On Settling of a fully-differential amplifier

Dear guys,
Due to the periodic drift of the output CM level, the outputs (Voutp & Voutn) of a fully differential amplifier accordingly drift even with sufficient SR and UGB, as shown in the figure attached. This is the typical case.
However, if we look at the differential signal (Voutp-Voutn), the influence of the CMFB vanishes. The differential signal seems perfect.
Comparing the single-end signals and the differential representation, we may conclude that the former are not well settled, whilst the latter is well settled. (Confusion??)

Actually, in my practise, opamp exhibiting this phenomenon still works well preliminarily by simulation (this conclusion has to be fully verified further) , however, I'm pretty worrying about the result of the final testing.

My question is, when we talk about the settling of a fully-differential amplifier, which one we mean, signal-end signals with effect of CMFB or the differential signal, providing that the drift of the output CM level is not severe.

Thank you!

I have difficulties to reproduce your description in the waveforms. The CM signal is relative small, the difference between single ended and differential waveform not very significant.

But even if the residual common mode signal matters, your post doesn't give a hint under which circumstances the "periodic drift of the output CM level" is observed. Is it due to input CM signal and imperfect common mode rejection? Or is the input signal purely differential and the amplifier producing unwanted DM to CM conversion (e.g. due to non-linearity)? Please clarify.

You should absolutely show the test setup.

abonic said:

... However, if we look at the differential signal (Voutp-Voutn), the influence of the CMFB vanishes. The differential signal seems perfect.
Comparing the single-end signals and the differential representation, we may conclude that the former are not well settled, whilst the latter is well settled. (Confusion??)

Actually, in my practise, opamp exhibiting this phenomenon still works well preliminarily by simulation (this conclusion has to be fully verified further) , however, I'm pretty worrying about the result of the final testing.

Click to expand...

Never fear! That's why we use as result the differential signal (Voutp-Voutn). It's the meaning of a differential amplifier to minimize all the common mode signals, like clock and voltage reference signals, which all wear their unavoidable peaks and ringings.

I had the same phenomena with opAmps and even an ADC, and both worked well in test and reality.

erikl said:

Never fear! That's why we use as result the differential signal (Voutp-Voutn). It's the meaning of a differential amplifier to minimize all the common mode signals, like clock and voltage reference signals, which all wear their unavoidable peaks and ringings.

I had the same phenomena with opAmps and even an ADC, and both worked well in test and reality.

Click to expand...

Thank you erikl, it is really a good news for me. I'll leave this problem alone, and go to tapeout.
when the testing results come out, I'll post some results.