Uh, when we were discussing previously the uh, typical valve topology or tube topology for a uh, condenser mic, we're looking at uh Bier Supply yeah, plus 60 volts or whatever going down to the this is the microphone capsule then and capacitor coupling off that into something a bit like that? Okay, we've got two high value resistors, two very high value and relatively expensive resistors Y and a coupling cap that, uh, it's not allowed to be leaky. Sure. So what type we talking about? Uh, usually. yeah, uh.

poly propoline. uh, polystyrene Poly. put the ketle on and what values did you need in there? A Nerad or so? Oh okay, yep that that's doable in those techn. Basically, when You' got 50 puff and a nanofarad, then that doesn't represent much series loss? Sure.

yeah. When we went to the fat-based circuits I wanted to try and get rid of one or the other of those resistors mhm I had a running plan. Oo, do tell, do yeah. which was this? First of all, get your J fit.

You'll probably want to fix its bias with you guessed it, a high V resistor. Yep, higher the B yeah, somewhere. why don't we simply connect that to our microphone? High voltage? Supply Yes, let's face it, the same. it's exactly the same.

Yeah, so let's say that's uh, let's say that's sitting at I don't know. 60 volts? Well, actually I think we took that up to uh, 90 volts just for the El of it. And let's say that's sitting here at 10 volts which and there's a good reason for that. sitting at 10 volts.

we get 80 volts across there. Yep, we get one resist only as long as we don't have any leakage in the microphone capsule. the gate is sitting at 10 Vol and there's we've gotten rid of an expensive resistor. a surplus capacitor.

Yep, and in principle, what we should have? Uh with a typical uh depletion mode. Jfet if that's a 10 volts, that might be say 10.2 volts or 10.5 volts or whatever. Let's use that as the front part of a closed loop Unity Gain voltage F Okay, let's do so. What we might do here is put a resistor there.

Uh, we'll put a PMP transistor there. Uh, as a voltage gain stage. Um, we'll U I'll come back to that in a second. We'll load that with a current.

Source Yep, All okay, what typically would you have made the current Source out of couple of a resistor? a resistor? Fancy, fancy. all right. Actually, it can go anywhere from a resistor to a proper current source and in various models. I used V various different topologies.

Uh, now in actual fact I was using about 15 volts there. about Uh 7 volts there that was sitting at about Uh 7 .5 Vols I that was slightly higher than that. Yep, Uh, then what we did there was this is this is going to be look disturbingly like a power amplifier. Let's put a basically a bias voltage in the middle there and take it off to a it is.

Yeah, it is. It's a power. an okay we've got a voltage amplification stage there. We've got a bias.

Incidentally, red leads are good for that. Yes, yes they are y Although leads can be noisy. um I never came across noisy one. Okay, uh, certainly the just you know.

red leads about 1.5 1.7 volts. Uh, which allowed you know? Well, 66 and a few of voltage acoss there. So that's our output and what do we do with that bang straight on there. Yep, that's to minimize the crossover Distortion Of course the bias is to minimize the crossover.

Distortion Uh now let's just say we've got say a 1 milliamp current Source down there. or maybe even a 6 K8 resistor. The value of resistance there will determine the fat current. Basically, let's just say that we have a a 1K resistor there.

That transistor is going to require about6 Vols to turn it on. So guess what we're going to get about6 you will milliamps down there if that tries to drive. If that tries to pull more than 6 milliamps, it's going to increase the voltage there. That'll pull that lot up.

It'll bring that point up. It'll shut that device off relatively speaking. So that's that's the resistor which sets the operating current for the Fet. Its operating voltage is set by the fact that that voltage there is6 of a volt less than the supply.

and that voltage there is going to be slightly more than the bias voltage. And it's the bias voltage here which sets the output voltage. All the queres and conditions explained. and boom, it's Unity Gain nice.

And because it's Unity gain closed loop. its linearity is exceptional and that's to want in a good M Yep. Now the other thing is I was discussing previously the role of capacitors and I said that we typically about 10 puff there, about 30 puff there. Yep, that 10 puff there is it bootstrapped, not on your Nelly because doesn't matter how big a signal we put there, the the signal voltage there is n all.

so our capsule is loaded just by that 10 puff. right? How about this this 30 puff cap down? Here It's a follower. There is no voltage across that cap. it's completely bootstrapped out.

Yep. so our 50 puff capsule is really only seeing a bit of stray wiring and 10 puff? Clever. So relatively lower attenuation. Very nice.

Now let's talk about that NT3 microphone. It's it was just a little bit different because the so what was this was this used in? Yes, that topology is used in for example, the uh, The Road Nt1000 okay or almost as shown right okay Yeah Bugger All difference. Uh, the fat. The Jfet that we were using in the NT3 Uh.

