,the lack of tolerancing on that internal 3kohm leaves a gain uncertainty in the signal channel, small effect and not of great importance for this Audio app.
the point 4 above, the PCM4202 PDS ckt is obsolete in the sense that the THS4551 is much lower HD (at much lower power) than the THS4131 or OPA1632 equiv. We actually remeasured that back on 2006 with this result, compare this to the THS4551 datasheet . The voltage noise of the older FDA's lis ower, but normally swamped out by the R's around the device. these updates never got into the datasheets. And remember the THS4131 is the same die as the OPA1632.
