Compressor, theory of operation

[amorris]

If you need a specific question, try this:

Where can I find some nice general info

specific is more like:

the trimmer that goes between the cathodes...what do I aim for with it?..equal currentdraw through the tube halves, same voltage on anodes? or is it something more complex?
also... I'm just starting to build this one..am I crazy if I ignore the GR metering and just go with hanging VU's on the output?...I mean, knowing the GR is good, but the aim with this one is "glue" the tracks together, right?...not that it would be a big thing wireing it up with a switch and and opropriate resistor...



http://www.conditionedresponse.com/DIY/C5/Comp5update.jpg

 

[guitarrock04]

Look mate, I'm not asking you to sit down and write out for me everything you know. I'm asking for experienced members to point me towards some knowledge sources that are already out there. Simple enough, I thought...

I appreciate the suggested links and books offered.

 

[amorris]

Totally,

hey, there is a great thread where CJ interviews rein narma. he designed the fairchild 670. very long thread, but totally worth the read.

http://www.groupdiy.com/index.php?topic=21531&highlight=rein+narma

 

[AnalogPackrat]

Guitarrock,

Other than the few links that have been provided, I don't believe you are going to find a lot of well organized "book-like" content on this subject. I know of no Art of Compressor Design by Smashwitz and Bump.

PRR's summary should be put in a META. That's the gist of the problem. The ways to implement a compressor are manifold. You've got feed-forward vs. feedback topologies, side chain design (which gets really interesting for stereo), application of gain control (opto, VCA, FET, vari-mu, etc.), makeup gain amp design, and more. It's more than a book...a trilogy at least.

That being said, there is a lot of practical and semi-organized information on this forum. The giant LA-2a thread in the Lab has tons of golden nuggets buried in it. There have also been some great threads on the VCA based SSL buss compressor (and various cousins of that beast like Gyraf's and some really good stuff from SSLTech, too).

Roger Foote and Wayne K have started some great threads about using THAT chips for compressors. There's one thread in particular (I believe it was in the Drawing Board forum) that was a great discussion of various summing methods for stereo compression. Then there's the 1176 discussions if you're curious about FET based designs. If you're into tube stuff there are several vari-mu designs (from PRR's elegant and simple one to Analag's massive and impressive effort).

Finally there are many threads which cover some of the less well known, but still interesting designs of the past (Gates Sta-Level and the like).

So do some searching here and you will find some of the best analysis, design, and implementation information in the world on this topic. And that, my friend, is a gift horse of massive proportions.

Regards,
A P

<edit--addendum> Actually, there is the (late) Dave Jahnsen's book on the LA-2a. I've never seen it personally, but I believe it may have a few errors.

 

[lofi]

[quote author="guitarrock04"]Look mate, I'm not asking you to sit down and write out for me everything you know. I'm asking for experienced members to point me towards some knowledge sources that are already out there. Simple enough, I thought...

I appreciate the suggested links and books offered.[/quote]

how about posting a schematic your interested in and we can work through it? i know the big boys on here will just look and follow the schem instantly but the rest of us could use it as an educational pointer.

if we ask people to either walk on by or tell us where we are going wrong then it might serve as a usefull primer for other newbes as they join?

thinking about doing this anyway

Iain

 

[guitarrock04]

Sure, I'd love something like that. I posted a thread some time ago, regarding a Volumax, and PRR gave me a great bit of advice, but I didn't get a whole lot of interest otherwise...

I have a CBS Volumax model 400. http://www.freeinfosociety.com/electronics/schemview.php?id=313

If someone would be so kind as to give a rundown of this bad boy, I would be greatly appreciative.

Also, I recently had a chance to work with a Spectra Sonics 610. Does wonders for snare... I found some old threads on here regarding the possibilty of a clone project, but it just kinda trailed off...

Has anyone attempted a clone of one?

 

[PRR]

> If someone would be so kind as to give a rundown of this bad boy, I would be greatly appreciative.

How appreciative? How many words do you want and how much do you pay per word?

But wait.... back in the 1960s some tech writer spent several days writing this up for the Volumax Service Manual. Probably got paid a couple hundred dollars. You can buy a dead-tree for $35. You can even find it on BAMA (hint: try the mirror site).

 

[lofi]

ftp://bama.sbc.edu/downloads/cbs/volumax400/Volumax-Mod400.pdf ?

 

[opacheco]

Well, I seach information of the pro design of autentics Compressors(like LA´s, Fairchild´s, 1176´s).....How they design these equipments????...What is the Theory behind of these units????

