Ultra low-noise mic pre?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
But the fact remains that the mike pre noise does contaminate my microphone tests to some degree.
Sorry, I'm probably too stupid for that one--with the graphs you showed that the mic preamp noise is 10 dB to 60 dB (depending on frequency) below the self-noise of the microphone, which proofs that the contribution of the preamp can be neglected. Can you elaborate?

Samuel
 
Sorry for not explaining it better.

One of these pre-amps gives more noise than the other in combination with a microphone. It is not a gain mismatch or a cell phone, since I get the same results over and over again.

It looks like the noise contribution is larger than expected from the second graph, which shows a very low noise for both pre-amps.

One could suspect that the reason for this might be either the much higher input impedance of the other pre-amp or simply the impedance combination. Remember these are not purely resistive impedances, and there's also a cable connecting the two.

But I wasn't necessarily trying to find the explanation for this. For me it's enough for now to note that there is a practical difference in microphone pre-amp noise contribution. Hardly surprising, it would be kind of odd if they were all the same in this respect.

What I need (sorry for repeating myself) is a pre-amp that gives the lowest possible noise from a given microphone, while sacrificing other qualities as needed. For instance, it could have a 50 or 200 Ohm input impedance if this helps to keep the noise down. I was hoping to get some hints from the much wiser people on this board, and I did!

Thanks.

Martin
 
One of these pre-amps gives more noise than the other in combination with a microphone.
Based on practical experience and listening, I guess?

One could suspect that the reason for this might be either the much higher input impedance of the other pre-amp or simply the impedance combination.
Sure--the one with lower input impedance is bound to have higher noise figure.

If you want further input I could sketch a preamp for your needs. Do you have acess to LSK389/2SK389 transistors (or don't mind matching LSK170/2SK170)?

Samuel
 
[quote author="Samuel Groner"]
One of these pre-amps gives more noise than the other in combination with a microphone.
Based on practical experience and listening, I guess?[/quote]

Yes, and the first graph seems to confirm the practical experience. If you think the input impedance difference makes the difference, that's probably it. The Portico has an unusually high 10k, but the Millennia is pretty high too?

If you want further input I could sketch a preamp for your needs. Do you have acess to LSK389/2SK389 transistors (or don't mind matching LSK170/2SK170)?

That would be fantastic, you are very kind! Have a bunch of 2SK170's I can match right here, but can try to order some LSK389's too.

Martin
 
Good point. If you look at the first graph you can see that the peak just below 20kHz (which is noise from the RF microphone) is pretty well matched for both pre-amps. Usually I also double check the gain matching with an acoustic source.

Martin
 
[quote author="Martin B. Kantola"](according to MKH-40 specs)[/quote]


This reminds me, the other day I was thinking about the prospect of a more modern RF condenser design using an ultra low noise PHEMT device in the front end. worthwhile, I dont know. at very least you could move the frequency up a good bit. DIY? probably not a good idea though there are some people around here with the experience to pull it off.
 
When going for low noise, keep in mind your power supply as well. Getting vanishingly low ripple would be important, I would think, and this requires more than just big capacitors. Just what it requires I don't know yet - I'm currently struggling with this problem (power supply induced noise) in my current build.
 
Here we go: low_noise_amplifier_r1.pdf

If you use single JFETs you might want to match the pairs for Vgs and Idss. If you use Linear System parts you might want to check their noise level as it has been reported that some samples show a 10 dB increase in voltage noise.

Not shown is the phantom power supply, but you'll easily add that.

Let me add that the mic preamp noise level differences you observe in practical use happen at 20 dB-40 dB gain. At these gains the contribution of the preamps may indeed be significant and show differences (I think I've seen a Neumann article on this). Not so at 60 dB.

[Edit: the Neumann article is the file lect0021.pdf--can't finde it on the site right now but it confirms the repeated statements that at high gain the preamp contribution can be neglected.]

Getting vanishingly low ripple would be important
Not if your preamp has good PSRR. I've built preamps with -145 dBu EIN (wich is almost 10 dB better than most preamps) using standad PSUs and without taking any particular precautions. Magnetically induced hum is much more of a problem and likely the cause of your problem in your current built.

If you look at the first graph you can see that the peak just below 20 kHz (which is noise from the RF microphone) is pretty well matched for both pre-amps.
I'd rather have said that the midband range of your plots actually show a mismatch of ~2 dB. What's your "acoustical source"?

Samuel
 
[quote author="crazytooguy"]When going for low noise, keep in mind your power supply as well. Getting vanishingly low ripple would be important, I would think, and this requires more than just big capacitors. Just what it requires I don't know yet - I'm currently struggling with this problem (power supply induced noise) in my current build.[/quote]
If the problems are not coming from what Samuel suggested (they probably are) or if you're just interested
in going further where normal regulators go, you might find the following of interest:

http://www.wenzel.com/documents/finesse.html

Regards,

Peter
 
[quote author="Samuel Groner"]Here we go: low_noise_amplifier_r1.pdf
[/quote]

Coooool! Thanks! I'll probably report back end of the week. Will try a single channel first.


