Why do electret microphone capsules need such high resistance?

[carboncomp]

Thanks to the amazingly helpful people here I have been looking at a bunch of DIY mic designs, and thinking I may be better off starting with a cheap electret build, to learn and break things. But I was wondering if someone could explain just why electrets need a 1G resistor across them. To be frank, I didn't even know they went that high till this week.

 

[gyraf]

on electret with built-in FET, you won't need the very-high resistances

 

[Chrisfromthepast]

Check out minute 15 of this video for a brief analysis of what happens without the 1G resistor, which relies on parasitics, and then the noise implications of setting that value as high as possible

Note this is for a capsule like tsb-2555, without the built in fet.

One takeaway that may not be obvious: on your first mic build, you may underestimate how thoroughly you need to clean the high impedance section of the pcb. If you have noise, check for flux residue.

 

[kingkorg]

Thanks to the amazingly helpful people here I have been looking at a bunch of DIY mic designs, and thinking I may be better off starting with a cheap electret build, to learn and break things. But I was wondering if someone could explain just why electrets need a 1G resistor across them. To be frank, I didn't even know they went that high till this week.

It doesn't have to be across the capsule, it is there to bias the gate, it can be connected to gate and bias voltage for example.
However to simplify things iet's say it goes to the ground. Capsule is a capacitor, usually about 50pf. You can just imagine AC(sound) goes through it. It forms high pass RC filter with that 1G cap. You lower the value of the resistor and cut low end.

 

[abbey road d enfer]

The main reason for having the highest value possible is not primarily to extend the LF response, it's to decrease perceived noise.
Math show that so-called KTC noise increases with the resistor value, so why would it result in less noise?
The reason is that although the noise energy increases, the audible part of it decreases. Noise density increases at a rate of 3dB for each doubling of the resistor, but, for a given capacitance, the response at audible frequencies decreases by 6dB, by shifting the cut-off frequency of the low-pass filter constituted be the resistor and capacitor.

Actually, built-in FET capsules have a rather large impedance. Most of the specs range about 600r to a few kohms. In fact, it's the power feed resistor that governs the specified impedance.
The capsules where the FET operates in their pentode region have a impedance of several tens of kiloohms, those that operate under very low voitage operate closer to the triode mode, with a slightly lower impedance.

 

[abbey road d enfer]

Check out minute 15 of this video for a brief analysis of what happens without the 1G resistor, which relies on parasitics, and then the noise implications of setting that value as high as possible

I'm a tad surprized at this. I may need to be corrected, but I have always assumed "1G-less" mics used a dedicated FET comprizing bias diodes on the same substrate. The 1G resistor being replaced by back-to-back diodes, the reverse leakage providing the necessary galvanic connection.
Relying on unspecified gate leakage seems to be a quite adventurous practice IMO.

 

[moamps]

I stopped watching that video when I saw that person using a wrong symbol for FET and how bad he draws schematics.

 

[Khron]

"Wrong" symbol? So a "rough approximation since everyone involved is fully aware what we're talking about" is... unsatisfactory?

 

[abbey road d enfer]

"Wrong" symbol? So a "rough approximation since everyone involved is fully aware what we're talking about" is... unsatisfactory?

I agree; this approximation is not cause enough for condemning the guy. It's the rest, the pomposity, the way the compère bows in admiration, that make me puke.

 

[moamps]

Sorry, but it shows me that person doesn't know exactly what he's explaining. So, for someone who probably knows what he is talking about there is no need to use "approximations" and especially in the context of explaining diodes leakage currents and uses a symbol (from another type of transistor) where the drain is isolated from the gate.

 

[kingkorg]

I would have not to agree, i don't think we should be nitpicking here. That series of videos is quite a treasure, Doug talks about a lot of things not mentioned elsewhere. Not sure if you know who he is, but his designs are quite unique in world of mics, and he did design the quietest readily available mic. He certainly knows what he is talking about.

I see it as two guys chatting about stuff, and camera just happen to be there. Doug has been away from mic business for quite a while, and seems to be putting things on the board off the top of his head as he remembers them.

I would love if some people from Akg, or Neumann or whoever, were willing to make something similar.

 

[moamps]

I would have not to agree, i don't think we should be nitpicking here. That series of videos is quite a treasure, Doug talks about a lot of things not mentioned elsewhere. Not sure if you know who he is, but his designs are quite unique in world of mics, and he did design the quietest readily available mic. He certainly knows what he is talking about.

Ok, I may have been a little harsh on him, and I honestly don’t know who he is. But I still don't think the mistake about the symbol should have happened to him.

 

[kingkorg]

Sure, he might be getting rusty. He was chief designer engineer at Røde. He came up with their most popular and designs, both tube and solid state. Coolest thing is he didn't rely much on previously existing designs. Also, one of the kindest guys i've ever interacted with.