That is ignoring the fact that the output impedance of the leg that is driven by the drain is higher than that driven by the source.
This difference is much higher than 1% tolerance on the 47 ohm resistors.
Typically the difference is between 5 and 15 ohms.
If one is anal about output Z balancing, one of the resistors should be adjusted.
Yes, when you're referring to a classic Schoeps, you're right. But you have to read this sentence in the context of the rest of my post, where I explained what I consider a good BJT output topology: CFP, preferably using matched transistors. A CFP will almost totally "isolate" the driving impedance from the output, much better than a single BJT will do. And if the cut-off frequency of the low-cut filter is determined by the input impedance of the output stage, then the cut-off frequency will also be better defined when using a CFP output stage. This is what I had in mind for one of the designs I'm working on (single output shown only, as you may have guessed already):
Yes you're right of course! Not sure why I had 1% in my head? So the 6k81 0.1% resistors can vary in value by up to c. 6.8Ω.
A quick test of a handful of components here shows it's not difficult to match the output electrolytics to within 1uF of each other .... so around 1.5Ω difference in reactance at 50Hz.
And quickly checking the values of around 10 x 2200Ω 1% resistors confirms that it's not too difficult to find 2 or 3 of them within a couple of Ohms of each other -- so around 0.1%
Should be able to get a fairly respectable CMRR by spending a few minutes selecting components, without actually needing to buy expensive 0.1% components?
Still don't know why I had 1% values for the 6k81 resistors in my head?......
I did 46.5/(46.5+10000) = 0.004628 (according to Excel), which by a very happy coincidence comes out just as -46dB. It won't work like that for other numbers, that's for sure
This is a sequel of a time when microphones had fully floating outputs and mic preamps had input transformers.
Analysis of CMRR in a balanced connection shows that both Common-Mode and output impedances alter the result.
In other words, even with 0.01% phantom-feed resistors, CMRR will be degraded if there is a difference between the output impedance of the two legs.
Absolutely. By definition, CMRR aplies to a connection that involves both a source and a receiver.
Balancing the output impedances is as equally important as balancing the CM input impedances.