Rumor has it was originally designed back in the mid to late 90s for use by The CIA in some of their little bug microphones. Y and Uh was manufactured by Siliconic. Uh, I wish I could remember the part number because it's cute little beasty consisting of the J fet and a pair of backto back diodes. Okay, now what do you do with them? Well, the first time I tried using that fet, I simply used it in the normal configuration where you ground that and you ground that and you cool current out the top.

Those are horribly nonlinear devices as You' expect yeah, more than about 10 or 20 M volts of signal there and it was basically nonlinear as okay, why would you use those in the first place Because they actually act very nicely as a replacement for a really high value resistor. Uhhuh, of course. yes, as long as you can keep the voltage cross those low, they act like a high value resistor. How the hell do I use those? Well, the first thing I did, uh, let's just whack a bit of resistance in there and you maybe a Meg or something like that and connect that to.

Okay, so these came out as a separate pin. Yes, it was a four pin device, right? Okay there there there and of course there nice. One of the selling features of them was really low input capacitance incidentally got it and I figured I'd use these because again, it gets rid of what was at that time a relatively expensive, very high value resistor. Yep, but how to use it? Bootstrap the buggers one Meg resistor to our buyer Supply and then connect that to a capacitor which goes you guested to? Yep, the well.

basically the uh, the uh Source con connection. Yes, so that it it vanishes. Yeah, yeah, so that was one of the really cool things. H Maybe somebody on the blog save cost H not only to save cost, but it was convenient little package right? and if we pull this out, you'll be able to see it living up the front just there.

Yep, that that little uh 5 Pin So 23? Yeah, right and look, maybe somebody can do a bit of research and tell us all what the hell the part number on that is because I can't find it anymore. It doesn't appear on data sheets anywhere H But that was an interesting little circuit tweak that saved us a resistor. the other one was and these aren't .01 Cent resistance. Yeah, they're like, uh, even in volume, we might be looking at 30 cents or thereabouts cost instead of you5? Yep.

So saving you pulling 30 cents out of the product cost is it's an exercise worthwhile doing absolutely. particularly if you can do it at the expense of you know, a 05 Cent resistor and maybe a 1-cent capacitor. Oh, let's close the loop. Yes, Okay, the initial prototypes of this microphone.

Um, well. they went out for beta testing. Comments: came back they're a bit lackluster up in the top end, right? Uh, they didn't have enough. kind of pizzazz.

Not enough excitement. Uh, and basically enough. Distortion Uh, the frequency response? Too flat it was. The frequency response looked a little bit like that, where those ripples were up in the kind of the 15 khz plus region.

your ripples aren't big enough exactly. you got to have big ripples to. Yeah, what? Big Ripples And basically what they wanted was a bit of extra excitement. Up here.

you a little bit more at about 6 khz or you added to the data sheet. Extra excitement? Yes, Well now we had to the new model excitement. We had to add some more some more qu quack. So we've got the perfect opportunity here.

H to do a little bit of finessing because it is a closed loop circuit. So I Kind of scratched my head and walked around the circles for a while and opted to put in a Bridg Te circuit. right? Yes in the feedback. Network Now Bridg Te network has a response that looks like that.

It's got a a rather minor Notch which when you use it in a feedback Network gives you. you. You guessed it. Yep, little.

So these microphones have that little uh Bridge te circuit in them just to wear a little bit of woohoo to the sound. How did you choose the frequency you were going to put that up? Uh, basically the knowledge that they wanted a little bit more woohoo and uh, you know, the sales guys who were there at the time were kind of canny enough to realize where it needed that little bit more a little bit more Pizzaz right? And we did a couple of different cut and tries to find out first of all, how much and secondly at what frequency. And yeah, it was around about the 6 khz Mark that they wanted the extra squirt. Okay, and I think about 4 DB thereabouts was so am I just a bloody amateur because I if I'm looking at the microphone data sheet I want to see this flat I'm going to buy the one that's flat as attack I'm an amateur clearly I don't know what I'm doing? no don't no.

uh. if you're after a mic microphone as a measuring instrument, of course you want dead flat. Absolutely. Uh, if you after a microphone to measure, uh, sorry to record uh, something like, uh, an orchestra or something whose sound was absolutely defined and known and you didn't want to muck about with it, y pick something flat.

Yep, that's not what happens in Studios studios are creative environments. You're trying to create the sound that winds up on tape, which is why you have this cupboard or room full of different microphones with different wobbles and different micro structures and different directional characteristics and different amounts of distortion. Because you want to choose a creative tool that actually kind of sounds like what you really want it to sound like. Are people still doing that in mics these days? Oh hell yeah.

yeah, they're not. They're not doing it in the the other end. No, Um, because look, everybody knows it's cheating to do it at your digital audio workst. Yeah, If you can't do it with a room full of $10,000 vintage microphones, What's the point? Exactly.

