B&K microphone preamp construction

Well, I use a '90's vintage 2144 dual channel analyzer for acoustical measurements and the thing went on the blink.
I think it's a disk drive, but no schematics are typically published. Called in and left a message with them, bu frankly i'm not in a position to pay them thousands for repair right now. The thing is horribly expensive.
( I used to use it for car NVH at the corporate acoustics lab and made a deal to trade consulting services
to buy it when I started my own business)

Really I just need the power supplies and preamps for the calibrated mics, so in the meantime I think i'll just build some up. I usually use a PC these days anyway rather than the clumsy 2144, which has no line outs. I only used it with the B&K mics and intensity probe.

From what little my manuals document, They list the pre as 28V supply, 0.25pF input capacitance FET, and 1G input impedance.

So I'll use a doubler on a 120V little Isolation transformer ( actually two PC board mount little 1VA 120:24V power transformers back to back) and  regulators for the 28 and 200v. Have some in the lab.

Have to find an appropriate FET (low input capacitance) so I don't spoil my mic calibration. Calibration is done "unloaded" acording to my mic paperwork.

I have a pistonphone but can't find where the #%& I stuck it.

Really need 2G to +200 and ground  (or divider) but may just use 1G.

Anybody know what's inside the things? Is it just a Fet Follower with Unbalanced output as I suspect?
Honestly I don't think there's anything particularly special about them other than the outrageous price.

I tried to disassemble one, but they seem not very takeapartable. Didn't push it.

Might machine some of those ridiculous B&K connectors they used in the '90s too. I'm not paying hundreds for them.

I wonder if anyone else here has a need or interest in shop made B&K type measurement mic pres and supplies?

Les
L M Watts Technology

Ultra clean classical recording pres ? sure

I use them for microphone testing, but sure...the capsules are just dead flat and pretty low distortion.
Great for some super clean misic recording.

As far as the head amp/power supply...well that's the question.

A single FET source follower (which I suspect) is good but not the ultimate in low distortion.

A 2 fet one with active load is much better.

And the ultimate antiseptically clean I think would be a LN Fet front end op amp with huge open loop gain and much feedback.They are not as low noise as discretes, but these 200v capsules have huge output. The mechanical pressure noise and bias resistor noise would tend to predominate over voltage noise of many FETS I think.

Again, that's the question. What's in there? I don't want to break out the abrasive saws like that silly science channel show to find out.

Oh, the mechanicals are fun too... on one preamp i'm measuring M20x.75 threads...seemingly a bastard size.


Les Watts
L M Watts Technology

Follow up:

B&K called. They have no compatibility information on the pres and power supplies. No schematics.
No information whatsoever really. Guess it has to do with understaffing due to the economic times.

So to get going again I have to shop build some pres/power supplies for these mics.

Already have a circuit cooked up...i'll just build it.

Don't need any exotic low noise Fets...I calculate the(thermal) pressure noise of the 4145 1"capsule as
10.9 nV/sqrt Hz, and the 4190 1/2" capsule at 17.4 nV/sqrt Hz. Their pres must have very high input noise voltage considering Ciss= .25 pF. That means low transconductance and high noise.

That's about it for this thread. If anyone else needs information on the pres, PM.

Now to build these up and get back to my measurements...

Les
L M Watts Technology

Good luck Les.  Thanks for the read.

Did you look at
http://www.bksv.com/Library/Primers.aspx
microphone handbook 1?

I have not looked at it.

Gus,
I do have the B&K literature.

Back on after a long lightning strike induced delay. After the great fun of machining all the bastard-threaded  B&K connectors,
i'm now building up a lunchbox power supply. It supplies +200 and +/-15 to the preamp. It's a little different from the unbalanced
B&K preamps. The biggest difference is that it doesn't cost 1100 euros!

Just using a fet phase splitter and a couple op amp followers.

Need multipin connectors. Have some 6 pin DIN laying around...might use those. Not the best, but I don't like the price i'm seeing for things like the LEMO ones.

Les

Interesting!

I am working on a pair of studio mics based on 4145 capsules. A machined brass thread mount would be extremely useful.

The 10% distortion limit of the 4145 is 150db. At a sensitivity of 53.8 mV/Pa that works out to 34V RMS (about +32dbu); 28V isn't enough buffer supply voltage for the high output capsules, so I wonder about that. I have a vacuum tube follower based preamp type 2613 that works with the 4145s that I think uses 160VDC , maybe the 28V is a tube heater voltage?

