elektret mike amplification

That's a peizo transducer, not an electret.

There are just 3 time honoured methods: a vacuum tube, a FET or an RF oscillator plus FM detector circuit.

The latter being far the best in terms of noise and insensitivity to moisture.

... Phil

The gap between the "plate" elements changes. dC/dt ~ V, roughly.

It's fine. They drive a circuit that has to work with dynamic mics.

Heh. May be. 3x better S/R is questionable - resistors up to gigohm range are common in mics. Just the shot noise alone ...

The point of those electrets are that they are cheap. Look at the SNR specs for a Schoeps mic - they're quite good and also very not cheap.

You'll find that room tone ( the noise from just the air in the room plus reflections ) is louder than everything else anyway.

It's not that it's spectacularly new... The same technique is used in photo-diode amplifiers sometimes. With a resistor as feedback element.

I think the right order would be to first see whether it works as expected. Then I'm sure I won't start a company, but be my guest...

joe

I'm afraid some people keep confusing piezo with electrets.

en.wikipedia.org/wiki/Electret_microphone

joe

Hehe, one probably could. :)

joe

Maybe I should retract that statement...

joe

}snip{

I understand the plastic is melted and mixed with polar molecules, the let cool down in a strong electric field, and then metallised. Would that be able to keep the contamination out?

joe

With electrets, this happens; it just takes a very long time. West and Sessler used a metallized Teflon foil for the first ones at Bell Labs ( yes, I cheated and looked - couldn't remember the details ).

The charge on the back plate doesn't actually move much. The other plate is just a dielectric, possibly also sputtered or otherwise metallized for reasons I'm not sure of.

I'm probably forgetting some detail or another. The JFET has really high input Z...

I'm not sure an external field is required... don't capacitors in general have 0 field?

At any rate, I'd still class them as condenser mics because there's two plates, one's moving and this causes the signal.

That is interesting, though.

Metaphorically, yes. Literally, no.

Capacitors in general have no external field, to my limited understanding.

But there exist electret microphones that are not PDVF microphones, correct?

Indeed, they appear to be generally deployed as contact mics. Thanks for the reference; I'd never heard of those. Those look interesting.

Those are "front electrets." Back electrets were more common because you could use anything that was a dielectric for the diaphragm, which opened up more design choices.

Or they could have buggered the drawing.

I recall back in the '80s repairing a Sennheiser MKH816T shotgun mic. Even after I traced the circuit, it still took a while to work out that the capacitance of the mic capsule controlled the frequency of an 8MHz oscillator, and this was followed by an FM discriminator. I have to say that coming across a Foster Seeley circuit (which looks a bit like a full wave rectifier with a diode on each leg of a centre- tapped inductor) when you're expecting to find an audio amplifier, is rather weird.

I can't find a schematic online, so I'm trusting my memory from almost 3 decades ago.

Regards, Allan

The mic capsule does not directly modulate the 8Mhz xtal oscillator, but rather the tuning point of the FM discriminator instead, with the same outcome.

Impedances are kept low and hence so is the noise level with no need for any DC bias on the capsule.

.... Phil

Thanks for that. I see it uses a ratio detector rather than a Foster Seeley discriminator. I guess my memory isn't as good as I thought it was.

Regards, Allan

And the capacity does not change in a linear way, IIUC.

I don't see a resistor in my proposal.

I was trying to approach it a little more fundamental than that. :)

joe

They depolarize internally, which is slow at room temperature. The air ion thing is pretty fast by comparison, but the charge has to have somewhere to go, i.e. an empty quantum state at the right energy. Metals always have th ose, but dielectrics and semiconductors often don't.

Cheers

Phil Hobbs

Where is that noise coming from?

My proposal involves using the very same JFET that's already in the elektret microphone so I simply don't understand why that would increase the noise. Can you explain how this works?

With respect to the RF oscillator I can see (and Phil Allison already showed it) how one could use that in a 'real' condenser mike, but not how this should be done with an elektret.

BTW I found a comparable approach in this paper:

and they do it much simpler than I had anticipated: the JFET _is_ the whole pre-amplifier and the feedback basically is a capacitor from drain to gate. I'm sure however that if you'd increase the loop gain by adding 2 more stages the voltage swing on the gate can be further reduced, thereby increasing the sensitivity a little more. Stability can then still be assured by placement of local poles and/or zeros. A technique Nordholt explains in his book.

They get a noise of 20 nV.Hz^(-1/2) @ 1 kHz. I get the feeling that if one would replace the PVDF/TrFE foil with a teflon, mylar of PTFE electret foil, one would get (almost) exactly what I was after (except for the loop gain).

How does moisture affect an electret microphone?

joe

That's not cheating, that is studying or refreshing.

I feel that as a little over-abridged. In a non-polarised condenser microphone like that Sennheiser MKH 105 they use the capacitor's change in capacity to modulate a frequency. In both the externally polarised condenser microphone as well as the electret microphone they use the varying voltage, but what exactly (physically/fundamentally) causes that voltage to vary, I wonder?

My take is that the electret's polarisation causes an electric field between the plates electret diaphragm and the opposite capacitor's plate. This field causes this opposite plate attract or push away electrons, hence it gets polarised also. When the diaphragm moves, the distance between it and the opposite plate changes and with that the electric field (E) changes, which in turn causes the number of excess electrons (or 'holes') on the opposite plate to change. That leads to a movement of charge. Either exclusively through the capacitor wires if it is shorted, or through polarisation of the opposite plate, including Cgs, leading to a change in voltage. Of course a combination is also possible if the electret is not really shorted, as in a nfb amplifier with low loop gain.

My hunch was that if you were able to move all the charge through the amplifier that you would get a roughly 3x better S/N-ratio and more sensitivity.

One could easily reason why that would be the case, at least with the components that I presented in my OP (electret 10 pF, Cgs 5 pF) because using a nfb amplifier with 5 pF feedback capacitor would immediately give a 3x higher voltage and the same noise (unless someone explains me why that would not be viable) if you use the same JFET as input stage.

joe

Right, sorry.. I'm thinking the model for them may be similar, so they get lumped together in my brain.

George H.