Dynamic Microphone preamp design help

[fcb]

ElectroIrradiator.  I would cut the nuts off an engineer who just chucks in 100uF electrolytics without good reason/experimentation, those things are (relatively) expensive to put on a board.  And yes, keep increasing the size of the capacitor it will lower the noise contribution from the PSU from that part of the design - why not 1000uF?  Personally, i'd rather put a gyrator in.

I'm great believer in experimentation, try out a few combinations - use whatever meets your performance & cost specification.

[atwoz]

I was searching and found this IC: MAX9814. It is a mic amplifier with Automatic Gain Control (AGC). What do you guys think?

[ElectroIrradiator]

atwoz, Pippy

The PC input may perhaps be an ordinary line in, which expects about 1Vrms, more if you want to saturate the ADC. Those line in connections are quite common on many motherboards and PCI(-e) audio cards.

A dynamic mic will have a signal output level of about -60dBV or even less. If the PC input is 1Vrms line in, then you may need 60-70 dB of gain to reach an acceptable level. Your current amp has about 61 dB of gain, max, so it might actually make sense if you need some more.

I had a very brief look at the MAX9814, and it appears to contain what we used to call a VOGAD: Voice Operated Gain Adjustment Device. If the AGC can be turned off, then the chip may be suitable, though I haven't read the whole datasheet yet.

If the AGC cannot be disabled, then I would be dubious about whether it would work for singing. Usually you would want a manual, variable gain amp for mic volume control, but not an automatic one. That could end up sounding extremely unnatural, as the pre amp would - intentionally - remove much of the dynamic power in the voice of the singer.

How bad this might be would depend on the actual delays involved, though.

[atwoz]

Pippy,

I recorded some voice using a good quality USB mic and the preamp + dyanmic mic I built and compared the waveforms using some software. You were right, the gain is not the problem. It seems like the waveform is distorted. I will keep doing some tests. Thanks for the help, each time I get closer and closer. I will try to post the waveforms later.

[atwoz]

I have the theory that the preamp is filtering some of the high frequency components of the audio, making it sound hollow. The waveform lacks some peaks that voice waveforms usually have. It's mostly sinusoidal. How should I calculate the ac coupling caps? I used 100uF standard electrolytic but I didn't give much though into those. Should I use lower values?

Thanks!

[ElectroIrradiator]

Which opamps are you using? The input stage may be running out of high frequency gain, if the gainbandwidth, GBW, is too low.

You might also try injecting a weak signal at various points in the amp to check where the high frequencies disappear.

AC coupling caps are way large enough, those aren't the problem. They would also cause low frequencies to disappear first, if they were too small.

Finally, did you use the 100pF compensation caps you planned?

[atwoz]

Hello ElectroIrradiator!

I'm using the MAX4477 with a clean 3.3v Supply. I did put the 100pF on both stages.

I will test the microphone on another amplifier to see if it is working, just to rule out the mic being the problem.

Ill keep you posted.

[ElectroIrradiator]

If the mic is OK, then try this: Look at the amp output, while the PC is connected. Then disconnect the PC and check output again on the scope, with no load attached. See if you can subjectively spot any difference in frequency content.

The problem may be that these opamp cannot drive much of anything, particularly not at a high gain setting, and the PC probably has an input impedance around 10K. The 27K feedback resistors might also be too low for these chips. I may even have to backpedal on the 100pF feedback capacitors...

[atwoz]

Could I add a voltage follower inbetween the preamp and the pc?

[ElectroIrradiator]

Yes, assuming *that* opamp can drive the 10K at a gain of 1.

However, I must say I find some of the information in the datasheet a bit confusing. Various figures seem to tell different things, so I wouldn't jump to conclusions just yet. Which is why I'd suggest you take things in a methodical order, as you intended.

[atwoz]

I think I found the culprit. The microphone. I compared the output of a cheap $1 dollar electret microphone with the output of my preamp and they were different. As I said, my preamp doesn't have those high frequencies. I then checked the output directly from the microphone, and the noticed the same waveform. So, I was amplifying a bad waveform to begin with.

The design seems to be correct, as you mentioned. I'm in the process of getting more microphones to test this, but I'm like 75% sure that the problem is the mic I'm currently using.

I'll post the results once I have them.

Thanks a lot!!

[atwoz]

I tested with another mic and the preamp works much better. I still need to tweak some parts but it's looking better! Thanks everyone for the help. I needed very little gain by the way, like 3V/V