But I found this document about the compressor design and funtionallity:

COMPRESSORS AND LIMITERS An overview of what they do and how

www.music.mcgill.ca/~wes/docs/Compressors.pdf

Comments Please!!!

Thanks a lot,
Opacheco.

 

[JohnRoberts]

Besides the papers published by THAT corp, a lot of the early DBX work in dynamics design was published in AES journal papers and preprints.

JR

 

[opacheco]

[quote author="JohnRoberts"]Besides the papers published by THAT corp, a lot of the early DBX work in dynamics design was published in AES journal papers and preprints.

JR[/quote]

Thanks a lot JR!!!....

Opacheco.

 

[StephenGiles]

Ah, I see you are in Honduras - was that formerly known as British Honduras? I'm sure I have some old stamps from there in my Dad's old album.

 

[dale116dot7]

This might not be that helpful, but maybe it might.

Aristotle (and 4-H) - Learn to do by doing.

The way I learn is to build one circuit based on either an application note or a kit or DIY project using something close to what I want to learn. Sometimes I'll even buy a commercial product that is similar to what I want to build or learn. For example, for a compressor I might choose the G1176 or a Pico. For a reverb, I might choose an Alesis AL3201 and just follow the application notes to build up something.

Then I make a signal - say, for a compressor, a CD with stepped tones. Various frequencies so I can see what the compressor does to 30Hz and 1kHz, and 10kHz. Then I take a 'scope and prod at the circuit, while staring at the schematic. At this point, a lot of questions come up. Why did the designer use that value of resistor or capacitor or whatever. Then I change it to see what happens. Prodding with a 'scope while doing some modifications really brings it into perspective. Also, when you concentrate on one part of the subcircuit, read and really understand the chip or transistor or tube that you are looking at. If you're doing this with a Pico, really get to understand how the RMS converter works and what it does. Don't look at the rest of the compressor for a while, just understand that part of it. Then once you really are familiar with it, move on to how it fits in with the remainder of the box.

Reverse engineering by 'scope is yet another skill. I managed to reverse engineer a DSP board by a certain well-known effects manufacturer that does not like to give out service information. The hardest part is to reverse-engineer programmable logic devices - PAL's and FPGA's. For those, a logic analyzer is better than a 'scope but I make do.

Once I am very familiar with reverse-engineered circuits, I find I have a much deeper understanding of how to design something, how to use that product, and how to avoid pitfalls found in existing work when designing a new widget.

Hint - do not publish reverse engineered circuits, they are for your own eyes only, if the company is still in business. Lawyers and all.

If you know
-Dale

 

[opacheco]

[quote author="dale116dot7"]This might not be that helpful, but maybe it might.

Aristotle (and 4-H) - Learn to do by doing.

The way I learn is to build one circuit based on either an application note or a kit or DIY project using something close to what I want to learn. Sometimes I'll even buy a commercial product that is similar to what I want to build or learn. For example, for a compressor I might choose the G1176 or a Pico. For a reverb, I might choose an Alesis AL3201 and just follow the application notes to build up something.

Then I make a signal - say, for a compressor, a CD with stepped tones. Various frequencies so I can see what the compressor does to 30Hz and 1kHz, and 10kHz. Then I take a 'scope and prod at the circuit, while staring at the schematic. At this point, a lot of questions come up. Why did the designer use that value of resistor or capacitor or whatever. Then I change it to see what happens. Prodding with a 'scope while doing some modifications really brings it into perspective. Also, when you concentrate on one part of the subcircuit, read and really understand the chip or transistor or tube that you are looking at. If you're doing this with a Pico, really get to understand how the RMS converter works and what it does. Don't look at the rest of the compressor for a while, just understand that part of it. Then once you really are familiar with it, move on to how it fits in with the remainder of the box.

Reverse engineering by 'scope is yet another skill. I managed to reverse engineer a DSP board by a certain well-known effects manufacturer that does not like to give out service information. The hardest part is to reverse-engineer programmable logic devices - PAL's and FPGA's. For those, a logic analyzer is better than a 'scope but I make do.

Once I am very familiar with reverse-engineered circuits, I find I have a much deeper understanding of how to design something, how to use that product, and how to avoid pitfalls found in existing work when designing a new widget.

Hint - do not publish reverse engineered circuits, they are for your own eyes only, if the company is still in business. Lawyers and all.

If you know
-Dale[/quote]

Thanks a lot dale116dot7,

Anyone have another Idea??

Opacheco.