HI,
I'm designed an ADC driver using the THS4541 as a single-ended to differential conversor. The THS4541 is configured as an attenuator with gain 0.5V/V (see figure below).
This is part of an acquisition subsystem circuit, that has a analog multiplexer, a buffer a FDA to convert from single-ended to differential and one ADC. The acquired signal is processed to return the RMS value.. However I was getting a very noise acquired signal, that is affected by temperature (figure below).
The curves above are the same signal applied to the acquisition circuit where only the THS4541 are heated. The expected readings must be 1.20 Vrms.
When I remove the caps C75 and C151, the signal have a great improvement, both in noise level and in temperature stability.
In theory the caps C75 and C151 in the feedback must estabilize the THS4541, increasing the phase margin, why removing them improve the signal quality? It is awkward to me.
