Dumb idea for a mic circuit?

[tk@halmi]

Forgive me for pointing this out, but to me the asymmetry is visible if you compare the top and bottom half of the output sine wave. I am referring to the shape, not the magnitude.

 

[OneRoomStudio]

Forgive me for pointing this out, but to me the asymmetry is visible if you compare the top and bottom half of the output sine wave. I am referring to the shape, not the magnitude.

That's fair. There may be some asymmetry resulting in some even-order harmonics, but the same could be said about almost every "classic" microphone. As long as it's subtle and no hard-clipping, I'm fine with it - especially since this is with a 500mV input, which is much higher than the circuit would normally see.

 

[inkster]

As a young engineer/tech/tinkerer I marveled at the beauty of phantom power...an elegant solution to the problem of powering a condensor mic without an inconvenient power supply to set up, only problem back then I couldn't afford any of the mics I dreamed of owning but I was able to get my hands on a Senn T-powered mic.
That helped jump start my fascination with mics once I built my own 48V to 12V adapter box thingy.

How would a +/- voltage potential safely and reliably work utilizing a 3 pin XLR?...seems to me an out of phase cable or connection would cause issues resulting in a need for protection diodes or a circuit to resolve the potential for external wiring errors inside the mic.

I'm a huge fan of well designed higher voltage circuits like those found in the 5088 console and some older solid state designs but my 2 cents would include a DC-DC circuit to bring the 48V up to the desired voltages, which of course is already widely used but there is always room for innovation and experimentation.

Build a working model, I would love to see the results!

 

[jp8]

That's fair. There may be some asymmetry resulting in some even-order harmonics,

Did you make an FFT of this waveform already? For most professional condenser mics, max input and output levels are specified at 0.5 % THD. On an oscillogram, THD often becomes visible only at levels above 1-2%. As the distortion is very obvious, I'd expect way higher THD. Could be exactly what you want if you'd like to add some "color", but if you want to compare apples to apples, you should specify the input and output levels at 0.5% THD.

I'd suggest to take a more realistic load to the circuit. Most preamps have an input impedance of 1-3k. A 10k load gives too rosy a picture of the distortion and max input- and output levels.

resulting in some even-order harmonics, but the same could be said about almost every "classic" microphone.

Sure, any circuit will generate lots of distortion if you drive it hard enough. But to say that every classic microphone generates the same amount of distortion at the same input level is just not true. There are many clasdic designs, from properly designed and biased Schoeps circuits to OPA Alice or Rode NT5 that generate less than 0.5% THD at 500mV input. Some can handle several Volts at their input. If you add a feedback capacitor as I mentioned in one of previous posts, I'm sure you'll improve the performance considerably!

Jan

 

[abbey road d enfer]

For some reason, I get different results.
Here I compare the outputs with 0.5V and 1V at the input?
With 0.5V distortion is visible, and with 1V clipping is evident.
Too much uncontrolled gain.

 

[jp8]

For some reason, I get different results.
Here I compare the outputs with 0.5V and 1V at the input?
With 0.5V distortion is visible, and with 1V clipping is evident.
Too much uncontrolled gain.
View attachment 140871

Yes, without any feedback applied, it's just an open-loop CS amplifier with Gain governed by gm and load impedance only. With gm all over the place ( device-to-device variations and as function of bias current) and a frequency dependant load, gain will be ill-defined.

Jan

 

[moamps]

View attachment 140871

Maybe I'm not seeing well or I'm already that old, or it's too late, but considering that IMO the gain of the circuit cannot be calculated if the active parts are not biased into the active region where they work as some kind of amplifiers.
So, if the gate voltage is 0V, for this type of FET, the source cannot be at a more negative potential if we want to place it in the active linear region.
At the same time, if the base of the BJT is at 0V, the emitter must be at a lower potential, around -0.65V.
These are two contradictory requirements.
This means that IMO this circuit as it is does not work as a linear amplifier at all. Not to mention what the input impedance of the circuit is if VGS is positive.
But I may be wrong... Also, forcing the current of that FET to about 13mA does not seem optimal to me.

 

[OneRoomStudio]

Well, if nothing else, I now know that the answer to my topic question is a resounding "Yes."

A few things to answer some questions though:
-I never intended to feed this power via 3-pin XLR. This would need at least 5 pins, like a tube mic.
-This wasn't intended to be as linear and distortion-free as an Alice or Schoeps (I'm personally a fan of mics with a bit more color)
-I wasn't shooting for "perfect," just something different.

Clearly not a direction worth pursuing though. Thank you all for taking a look.

