Side chain compression pedal??

[jamierichards1234]

Hey,

I'm looking for some advice.

I'd really like to make a pedal with an external sidechain input to do daft punk style side chain compression. I know the empress has this feature but since I've started playing around with DIY I have completely lost my desire to spend £250 on a pedal..

I'm pretty sure that with time I could figure out how to modify a simple optical compressor to do this, but for it to be any use I'd really need some kind of release control. Attack would be nice too but not so important.

Are there any optical compressor circuits out there that have a release control which could be easily modded to my purpose? I'm not even fussed about it having internal sidechain, as I just want it for this effect, so I think modding it wouldn't be too hard.

I built an Engineer's Thumb compressor (which sounds great) recently and I feel like the absolute best case scenario would be to modify that for an external sidechain so I could have controls for attack/release/threshold/ratio/makeup gain, but I wouldn't have a clue where to start.

If anyone could give me any clues on how I might be able to modify the engineer's thumb, or suggest a suitable candidate for modding I'd be very grateful.

Cheers

Jamie

[Phoenix]

A side-chain compressor is just a regular compressor, but with a different signal being sent to the envelope detector circuitry. So taking the Engineer's Thumb as an example, all that's needed to convert it to side-chain use is to disconnect the non-inverting inputs of op amps A and B, replicate the input circuitry for both (so the 10Meg, 1k, red LED's and 10n), use a 1Meg pot on the input of op amp B for threshold control, a plain 1Meg resistor for op amp A, and Bob's your uncle. Op amp A is for your signal input, and op amp B is for the side chain input. You can add tone controls, gain controls, whatever you want for pre-treatment of the side-chain input.

[jamierichards1234]

Thanks for getting back to me.

Do you mean like this? (excuse my dodgy photoshop job)

[Phoenix]

Yes, that's exactly right.

[jamierichards1234]

Thanks for the help, can't wait to get this up and running!

[jamierichards1234]

Ok so I've got this up and running but it's not quite 100% there. It's more or less what I was after but I'm still having a couple of problems..

a) I imagine this is a slightly more technical problem than my last one - I'm getting a little bit of the sidechain signal bleeding into the audio path. Any ideas on how I might be able to reduce/stop this?

b) I could do with more compression. I'm guessing I can just boost the sidechain signal to get there, but that will probably make 'problem a' worse..

Thanks

[Phoenix]

Regarding the bleedthrough, it could be related to layout, try to keep the sections away from each other.
Shouldn't get through the buffered Vref.
If you're using a TL074, it MIGHT help to split that into two TL072's, and decouple the power supplies between them, use one for opamps A and D, the other for B and C, add a 47ohm resistor between 9V and Vin (pin 8 ), along with a 330uF cap to ground from Vin of both TL072's. You could omit the 33ohm resistor from the original design.
If you still can't get rid of the bleedthrough, you could try injecting some of the sidechain signal into inverting input of opamp A, this would require extensive experimentation to determine how much. Note the bleedthrough may also be inverted phase (depending where it originates in the circuit), so adding the raw signal may make it worse, if so, an inverting stage will correct this.
That's all I can think of off the top of my head.

For more ratio? Jeez, the Engineers Thumb is already capable of limiting, not sure how you could need more, but I suppose if you're able to find a higher value pot for the ratio control, that'd do the job.

[Kipper4]

I has to be a layout thing surely.
The rectifier section will only output DC. (U1b U2a)
Have you tried a shorter release.

BTW What kind of signals are you using with this?

[jamierichards1234]

Thanks for getting back to me.

I've been testing it so far using a guitar as the input and some drum sounds from my OP-1 as the side chain. I added a boost to the sidechain signal (first thing I've actually built on veroboard) and it's definitely pumping hard enough now. It was nothing to do with the amount of compression from the pedal, just the relatively weak OP-1 output being fed to the sidechain, so that's all sorted.

As for the bleed through of side chain signal, I suspect that you are right that it's a layout thing as it's a bit of a mess right now! I think I'll try and draw up a nice tidy PCB, taking into account all of Phenix's suggestions and see how that goes.

In answer to Kipper4, Yes I have tried a shorter release. I have a 500k pot in place of the 470k resistor and I'm actually finding that the sounds that I'm after are the shortest release times on offer (before everything goes mental when I turn the knob too far). Valvewizard's instructions say not to go less than 100k and the pedal clearly doesn't like it when I do. I'm just wondering, before I start to draw up a new pcb, whether there is any relatively simple way I could modify it further to allow even shorter release times?


Cheers

Jamie

[Phoenix]

Just a crazy idea, not 100% that it would work without further experimentation, but a non-inverting op amp in the feedback loop of op amp B might just do it. Take the signal from the junction where the 1k, 1u, and 100kA pot's wiper and clockwise lug meet with the 470k resistor/500k pot and snip the 470k resistor/500k pot away. Then add a non inverting op amp, with non inverting input taking the signal from that junction, and place the 500k pot on the output of the new op amp. If it works and doesn't cause instability, you should be able to drive it down to 10k or maybe a bit lower if using a TL07x, or 2k if using something like an NE5534.

