Op-Amp Mic Pre Trouble

Hello, I have a LM4562 op amp circuit set up as shown in the microphone preamplifier section of the data sheet in the following link

formatting link
. The performance of the pre is very good from 10 Hz to 50 kHz on a line input. But when connected to a balanced microphone and set to a decent amount of gain from 40 dB to 60 dB I get a 10 kHz peak about 20 dB above the noise floor. The microphone is an industry standard shure SM-58. The power supply is made of batteries so I don't think the source is the power supply. There are no capacitors or inductors in any of the signal path section to cause oscillation besides the microphone itself. Anyone have any clue why an op amp circuit would have a peak at 10kHz. There are 47uF caps on the power inputs to the op amps. Any other suggestions for capacitor values I should try here? Any help is welcome thanks.

Reply to
nirvanabt4
Loading thread data ...
** Desperate Groper from HELL alert !

** You get a 20 dB " peak " in WHAT exactly ??

The frequency response ??

The background noise level ??

What is you measurement method ??

BTW:

The LM4562 is a VERY NOISY op-amp.

...... Phil

Reply to
Phil Allison

Hey, there is a peak at 10 kHz 20 dB higher than the rest of the signal. It is a clear 10 khz tone you can hear in headphones.

I'm using a echo layla 24/96 audio interface into FFT analyzer software with a hanning window at a 24 bit 96 kHz sampling rate. I am comparing a direct loop (in and out of the interface) vs a loop through the pre amp with 40 dB attenuation before the pre. That test comes out perfect even with the extra gain. The test material is 24bit 96 kHz white noise.

It is only when I connect a microphone to the preamp that a 10 kHz peak appears. To clarify the peak is 20 dB higher than the rest of the signal. This is also very audible in headphones when listening.

Any tips if a INA217 would be a better for a mic pre, the circuit is nearly identical in that op amp in comparison to what I am using now.

Reply to
nirvanabt4

A gain of 1000 puts output signals that are very much larger than input signals in close proximity with each other.

You may not have any intentional capacitors in the circuit, but you certainly have unintentional stray capacitances. Do you have a photo of your layout available somewhere in the net? Or you can email it to me.

The 3 opamp differential amplifier (instrumentation amplifier configuration) has two high gain, non inverting amplifiers in it, each of which can produce inadvertent positive feedback. Combine that with a 55 MHz gain bandwidth product and you have a lot of potential for trouble.

Reply to
John Popelish
** Top Posting GROPER ALERT !!

** Cut out the smartarse use of "hey" - OK .

** Then it is an OSCILLATION - not a " peak ", you goose.

Learn what words mean before you go using them !!!!

** That is an OSCILLATION - not a " peak " you goose.
** What value input resistors " R" have you used ??

Is the OSCILLATION at full output level - ie rail to rail ?

Have you used normal shielding practices ?

Is mic pin 1 grounded?

BTW:

DO NOT TOP POST AGAIN !!!

........ Phil

Reply to
Phil Allison

You would almost wonder if it is picking something up somewhere when it is oscillating like that. Is your circuit in a shielded box? If you can't get rid of the oscillaition, maybe you could try just connecting it in a standard inverting or noninverting configuration, or doing a 2-stage setup using each opamp in the package for each stage, and do a gain of 20 in the

1st and 50 in the 2nd, and check the performance of it then. It may not be much different than in the instrumentation type setup they have in the app note.

You might also want to check out:

formatting link

Much lower noise floor.

Reply to
tempus fugit

"Phil Allison"

Correction:

** Should be the MAX4469 .......

....... Phil

Reply to
Phil Allison

Yeah I just figured I'd try them out (samples) they really are noisy though. Also, I stopped the oscillation with two 400 pf capacitors on the inputs terminals connected together to a 47 pf capacitor to ground, now it has a nice flat frequency response...with lots of noise, but I am adding 60 dB of gain so it could be anything in the room making it noisy too. Thanks for your help everyone! I'll have to try out some other op amps and see how they change the performance.

Reply to
nirvanabt4

Yeah I just figured I'd try them out (samples) they really are noisy

** Nonsense.

At 2.7 nV/ rtHz EIN they are very low noise op-amps.

** That is only RF suppression - not likely to affect 10 kHz much.

Perhaps 10 kHz was an alias frequency from a much higher oscillation frequency.

** Bollocks.

** You must replace that SM58 mic with a 270 ohm resistor when noise testing.

Same thermal noise as the mic with no room noise.

** Waste of time

You won't beat the one you have with ANY normal op-amp.

....... Phil

Reply to
Phil Allison

I bet you got bitten by a layout problem and this amp is oscillating. The observed peaking is typical. Long traces to the gain setting pot will cause this. Try replacing them with a single short resistor between the inverting inputs for testing. Also heavy capacitive loading might cause it. a 100R in series with the O/P is always good. The chosen opamp is good for filter and line-level apps, but for a pre I would try the new AD8599 as frontend followed by a difference amp with integrated resistors. Try to have a fixed gain of 20dB on the second stage and use an attenuator for line level. At 60dB no single stage is performing very well. You are talking about battery, how are you creating a split supply at what voltage?

--
ciao Ban
Apricale, Italy
Reply to
Ban

The trace wires for the first gain stage are fairly long I'll have to redo that at some point and see if it will fix the 10k problem. I don't have one op amp doing 60dB, I have a 40 dB gain stage adjustable followed by a fixed 20 dB gain stage. I do have the 100 ohm on the output also.

And it's fairly easy to make a split supply with batteries and best part is no power supply noise. I am using three new nine volt batteries to get 27 volts total, then using two caps and two resistors, one each to the positive supply and one to the negative supply. You create a ground in the middle with +/- about 14 volts which is what I am running the op amps at.

Ban wrote:

Reply to
nirvanabt4

snipped-for-privacy@gmail.com wrote:

OK

Hah, that really sounds for trouble! Better to have nice low impedance gnd also at DC. Take only two batteries with the common connection soldered to GND. Use a different pair of +/-18V batteries with the diff.amp (chosen for

+/-22V max.) Have also this battery array soldered to gnd.
--
ciao Ban
Apricale, Italy
Reply to
Ban

It's a fairly rubbish mic pre-amp so I wouldn't lose any sleep over it !

Graham

Reply to
Eeyore

Phil is right in these areas. I earlier recommended the THAT1510 as a lower noise alternative (I would think it is about as low as you're going to get), but I didn't mean to suggest that this one is noisy. Most opamps are much higher EIN than 2.7 nV. You are also right about the room noise - that's probably what you're hearing. Phil's suggestion of using the 270ohm resistor instead of the mic will give you a more accurate idea of the noise level.

Reply to
tempus fugit

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.