This Week on the Breadboard: The Stupid-Simple Compressor

[Chuck D. Bones]

I just had to try this to see what happened. A simple circuit that works quite well!


Here's the breadboard. I used a pair of generic Vactrol-type opto-couplers I bought from EG. Makes it easier to test in a lit room compared to an LDR illuminated by two LEDs. LDR (opto) selection is critical. We want an LDR (opto) with a large resistance range and a medium response speed. Too fast and we get distortion at low freq. Too slow and it breathes. Any single FET-input opamp will work. I tried a TL081 and an LF351, both worked well. I originally powered this circuit with ±9V because I already had a charge pump on the breadboard. Then I switched to +9V power, which is how it's configured in the pic below. No difference in performance.



Here's the harmonic spectrum with a large input signal and SUSTAIN dimed. 2nd & 3rd harmonics add up to about 2% THD.



Here's the response to a tone burst. The first cycle is clipped just above the compression threshold by the LEDs. After that, the LDRs take over and keep the signal level just below the clipping threshold.



The next rev has BLEND & TREBLE controls. It's still undergoing testing.

 

[benny_profane]

This is interesting. Does the initial attack produce an audible clip, or is the LDR response quick enough that the transient doesn't register to the human ear?

 

[Chuck D. Bones]

I don't hear it as distortion; I think the duration is too brief. Bear in mind that all optical compressors have a slow enough response time that the first cycle (or cycles depending on attack speed) can easily drive the opamp into saturation. That makes an audible POP which I don't hear with this compressor.

 

[Coda]

Will it make me sound like Mark Knopfler?…

 

[benny_profane]

Bear in mind that all optical compressors have a slow enough response time that the first cycle (or cycles depending on attack speed) can easily drive the opamp into saturation.

Right, that’s why I was asking. The burst waveform makes it seem like it’s got a rather fast attack time. I’d be interested to see the graph with a slightly larger signal.

 

[Chuck D. Bones]

How much larger? 450mVp-p is a pretty healthy pickup signal.

 

[benny_profane]

How much larger? 450mVp-p is a pretty healthy pickup signal.

You’ve designed the block to response well you anticipated signals. My curiosity is more to see how it responds outside of the design parameters to compare the initial attack region. Feel free to dismiss if that’s pointless.

 

[fig]

Stupid-simple is my middle name. To the breadboard!

 

[Chuck D. Bones]

You’ve designed the block to response well you anticipated signals. My curiosity is more to see how it responds outside of the design parameters to compare the initial attack region. Feel free to dismiss if that’s pointless.

I just ran the tone burst test with 1Vp-p input. The output looks the same.

 

[swyse]

What sort of ranges for the LDR would be reasonable? I have an order in for some GL5516, but I think they don't fit the large resistance range criteria and I have no idea what is considered fast for these, they are pretty new to me.


GL5516
Light Resistance: 5 – 10KΩ
Dark Resistance: 500KΩ
Rise Time: 30ms
Fall Time: 30ms

 

[Chuck D. Bones]

I just looked at the GL5516 datasheet. They do not define the test conditions of Rise Time and Fall Time, so to me those numbers are meaningless.
The LDRs I've tested respond very quickly to increasing light. Some respond almost as quickly to decreasing light, but the vast majority respond much slower to decreasing light. Slower response time is better in this case.

Best to socket or breadboard them and listen to the distortion level when playing low notes with SUSTAIN dimed.

The GL5516's resistance range is acceptable, but the sustain will be limited by the 500KΩ dark resistance. As a comparison, all of the Opto-Couplers I'm using have dark resistance over 2MΩ and most of them go well over 10MΩ. In this circuit, 5K - 10KΩ light resistance will work fine.

 

[fig]

VTL5C4s show an off resistance of 400MΩ

 

[fig]

I've got a bunch of different VTLXXX ...I'll try a few in this breadboard and report back.

 

[Chuck D. Bones]

Slightly off-topic, just how many breadboards do you have?

I have eight, which I used to think was excessive. Now I'm not so sure.

 

[BuddytheReow]

Can you share the link for the Vactrol? This is on my breadboard list

 

[Chuck D. Bones]

Can you share the link for the Vactrol? This is on my breadboard list

Who are you asking? Fig or Chuck?

 

[BuddytheReow]

Either. I've got a LDR and led stuck together with electrical tape and I figured I need an upgrade

 

[Chuck D. Bones]

Wait for these to go on sale at EG. Don't pay full price.

https://www.goldmine-elec-products.com/prodinfo.asp?number=G15396

The advertised specs are "optimistic." You have to test them because their actual specs are all over the place and you might get a few duds. I paid $2.95 for a bag of 12 last month and had one "dud." I always find enough good ones to make the total cost acceptable.

You can also try some black heat-stink tubing to join an LDR to an LED in a more robust fashion.

 

[fig]

Slightly off-topic, just how many breadboards do you have?

I have eight, which I used to think was excessive. Now I'm not so sure.

Just a guess... 30ea of the 400/800s...I stripped a couple dozen last week.

It was tight on a 400 using the stupid zif, so I got creative jumping to make room for the LEVEL trimmer.

 

[Chuck D. Bones]

I prefer the 1660 point ProtoBoards. Enough room to stretch out. Sometimes I build two different pedal circuits on one.

My ZIF sockets tend to pop out of the protoboard, I have to keep reseating them.