Inside an M-Audio USB Mic

[jaykeMURD]

I decided to open up my M-Audio Producer USB and see what's going on behind the grill.

The mic is sometimes advertised as a Large-Diaphragm Condenser......I don't think this qualifies as such.

I've provided some basic info on the picture.

I'm curious as to what that FET is doing?

I can't find any info on that particular FET...

I wonder if an actual large diaphragm capsule could be fitted to the mic? What are the power requirements of your run-of-the-mill chinese LD capsule?

 

[mshilarious]

The FET is 2SK596, it is doing what FETs usually do in microphones--impedance conversion. It also has voltage gain. That is a 16mm electret capsule, so there is no polarization voltage. I would guess the FET is running off the +5VDC USB rail, which means no step-up converter to generate a polarization voltage for an externally biased capsule. Post a picture of the circuit to be sure.

The XLR-->USB converters on the other hand do have step-up circuits to generate +48V phantom power. If you want to use a large diaphragm, externally biased capsule into USB, those would be your best bet.

As for this mic, the capsule isn't bad, you can improve SNR by replacing that FET with any of the usual lower noise suspects (even 2N5457 works fine). 2SK596 is very noisy by comparison. You will also need a high-value resistor (at least 100M) to bias the FET. The easiest method is to ground, and given the limited supply of 5V that works fine here. Also, the FET source should not be grounded, as is stock, you should have a resistor (not a critical value; 3K3 to 10K will work) from the source to ground. This will reduce distortion/increase max SPL handling.

That will also reduce the gain at the FET, which could conceivably drop the level enough that the USB chip becomes the noise floor. You can fix that with more gain--try a lower value for source resistor (this will limit max SPL handling), or add an opamp gain stage, maybe with a trim. Make sure the opamp is specified at 5V single-supply. Or you could just add another FET gain stage--but first post the rest of the circuit, it might already have another gain stage, or an opamp buffer for the USB chip that can be tweaked for gain.

 

[moresound]

I've been waiting for a thread like this.
MSH .... is this this the first thread ever on modding a USB microphone?

 

[jaykeMURD]

Holy moly MSH! Let me re-read that again before I articulate anything!
Man, you know your stuff.

I hope these pics will suffice.

I'd be very interested in swapping the stock FET out for a better one.
I already did one "mod", I removed the felt liner from the grill as well as the secondary mesh, leaving just the outer grill. I did change the sound, but I can't decide if I like it.

 

[mshilarious]

Yeah, not much there, the quad chip would be the USB CODEC although I can't read the label. The SOIC-8 on the other side is probably a buffer for the mic or the headphone out? Can't read that either. Some of these CODECs have direct-drive circuits for both functions, although it looks to me as if the two big electro caps are headphone coupling caps. It's really hard to follow those traces in the picture though.

And in case you were wondering, the silver oval is a 12mHz crystal that the USB chip uses for a clock.

I would start by measuring the stock mic's noise, get the room as quiet as you possibly can and stick it between a couple of pillows, and post a 10 second wav (not mp3!) of the result. Also, figure out what level 0dBFS is by playing back some source (a 1kHz sine wave works well) until you hit that (or another reference level if 0dBFS is too loud, say -10dBFS). Put the mic 1" from your monitor's woofer for that test, the idea is to keep the playback level low to minimize distortion in the playback system.

Then change the FET (which is a $0.10 part, unfortunately a 100M resistor is a few dollars ), including a source resistor, and measure noise & level again. Those three components (FET & two resistors) can all be done point-to-point at the capsule, as with stock.

If you don't have experience with SMT, don't touch that board. I mean *if* the opamp buffers the mic, and *if* it is the noise floor (after FET-swapping), then it might be worthwhile to replace, but I kind of doubt that . . . best left alone, probably. And that might not even be an opamp!

 

[mshilarious]

Actually if you want to be really clever, instead of a source resistor add a 10K pot from source to ground. Now you have a gain control

 

[jaykeMURD]

Interesting....hmmmm. I've got a few too many projects going on right now, but in a few days I'll get back to it and record some stock samples.
So, if I swap the stock FET with a 2N5457, I'll need 2 separate resistors to "balance" it out?
Forgive me, but without being shown (drawn out), I don't quite understand where they'd go?
I do plan on doing all this for shits and giggles, but like I said, it'll be a bit.

Thanks again.

 

[mshilarious]

Like this:View attachment 64488

 

[jaykeMURD]

That's more like it, haha.
So would the line connect before or after the 10k resistor? Does it even matter?
I might be over-analyzing that.

Are 100m and 10k the values you'd recommend? What type/voltage?

 

[mshilarious]

Any >100M resistor you find will be rated at 250V or higher. And the source resistor or pot, well the circuit sees 5V, so you can solve for that. Generally resistors are rated for power rather than voltage. If you are going with a fixed resistor rather than a pot, 10K could be too high, start with 3k3 or so and adjust to taste.

I don't understand the question about where the line connects. The mic in to the PCB goes to the drain of the FET.

 

[Lasec0203]

Hey, im trying to work this mod. could you repost this picture of the two resistors and FET setup. the current link is invalid. thanks

 

[Lasec0203]

Like this:View attachment 64488

Hey, im trying to work this mod. could you repost this picture of the two resistors and FET setup. the current link is invalid. thanks

 

[jonnyexoh]

I'm bring this back to life. I would love to how this mod worked out. If it was positive i'd love to do this as well, but the attachment isn't there anymore. Any electrical/sound engineers out there willing to help out?

 

[Chater-La]

I'm bring this back to life. I would love to how this mod worked out. If it was positive i'd love to do this as well, but the attachment isn't there anymore. Any electrical/sound engineers out there willing to help out?

Try to contact the user who posted the attachment to put it up again...

 

[PRHunt]

That was mshilarious. Don't think he hangs here any more

 

[Chater-La]

That was mshilarious. Don't think he hangs here any more

OK but he can be contacted via PM/email...

 

[kity92]

Hi there I'm sure still you guys are online but I have the same microphone and its broke. Is there any chance to diagnose or fix it ? Problem is when I plug mic there is no sound, just can hear hiss and buzz sound. I'll be appreciated to any tiny helping method.

Thx.

 

[BroKen_H]

@kity92 Just to be clear, you are asking about repairing a $120 microphone so I assume you are handy with soldering, and capable of (and of course comfortable with) breaking it down? Sending it to a shop if you start will cost more than it is worth.
I am in no way, shape or form capable of helping you diagnose or fix a microphone, but I want you to be certain that is the path you want to head down.

 

[rob aylestone]

What's the back story? Did it suddenly stop, or has it been in a gig bag for years and you just refound it? Sure it's the mic, and not your computer using the wrong driver? One of my USB devices suddenly appears to be a printer, according to windows. The latest Windows update reclassified a data cable as a printer driver.

Just let us know how it broke, what you were doing and how you currently connect it.

 

[junplugged]

wild, from 9 years ago, interesting tho, I'm only now just learning my transistors, some lecture on YT got me started...