Simple transconductance amp powered by 48VDC@2mA?

Hi, I am trying to hook up some telecoms headsets to an audio mixing desk to make a 4-person intercom system on the cheap! I have a pro audio engineering background, but my analogue electronics is pretty rusty.

I've made some interface boxes that simply convert the 4-pole TRRS headset jack to pro-audio connectors (XLR for mic, stereo jack for headphones), but the electret(?) mic in the headset can't drive the ~6m XLR cable into a low impedance mic input on the mixing desk (but headphones work fine).

Having skimmed back through my copy of The Art of Electronics (bought in 1994!), do I need a 'transconductance amp' to convert the high impedance electret output to low impedance for the balanced cable run, and will the electret be destroyed if I remove the DC blocking capacitors that are isolating it from the available 48VDC from the desk (aka phantom power)?

If a preamp is needed to drive the cable, can I power it from the phantom power (48VDC@2mA), and how simple (parts count) can it be?

Many thanks for your attention!

This is an interesting article - I didn't realize the electret case could be clipped to separate the internal FET so easily... the end application is different. Also, there are several posts around the interweb if you search electret and phantom power.

This is a circuit I designed many years ago to do exactly the same thing, but for speech recognition mics in a home automation system.
[And thanks to Willen for redrawing it in legible form, from the pencil back-of-a-beermat grade original sketch I sent him!]

Ignore "Gnd 1" and the F terminal, they were from an earlier version.

It's for a three-terminal electret mic with internal FET preamp; I got a batch of very nice Shure mics with long cables but bare capsules, at the time.
The capsule connects to A for positive / fet drain, B fet source, C ground/ casing.

The balanced output is D & E with C again as ground.

The concept is that the signal is fully balanced (signal-wise) at both the mic and the preamp output. It turns out to be very similar to the Schoeps condenser mic circuit, though I had not seen that then.

The big difference is that I used a balanced common emitter output stage, effectively modulated current sinks, rather than emitter followers.

The work very well, a mic capsule set flush in a wall can pick up voice commands from 20ft..
If it's a two terminal mic capsule, just ground terminal B; it should still work reasonably well.

[And just to be clear, it is designed for a 48V phantom powered system].

gophert said:

This is an interesting article - I didn't realize the electret case could be clipped to separate the internal FET so easily... the end application is different. Also, there are several posts around the interweb if you search electret and phantom power.

https://levysounddesign.blogspot.com/2010/09/electret-condenser-microphones.html

Click to expand...

Great find, thank you! I'm looking at fitting this into the 'adaptor box' that houses the different connectors, but haven't yet worked out how to include a SPDT switch for momentary/latched talk (ie. a switched 'mic on' signal)... Any suggestions would be welcome.

rjenkinsgb said:

This is a circuit I designed many years ago to do exactly the same thing, but for speech recognition mics in a home automation system.
[And thanks to Willen for redrawing it in legible form, from the pencil back-of-a-beermat grade original sketch I sent him!]

Click to expand...

rjenkinsgb I can't see your circuit diagram here - am I being dim?! Please let me know if you're referring to the Levy Sound Design circuit posted by gophert above, or do repost your own schematic if it didn't come through the first time, thanks.

rjenkinsgb said:

The concept is that the signal is fully balanced (signal-wise) at both the mic and the preamp output. It turns out to be very similar to the Schoeps condenser mic circuit, though I had not seen that then.

Click to expand...

I'm familiar with the (rather lovely) Schoeps mics, but have never disassembled one!

Try again? I inserted the original image directly from a private conversation on here, but thought it would be copied in to the post.

I've now downloaded it and added it back in to my previous post as a separate image - hopefully that should be properly visible...

rjenkinsgb said:

Try again? I inserted the original image directly from a private conversation on here, but thought it would be copied in to the post.

I've now downloaded it and added it back in to my previous post as a separate image - hopefully that should be properly visible...

Click to expand...

Yes, that worked, thanks!