 

[k brown]

I've really liked this circuit for "something different" in a mic (constant current source).

B1 Buffer, designed by Nelson Pass as a no-gain HiFi 'preamp'.
https://www.passdiy.com/project/preamplifiers/b1-buffer-preamp

I adapted it for use in microphones by replacing R103/203 with 1G, eliminating everything before the FETs (replacing with 1G +820pf, for capsule polarization) , adding an impedance-balanced output, and providing power from a pack of five 9v batteries supplying polarization, and tapping off at 18v for the FETs.

Despite many saying that 2SK170 is ill-suited for use in microphones, this circuit has performed beautifully with Primo EM21/23, and Oktava '012 capsules.

 

[tk@halmi]

There are changes you could make to have better control over biasing and gain.
- Bias Q2 using a common two resistor voltage divider.
- Tame gain with partial bypass of the source resistor (just add a smaller resistor in series).
- Reduce current a bit to give those transistors a little break.

 

[ricardo]

The nice part about using +/- 60V is that it would make polarizing the capsule a piece of cake. The backplate could be grounded, +60 could be applied to the front capsule, and the rear capsule could be switched between +60V, 0V, and -60V for omni, cardiod, and figure-8

If you want +/- 60V, the voltage multiplier in the usual Vp circuit lends itself to generating +/- Vp easily.

But what is the advantage? There are better methods for variable pattern which, because for cardioid only one diaphragm is active, that pattern can have less noise.

One better simple circuit is Zapnspark's Simplified Switched Pattern Condensor Microphone S2PCM circuit v2.31 in MicBuilders. Zephyr.doc has detailed measurements of such a beast in https://groups.io/g/MicBuilders/files/Mic Measurements. You have to join.

 

[abbey road d enfer]

There are changes you could make to have better control over biasing and gain.
- Bias Q2 using a common two resistor voltage divider.
- Tame gain with partial bypass of the source resistor (just add a smaller resistor in series).
- Reduce current a bit to give those transistors a little break.

Biasing the FET's gate to a negative voltage achieves this. I've simulated it. It doesn't fully solve the issues of uncontrolled gain (too dependant on FET's characteristics), limited headroom (which is ironic for a high-voltage circuit), and poor distortion performance.

 

[Cqwet Dbdfte]

Many mic circuits are voltage followers, with no gain.
Gain can be added later. Low noise would be a design priority.
The transformer values in the simulation should be for inductance not impedance.
645 Henry would be hard to fit inside a mike.

 

[abbey road d enfer]

Many mic circuits are voltage followers, with no gain.
Gain can be added later. Low noise would be a design priority.

+1

The transformer values in the simulation should be for inductance not impedance.
645 Henry would be hard to fit inside a mike.

I believe they are inductances. Simulation softwares understand only nominal inductance.
Anyway 645 ohms would be rather low for the primary.
OTOH 645H suggests a nominal impedance of >50kiloohms, which is unnecessary and very hard to produce in reality.

 

[OneRoomStudio]

They are inductances, and I took them from the product page: https://utmindustry.com/utm0547/

 

[Cqwet Dbdfte]

They are inductances, and I took them from the product page: https://utmindustry.com/utm0547/

Nice!

 

[abbey road d enfer]

They are inductances, and I took them from the product page: https://utmindustry.com/utm0547/

There's something wrong with these figures.
645/8= 80.6, which results in a voltage attenuation of 18dB, which is correct, and voltage ratio of 9:1 and an impedance ratio of 16.6k:200, which is not what the specsheet says.
Probably a copy-paste error.

 

[ccaudle]

especially since this is with a 500mV input, which is much higher than the circuit would normally see.

That is only true for distant placement. Close mic'ed instruments can produce output much higher than that (see KingKorg's posts on a capsule buffer amp using an op-amp running on 36V).

 

[OneRoomStudio]

That is only true for distant placement. Close mic'ed instruments can produce output much higher than that (see KingKorg's posts on a capsule buffer amp using an op-amp running on 36V).

Yes, it is possible to generate much higher voltages....but most microphones will see (on average) much lower voltages. There aren't many (any?) mics out there than can swallow a multi-volt signal without distortion. Again, the goal here was never to design a perfect microphone capable of any and all things.

 

[OneRoomStudio]

There's something wrong with these figures.
645/8= 80.6, which results in a voltage attenuation of 18dB, which is correct, and voltage ratio of 9:1 and an impedance ratio of 16.6k:200, which is not what the specsheet says.
Probably a copy-paste error.

@igs Are the Primary/Secondary inductances for UTM0547 correct? The webpage states 645H & 8H.