Help: Amplify AC microphone, output DC 5v

Correct. You need an isolated 12v supply. The ground is generated internally, and is NOT connected to the 12v -ve.


Ahh, get you now.
A floating 12volt supply is connected to +ve and -ve.
+7volt and -5volt is generated with respect to the amp's ground (with R11, R7, R8 and the TL431).

If you build an inverting amplifier, the impedance of the mic gets added to Rin to determine the gain and since the microphone impedance is not generally constant over the frequency range, you can get frequency response variations. To minimize that, Rin should be much greater than the mic impedance.

You normally would not add series resistance to a microphone signal as that will significantly increase the
noise floor due to the resistor's Johnson noise. Even 1k in series will generate 4nV/sqrt(Hz)

Use only the non-inverting circuit with no series resistance for good noise performance, and use a low noise
opamp (4nV/sqrt(Hz) or less is good)

Opamps - a LM358 might do at a pinch, but a NE5532 would be much better.

The LM358 is not suitable for audio signals, the bandwidth is about 5kHz for large signal swing
for one thing. The excellent NE5532 isn’t going to work from a single 5V rail alas, there are
modern 5V rail-to-rail opamps that are great audio performers, I’ve used the AD8656 dual
opamp which not only has low noise and very good performance for audio, it can drive high
current loads such as headphones directly.

The ultimate microphone opamp is probably the OPA1612, more expensive of course, but
you can’t fault 0.000015% distortion specification!

my circuit provides a 12v rail for the NE5532. Should work.

There are indeed much better devices around these days - but the OP isn’t after super HIFI…

And he has an engine screaming away at >100dBA next to his mike, so I doubt that super low noise is necessary either.


Maybe i don't have the power connect correctly. I have it connected at the +ve and - ve.
Should it instead be connected to +ve and where the "zero w/ arrow" is?

Would someone confirm i’m using the correct pins for the opamp and the TL431?
And anything else that isn’t obvious for this noob?
Please see attachment.

Rewired the circuit this morning. Results below are w/ power connected to +ve and -ve:
No input connected:.35 volts
Either lead from microphone connected to upper input by itself: 1.45 volts
Everything connected, microphone power off: .24 volts
microphone connected to 1K frequency 114db source: .40 volts.

As far as i can tell it's wired correctly. Quite possible i'm not understanding how it should be wired.
Any help in figuring this out would be appreciated. Feel free to ask me question and i'll try to answer them for you.

Your pinouts seem correct.

But the voltages you measure are wildly wrong - so there must be some mistake.

Check :

1/ Voltage between -ve and '0V' - should be 5V .

2/ Voltage between '0V' and +ve - should be 7V.

3/ Voltage Between A1 pin 1 and '0V' - with no input should be 0V +/- a few mV.

that'll do for a start..


Can you post a picture of your project so we can see your component layout?
Thanks… Tom… :slight_smile:

A thought…

Maybe I’m asking a bit much of the NE5532 output stages. They should be good for +/- 14mA… but…

Make R12 and R13 = not fitted or eg 10K. Added to increase +ve swing.


Would someone confirm i'm using the correct pins for the opamp and the TL431?
And anything else that isn't obvious for this noob?
Please see attachment.

Hmm, I am confused by some of that circuit... C1 is clearly wrong as it should go to signal
ground, not the negative rail, and the whole feedback network is over-complicated, you just
need a capacitor in series with R4 to reduce the DC gain to 1.

Why are you connecting any part of the signal path to the +ve rail via resistors? The signal
path is independent of the rails, it should only interact with signal ground. R13 and R12 are
serviing no purpose (unless the opamps in question have very asymmetric output stages,
which modern devices don't suffer from)

I'd have used the standard opamp method to setup a virtual ground. precision shunts are
for when you need precision voltages!

Hi Mark ....

C1 to -ve to dc bias it correctly if a polarised type. I could have specified a non-polar type, I suppose
and connected one end to 0V.

Asymmetrical supplies as we want +ve going output voltages compared to 0V.

TL431's are cheap and good - What if the supply isn't 12 but 11 or 14? I want just enough neg ( -5 wrt 0V) to get the NE5532 to work correctly and as much +ve headroom as poss.

A1's feedback paths: This is a crib from the front end of AIR Montserrat's console , which I did years ago at Neve. It's still in use today.

You need to supply the bias current and dc feedback to the -ve input of A1 independently of the ac feedback - otherwise the +/-4mV input offset is multiplied by the preset ac gain of 101 X. A cap in series with R4 also achieves this. But my method is not much more complex, and derives from the need to get 11 steps of gain in 5dB increments. Probably not necessary here.

You're right about the probable redundancy of R12,13, as I mentioned above. The idea was to get a bit more +ve headroom. I want 5V max wrt 0V at the output and the NE5532 will only go to within 2V of the pos rail, so a bit close.

But by all means post an improved circuit.


Guys, thx for you help on this.
-ve to “0V” is 5v.
“0v” to +ve is 7v.
Not sure what A1 is. I took it to mean pin 1 on chip. Voltage fluctuates between 0 and -20mV.

See pic of circuit. Will be difficult to see what’s what.
Of course, let me know if you have more questions.

Forgot: both 10u caps are electrolytic. I have negative legs going to negative power rail.
And will try changing R12/13 to 10K.

With R12/R13 changed to 10K, results about the same as post #27.

Not surprised.

PSU values sound fine.

Chip orientation in your pic - where's pin 1? Bottom left or top right?


bottom left.

Thought so.

In which case you've got the power reversed to the NE5532.

Should be :

Pin 4 -> -ve.

pin 8 -> +ve.

You've probably blown it up.


-ve is negative power and +ve is positive power, yes?
In picture yellow jumper goes from lower right corner (pin 4) to negative power rail, and upper left corner (pin 8) goes to positive power rail.
Isn't that correct?