Compressor design

[Don Symes]

Hey there, smart and helpful people -

I've decided that, for similar money, I can do an all-tube version of the new Princeton Recording (gonna call it the Trenton Studio). The plan is to start with a Weber 6A14 or 6A20 with a Dumble ODS style overdrive and a Weber MASS as an attenuator.

The part that has me stuck is the compressor. The tube compressors I have schemos for (Altecs, mostly) use funky tubes and transformers I don't really want to source. Most of the stompbox types use the CA3080 Operational Transconductance Amp. The compressors that use regular op-amps and LDRs all have the LDRs configured as input attenuators rather than as gain fiddlers, so I'm wondering why that is.

My idea is to use a 12AX7 for a Philbrick tube op-amp circuit as the op-amp in the compressor, with a regular solid-state sidechain.

Does that sound workable?

[Rob Mercure]

Hey Don,

Most classic tube compressor designs split the input signal feeding part to the control amp and part to the controlled amp (for want of a better term). The control amp amplified the signal then rectified it and applied it to a grid of the controlled amp. This could be the supressor grid of a pentode (or the screen grid) or just biasing the control grid of a triode. I guess you'd consider the 6L7 "funky" but it's my favorite candidate - specifically designed for such purpose - but any pentagrid converter type tube will do and you can pick a new one up at AES for pennies. While most of these designs use a transformer the control amp circuit could be an op amp using a tube for the controlled part with a small parts count.

Rob

[Don Symes]

This could be the supressor grid of a pentode (or the screen grid) or just biasing the control grid of a triode. I guess you'd consider the 6L7 "funky" but it's my favorite candidate - specifically designed for such purpose - but any pentagrid converter type tube will do and you can pick a new one up at AES for pennies. While most of these designs use a transformer the control amp circuit could be an op amp using a tube for the controlled part with a small parts count.

Thanks, Rob, but I'm a little worried about how microphonic small pentodes are supposed to be (EF86, for example). That's why I was looking at a tube op-amp in a non-OTA circuit.

I can understand the sidechain (control amp) in the solid-state designs. The tube designs are causing me some discomfort ...

[Steve Conner]

Hi Don,

The only real "magic" in a compressor is the variable gain element. Tube compressors have to use one of three things: a LDR, or vari-mu tubes, or triodes configured as a voltage-controlled resistor.

The LDR is by far the easiest. It doesn't require a tube op-amp, any more than the trem bug in a Fender does. In fact, it works the same way as the Fender trem bug.

The vari-mu tubes are often used in push-pull pairs to cancel control voltage feedthrough while maintaining bass response, but that may not matter in a guitar application. A P-P pair of vari-mu tubes is what's in that vintage Fairchild compressor that sells for as much as a small house.

Fred Nachbaur uses a compressor circuit with a single vari-mu tube in many of his amp designs. The EF86 is not vari-mu, but it might be worth trying anyway.

A push-pull power stage with poorly regulated screen voltage gives a similar compression effect to a push-pull pair of vari-mu tubes, FWIW. You could control the screen voltage from your sidechain, or play with resistors and capacitors to set the attack and decay times on the screen node, essentially making the screen supply your "sidechain".

The voltage controlled resistor is just another one I threw in. I've seen an amp schematic where a pair of back-to-back triodes were used in the vibrato instead of a trem bug, but I can't remember which one it was.

Finally there is the compressor from the Trace Elliot V8 bass amp. It uses two sections of a 12AX7 and some solid-state diodes. I have no idea in hell how it works, but it seemed to get good reviews for its tone.

[Rob Mercure]

Don, Steve pretty much filled in the blank I left - and then some - so I just wanted to address your "microphonic tube" concerns. First off the 6L7 is a metal or glass octal so you don't have to worry about any miniature issues - but these are becoming more and more scarce by the day. There are scads of small pentagrid converters such as the widely used 6BA7 available with no current audio use making them quite cheap. AES has them for less that $7 each and I suspect that you can find them, or other p-convs, for half that or less if you search the web (for example the 12 volt cognate - 12BA7 - in only $2.13!) - cheap enough that you can buy a dozen or so and cull out any that are microphonic. But as these tubes were widely used in small "all American Five" and similar type "table" radios they had to be fairly rigid as 4" speaker was only inches away mounted in a cabinet with little mass so I wouldn't worry about microphonics. While there are some notable microphonic NOS tubes this is more of a problem, IMHO, with current production tubes.

I cobbled up a tube compressor using a 6L7 about 8 years ago and generally enjoyed fooling with it but in the end I tend to use a fairly simple stage rig so that I don't have to pay too much homage to Murphy.

Rob

[The Radium King]

http://satanda.aspweb.cz/schem/compr...compressor.jpg

http://www.dogstar.dantimax.dk/tubestuf/4x4compr.htm

[Don Symes]

Hi Don,
The only real "magic" in a compressor is the variable gain element. Tube compressors have to use one of three things: a LDR, or vari-mu tubes, or triodes configured as a voltage-controlled resistor.

Triode-as-vcR? Hmm ... with the cathode at some fixed voltage (reverse-biased diodes?, zener?) grid grounded::maxR, grid approaches Vk::minR. This seems to be what the Altec 436c does, but I'm having a hell of a time understanding the circuit.

The LDR is by far the easiest. It doesn't require a tube op-amp, any more than the trem bug in a Fender does. In fact, it works the same way as the Fender trem bug.

It does require some gain element, though. I started out expecting to find the LDR in an op-amp's feedback loop, but since that's not the case all I need for that approach is a relatively low-gain triode stage. Maybe go for a hi-fi stage there.

The vari-mu tubes are often used in push-pull pairs to cancel control voltage feedthrough while maintaining bass response, but that may not matter in a guitar application. A P-P pair of vari-mu tubes is what's in that vintage Fairchild compressor that sells for as much as a small house.

Fred Nachbaur uses a compressor circuit with a single vari-mu tube in many of his amp designs. The EF86 is not vari-mu, but it might be worth trying anyway.

I mentioned the EF86 as an example of a problematically microphonic pentode, hoping to hear that some other choice was better, as Rob so kindly did.

A push-pull power stage with poorly regulated screen voltage gives a similar compression effect to a push-pull pair of vari-mu tubes, FWIW. You could control the screen voltage from your sidechain, or play with resistors and capacitors to set the attack and decay times on the screen node, essentially making the screen supply your "sidechain".

Interesting idea. More that I'm willing to take on at this knowledge level, though.

Finally there is the compressor from the Trace Elliot V8 bass amp. It uses two sections of a 12AX7 and some solid-state diodes. I have no idea in hell how it works, but it seemed to get good reviews for its tone.

[/QUOTE]

Thanks, I'll look that up.

And many thanks to both you and Rob in general.

[Don Symes]

So, does this look sane? (already thinking audio taper rather than linear)

Much obliged.

[tbryanh]

I was wondering what editor you used for the schematic?

Is it open source (free)?

Thanks

[Don Symes]

I was wondering what editor you used for the schematic? Is it open source (free)?
Thanks

I used the 'Draw' section of OpenOffice. As a schematic tool, it SUCKS, but one uses what one has.