Photos of one of the originals, if you want to build one; I made a batch originally and had a couple left over.
They were made to fit in flanged die-cast boxes for protection and screening, around 90 x 35 x 35mm not counting the mounting flange.
The BC107 is long discontinued. I'd use ZTX851 now, they are superb ultra-low-noise transistors when used in such circuits.

Photos:

Both of the circuits suggested above seem to assume access to the electret capsule (with three points of connection: positive/FET drain, FET source, and casing/ground), potentially requiring tracks to be cut to separate the FET source and ground from one another, but I am looking at the two contacts on the headset TRRS jack plug that relate to the mic alone... I could conceivable break open the headset mic boom to access the electret inside but it will be tiny, not the fingernail-sized capsule you'd get from a Panasonic/Shure/Schoeps mic.

Can I simply ignore the electret 'ground/case' connection shown in these circuits and connect the mic's 'positive' and 'negative/common' (as found on the headset jack plug) with the points labelled 1 & 2 (in the Levy Helmke design) or A & B (in the rjenkinsgb design)?

I'd use A for signal and C for mic ground, in my version. Also connect B to ground.

Having the mic screen connected to an input does not seem to be a good idea to me; it would probably be OK with a bare mic capsule and direct connections but not with a cable in between, picking up noise.

In my version, you could also try directly cross-connecting the transistor emitters? That would convert it to a "long tailed pair" type configuration and it should then still give a reasonable balanced output from the single-ended input.

Looking at the bottom of the stripboard, the rightmost used track is ground [C], the two connections immediately to the left of that and just above / below centre are A & B and the leftmost used tracks above/below centre are the outputs.

rjenkinsgb said:

Photos of one of the originals, if you want to build one; I made a batch originally and had a couple left over.
They were made to fit in flanged die-cast boxes for protection and screening, around 90 x 35 x 35mm not counting the mounting flange.
The BC107 is long discontinued. I'd use ZTX851 now, they are superb ultra-low-noise transistors when used in such circuits.

Photos:
View attachment 125975 View attachment 125976

View attachment 125977

Click to expand...

Thank you, these are very helpful. Oddly, BC107 is still available (and in stock) at CPC Farnell!

Yes, you need a preamp circuit for the mic. Think of the mic capsule as a 10 mV voltage source with a relatively high output impedance, that needsa few volts of DC bias to function. There are a zillion mic preamp circuits on the innergoogle, ranging from basic ( 1 transistor plus 2 resistors ), to a full-blown low noise "audio opamp" based audio console input stage.

I recommend against opening the mic capsule; partly because of your skill set, but mainly because there just is no valid reason for it. Let the mic be the mic; dealing with its quirks is not at all difficult.

Here are grabs from a search for "electret peramp circuit". On the left, note that R1 and C1 will be a part of any electret preamp, to bias the cartridge with DC and then remove that DC from the audio signal. The right circuit is more "audio quality", and takes a slightly different approach to DC biasing.

ak

rjenkinsgb said:

I'd use A for signal and C for mic ground, in my version. Also connect B to ground.

Click to expand...

Got it.

rjenkinsgb said:

Having the mic screen connected to an input does not seem to be a good idea to me; it would probably be OK with a bare mic capsule and direct connections but not with a cable in between, picking up noise.

Click to expand...

Yes, that occured to me, but there's no space inside the headset to mount the preamp, so if i can find mic +/- and a separate screen then I will use those three connections, although it means butchering the rather delicate wiring inside headsets like this:

rjenkinsgb said:

In my version, you could also try directly cross-connecting the transistor emitters? That would convert it to a "long tailed pair" type configuration and it should then still give a reasonable balanced output from the single-ended input.

Transistor Long Tailed Pair » Electronics Notes

The transistor long tailed pair circuit used to provide a differential amplifier function within many electronic circuit designs.

www.electronics-notes.com

Looking at the bottom of the stripboard, the rightmost used track is ground [C], the two connections immediately to the left of that and just above / below centre are A & B and the leftmost used tracks above/below centre are the outputs.

Click to expand...

That's a good shout if i'm unable to separate the mic's negative signal connection from the ground/common/screen connection.

