Microphone noise (hiss) -Question for the mic-guys.

[SSLtech]

This may be a theoretical question , but it's sparked by a fault/repair issue, so i thought I'd try here instead of the drawing board... -here goes:

I have a couple of MXL 603S's here. they're pretty reasonable for small-diaphragm condensors; I can certainly always find a use for them!

However, interestingly, there are two slightly different versions. Electronically they look to be the same, and they sound similar, apart from one really interesting difference: the (hiss) noise floor is about 15dB-20dB higher on one than the other. The noise follows the mic body and not the capsule. (capsules are interchangeable.)

The schematic on the omnipressor site looks close enough, and it seems to be yet another variation on the Schoeps circuit. Essentlally a FET impedance converter, polarity-splitting the output into two phantom-fed PNP output transistors.

The Trim pot from the FET source to ground, I assume is to set the quiescent current and therefore the voltages at the drain and source, I assume? -I wondered if the FET was the source of the noise: the first things I thought of were the FET and the 2x 1Gig resistors.... I was considering swapping them from mic to mic and seeing if any of them "moved" the noise over.

-Am I right in assuming that I should be able to measure the voltages at the source and drain, and -after putting a new FET in- twiddle the pot to set them the same as they were with the old FET? -Or is there more to it than that?

Cheers,

Keith

 

[stefanzap]

is the noise floor higher, or does one (the noisier one) need more gain than the other in order to get equal output?

The reason I ask is that i had a similar thing going on with two schoeps style circuits I a/b'ed and after going through through the whole thing, finally figured out that half of the balanced output wasn't happening (due to a bad bypass cap). Thus, it was only outputting half of the signal. Also had this happen with a bad transistor. Easy way to check this out is to disconnect pins 2&3 - one at a time, of course.

 

[SSLtech]

Gain (acoustic) is the same. At the same gain setting, either combination of capsule & amplifier gives pretty much exactly the same signal sonic character and level, so there's no additional mic pre noise or whatever... Basically they sound the same other than the soft "Hsssssssssssssssss" in the background.

Keith

 

[stefanzap]

Good chance it's one or both of the transistors.

However, I'd still try the pin 1/3 and pin 2/3 test to see if the mic is not outputting both sides of the signal.

 

[NewYorkDave]

Is the noise still there with no capsule attached and the gate of the FET AC-grounded through a large cap?

 

[Gus]

Keith

Check clean the around the resistors input cap and fet gate flux can/ will cause noise. 91% isopropyl alcohol and/or distilled water.

FWIW I had a noise problem with an early microphone build. It turned out that finger oils or dirt across the input cap was causing enought leakage to cause a noise problem. I had handled the cap with my fingers. I now use tools to mount the input high Z parts.

I have not had a high noise 170 in any of the china microphones 30 or more last I counted I have bought, so I am thinking dirt for your 603..

 

[SSLtech]

Cool. I looked at a bottle of 91% last night and thought about just giving the board a "dunk-&-a-brush"... maybe this evening I will.

Keith

 

[Guest]

[quote author="SSLtech"]
from one really interesting difference: the (hiss) noise floor is about 15dB-20dB higher on one than the other. The noise follows the mic body and not the capsule. (capsules are interchangeable.)
...
Essentlally a FET impedance converter, polarity-splitting the output into two phantom-fed PNP output transistors.
[/quote]
If it is polarity splitter, source and drain resistors are the some or near value (in one microphone)
Are values the some between both mics?

xvlk

 

[BBob]

The bodies are the shield in these mics. The top of the shield contacts the deck thread and the botton the retainer on connector. The alloy tube is usually plated. (MXL. Feilo, etc)

Hiss can and will vary with RF ingress. Is there a fairly strong RF source near you? {perhaps an FM broadcast station? If so, a variation in the shield would explain the increased noise floor when bodies are swapped.

The fix might be to clean where the metal parts touch.

Bbob




http://microphonium.blogspot.com

 

[JohnRoberts]

[quote author="SSLtech"]This may be a theoretical question , but it's sparked by a fault/repair issue, so i thought I'd try here instead of the drawing board... -here goes:

I have a couple of MXL 603S's here. they're pretty reasonable for small-diaphragm condensors; I can certainly always find a use for them!

However, interestingly, there are two slightly different versions. Electronically they look to be the same, and they sound similar, apart from one really interesting difference: the (hiss) noise floor is about 15dB-20dB higher on one than the other. The noise follows the mic body and not the capsule. (capsules are interchangeable.)

The schematic on the omnipressor site looks close enough, and it seems to be yet another variation on the Schoeps circuit. Essentlally a FET impedance converter, polarity-splitting the output into two phantom-fed PNP output transistors.

The Trim pot from the FET source to ground, I assume is to set the quiescent current and therefore the voltages at the drain and source, I assume? -I wondered if the FET was the source of the noise: the first things I thought of were the FET and the 2x 1Gig resistors.... I was considering swapping them from mic to mic and seeing if any of them "moved" the noise over.

-Am I right in assuming that I should be able to measure the voltages at the source and drain, and -after putting a new FET in- twiddle the pot to set them the same as they were with the old FET? -Or is there more to it than that?

Cheers,

Keith[/quote]


The circuit looks like it will be most sensitive to noise in the FET.

It looks like FET trim sets DC voltage operating point at drain for good dynamic range (4.5V if I'm reading the schematic correctly). I'm not sure it will impact noise floor if misadjusted just headroom.

Fault or incorrect operation of later bipolar stage could also impact noise but first FET is likely candidate. It seems you might be able to probe with a scope with a sine wave playing to confirm proper operation of all stages if swapping FETs doesn't localize noise there.

I'm not very familiar with 1G resistors so don't know if they ever get funny.

JR

 

[bcarso]

[quote author="JohnRoberts"]
I'm not very familiar with 1G resistors so don't know if they ever get funny.

JR[/quote]

A few, given an empathetic audience, achieve moderate success in standup, being naturally accustomed to a lot of resistance---but they are the exceptions.

I'm sorry, it's the holiday cheer typing.

Seriously, the voltage noise coefficient for high-meg Rs is pretty bad at best. Add some contamination of the ionic variety and you can have enormous amounts of pinkish noise.

 

[SSLtech]

[quote author="bcarso"]I'm sorry, it's the holiday cheer typing.[/quote]
More Holiday cheer, please! :green:

Actually the same basic thought (just differently-worded) was what sprang to my mind also!

I've just got a few other dead mics here, so I might like to swap out 1-gig resistors and see what happens.

I'd concluded the same thing about the trim-potentiometer, and if I got round to swapping out FETs, I was going to set it to keep the drain/source voltages about the same after swap-out.

I'll probably get to this later today, -Will keep everyone posted...

Keith

 

[Gus]

Keith
The Schoeps circuit uses a NS Fairchild process 50 type fet if I am remembering correctly. The 603s use a sk170. Some of the china stuff uses a J305s, a P50 fet.

The 603 fet phase inverter drain and source voltages might be set at a different voltages than the ones marked on the Schoeps.

I have not had a noise problem caused by a 200meg to 1 gig resistor in a microphone circuit yet FET or Tube . I have had noise caused by dirt and leakage paths. I forget the site but I seem to remember a post about some old tube microphones high value input resistors being made by a drawn line of China Ink on a substrate between two wires.