Optical compressor Help?

[Havaden]

Hello builders
Im a novice pedal builder and only have built the bazz fuss and noisy cricket.
Im going to order a bounch of 1590a enclosures and make myself a small selection of "Must Have" effects.

I allready have circuits for a fuzz, distortion, tremolo and boost. But i want to make a compressor as well.
I have seen the "orange squeezer" and still think it looks a bit out of my league.
So i learned a bit how the compressor works and think that i have made a somewhat compressorlike circuit
so how bad is it? how can i improve it?

[Kipper4]

I cant help as I'm no EE.
If you want a low parts great compressor. Check out Jon Pattons Bearhug compressor.
Good luck buddy and welcome aboard btw.

[duck_arse]

have a look at this circuit, the colorsound/roland sustain:

http://www.generalguitargadgets.com/diagrams/cssussc.gif

it might give you some ideas. you can use 2n3904 instead of the BC's.

[midwayfair]

Did you make that, or did you find it somewhere?

There are a host of problems that you will encounter with such a design, though I will say that it's pretty cool looking and has some neat ideas. So I'll just run through the issues:

1) Q2 and Q2 need resistance on between the Collector and +9V. Right now there isn't any. Result: All signal is dumped to the 9V power rail. You'll want something like 10K.

2) Smoothness of compression will depend VERY VERY heavily on the properties of the light dependent resistor. You need a very, very slow turn off time for the LDR or it will flutter like crazy. This isn't an insurmountable problem ... the VTL5C10 is used in the Demeter Compulator. But frankly, it's just bad design -- Demeter could have just used an envelope circuit without ripple and saved themselves the trouble of relying on a fairly extraordinary part.

You'd be much better off using an envelope circuit to turn on the LED, rather than relying on the guitar signal all by itself. It's not that many parts, and it can be just a diode, capacitor, and resistor or two. You could tailor the attack, decay, and strength of the envelope to control the compression FAR better than relying on a "magical" LDR.

3) Ratio of compression. You've specified a 100K LDR. These aren't common. In any case, with the 50K "attack pot" (this is actually more like the minimum compression", you have a very limited range of compression available. If you want to use the same method of gain reduction, consider an LDR with a wider swing. Unfortunately, you'll need to do some math to determine the total ratio by calculating the gain at max light and minimum light. I can't help you there, since I suck at math. I can tell you off the top of my head, though, that you're going to run into an issue where the gain reduction is not linear and your LDR is doing very little. I'm not sure I can explain this in a short paragraph, but volume changes are logarthymic and voltage dividers need to reflect this. This design has a voltage divider formed by the [LDR, Attack, and bias] and [some "hidden" large resistance to ground at the base]. In this design, your LDR swings about 99K in a system with a max of 1,150,000Ohms. But a change of 100K when the bias resistor is near 1M isn't as drastic as a change of 100K when the bias resistor is near 100K. In other words, you'll need to run that bias trim way down for the LDR to have much of an effect. Then there's a bunch of other issues. Because the bias is so low, you don't get much amplification to begin with. This means your signal is really quiet, probably below unity (requiring a make-up gain stage) and possibly too low gain to turn on the LED on the emitter of Q1 even when it's not compressing.

This isn't a horrible place to put the LDR. It's not so different from putting it Base-Ground, which is how more than one compressor does it.

This is one place where op amps have a big advantage over discrete transistors. If you put your LDR in the feedback loop of an op amp, if it drives resistance to 0 you still have a unity gain buffer. That's not the case here. If your LDR reduces the resistance far enough in this design, you won't have unity gain. You can probably tinker with it to ensure that it's never below unity (with the bias trim), but it's more work.

In any case, you definitely want a larger LIGHT-DARK range on that LDR.

4) Sound quality. Your gain reduction affects the input loading. This means that while compressing, your circuit could load the pickups and lose treble content and produce distortion. Depending on how high the gain is, you could overdrive Q2 with just a humbucker's signal.

The best thing to do here is stick a buffer in front of the circuit.

5) Some other things worth looking at ...
The Orange Squeezer is NOT a difficult project. Look at it closely and you'll see that it's really just a couple sections: There's a single gain stage from an op amp, and then an envelope that puts some DC into the gate of a FET to accomplish some gain reduction at the op amp's input. That's it. It has a few tone shaping bits, but it's not difficult to get working (lots of instructions about dialing in the bias). It was one of my first pedals and I've made several over the past couple years. They're like candy. Plus there are multiple PCB projects out there and TONS of posts about it. Madbean's PCB for it (Cupcake) fits in a 1590A.

-If you want a really good, simple compressor, check out John Hollis's Flatline. Madbean Pedals sells a 1590A-specific PCB project for it called the Afterlife. It performs extremely well. There are some other simple optical compressors out there, but they're either a little more complicated (Samhay's "Another Optical Compressor"), use less common parts (Anderton), or have some issues (Mictester's Really Cheap Compressor, which distorts a lot).

-If you want a compressor that sounds and works a lot like the Orange Squeezer but has a transistor-based dry path and an optical device, look up Ray Ring's opto-FET compressor. I made a 1590A layout for perfboard, but a guy over on Madbean put together a PCB for it as well. I really like that compressor; there's a demo on my YouTube.

[Havaden]

Thank all of you for the circuits!

Midwayfair: Yes i made the circuit (in Circuitlab) in a few minutes based on the information i got from how a comressor works.
And the LDR tremolo im going to build gave me the idea.
Thank you for all the hints and alternative circuits! 


So this is a simple idea that might work if somone mod/fix and tweak it a bit.
If so, anyone should feel free to do it and post it in teir own name.
just please mention this thread when they do it.

[armdnrdy]

If so, anyone should feel free to do it and post it in teir own name.
just please mention this thread when they do it.

I don't mean to let the wind out of your sails but.....I'm not sure if you are aware that there are many compressor designs using optocouplers that have been manufactured for years. These range from stompbox size to studio grade, high dollar rack mount units.

Here's an application note from one optocoupler manufacturer:
http://www.silonex.com/audiohm/compressor.html

The above link contains enough information to build a working optocoupler based compressor with all the bells and whistles. 

You can delete parts of the circuit that you deem unneeded for your application.

[Havaden]

Hahah! ok, as i said im a novice builder so i didn't know 
But thanks for the info