Designing a High-Quality Universal XLR Mic Circuit

[windorey]

Hello everyone! This is my first post here, apologies if I'm not up to speed with the forum standards.

After discovering wonderful low-noise capsules like the EM272, I set out to build an XLR microphone using them. However, while researching how to connect a 3.5mm mic or a 2-pin electret capsule (common source configuration) to XLR, I noticed that most existing solutions fall short in several key areas:

• Compatibility with 2-pin mic capsules
• Phantom powered operation
• Properly impedance-balanced outputs for CMRR
• Low output impedance to avoid signal attenuation
• Low self-noise
• High headroom
• Low distortion

Turns out, a circuit that meets the afformentioned criteria would also serve as a pretty solid universal mic circuit for:

• Mics with an integrated JFET (e.g. EM272, cheapo 3.5mm mics)
• Internally polarised electret microphones with an external JFET
• Condenser microphones with external JFET (with the addition of a polarisation voltage generator)

A good example to illustrate existing products pitfalls is the RODE VXLR+:



Neglecting minor details, the output impedance of the hot line is 2.2k. A typical phantom pres input impedance is ~2k, attenuating the signal by ~6dB.
Furthermore, the impedances at XLR 2/3 are not matched, making the balanced XLR connection pointless. The self-noise is hard to judge without owning it, so I won't say anything.

I'd love to hear any ideas! Please don't be put off by my long-winded post, any and all replies are welcome.

 

[Khron]

If we're talking electrets with built-in JFETs, hard to go wrong with "ye olde" Alice:

https://audioimprov.com/AudioImprov/Mics/Entries/2015/4/23_Basic_FET_Microphone_Circuits.html

 

[windorey]

If we're talking electrets with built-in JFETs, hard to go wrong with "ye olde" Alice:

https://audioimprov.com/AudioImprov/Mics/Entries/2015/4/23_Basic_FET_Microphone_Circuits.html

Very thrilled to have such a quick reply, thanks a lot!
What is the noise performance like?
Sadly, this circuit seems to only work with 3-pin capsules properly, without having some sort of special shielding.
Any idea for a better circuit?

 

[k brown]

Umm - the 'Alice's TSB-165A doesn't have a built-in FET. The other circuit on that page is for 3-wire built-in FET electrets.

Also, to the OP: it's much more productive to put your efforts behind 3-wire capsules like EM273; grounding the source of a FET (2-wire capsules) is the worst way to bias it.

Are you only interested in making an omni (EM272/3); if not, the 3-wire Primo EM200 is one of the best built-in FET electret cardioids out there.
https://micbooster.com/22-primo-microphones

While suppliers' stocks last; it seems Primo has discontinued it.

 

[pmfalcman]

For built-in JFET electrets there's a minimalistic Simple P48. Works great with EM272, AOM 5024 etc.
There are its versions for 3 wire capsules too.

For external JFET electrets there's nice, simple, quiet OPIC circuit.

 

[micolas]

Hi @windorey ,
Welcome to our group!
Follow the suggestions above, or do a search on Groupdiy for "EM272"
And read, to really understand what it is about.
All this has been debated many times.
For more subtle details, colleagues are here.

 

[rogs]

I've found the EM204 (the 2 wire version of the EM200 mentioned by K. Brown in post #4) to be the best electret internal FET capsule I've come across.
I used the attached schematic to create a '2 wire electret' to balanced phantom powered interface, and find it works well.
It should also work well with the EM272 you mentioned - although (IMHO) the EM200/EM204 capsule is in a different class....

 

[windorey]

I've found the EM204 (the 2 wire version of the EM200 mentioned by K. Brown in post #4) to be the best electret internal FET capsule I've come across.
I used the attached schematic to create a '2 wire electret' to balanced phantom powered interface, and find it works well.
It should also work well with the EM272 you mentioned - although (IMHO) the EM200/EM204 capsule is in a different class....View attachment 146978

Thank you for your reply! The schematic is very interesting, it's like a schoeps circuit but with BJT phase splitter.

Might actually build up this circuit. What is the noise performance like, and how good is the headroom? I assume the circuit can very easily be modified to work with a 3-pin capsule by just using the capsule itself as the phase splitter.

 

[Khron]

I assume the circuit can very easily be modified to work with a 3-pin capsule by just using the capsule itself as the phase splitter.

Ahem...

If we're talking electrets with built-in JFETs, hard to go wrong with "ye olde" Alice:

https://audioimprov.com/AudioImprov/Mics/Entries/2015/4/23_Basic_FET_Microphone_Circuits.html

 

[k brown]

I think @Khron is directing your attention to this schem on that page (which can be used as-is with the EM200, though the two 5.6k resistors should likely be optimized for it's FET):

 

[windorey]

I think @Khron is directing your attention to this schem on that page (which can be used as-is with the EM200, though the two 5.6k resistors should likely be optimized for it's FET):

I understood it, thanks for the clarification though.
My thought process was that if I were to build this circuit for 2-wire mics I could easily modify it with minimal effort to work with 3-wire mics if I were to get such a new capsule down the line.
About the resistor optimisation, I guess it’s important to keep the drain-source voltage high enough for it to not go into low voltage gain rolloff while still giving it enough room to swing. What would be the highest DS voltage such a capsule JFET could handle?

