# elektret mike amplification

• posted

In the past I've tried to (and actually did it) design and build some pre-amplifiers for use with a elektret microphones and found out that inside the electret there's one ingeneous little JFET which 'makes it all work'.

Lately I was wondering however in how far this JFET is the optimal solution and whether those mikes could be improved by using a slightly more complicated circuit inside.

The case is that some elektrets

have an internal capacitance of about 10 pF and some JFETs, used inside, have a Cgs of 5 pF (or centimetre ;).

Now the way these mikes function seems to be that a pre-polarised dielectric membrane is moving to and from its backplate, attracting more and less charge to it.

This charge (the amount of it) I think contains the real 'information' that one wants to amplify. But the JFET is merely amplifying the voltage over the mike's + JFET's capacitances.

With a mike internal capacitance of 10 pF and a Cgs of 5 pF, this would mean that 2/3 of the moved charge is sticking in the elektret itself, and only 1/3 is moved to the JFET's gate, generating a voltage about 67% lower compared to if all charge would have gone into Cgs.

Wouldn't it be better to incorporate the JFET in an amplifier with a negative overall feedback loop? Then the input voltage of the JFET, and also the voltage over the elektret would be practically (for engineering purposes) zero and all the charge would be absorbed by the feedback loop. If for instance the component in the feedback loop would be an external capacitor of 5 pF and the input stage of the amplifier would be the same JFET, then wouldn't we have a true charge amplifier and a 3x times better S/R?

Actually I think the answer is yes. Do these mikes actually exist?

My guess is that you'd only need to put 1 small IC plus a 5 pF capacitor (or whatever value one might specifically prefer) inside the elektret to make this work.

joe

• posted

I believe you are mistaken on how electrets work -- I'm pretty sure that there's a little chunk of piezoelectric material in there that generates a voltage when flexed or compressed or whatever.

Beyond that, though, you may have something. Care to start a company and see if you can make it fly?

```--
www.wescottdesign.com```
• posted

Hi,

I think that an electret is different from a piezoelectric. With the electret you have a chunk of material with a fixed polarization (the equivalent of a magnet just to make an example). The electret doesn't undergo deformation and doesn't "generate" charge. It is equivalent to a voltage source which polarizes a capacitor. One plate of this capacitor moves due to the acoustic pressure and because of the polarization of the electret a signal is generated. A piezoelectric material on the other hand generates a charge when it undergoes deformation. Bye

Marco

• posted

No, sorry - you are the one who is mistaken. An electret is basically a permanently charged capacitor with a movable "plate". Sound waves move the plate which causes a varying voltage, due to the fixed charge. The charge is permanent by being embedded in a polymer.

Bob

• posted

Correct! We are now able to make electret microphones directly on a Silicon chip. ...Jim Thompson

```--
| James E.Thompson                                 |    mens     |
• posted

Back to Joe's issue... The moving plate generates a variable charge that is spread across the capacitance of both the microphone plate and the transistor gate. Joe has his numbers a bit off in that the small capacitance of the FET lowers the voltage by 1/3 leaving 2/3 of the original voltage available to operate the FET.