[quote author="Samuel Groner"]
If you use single JFETs you might want to match the pairs for Vgs and Idss. If you use Linear System parts you might want to check their noise level as it has been reported that some samples show a 10 dB increase in voltage noise.[/quote]

OK.

[quote author="Samuel Groner"]I'd rather have said that the midband range of your plots actually show a mismatch of ~2 dB. What's your "acoustical source"?
[/quote]

The noise in the midband was not a mismatch and not from the mike either, but picked up by the preamp, found the source yesterday! It turned out that the power supply for the adjacent channels made the noise, but only when the preamp was turned off. In other words, nasty ambient electrical noise. Will do a new set of measurements as soon as I can. Terminating the input with a 200 Ohm resistor naturally killed the noise too.

Martin
PS. For "acoustical source" I use various stuff. Mechanical and electrical metronomes as well as tone generator through a speaker. Matching the levels is easy as long as the mike is the same...
 
[quote author="StephenGiles"]Where does ambient noise come into the equation I wonder?[/quote]

All very well to ignore my question, but a combination of bad ears and general ambient noise can surely fool the listener into hearing noise in a preamp that may not be there can it not?
 
[quote author="StephenGiles"][quote author="StephenGiles"]Where does ambient noise come into the equation I wonder?[/quote]

All very well to ignore my question, but a combination of bad ears and general ambient noise can surely fool the listener into hearing noise in a preamp that may not be there can it not?[/quote]

Not sure what you are trying to say with this, so please forgive me if I've misunderstood something. Never ignored your question, but my studio is designed with great care for very low ambient noise, both acoustically and electrically. Computers are running on a separated power with galvanic isolation from the audio equipment, and the audio power has its own dedicated and clean ground, separated from the electric company ground lines (which can contain a lot of noise in some cases)

In this case, by turning off the other pre-amps a new noise source was introduced, which also showed up on the graph. You might call it a design error in the pre-amp. There should probably be some kind of dummy loading when the unit power button is in the off position. It has not been there before, since I've always used all pre-amps powered on. Initially I was worried about having these switching PSU's, and now I might indeed have to find another solution.

And my ears are fine, thank you.

Martin
 
Noise?

With Queen Townsend in the studio, whats the point?

V76 is best spy chassis on the scene, omit feedback, open loop is s/n heaven, all in the in put, it's always the input balance for s/n, look at WE 117a torroid, awesome rejection

low noise = mu metal

buy old truiad or peeless triple shield moving coil


i have a dc household, compuers are ultra noisy i have noticed since goinfg dc.

refrigerators are ultra noisy, so i un plugged it.
nice ti not have that cycling BS going on, what a difference.
 
You can try super high CMRR transformer and 2SA1191 or THAT3xx building a mic preamp if you wanner "ultra low noise".

or , you can try all transistor but opamp (except AD797)

you must building high quality PSU if your circuit have low PSRR.

easy to test : measure output noise at 600(200)ohm load in 60dB gain, no need source. I misdoubt EIN lah```

Enjoy :grin:
 
[quote author="CJ"]Noise?

With Queen Townsend in the studio, whats the point?

V76 is best spy chassis on the scene, omit feedback, open loop is s/n heaven, all in the in put, it's always the input balance for s/n, look at WE 117a torroid, awesome rejection

low noise = mu metal

buy old truiad or peeless triple shield moving coil


i have a dc household, compuers are ultra noisy i have noticed since goinfg dc.

refrigerators are ultra noisy, so i un plugged it.
nice ti not have that cycling BS going on, what a difference.[/quote]


old permalloy noise higer than now . :wink:

Enjoy
 
Transformers now? Interesting, I suspected that might be one way to go. It's been a busy week and I haven't been able to get to work on this project as I had hoped, but any day now...

Will certainly try that MC transformer idea, should be a good fit with regards to the impedances. The more I work with microphones, the more sense it makes to have a pre-amp designed for one thing only, having a LOT of clean gain. Fact is, even 90dB or so would be useful.

Combining the teachings of you guys, it might be an idea to run the pre-amp on batteries. Only need short tests anyway, not hours of use.

Then there is this other useful special pre-amp, which is exactly the opposite, able to deal with huge levels without distortion. But wait, that's only a line level pad with phantom power... :)

CJ, I unplugged my parent's freezer for a recording session some years ago. Forgot to plug it back in... Be careful, it was a mess. Are you talking about Brian & Pete? The Ox aka Thunderfingers was my hero.


Martin
 
Transformers now?
They don't work well for instrumentation because the frequency response and exact gain is unknown (especially with varying and/or complex source impedances). In addition to this, they allways add thermal noise (though the contribution might be low) and have low input impedance (which is not what we want).

Fact is, even 90 dB or so would be useful.
Simply use a second pre-amp in series--it won't add any significant noise. Or add a 20 dB noninverting gain stage at the output if you want to use my suggested design.

Samuel
 
Back
Top