Okay, we're up as far as describing a a typical old style valve microphone. yep with uh, modern Edition In the case of the Classic 2, let's go back to basic electric mics. yep as found everywhere which don't have bias, which don't have tube amplifiers. they do actually have buffers, but they don't have output Transformers Nope, none of that rubbish.

Yep, and in fact, all so these are our little 10cent mics that we find in, uh, everything. So we'll uh do uh. Now the the solid back electrode might be at ground or it might be the actual uh, hot electrode. If you like right, the diaphragm might be at ground or it might be the hot.

We don't know. the electrc to polarize the electric M mic might be on the diaphragm or it might be on the back plate. Doesn't particularly matter, we'll just draw it that way for convenience and what we've got. There is a thing where when we talk at it, we get volts out of it.

Yep. But as trans users, what what type of level we talking about? Uh, let me think, typically they're around about minus let's call it - 45 Dbv per Pascal which translates to let me seeus, 40 is 10 Mt So minus 45 is probably about I don't know 4 molts or something like that. So maybe 4 molts. Uh, get it right lad, 4 Ms at 94 or DB s P L give will take a bit of poetic license.

That could beus 35 could be- 50. Got it? Okay, so uh, if it was, say minus 40 that would be 10 m volts at 94 DB 94 DB corresponding to one pel. Incidentally, that's the usual format for mik manufacturers to specify sensitivities, so have to do the translation of that to there and that to there. Yep, Me: Okay, now what can you do with a transducer that's putting out well 10 10 on M volts with the source capacitance of maybe about five puff? Well, let's kick off by sticking it into a jet because that's what everybody does.

That's what they do. Stick them right on the back there. Y So J fet and that's pretty much all that's just about it. Yeah, now in the good old days, we actually see a discreet Bias Resistor there to hold the gate near zero.

Yep, that would be a gig. Omish resistor right? there are, well, how do you put it? two different basic electrical styles of uh, electr mic in one of the styles that that and that are all bought out as separate electrodes. Yep, the idea being that you can put a voltage on there, hang a load resistor off there, and there's your uh, ac voltage with some DC superimposed coming off there. Sure, rarely done, even though generally that particular topology offers better uh, linearity and higher SPL handling and all of that kind of stuff.

usually not. What happens usually what you get is are they using electric mics in any pro applications? Yes, yes, whereabouts uh. There are a few stage microphones which are back Electric types. Uh.

even companies like Senheiser and AKG are offering uh, stage and recording microphones which are electric. Uh to what advantage? Um, the I Think the main advantage is they're cheaper to manufacture. There's less electronics in there to go wrong. Yep, you know the bio circuitry Etc U and you can use them over a wider range of uh, incoming DC Supply Voltages Got it? Uh, One of one of the areas where you'll see lots of electrics used in preference to uh externally polarized is whenever you see a skinny goose neck with a little head on it as used in courtrooms panels uh, Parliament that kind of thing.

they're all going to be electric. Yep. and the usual players such as Shaw and saniser and AKG they are all going to be manufacturing that style of thing, right? Most of these look like that. Yep, that beasty there.

jfet uh becomes a voltage to current transducer. so it's a trans conductance device. y So all of a sudden now what you're seeing here is uh, not a given number of Uh volts per pressure. It's a current modulation per pressure that only gets turned back into a voltage once you add a load resistor doesn't work without it correct and then you AC couple off that.

Yeah, Yep. And in fact for some reason one one of the standards is to when they're specifying the sensitivity of these mics. Uh, their standard is to use a 2K I've noticed that Mhm? uh simply because it forms a common base for everybody. y with a 5 Vol Supply Yep and it just gives everybody a A A Common, a common basis, a common platform from which to work although you can operate them much lower.

Supply voltage. Yeah, now let's have a look at 1 And2 Vols Yeah with caveat with caveat: I've got one right here in my pocket runs off a single single coin cell. Mhm. That's it.

Mind you. one uh, 1 and a half or 3 something volt. No one and a half. Okay, one and a half volts.

There you go. Much depends on the characteristics of the fat. Yep, let's have a look at the current versus current versus voltage characteristics of a J Fet measured voltage. there.

Mhm current through there, zero bias on the gate, right? Uh, most of them will have a characteristic that looks like that once you get above a certain voltage about there. that thing produces a constant current. Source right? And it's in this region here that you want to operate the fet. Yep, Why? Because as soon as you start coming down into this part of their characteristic scks the gain.