AnalogKid said:

Yes, you need a preamp circuit for the mic. Think of the mic capsule as a 10 mV voltage source with a relatively high output impedance, that needsa few volts of DC bias to function. There are a zillion mic preamp circuits on the innergoogle, ranging from basic ( 1 transistor plus 2 resistors ), to a full-blown low noise "audio opamp" based audio console input stage.

Click to expand...

Lesson #1 was to learn that electrets need a few volts to bias them!

AnalogKid said:

I recommend against opening the mic capsule; partly because of your skill set, but mainly because there just is no valid reason for it. Let the mic be the mic; dealing with its quirks is not at all difficult.

Click to expand...

That's good news to me because the sort of headsets I'm working with are not designed for ease of access/servicing/modification.

AnalogKid said:

Here are grabs from a search for "electret peramp circuit". On the left, note that R1 and C1 will be a part of any electret preamp, to bias the cartridge with DC and then remove that DC from the audio signal. The right circuit is more "audio quality", and takes a slightly different approach to DC biasing.

Click to expand...

I must admit I am far more au fait with opamps than transistors, so I 'get' the second circuit!

Thanks for chiming in, it's appreciated.

andrewclark1 said:

Oddly, BC107 is still available (and in stock) at CPC Farnell!

Click to expand...

Amazing - I thought they were virtually extinct now!

Re. the mic ground.

As it's part of a headset and the ground is shared with the earpieces, it will already be connected to an equipment ground.

Using the A and B connections could be advantageous in that case, as you are then getting the mic signal without adding possible hum loops or cross-connecting grounds from different items.
Try the linked emitters as well, as that may improve the output.


This circuit does provide some voltage gain as well, unlike the Levy PNP Emitter follower buffers; roughly 7x assuming the mic input on the equipment uses the standard 6.8K load resistors to feed phantom power [the ratio of the emitter resistors to load resistors].

rjenkinsgb said:

Re. the mic ground.

As it's part of a headset and the ground is shared with the earpieces, it will already be connected to an equipment ground.

Using the A and B connections could be advantageous in that case, as you are then getting the mic signal without adding possible hum loops or cross-connecting grounds from different items.
Try the linked emitters as well, as that may improve the output.

This circuit does provide some voltage gain as well, unlike the Levy PNP Emitter follower buffers; roughly 7x assuming the mic input on the equipment uses the standard 6.8K load resistors to feed phantom power [the ratio of the emitter resistors to load resistors].

Click to expand...

Again, that's very helpful, thank you. It looks like I have a few variants to try out, so I have just ordered parts online.

andrewclark1 said:

That's good news to me because the sort of headsets I'm working with are not designed for ease of access/servicing/modification.

Click to expand...

Everything you want to do can be done at the connector. No need to cut open anything, splice into a cable, etc.

A typical connection for TRRS is shown in the left-side schematic. R1 and C1 both connect to the Sleeve (the mic connection). The other end of R1 goes to 3.3 V or 5 V, and the other end of C1 goes to the preamp input. For 3.3 V, I'd decrease R1 to something in the 2.7K - 4.7K range.

ak

AnalogKid said:

Everything you want to do can be done at the connector. No need to cut open anything, splice into a cable, etc.

Click to expand...

That's music to my ears!

If you don't want to hand solder a circuit, Adafruit has microphone+preamp modules, one with AGC and one without. You can remove their mic and run (short) wires to the TRRS jack.


ak

AnalogKid said:

If you don't want to hand solder a circuit, Adafruit has microphone+preamp modules, one with AGC and one without. You can remove their mic and run (short) wires to the TRRS jack.

Electret Microphone Amplifier - MAX4466 with Adjustable Gain

Add an ear to your project with this well-designed electret microphone amplifier. This fully assembled and tested board comes with a 20-20KHz electret microphone soldered on. For the ...

www.adafruit.com

ak

Click to expand...

That's good to know, but personally I find most of the enjoyment is found in understanding the principles and building a solution, albeit needing a push in the right direction from those with far more expertise! Thank you for the tip, I might fall back on the Adafruit options if I need to expand the number of headsets in future.