Thanks olafmatt for thinking about this project.
I anticipated output coupling caps internal to the microphone, but I couldn't find an example of that topology in a real microphone that I had available to test.
I had to test for that possibility using a capacitor-coupled generator.
With capacitors coupling the input, the topology is very similar to a conventional input with a differential servo.
There are some scaling differences in the value of the injection resistors: 100K vs 1M2 (typically).
It appears to me that the Thevenin equivalent of the injection resistor in parallel with the bias resistor along with the "input" capacitor sets the time constant of the input RC network.
Comparing a conventional AC-coupled reference design with (1M2||2K) to the DC-coupled input with (100K||5K) with added input caps and they are close.
The relationship of the servo's response to the input RC versus gain is what could be different.
Its been awhile since I looked at this but as I recall I didn't see any adverse effects with input caps in place and saw no need to re-tune the servo.
The damping seemed correct at all gains both with and without caps.
I tested this two different ways one intentionally and one unintentionally.
The first time I went looking for it. That was not with the Deboo.
At least one other time I recall with the Deboo looking at the effects of input capacitors on 1/f noise vs. no input capacitors and the resulting 1/f.
If there were a problem I should have seen it here:
viewtopic.php?f=6&t=14&start=201
I was switching gain that day (and a lot of others) like a maniac.
So to answer your question I did test it and didn't see servo instability but it's been a very long time since I did.
If there were ultimately a need to have the servo "switch gears" with gain to maintain stability it's very easy to do with the Differential Deboo by switching a single capacitor.
When I get a chance I will check for this: The worst combination would be at high gain with input coupling caps.
But I should have seen it by now.