I'm not sure what Joe is thinking he is going to improve. The FET provides enough gain that the small loss due to the gate capacitance is not an issue that I can see. More importantly, the FET provides impedance matching from the relatively high impedance of the microphone to the amplifier input.

```--

Rick```
• posted

You can't produce a persistent electric field from a metal surface without applying power. Air ions and leakage cause current to flow until the external field goes to zero.

Insulators and even semiconductors are a somewhat different matter, because all those field-neutralizing electrons need quantum states to sit in, and there sometimes aren't enough.

However, to make an electret microphone you use PVDF film with metal on both sides, so there are lots of states available and therefore no external field. PVDF mics are piezoelectric, not capacitive.

Cheers

Phil Hobbs

• posted

As I understand it, here's a long tradition of modifying these electret capsules, rewiring the JFET circuit configuration (converting from a common-source configuration to a source-follower configuration).

The advantage is significantly reduced distortion, especially at higher output levels.

The disadvantage is lowered output and a poorer SNR.

It's commonly referred to as the Linkwitz modification; Linkwitz credits it to his former colleague Lyman Miller.

Unfortunately it doesn't appear that these capsules are constructed in a way which would provide easy, direct access to both sides of the electret element... although I suppose one might run them on a reversed voltage polarity so that the JFET gate was forward-biased?

• posted

Because the PVDF mic is piezo & not capacitive, it is not an electret. An electret is capacitive.

Bob

• posted

Thanks :)

Exactly, and that's what carries the information: the variable charge. If you could lead all that charge through feedback capacitor you would have:

1) a 3x higher S/N-ratio (sorry, I miss-typed that wrong in my OP), 2) a much lower output impedance, 3) less distortion from non-linearity of the JFET's junction.

You have to think it the other way around: The elektret's internal capacitance lowers the voltage by 2/3, and 'keeps' 2/3 of the charge from moving to the JFET's gate.

a 3 x higher output signal with the same feedback capacitance as is Cgs, (roughly) the same noise--the additional capacitor and the noise on the output of the circuit (OpAmp?) add noise--hence a 3x improved S/N-ratio.

If I assume myself correct, then the loss is twice as much as you see it. And it's not about 'enough gain to make it workable'. Of course it works as it is done now. I'm only thinking that it might be able to be improved by a factor 3, both in signal as well as in S/N-ratio.

So does my proposed circuit. :)

joe

• posted

Go to the 18 minute mark for a bit about how and why the noise rolls off in the (fet/driven capacitor/resistor) eh, mic circuit.

Mikek

• posted

An 'electret' is supposedly the electrical analogue to a permanent magnet, but it isn't so.

There are no free magnetic 'charges', aka monopoles, so a magnet can just sit there producing a large B field in the surrounding space. The electric-field situation is quite different, because the world is full of electric charges. These are free to move around to minimize their energy, i.e. to neutralize all those external fields.

It's quite true that an electret has a frozen-in polarization, much like a permanent magnet. What isn't true is that this produces a persistent E field in the surrounding air/vacuum/whatever.

As I said earlier, this has to be modified somewhat in the case of nonmetallic surfaces, but AFAICT all PDVF microphones have two completely metallized surfaces.

They're piezo mics, not capacitive ones.

It's funny that this is such an emotional issue for some folks, as we've seen before on this NG.

Cheers

Phil Hobbs

```--
Dr Philip C D Hobbs
Principal Consultant ```
• posted

My proposed modification would involve more components (1 IC and 1 capacitor).

The charge-to-voltage amplifier that I propose would do that even better...

...with a higher output and a better S/N-ratio.

You can call my proposal the "joe's wits modificaton" :)

No idea whom to credit with this proposal. Maybe Ernst H. Nordholt, who wrote this book that I have studied some time ago...

He had the funny and a bit extravagant but proven-to-be-correct idea that the best pre-amplifier (at that time) for a moving coil pick-up element would be obtained with a power transistor as its input stage. His reasons were that this gave the lowest base resistance, hence lowest input noise voltage and a low input noise current contribution due to the low hfe. Of course his circuit also involved an overall negative feedback loop. Disclaimer: All of this is 'IIRC'.

joe

• posted

One common electret mike has a thin metal diaphragm that moves, near a rigid metal-electret backplate that doesn't move.

That doesn't look like it can be piezoelectric to me.

```--
John Larkin         Highland Technology, Inc
picosecond timing   precision measurement  ```
• posted

If the top layer of the PVDF isn't metallized, the situation isn't as clear, because of the lack of states for the neutralizing electrons to occupy.

Did that picture come with any detailed discussion?

Cheers

Phil Hobbs

```--
Dr Philip C D Hobbs
Principal Consultant ```
• posted

I've got two simple models of a piezo, the first is a (pressure dependent) voltage source, with a series capacitance. The other is a current source w/ parallel C. (Thanks to Norton.) To me the current source is more physical, but I have a better feel for voltage sources.

George H.

• posted

No, but that structure seems to be common.

and the type I posted the pic of looks like the "back electret" type.

and here's a patent for a rigid back plate electric mic.

I don't understand how the electret maintains a field in the gap, but it seems to.

```--
John Larkin         Highland Technology, Inc

lunatic fringe electronics```
• posted

Why don't your build your circuit and test it? You can buy electret mics with no FET.

```--

Rick```
• posted

Methinks there are two different types. Capacitive and piezoelectric. See illustrations: Also, piezo mics generate their own voltage and would not require being pre-charged as in the capacitive flavor.

```--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com ```
• posted

Use the electret mike as a variable capacitor, to FM an LC oscillator.

```--
John Larkin         Highland Technology, Inc

lunatic fringe electronics```

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.