The trans conductance of that just goes to put to SH Now I Went through a stage of measuring rather a lot of different electric mics that uh gain reduction Point uh varied from fat to fat from fet type to Fat type. uh, most of them, uh, were good, down to around about 1.1 or 1.2 volts. and with then plummet, there were some uh, better fets that would come down to maybe 8 Vol before they started rolling off in game. So you could run those off a single alkaline cell until they're dead.

Yes, practically. So for example. Um, let's take a typical Jfet that might, uh, have a constant current region where it draws about 350 microamps and that's a very, very, very typical figure. It is for the Jfets used in electric mics.

Let's say up here, we ' got a single cell at 1.5 Vols and let's say that that particular fit uh would maintain its gain down to say one volt just to make the numbers easy. Yep, okay, what value of resistance can we reasonably put there in order to get those conditions? Well, we've got half a Volt across it. We've got 350 microamps through it. Uh, R equals .5 Vol over.

35 milliamp. and at a rough guess, that's probably going to be about 1.4k something like that. So you might actually toss in a 1K resistor. which means that that battery can be depleted to 1 Vol +35 volt.

Yep, before it starts losing gain, Got it? Or you'd pick a fit that instead of uh, having a knee point at one vol went down to say8 Vol and maybe Drew less current Mhm. Now you can make that resistor smaller. That will allow you to flatten the battery further before you get gain roll off. Yep.

penalty. The lower the resistor there, the lower the signal gain to that point there of course. So it's all a compromise. And of course, if you got a few more volts available Beauty you don't have any problems anymore.

Absolutely. but any. I Guess anybody who's using an electric mic uh, in a new piece of equipment needs to know what voltage they've got available so they can plan what resistor they use. Yep.

So for example, if you've got a piece of Kit with a 5vt supply. okay, let's just do this again, but this time with a five volt Supply you T of head, you've got a ton of Headroom With Cates you might want to maximize the gain out of that circuit there so that you don't have to have as much processing or so that you can maximize the uh dynamic range of your ADC That follows, so you might think okay, let's make that resistor as big as we possibly can. Okay, 5 Vols 1 volt. We've got 4 volts across there.

4 volts divided by. 35 is probably about uh 12K vaguely speaking. Okay, Beauty So you go to print with that circuit. Yep, all of a sudden you come along with you know most of your microphones are sitting there.

At 350 microamps you come along and you got a Fed in there. that draws 400 microamps. Guess what? that's trying to suck? 4.8 volts across that M Okay, 400 * 12 4.8 Vols insufficient voltage there. So instead of getting all of that lovely gain, it goes yeah.

Bonk Oh okay, I'll cover my ass instead of putting in 12K I'll put in 1K Well, you can do that and it'll cover all of those, but you've robbed yourself of so much gain. Yep, So maybe pick a resistor that allows the typical voltage there to sit at about half of your rail. This almost goes back to good old bias Theory where you try and bias things at half rail half rail. That's it.

Not only will it give you the best dynamic range as far as signal Excursion goes d y but it also allows for production variations in that F current. And you want the flattest battery chemical uh, chemistry possible? Yeah as well. Now the voltage rating of those fats is generally, uh, about 10 volts. So if you were using a, say, a 12vt or 15v Supply you might want to either regulate it down a bit or resistively drop it down a bit.

Yep, or do something to ensure that the voltage there stays below around about that 10vt level because otherwise they can start avalanching or getting noisy or undesirable effects. And this is how. uh, this is like the Phantom voltage applied on like a PC microphone input. or something like that.

They will. They will have that voltage superimpose on there and you're supposed to hook up just the two wire. Not only they have the voltage which is usually around about three and a half to 5 volts and they have the resistor built in. Yes, Okay, so they've done that bit Inside the Box inside the pie.

They only want you to plug that much in. Yep. Now there's yet another caveat that you have to be aware of. Okay, how which is noise on this rail here? Yep, if this thing, if if this J Fet is within that working within that constant current region, it's presenting a very high impedance there.

Which means that any noise there gets presented immediately to there abely. So anybody who's using these from anything that might have a bit of noise there like a PC No, they're quiet. Uh, make sure you decouple the hell out of it. Yep, that's why PC mik are so crap.

They really are. They're pretty God Awful. Yeah. and sometimes it's simply because that ground is not at a quiet place.

Absolutely? Yeah. Okay, so that's care and feeding electrically of. um, electric mics. electric mics.

It's it's easy. Yep, Get the manufacturer spec on the mic to find out roughly what The Quon current is going to be Mhm and your resistor around Bat and your supply rail? Too easy. Yeah. Bob's your uncle not rocket science? No, not at all too easy.