 

[rogs]

Thank you for your reply! The schematic is very interesting, it's like a schoeps circuit but with BJT phase splitter.

Might actually build up this circuit. What is the noise performance like, and how good is the headroom? I assume the circuit can very easily be modified to work with a 3-pin capsule by just using the capsule itself as the phase splitter.

The noise performance is not bad. The capsule has a S/N ratio of almost 80dB, and the BC549 is a low noise device. The transistor is biased so that there are c.3 volts each across R4, R5, and the transistor itself.... So each leg can swing almost 3v p-p. The max SPL for the EM200 is about 127dB, so I suspect that will be the limiting factor.
There's a short (unprocessed) spoken word sample of the configuration HERE (the file is only about 3MB).

 

[MicMaven]

build this circuit for 2-wire mics I could easily modify it with minimal effort to work with 3-wire mics if I were to get such a new capsule down the line.

Risky Novice Reply No. 293864

Although I am surely playing out of my league, I believe one can wire three terminal capsules as indicated in the attached photo and simple schema I once received from Mr. Jules R who was heavily involved in OPA Alice development. The S and D (signal and drain) are connected to the same lead, with G wired separately.

I wonder, without knowing for sure, whether the simple test circuit designs which are part of the cut data sheets for capsules sold by www.jlielectronics.com might provide some clue how to connect and power FET-inclusive capsules through and XLR connector. (Yep, I am mostly guessing here - I may learn something about this, myself, and even hounded out of the principality before we are done, but I am taking the chance in the interest of personal edification!)

For what it may, or may not, be worth, I purchase little $5-8 circuits on eBay.com which have XLR connectors and work well with FET-inclusive condenser capsules. While they work best with 48v phantom voltage, they also work with 5-10v bias voltage found on computer sound cards and amateur radio transceivers, etc. (See third photo below.) Crawling even farther out on a limb, I understand BM-700/800 circuit boards are powered through XLR connectors and also work with either 48v phantom or 5-10v bias voltage - so I wonder if you might learn something useful from how they are laid out. (Yep, I am really reaching beyond my station now!)

MY POINT is, you might, maybe perhaps, learn something about powering these capsules through XLR connectors from these various circuits found in the larger marketplace. Um .. ahem ... at least I sorta think so. (The rest of the gang will either confirm or torch this message whilst I ride a rail to another venue ....)
Good luck. My road to Heck is paved with all of my good intentions! / James

 

[windorey]

Risky Novice Reply No. 293864

Although I am surely playing out of my league, I believe one can wire three terminal capsules as indicated in the attached photo and simple schema I once received from Mr. Jules R who was heavily involved in OPA Alice development. The S and D (signal and drain) are connected to the same lead, with G wired separately.

I wonder, without knowing for sure, whether the simple test circuit designs which are part of the cut data sheets for capsules sold by www.jlielectronics.com might provide some clue how to connect and power FET-inclusive capsules through and XLR connector. (Yep, I am mostly guessing here - I may learn something about this, myself, and even hounded out of the principality before we are done, but I am taking the chance in the interest of personal edification!)

For what it may, or may not, be worth, I purchase little $5-8 circuits on eBay.com which have XLR connectors and work well with FET-inclusive condenser capsules. While they work best with 48v phantom voltage, they also work with 5-10v bias voltage found on computer sound cards and amateur radio transceivers, etc. (See third photo below.) Crawling even farther out on a limb, I understand BM-700/800 circuit boards are powered through XLR connectors and also work with either 48v phantom or 5-10v bias voltage - so I wonder if you might learn something useful from how they are laid out. (Yep, I am really reaching beyond my station now!)

MY POINT is, you might, maybe perhaps, learn something about powering these capsules through XLR connectors from these various circuits found in the larger marketplace. Um .. ahem ... at least I sorta think so. (The rest of the gang will either confirm or torch this message whilst I ride a rail to another venue ....)
Good luck. My road to Heck is paved with all of my good intentions! / James

View attachment 147649View attachment 147650View attachment 147651

Hello, thank you for the reply

I think these cheap XLR circuit boards are very similar to the BM800 circuit, and are also very similar to the circuit found in post #7. If you’d like to bother, it would be great if you could try to trace out the parts and make a schematic for it!

From what I know:

The “test circuit” in the JLI datasheets is a standard JFET “common source” configuration. For 3-wire capsules, they also sometimes use “common drain”.

The Alice circuit (found in post #10) works in a rather nonstandard way, it is kind of common source and common drain combined to create a phase splitter.