[jamierichards1234]

So I've finally got round to trying the opamp in the feedback loop but it doesn't seem to be having the desired effect..

It seems to be stable, but it's actually greatly increasing the release time rather than decreasing it!

I can get away with turning the pot down much lower than before, but the overall release time has gone up so much that shortest release time is about 5 seconds. Just to check I'm understanding your instructions correctly, I've included a doodle of what I've done.


At the moment the best setup I've managed is without the opamp in the feedback loop, with a 47k resistor in series with the pot (so I can get away with turning the pot the whole way without everything going nuts). It's just about stable like this, but if I reduce the resistance any further it isn't happy. The release time is probably quick enough for most uses, but it would be really nice to have the option of going quicker (also valvewizard's instructions said don't go below 100k and I'm slightly uneasy about using something that is 'just about stable').

Any thoughts on what else I might be able to try to reduce release time?

Thanks

Jamie

[jamierichards1234]

.. I should also mention that I don't need a wide range of release times, so if there's some sort of mod that can be done to reduce the overall release time (at the expense of losing longer release times) then that would still solve my problem.

Cheers

[Phoenix]

You haven't drawn the release pot. Where have you placed it?

[jamierichards1234]

oops sorry!

[Phoenix]

Wait, are you connecting Vr (4.5V) to the non-inverting input of the new op amp? That will definitely cause this issue, I'm surprised it would work at all! Disconnect that and it should work correctly.

[jamierichards1234]

Ah yeah sorry, you may have guessed by now that I don't exactly know what I'm doing!

I actually started out without the 4.5v connected, but it didn't seem to work. The guitar went quiet (maybe half the volume) and the compression didn't seem to have any effect. So I thought I'd try connecting 4.5v like on the input opamps.

The really strange thing is that when I connected 4.5v to the noninverting input, the compressor did actually start to work (albeit with a really slow release).

I've disconnected it again now and I'm having the same problem of low level and no compression. I swapped out the opamps too just to check that I hadn't cooked anything. Now my non-technical mind is wondering if maybe the compressor is effectively stuck 'on' and that's why the volume is low?

Any ideas?

Thanks for your help on this!

[merlinb]

whether there is any relatively simple way I could modify it further to allow even shorter release times?

Use smaller storage caps in the side chain (the 1u caps). Halving them will halve the attack and release times, etc.
But bare in mind that the shorter the release, the less smooth the control voltage becomes, eventually approaching raw rectified audio.

Your opamp mod won't work, btw, since there's no path for current to get back into the side chain caps after the rectifier has sucked it out!

[jamierichards1234]

Thanks, I'll give that a go. 

[jamierichards1234]

So I've finally come back to this having spent most of the summer distracted by recording/festivals/getting married etc and I'm still having some problems.

I've tried to make it switchable between internal and external sidechain (see pic). When it's set to internal sidechain it works fine and sounds great, but when I switch to external sidechain, the audio gets very loud and distorted (without feeding any sidechain signal). If I adjust the ratio control all the way to no ratio the distortion goes away. When I feed in a sidechain signal it works as expected on the distorted audio and compresses it (it cleanly compresses the distorted audio if you know what I mean). Any ideas? Should this input/sidechain setup theoretically work (see pic) or should I be looking for problems elsewhere in my build?

There's a few things I should probably mention:

• I'm actually running 2 of these circuits at the same time from one power supply circuit, and switching both with a dpdt switch. I haven't got as far as detaching one of them yet as I wasn't sure if that was likely to be the problem and I didn't want to start pulling stuff apart unnecessarily.

• I am using a 500k pot for ratio as I didn't have 1m, but I tried it with a 1m resistor and it didn't seem to change anything.

• Also I haven't fitted the clipping LEDs as I didn't have any but I don't think that's the problem as it sounds fine when set to internal sidechain so I don't think the levels are too hot..

• I have fitted a pot for attack (100k) and release (500k in series with a 100k resistor to stop me going too low).


Reading my previous comments it sounds like I had it sort of working but I seem to remember it sounded a bit distorted at the time, and I was probably just setting the input level way low to compensate. Having made it switchable to the regular internal sidechain I now realise how much cleaner it should sound and messed up it probably was! 

If anyone can help me figure this out I'll be eternally grateful!

Thanks

Jamie

[samhay]

>when I switch to external sidechain, the audio gets very loud and distorted (without feeding any sidechain signal).

This is to be expected when the ratio pot is set high enough. U1a has a gain of ~ 100 when the ratio is turned up. This is easily sufficient to cause gross distortion of a typical guitar signal.

>If I adjust the ratio control all the way to no ratio the distortion goes away.

See above.

>When I feed in a sidechain signal it works as expected on the distorted audio and compresses it...

If you are feeding the side chain insufficient signal, this is expected behaviour.

>(it cleanly compresses the distorted audio if you know what I mean).

I'm not sure I know what you mean. The distorted signal gets quieter and/or you can hear the compression kick in?

If you set the ratio just just high enough to cause distortion with no side chain signal, what happens if you then feed it side chain signal? If the distortion goes away then you need to add a boost to the side chain input.