Gold plated. So anyway, uh. that that might be an interesting little uh trick for anybody who's playing with these. Oh, there's one vital component.

What are we missing? That I Missed out of this circuit. It's a closed loop system. Is it not? It is. It needs dominant pole compensation.

Oh, there you go. And without that, trust me, it oscillates like a banshee. Yep, how did you pick it? It screams, how did you? How did you pick that? Uh, actually basic, stick it in on works or did you? Ah, there is actually a perfectly viable technique that I use, which is uh uh. from that, onto work for any closed loop uh, system that you would design pretty much Yeah, yeah.

okay. um. connected to a Sig Jenny Yep. Square wave.

Okay, have a look at the output and find out whether it's going. Yep or yep, that's it. Yeah, Oh, bear in mind and you want it to optimize? Yeah, ideally you're after something a bit like that. Just tape.

Yeah, just rounds off. You ought to get it up to the point of maybe a small amount of overshoot Y In which case, you know that Yeah, it's pretty good. It's pretty stable, right? Uh, that is not on No, because production variations it's likely to turn into. and an oscillator that, yes, y uh trap.

For young players though, there's two kinds of stability issue that you need to worry about in any of these power amp circuits. Uh, one is the loop stability. Yep, as governed by your dominant pole compensation. or you can go for split pole.

There's many different, but the other one involves the simple fact that emitter followers voltage followers are renowned for oscillating by themselves as a form of negative impedance oscillator. whenever you hit them with a capacitive load. A big capacitive load. Yep.

and when you get that happening in for example, an audio power amplifier, a big one. Uh, you can first of all start isolating the inputs with Ferite beads to raise the source impedance at high frequencies. Uh, you can add Emitt degeneration. You can do that with Fite beads.

They're up in the hundreds of mehz. Oh yeah, look I've had uh, four mehz bandwidth transistors sitting there. One? Yep, 4 MHz bandwidth output transistor sitting there on just mounted on its own on a chunk of heat sink. and just with I Think it was two resistors and a capacitive load.

I turned that into about a 30 or 35 Meg oscillator and I could tune it by getting my source wire and just running it up and down the the length of heat sink. Nice. Yeah, so yeah. nasty.

And that's just all about parasitic inductances and stuff like that. Any parasitic inductance up in here is to be avoided Mhm, whereas any parasitic inductance here is to be appreciated. Yes, uh, sometimes even some uh, local degeneration. Yeah, there it depends.

But there are these two forms of stability. One involves the loop, the other involves just the output stage acting as a follower with stray inductance here, and low impedance at the bases and it forms a I think a cal pits or a heart, the oscillator or something like that once you hit it with some extra capacitors over there. Yep, Okay, so that's the uh. the basics of a lot of the Uh microphones currently being manufactured by Road road.

Yeah, terrific. Now let's go one step further. Okay, we got enough time. Yep This is really cool.

Okay So we've pretty much got the same circuit that we had before. Yep, uh with the bias network of your choice in the output stage and y y y uh, what we're actually going to do here is something dirty dirty dirty dirty. like the sound of it. Let's ground that.

Let's take this down here to Min - 120 volts. Yep, Okay, um. you you changed anything so far? No, we've put in some uh, get rid of those because we don't need them. Uh, we do have, however, need that one.

Um, do you know of any high voltage fets that'll go in there that'll run at like 120 volts? No. I Do not off hand. I Do they've got pilot lights inside them? Oh, there we go. Oh oh yes, you're a tube fan.

Well, this was for the Nt1000, right? And yeah, sorry people, it's another Heretics microphone. It's got a tube front end, but it's riddled with bipolar and there's a very sensible reason why you did this. Oh yeah, uh. First of all, um, okay, that that goes off to the output through capacitive coupling.

y y y okay I Want something that gives, uh, this is a non-inverting of course of course of course. and I wanted this particular circuit to, um, react such that how do you put it positive positive pressure into the microphone? M gave a positive output right. Best way of doing that was was there any absolute need for that? Uh, a distinct preference, right? Okay, let's put the microphone there. Yep.

and by that through our 5 gig resistor to a voltage of about - 60 mhm. Okay, guess what? Any positive pressure input there Okay. moves the plates closer together. voltage goes up.

Clarity Beautiful. Okay, why these inverted voltages? Why ground up there and - 60 - 120? First of all, it keeps the uh metal parts of the microphone mhm substantially at 0 volts. Okay, uh, just quietly. There is one other microphone that we did where that's actually sitting at plus 60.

and if you operate it with the cover off and put your lips on it, yeah, you will get a belt. Mind you, who'd be stupid enough to run it with the cover off? I Don't know, but this this is this circuit was absolutely fearsome in its dynamic range, right? First of all, with those supplies, we could get an output voltage there of about 35 volts. RMS Nice. Not bad for a microphone preamplifier.

