Transformerless 990 micpre

[Samuel Groner]

Unless we're looking at different schematics R10 and R11 are collector loads.  The schematic you linked too also provides input device current from the feedback resistors but in your link the input devices are biased -5 V or so. This schematic looks like inputs are biased at 0V so opamp output will need to sit at negative voltage to make required current.

If DC coupling to the opamps is (as usually) employed bias of the input pair is established by setting the collector current, not emitter current. This is because global feedback forces the - input of the opamp to the same potential as the + input, and the latter is fixed at a couple of volts below one of the supply rails. The collector current of the input pair is then given by I = U/R.

The -5.1 V bias in my schematic or some sort of CCS to the emitter (as typically used in mic preamp ICs) helps to center the DC output voltage around ~0 V for best output swing, but does not alter the quiescent current in the input pairs.

Samuel

 

[JohnRoberts]

This is dissembling to semantics..  The operating current is "established" or set by the voltage across R10/R11, but that same current must be pulled from the emitter. That is the path I am concerned about.   

In this schematic the opamp uses negative feedback to make the voltage at at the collector (it's minus input) the same as it's plus input.  As I have been saying, this current must be supplied by the 10k feedback resistors. This will result in a significant (negative) DC voltage at the opamp outputs. So not only are the output blocking capacitors backwards, the output signal swing before clipping is dramatically reduced.

As I also said I have actually done this before, on purpose, but not with this much DC offset baked into the output operating point. From a quick inspection it looks like both 990 outputs will be sitting at more than 12V negative.

I repeat that the OP should study some of the many variants of this topology published and available. Regarding this specific schematic I would suggest:

A) provide the input device operating current from some other path than output feedback R. (like current sources or filtered R to minus rail.)

B)  and/or, use a smaller value for feedback R (I did a lot of mic preamps using circa 3K feedback Rs).

C) and/or, bias the inputs at DC similar to your link example, but note this then reduces the input signal swing dramatically for values used here.

Sorry but this seems painfully obvious to me, I spent decades messing with this topology in the good old days. There are many ways to skin this cat but IMO this schematic isn't finished yet.

JR


PS: FWIW I have seen, and have also used this topology myself with the inner loop AC coupled to insure a reliable DC operating point for the opamps. There are many ways to skin this cat and get good performance.

 

[Samuel Groner]

Nothing I'd disagree with--the feedback resistors are much too large indeed. That's another reason why I pointed to my implementation.

Samuel