The spec for the thermal noise of the 4145 is 10db equivalent A-weighted ref. 20uPa; I guess that's how you worked out the noise voltage...

Typical B&K preamps load the capsule at tens of Gohms according to spec, perhaps they use a bootstrapped input follower. But how would the polarization work? I don't feel like sawing mine open either.

I'm also building a power supply and balanced line buffer for my 4003s. These are also sensitive (hot) enough to no need any front end gain. The power supply is 130V (they are electret capsules) and max output is also over +30 dbu. The preamp in these is powered by 130VDC and takes a constant 4mA. They call it a class A totem pole, as it can source or sink 4mA on it's single ended output, which is biased at +65V DC. Fun... Maybe that topology would be good for the preamp of the 4145 mic.

Thanks!

Michael

PS the 4145 is ~65pf capacitance so I suppose a 1Gohm polarization resistor will only roll off the subsonic frequencies

Hi Michael. Sounds like you're doing similar.

I haven't built up the tronics yet....just machined the case and connectors, and sketched up a quick schematic.

I could up the supply voltage a bit. (the B&K I have is 28v Vdd according to them.)

Also, I thought a little about the topology. The typical fet phase splitter should have a good bit
of low order harmonic distortion at high voltage swings...fine for color in a typical music mic, and ok for my
measurements since they are at low level.

But what if one used something like an AD743 Fet op amp (about 3nv/sqrt Hz ) as a voltage follower with a couple low noise bipolar op amps to balance...it would be dead clean up to clipping, and the capsule pressure noise would still dominate.

Thing is...I can't find what the input capacitance of the 743 is. I'll bet it's high. Not a big deal since it would pad the capsule a little and allow higher SPL.
But why not? Granted, it's expensive, but nothing compared to the capsules. Might be neat to have a couple
mics in the cabinet with vanishingly low distortion at moderate levels.

So I might do that.

As far as the bias resistors...Yeah the gate resistor can (and often is) bootstrapped a lot. But the polarization
resistor is in parallel with it (ac) and would seem to short out the bootstapping unless it was much higher value.

I agree the resistors don't have to be huge with that much capsule capacitance. And as we have discussed on micbuilders ... bootstrapping doesn't lower the 1/f bias resistor noise, although it does lower the LF corner.

Les
L M Watts Technology

> 10% distortion limit of the 4145 is 150db. At a sensitivity of 53.8 mV/Pa that works out to 34V RMS (about +32dbu); 28V isn't enough

True.... but where do I get a 150dB SPL source? How far away should I stand? (And how would I know the mike distorted it 10%?)

28V ideally allows 10V rms; that will "only" make it to 139dB SPL.

Isn't there a trick, padding the capsule with a capacitor to reduce output voltage?

Near-field of a hi-power cone speaker is not going to be over 130dB SPL.

1 Acoustic Watt in a 2" tube is ~~142dB SPL(?), so a 4"-coil driver just might touch 150dB SPL. When you get a contract for such study, then you can build a 100V-200V preamp. It isn't that hard; just rarely needed.

Gunshots and explosions exceed 139dB even 150dB SPL. But if needed for "dramatic effect", such impusles have to be clipped, limited, and reverberated to give the demanded impact on practical (<120dB SPL) playback systems. A 2-buck dynamic is good enough for capture.

Isn't there a trick, padding the capsule with a capacitor to reduce output voltage?

Click to expand...

Right, this is the most common form of attenuator. A cap across the capsule forms a capacitive divider.

So does stray capacitance and fet input capacitance.

With the 4145 putting out about 11 nV/sqrt Hz pressure noise you could pad quite a bit before fet voltage noise
and bias resistor 1/f noise come in to play, depending on bias resistor value.

So I imagine if you pad 6 db it's like doubling the supply voltage from say 28 to 56 with almost no dynamic range penalty from the fet and resistor noise.

Again the bias resistor noise goes DOWN as resistance goes UP. Why?
The low pass formed by Rbias C capsule  attenuates faster than the thermal noise goes up as sqrt R.
But there is a limit to that resistor value...I don't like to go above 5 gig or so without fears there will be
crackling and other noises in a humid environment. That's why the B&K preamps have heaters in them.

Les
L M Watts Technology