Pretty good. 35 volts. RMS is that's RMS that's not Peak Probably about two orders of magnitude higher than most line inputs will cope with. That had implications for how do you put it the maximum SPL that you could apply to the microphone capsule.

In fact, you can put put I think it was about 150 something Dbsl on the M capsule. So you're physically limited by the M capsule, not the yeah yeah. Secondly, uh, if you pick the right operating current again, the operating current is determined by that resistor there. pick the right operating current for that valve and the residual noise of the system corresponded to a how do you put it a a microphone A noise at the microphone of about 12.

DB You can express the system noise in terms of how much acoustic noise does that correspond to MH So we had 12db SPL equivalent noise we had I Think it was about. Let's call it 155 DB SP at and this wasn't a clipping. this was at like one 1% Distortion Wow. Okay, uh, subtract that from that and you get a dynamic range of what's that? 143 DB Dynamic range out of the one circuit.

Now there's not too many semiconductor circuits that'll do that. and here's a tube circuit with that kind of dynamic range. Awesome! Hey, and one of the tests. I Did you have to do this? You have to test it I Simply hung a set of headphones off off the output there and it worked bizarrely well.

It was loud in headphones, right? Incidentally, Uh, because I wanted this thing to be capable of feeding 600 ohm loads. Uh, it's actually got two sets of paralleled output devices so that it can actually drive a 600 ohm load at those kinds of levels. So from the from the M that's disgusting. Did anyone actually use it? Did anyone actually use it for this or design? Overkill It's pure design Overkill It might be pure onism on my part, but oh, incidentally, that noise figure there.

That was based on optimizing the operating voltage and current so that it would suit a very wide cross-section of tubes, right? Okay, I could have dropped that noise probably about four or 5 DB if I'd tuned the operating current for the valves. but H if you do that, you make the thing kind of unproduced. except. but yeah, getting that down to about, we could get that down to about 8db.

Wow. The best, uh, commercial, uh, semiconductor mics uh, are probably the Road Nt1a which is about 3 and a half or 4 DB noise. Uh, the Uh. Nyman Tlm 103 I think it is which it's around about 5 and 1/2 or 6 DB that kind of order of magnitude.

We could get this tube thing down to pretty close to the best that semiconductors had to offer with that ridiculous that kind of output level. That's ridiculous. How did you get these in production though? Uh, how now who did you get them from? How can you still get them? Nobody's making tubes. Yes, they are.

Yeah, they are like there's a few making tubes. Uh, and there's two different types. You can get the New Old Stock it's probably been manufactured Back in 1970 and just getting the warehous. Russ And then there's the actual New tubes, right? And uh, I think what road are doing that? They're using New tubes, but they age them quite viciously.

Uh, they operate them at quite High uh, voltages and currents to burn them in so that when they actually put them in a mic, they're going to be stable for a very long time. Got it? It's something that they do quite well. How much power will your pissing away heating this thing up? Um I was slightly R underrunning the filament. Uh, they normally a 6.3 volt filament I think I was running them at about 5.5 volts or thereabouts, just slightly under.

If you run too far under, you start stripping the thorum or whatever off the cathode, right? Yeah, yeah, um, and can't remember about 300 500 milliamps. They don't draw huge amounts of power, but how are you getting this from The Phantom Oh, not not from Phantom No, not not not. this valve. one this particular specific power supply, right? You're right.

and even ow God even the power supply had some uh quirks to it. the topology for the power supply just very quickly. We'll finish up on this I think okay again. negative power supply.

Yep, so ground up the top and a negative unregulated voltage down here. The first thing we created was a current Source Well, actually a current sink. Okay, so over here we've got our main reservoir caps off the bridge rectifier a current sink consisting of basically an N Channel fit. yep, and connected across it a Amplified Xena again, consisting of an N Channel fit.

So what do we have out here? First of all, that's low impedance. That's high impedance. any. Ripple there is severely attenuated there.

Got it? We had yeah, maybe hundreds of Mill volts Ripple There, We were lucky if we got tens of microvolts Ripple over here. And of course you can picture that this topology probably looked a little bit like uh, actually sorry, uh uh Amplified Xena There you go. And this one of course consisted of oh I might as well draw him down here. uh one with them I'm sure you've seen that one before and a bias resist up to who cares.

Yeah, too easy and that creates a you know current Source Mhm Amplified Xena Yep, nice little Supply which is intrinsically short circuit protected because that simply soaks up any shortcircuit current and has devastatingly low noise out there. Fantastic. So yeah, around one of you out there somewhere might have called to use that kind of topology one day. There you go I Hope and All road mik is still designed and built this way.

Yeah. Fantastic to the best of my knowledge. Mind you, it's what 13 14 years since I've been there. but uh, to the best of my knowledge, any of their high-end Studio stuff y still uses these techniques.

Beautiful! Thanks He.