Interesting mic preamp topology

[gyraf]

Recently came across a Studer 961 mixer - ya'know, the type with very-very transparent-sounding micpres.

Found the schemo - and although 5534-and-transformer based, it's very different from anything I've seen before:

Anybody care to explain it to me?

:razz:

Jakob E.

 

[jensenmann]

..note to self: watch this topic for replies :grin:

 

[jdbakker]

At first glance, it looks like a distant cousin of this configuration (some more discussion can be found here).

JDB.
[not that I expect this to be news to Gyraf]

 

[clintrubber]

The last digit in the pic was missing but it appears to be this patent, if you hadn't collected it already:

http://www.pat2pdf.org/patents/pat4567443.pdf

Bye,

Peter

 

[Animatic]

Transformer bootstrapping? Nah.
Looks non-inverting in the output.

T2 and T3 are pretty odd ducks.
No reference to how the're wound?

Looks like 2 phases are rejoined in one transformer,
and then after as a single signal passed to the following transformer.
But it isn't cancelling out? Or has the inductance re-timed it?

1st one gives hotter signal, then it's inverted
then gets recombined with 1st at lower level
and then THAT exits at still lower level?..
and is reinverted.. uh uh..

DANG this is one twisted unit.
So many inductors, so little time

Where's PRR when you need him....

 

[Svart]

Hmm I'm not sure about the transformer setup but the JFET volume ladder is kinda cool.

 

[JohnRoberts]

That seems like a lot of work to use a smaller (cheaper?) transformer when modern discrete devices (and some ICs) are now quiet enough to eliminate the transformer entirely.

Since low noise discrete devices were available in '85, and cheaper than even a simple transformer, I suspect this was designed by an engineer more fluent in transformer design, than up to speed on Japanese device developments. I won't pretend to understand all that's going on in this circuit and I'm not going to try. Compensating 5534 for unity gain scrubs off some slew rate available in a conventional gain stage but it's surely adequate for the application.

Interesting in a historical sense. I recall looking for some way, years ago to use a small transformer to just pass the error signal and not the full audio signal, but bringing feedback around the transformer isolation defeated much of the benefit of having a transformer in the first place and increased complexity, besides complicating stability. Another of my sundry ideas that I abandoned. KISS

My sense is most modern mic preamps still using transformers is precisely to have more of that flavor distortion. YMMV

JR

 

[edanderson]

this studer was designed as a relatively small and light weight "location" console. installed in trucks adjacent to transmitters, with mic line feeds of unknown and changing provenance, the benefits of input transformers are more apparent. when you have a console that needs to be smaller and lighter, but retain input transformers, this approach starts to make sense. less expensive probably wouldn't hurt, though i doubt this two transformer approach was much cheaper.

i'll agree that from a theoretical point of view, many of the essential issues that audio transformers were originally included in audio designs to address have long since been solved with cheap handfuls of silicon. some of the adjacent "bulletproofing" that came along with transformers, however, is still important in some situations (such as location recording). and the sonic "color" can be an important marketing point. so i still use them, despite the expense, when applicable.

ed

 

[PRR]

> Anybody care to explain it to me?

Transformer and IC2 distortion and bandwidth are potentially improved by the NFB. The signal at T2 6-8 is mirrored to T3 6-10, so IC3 sees mike-signal boosted with little error (we hope and pray). IC3 still colors the sound.

A sub-advantage(?) is that when the mike comes un-plugged, feedback voltage at T2 5-9 does not pass (much) to IC2.

What I just don't see: where is IC2's DC stability??? R32=22K and T2 3-2 may be say 50 ohms... DC gain is 400, there's maybe 1mV of input offset and a half-Volt output offset??? OK, but that drops right onto another winding which may not like a big DC smack.

It's a transformer-theory geek's wet dream. And unless you are that T-T geek and can supervise the winding, it isn't much use to you.

 

[bcarso]

[quote author="PRR"]

What I just don't see: where is IC2's DC stability??? R32=22K and T2 3-2 may be say 50 ohms... DC gain is 400, there's maybe 1mV of input offset and a half-Volt output offset??? OK, but that drops right onto another winding which may not like a big DC smack.
[/quote]

That looks like a HUGE problem. One wonders if there might be something missing here.

Also, the FETs, cute though they are, are going to contribute substantial distortion in shunt with those kinds of series R values. But maybe that predominant second harmonic is preferred subjectively.

 

[PRR]

> the FETs... substantial distortion in shunt with those kinds of series R values.

Are they in series? You can do this:

 

[JohnRoberts]

The secondary winding of T2 looks like an auto-former where there would be a multiplication of voltage at 3 by ratio of winding 4-2 divided by winding 3-2. But bias current and offset are DC so no help there at DC.

Instead of autoformer help, opamp must pull node 3 all the way up/down to match IC2 DC offset voltage +/- the opamp bias current times winding 3-4 IR drop.

It seems winding 2-3 would be low wrt 22k. but if transformer 1:1:12 perhaps secondary winding not uber low Z?

There are no blocking caps anywhere. Can we be sure this hand drawn schematic is reflective of actual production design? The lack of DC path for JFET gates (on) would make me nervous about PCB contamination or leakage paths interfering with reliable action. Maybe they have adequate reverse leakage current in the steering diodes to keep everything happy.

Whatever... I'm no transformer guy..

JR

 

[bcarso]

BTW IIRC the old trick of sending 1/2 the drain-source voltage to the gate helps JFETs even when used as analog switches. Adds a lot of complexity though.

Yeah I worry too if this schematic is completely accurate. Don't have time to look at the patent right now, although sometimes they are intentionally inaccurate!

 

[gyraf]

Thanks guys - I hadn't seen the reference to the patent number

At a quick glance, this schematic matches the actual PCB for the mic input section area.

Jakob E.

 

[ioaudio]

 

[ioaudio]

a circuit description can be found in the manual (96MB!)

...the secondary winding (T2-2-4) leads to the inverting input of the opamp (IC1-2/3) and is terminated by a virtual short circuit (IC1-3).
As a result the voltage output on the transformer is nearly zero and causes negligible harmonic distortion.
(C17-C16-R31) determine the square-wave behavior, and (R32) the gain of the opamp (IC2).

The output of the opamp (IC2-6) drives the transformer (T3) with low impedance. As a result the opamp´s output voltage is returned into the primary circuit [series negative voltage feedback].
Negative feedback minimizes the distortion of the output voltage induced in (T3). This output voltage is decoupled via the third winding of the transformer (T3-6-8) and taken to opamp (IC3) connected as a voltage follower. (C18-R34) linearize the frequency response of the transformer (T3)
...

 

[clintrubber]

[quote author="gyraf"]Thanks guys - I hadn't seen the reference to the patent number [/quote]
O yes you did, and then even decided to cut off the last digit :wink:

 

[tv]

Yeah. Naugty.