Capsule amp

[bitman]

Hi all,

I am subscribed to a surround sound mail list and someone posted this schematic of a capsule amp.

I thought I'd let my buds over here have a look at it.

http://www.ambisonia.com/Members/umashankar/cascode-amp.jpg/image_view_fullscreen

:Ron

 

[SSLtech]

Interesting... Phantom powered, -evolved from the Schoeps topology, but with the 6V2 zeners and that odd crossed-FET at teh front replacing the single FET.

Also, it seems to be for an electret capsule.

4µF stage-coupling caps... a bit larger than essential, perhaps?

Incidentally, if the Shuttle had landed at Kennedy instead of Edwards yesterday, I'd have posted a B-format file of the double-sonic-boom flypast sound yesterday to Ambisonia... I had the entire rig in my car, was regularly checking the NASA site and everything... oh well.

Keith

 

[PRR]

> the 6V2 zeners and that odd crossed-FET at teh front replacing the single FET

The double FET is a self-bias cascode. Keeps the voltage across the lower FET constant.

The 4uFd caps may not be over-large. The BJTs pull significant base current and base current noise. They work at capsule level. A too-small cap may give rising LF and subsonic noise.

The 6V2 Zeners do nothing in normal operation. The 100K on the BJTs makes them bias roughly like 1K-2K resistors, or 500-1K for the pair. The DC current is then 48V-12V over 3.4K + 1K + 10K or 36V/14.4K or 2.5mA. The collector-emitter drop in each BJT is around 2V. We should throw a Monte Carlo at it to see how Hfe variations will upset it, but it is unlikely the 6V2 do anything in normal operation. They may be startup protection or paranoia (30V parts are cheap, there IS 48V at the jack).

> had landed at Kennedy instead of Edwards

Call NASA and complain.

 

[bcarso]

[quote author="PRR"] We should throw a Monte Carlo at it to see how Hfe variations will upset it.[/quote]

I'm curious what the sentiment on the street is about how much d.c. current in a console input transformer we can tolerate.

 

[Gus]

If one looks at the schoeps schematic there is a note about matching and selecting the output emitter follower pair. Also the bias resistors are selected for current draw. The schematic also has the 6.2 zeners at the output section.

There was the THE patent circuit here not long ago with what looks like a "better" way to bias the output pair.

 

[persistentsound]

I had always assumed the zeners were there to clip the outputs when the signal is too hot, and also to protect the drive transistors in the case when the 10K resistor is strapped out (as it is when configured for P12 phantom power) and accidentally connected to a P48 phantom power source. Would they not perform both of those functions?

And now, what might be a stupid question: how does the capsule provide a signal to the JFET base if it's not charged? Is this a dynamic mic?

 

[persistentsound]

There was the THE patent circuit here not long ago with what looks like a "better" way to bias the output pair.

Care to provide a link to those of us who are not in the the know?

 

[clintrubber]

[quote author="persistentsound"]

There was the THE patent circuit here not long ago with what looks like a "better" way to bias the output pair.

Care to provide a link to those of us who are not in the the know?[/quote]

That circuit from 'THE' was discussed as an somewhat OT-thing in this thread:
http://www.groupdiy.com/index.php?topic=22301&postdays=0&postorder=asc&start=39

and here again steering away a bit from the original subject:
http://www.groupdiy.com/index.php?topic=20976&postdays=0&postorder=asc&highlight=zuehsow&start=61

and finally here's the patent-#: 7072478
http://www.pat2pdf.org/patents/pat7072478.pdf

Bye,

Peter

 

[bcarso]

Thanks Peter. I thought maybe it was that thread.

In essence they are just making the output e followers "Vbe multipliers", probably not part of the patent (which site I couldn't get to btw). That's still not really that stable, and as I asked a bit back there I would like opinions on how much d.c. current is permissible in console input transformers.

 

[clintrubber]

[quote author="bcarso"](which site I couldn't get to btw).[/quote]

Hi Brad,

Site down from here as well, but must have the pdf here on the machine. So if relevant I can upload the pdf somewhere else if pat2pdf keeps quiet.

Bye,

Peter

 

[persistentsound]

Alternate location for that patent pdf:

http://www.google.com/patents/pdf/Microphone_preamplifier.pdf?id=I6t6AAAAEBAJ&output=pdf&sig=YrS-X-L6wO_WOtYnODqRzaMB57Q

 

[bcarso]

That patent is pretty odd, although the central claim 1 has to be understood I believe as requiring ALL of the details cited. So many of the items under that claim are prior art, it is still surprising that it got through, although the patent process is in shambles these days.

 

[bcarso]

Another comment: as per some of the remarks about Rossi's circuits made in that thread, the output Z of each line is not equal, since the source impedance for a given emitter follower is very different from the other. This is also true for the basic Schoeps topology and the THE circuit.

Clearly we've been living happily with this situation, so it's not cause for panic---just something to know. I'm a bit amused to think that to alleviate the mismatch effects of the THE (and their ilk) we might employ Whitlock's THAT part. THOU art cautioned: THY mileage may vary.


Another observation: the noise at the inverting polarity output is higher than at the noninverting for these designs. This suggests that, despite the appealing simplicity and apparent symmetry, one could do better using an independent stage for polarity inversion.

One could do better still if some voltage gain could be gotten out of the input amplifier, so later stage(s) noise would get less important. This is hard to do, optimally, with a non-inverting stage.

 

[bcarso]

[quote author="persistentsound"]
And now, what might be a stupid question: how does the capsule provide a signal to the JFET base if it's not charged? Is this a dynamic mic?[/quote]
[quote author="SSLtech"]
Also, it seems to be for an electret capsule.
Keith[/quote]