Help with noisy preamp - replacement parts?

[JohnRoberts]

Really? But a mismatch is going to cause an offset error that could limit headroom pretty significantly no?

no....  the gain pot is capacitor coupled so the DC gain of that first stage is modest. The input devices are biased up at v/2 (6V?). The op amp inputs at v/4, but output is likely v/2.   

Another thing I just noticed about this circuit is the 7812 regulator after a diode is going to require more like 15DC supply and not 12V.

Looking at this closer I am inclined to abandon it for a better topology.

The input devices are running open loop so will distort pretty easily.

A single 12v rail will not deliver much headroom (but that is relative), so usable as long a gain is kept modest..

Better designs extend negative feedback to around the input transistors so they are effectively operating constant current and deliver much lower distortion. To do this requires at least two op amps (one for each input transistor). The second opamp could be repurposed to deliver this but it would involve a major cut and point to point kluge.

I wouldn't invest too much time, effort, or money into trying to improve this design. It's one saving grace may be it's relatively high distortion that some may consider euphonic (I don't).

JR

PS: My apologies to any who think my tone is "harsh" just calling it like I see it, after decades of designing mic preamps.  8)

 

[Bo Deadly]

no....  the gain pot is capacitor coupled so the DC gain of that first stage is modest. The input devices are biased up at v/2 (6V?). The op amp inputs at v/4, but output is likely v/2.   

Ahh, I see. I didn't see the separate emitter resistors and thought it was LTP with feedback. I have to start paying attention one of these days.

I wouldn't invest too much time, effort, or money into trying to improve this design. It's one saving grace may be it's relatively high distortion that some may consider euphonic (I don't).

It could be good for a solo in-your-face voice or for miking an amp. If you're going to deliberately overdrive something it might sound better if it clips a little softer which this circuit will. Otherwise multiple tracks with distortion like that probably don't mix well. You wouldn't want to record and opera with it.

So I agree. I would just try yelling into it for a bit and either keep it as-is or sell it to someone on gearnutz for $400.

 

[Bo Deadly]

Better designs extend negative feedback to around the input transistors so they are effectively operating constant current and deliver much lower distortion. To do this requires at least two op amps (one for each input transistor). The second opamp could be repurposed to deliver this but it would involve a major cut and point to point kluge.

Like this:

 

[Newmarket]

Like this:

Not quite as that depicts an "equivalent circuit" indicating current sources but not necessarily what is actually inside the IC.
The feedback loop as shown doesn't include the transistors.
Easier to see if you 'erase' the output protection diodes from the diagram.

Worth a read through...

http://www.thatcorp.com/datashts/AES129_Designing_Mic_Preamps.pdf

 

[JohnRoberts]

Like this:

Exactly... that is what is affectionately (cough) called the Cohen topology after an Australian engineer but it has been in use well before his "discovery".

That is the classic three op amp version, while there is a two op amp version that works almost as well.  Instead of identical op amps gain stages closing the loop between each collector to emitter (via feedback resistor), imagine a first op amp looking at both collectors differentially, and the output of that op amp run through a simple unity gain inverter to provide the opposite polarity feedback to the other emitter.

This two op amp version is the one that could be cobbled on top of the OPs circuit.

I preferred the three op amp version, but when I joined Peavey in the mid 80s they were already using the two op amp version, and I couldn't argue a benefit that justified using the third op amp (Peavey is a sharp pencil cost management organization).

JR

 

[Newmarket]

Exactly... that is what is affectionately (cough) called the Cohen topology after an Australian engineer but it has been in use well before his "discovery".

I see - I had overlooked the two blocks depicted as '-Av'.

IIRC Cohen wrote an AES paper on the topology hence the association with his name. But I don't think he claimed it as original work ?

 

[PRR]

Those who repeat history are doomed.

Those who apply Cohen's name to the things we mostly use for mike-preamps have never digested Cohen's teachings.

The eye-opening detail is the 300 Ohm resistors where nearly all implementations use 4k7 or 10k. The drawback is that for high gain the "Gain Set" resistor must be uncomfortably small.

A more obscure detail is that "Bias Set" allows choice of input transistor current, and the suggested values allow up to 7mA in each side of the input stage, pushing OSI below 100 Ohms.

http://leonaudio.biz/cohen.htm
http://leonaudio.biz/double.balanced.mic.amp.notes.pdf

 

[JohnRoberts]

  Graeme Cohen RIP 2015
=======
I have briefly considered optimizing mic preamps for real world A/D's only modest input voltage swing.

While typical op amps can drive lower value feedback resistors to modest voltages (they are current limited), some of the modern uber op amps can easily drive low resistances without raising a sweat.

My inclination to pursue this also passed.

JR 

 

[scott2000]

Those who repeat history are doomed.

Those who apply Cohen's name to the things we mostly use for mike-preamps have never digested Cohen's teachings.

The eye-opening detail is the 300 Ohm resistors where nearly all implementations use 4k7 or 10k. The drawback is that for high gain the "Gain Set" resistor must be uncomfortably small.

A more obscure detail is that "Bias Set" allows choice of input transistor current, and the suggested values allow up to 7mA in each side of the input stage, pushing OSI below 100 Ohms.

http://leonaudio.biz/cohen.htm
http://leonaudio.biz/double.balanced.mic.amp.notes.pdf

What's the purpose of the matching resistors in these locations???? Benefits???Is it only useful in certain circumstances???


https://groupdiy.com/index.php?action=dlattach;topic=68920.0;attach=59723;image


I noticed my older Ashly limiter uses the matching resistors in a similar way....



Sorry for the ignorance...

 

[JohnRoberts]

What's the purpose of the matching resistors in these locations???? Benefits???Is it only useful in certain circumstances???

I can only imagine a tiny benefit to CMRR wrt stray capacitance at the gain pot (?) but when dealing with such low resistance typical cable capacitance should not be a big deal.  That said the matching of the 2k resistors in the output differential amps are far more important.

Back last century I used a similar topology in a console and the mic gain pot was over a foot away from the preamp requiring a shielded cable to the gain pot.  I developed a shield driver for the gain pot cable so any capacitance at those nodes (emitters) was common mode.

https://groupdiy.com/index.php?action=dlattach;topic=68920.0;attach=59723;image


I noticed my older Ashly limiter uses the matching resistors in a similar way....

Um good, I guess...

Sorry for the ignorance...

Don't feel lonely, we're all ignorant.

JR 

 

[abbey road d enfer]

What's the purpose of the matching resistors in these locations?

The 1k's  are matched for equal voltage drop so DC offset is minimized. The 2k's, as JR says, helps CMRR. However, in the actual Philips product, there were two trimmers, one for DC offset and one for CMRR!