It's Not Unusual These days to see electric mics that don't have that resistor at all? Mhm: How do you? How do they work without being biased? I'm not quite sure I Still haven't touched this myself. I've got a suspicion that they're actually relying on things like PCB leakage which is up in the tens of gigs. Hundreds of gigs. Who cares? It vaguely maintains that somewhere within haling distance of zero.

This is like trying to have an A uh opamp like this and trying to AC couple inter. where's the bias and there are any real sources of bias Are leakage? Well, actually, there's a couple of sources. First of all, there's leakage down through the gate there Y and that, in fact is a source of noise as well as bias corruption. Now, typically, if you got a little fet with little Junctions and low Capac, you're looking at PCO Amp, right? Leakages? Yeah, mind you, something to keep in mind about uh J fets is that leakage current will increase around about at the rate of about a decade per 10 Dee rise.

Oh okay, so when temperatures get high, fets get leaky. Yep. but uh, It's Not Unusual to be talking about PCO there. So if you can arrange for Uh T or hundreds of P amps leakage to ground there by having soggy circuit board or I don't know TR soggy circuit board, then you can maintain the DC voltage there.

Somewhere within Hing distance is zero. Yeah, the other thing to watch out for of course is that's at zero. We've got signal excursions either side of zero. Due to that, the negative going see a high impedance.

any positive going signal greater than maybe 10 20 30 MTS starts to see a four diode Junction down there and winds up being nonlinearly clipped. ouch. So the larger signal, the worse it Clips there. Mhm.

So that we're discussing about Headroom is a Headroom limitation right now. The other downside of leakage is it's a noise. Source Mhm. Uh, can't remember the the equation off the top of my head for uh, noise current.

but basically every dribble of electrons that goes down through that that's a genuine unit of measure. Yeah, a dribbles worth. Uh, every electron that jumps that barrier y represents an Impulse. It does.

And uh, the thing is, if you got zillions of them going at once, they kind of ey out. Uh, which is why the noise of a DC current is proportional to the square root of the current. Double the number of electrons flowing through there. and the noise part of all that, those rushing ball bearings only increases by a factor of root2 right? Uh, so it's a noise Source MH That current gets converted to a noise voltage by the impedance at the gate, which consists of that capacitance mhm that cap capacitance, that resistance.

Yep, which you change with frequency. Yeah, it is, But by far and way, generally the most important part of the noise is that resistor or pseudo resistor. Yep, Interesting thing. the higher the value of that resistor, not the higher the noise, the lower the noise.

That is counterintuitive. Counterintuitive. And the reason for that is okay. let's plot noise voltage versus frequency mhm.

For a resistor, it's basically flat. Okay, equal noise voltage per unit frequency. Okay, if we double the value of resistance, we increase the Noise by < Tk2 or 3db right? Okay, where that's 3db. Okay, why? Because uh, noise voltage equals root of KT r Yeah, so it's the fact that it's proportional to root resistance Mhm.

Okay, double a resistance increased by 3db only? Okay, but here we've got our resistor in series with its noise voltage source. Okay, and that voltage there increases by 3db there. But it's got this low pass filter attached to it, so that will only be true up to a a cut off frequency, a corner frequency determined by the value of that resistor value of the capacitor. So let's just say this one.

Okay, cut off there. Mhm. Okay, so it's Corner frequencies there. When we double the value of resistance, we shift that down to half of what it was.

So all of a sudden our noise profile looks like that. Guess what? 3db lower M So as long as we're only looking at frequencies significantly above that corner frequency. Yep, whenever we double that. resistor, we KN 3db off the noise because we're in this rolloff region.

Very nice. T And that's part of the secret to making a load Noise mik is: make any resistance as high as you possibly can yep, and make the capsule capacitance as high as you reasonably can. Totally counter insured. Who would have thought.

Yeah, so higher resistances equal lower noise Lower noise? Yeah, it's a bombshell folks. Yeah, And that's why at Road microphones I think I was the first one to go from 100 Meg type resistances up to 5 gig type resistances. and it really paid off in making low noise preamps. Awesome! The other thing to watch out for is that noise current.

which as I said noise current appears across those impedances as a noise voltage. Those impedances. Well, guess what? the the impedance looks like. uh, a resistor in parallel with a capacitor.

the lower you can pull that down, the happier everybody is. And the smaller you can make that uh noise current, the happier everybody is. As far as keeping the noise down is concerned, for.