Hello,

I'm trying to build a preamp to connect a (balanced) microphone to an unbalanced amplifier, or to the microphone input of my smartphone.
This is the schematic:

A first LM741 is used to split the voltage from the battery to create a ground reference. Then I use one of the op-amps in a NE5532 as a differential amplifier, to 'convert' the balanced input to an unbalanced signal. The output of this op-amp goes to a passive low-cut filter, and to the input of a second LM741, that drives an analog VU meter (just as a non-inverting amplifier). Since the input impedance of the main amplifier would change the low-cut frequency, I used the second op-amp in the 5532 as an output buffer.

Some of the component values:
R2 = R3 = 33kΩ
C7: 47nF
R9: 6.8kΩ
R10: 50kΩ (sets the low-cut frequency)
R11: 1MΩ
C8: 4.7µF

I have two questions about this design:
1. Why R11? If I disable the low-cut filter (using S2), the output is very weak and distorted. By adding a 1MΩ resistor between the input of the op-amp and the ground, this problem is solved. But what is the explanation?
2. When I connect the output to the input of my smartphone, everything works fine, but when I connect the output directly to the line input of an amplifier or an analog mixer, the VU meter stops working. By adding a resistor (1.5kΩ) either in series with the output (between C8 and J3) or in parallel (between J3 and ground) this problem is solved. But this doesn't seem like an elegant solution, and I do not have a clue why this happens.
Could anyone explain this (seemingly) strange behaviour?

Of course, any other remarks on the design are welcome as well.

Thanks in advance !
Pieter

 

R11 maintains a DC voltage reference for U2.2's input when S2 is open. Without it, U2.2's output will saturate due to it's input wandering around due to differential pair current imbalance in the input stage. In analog as well as digital, *no* floating inputs.

59.6 Hz to 500 Hz seems like a very high highpass frequency range for what clearly is intended to be a high quality microphone channel.

Side comment - what is the gain range expected by adjusting R8? No matter how expensive the dual pot is, the resistances of the two halves will not track nearly well enough to maintain reasonable common mode rejection in the differential stage - and that is the only reason to have a differential stage. Better to fix both R8's at one-half the value for your minimum required system gain, and then make U2.2 a variable gain stage for the rest.

Side comment about my side comment - consider an integrated instrumentation amplifier in place of U2.1 and its resistors. You never will find precision matched resistors that are better than the laser-trimmed wonders in a diff amp from Analog Devices or Burr-Brown. Both make mic preamp parts that are equivalent to the classic SSM (Solid State Music) parts, and you get single-pot gain adjustment.

Don't know the exact reason for the output resistor issue. It is common practice to build out the output connection with 100 to 1K ohms, mostly to protect the output stage from accidental shorts. It also can prevent oscillation in the following input stage if that stage is "delicate" and doesn't like to see a very low source impedance.

ak

 

Thanks a lot for your answer!
I'll change the values for the low-cut.
Would an AD623 do the job as the input op-amp (instead of U2.1)?

 

Should be ok, but I didn't plow through the whole datasheet. Generally speaking, anything AD and BB are safe bets.

ak

 

Alright, thanks a lot for your time and helpful answers!
(One last thing though: logarithmic or linear pots for gain and low-cut pots?)