Lovesqueeze comp - howzit work?

samhay · January 31, 2013, 11:31:52 AM

OK - if I turn the input amplitude right down, I can see the drop in signal over ~5 ms. Still a little fuzzy about calculating the compression ratio though as I can't find the threshold and while I have remembered to turn amplitudes into dB, I am not sure how to deal the the op-amp gain. Doesn't really matter - It definity does some compression.
Looks like the recovery of the gate voltage can be tweaked by changing the value of the cap between the rectifying diodes.

midwayfair · January 31, 2013, 11:55:31 AM

Quote from: samhay on January 31, 2013, 11:31:52 AM
Looks like the recovery of the gate voltage can be tweaked by changing the value of the cap between the rectifying diodes.

Yes, in conjuction with the parallel resistor. Works out to uF * K = ms (according to PRR).

samhay · January 31, 2013, 12:20:26 PM

Quote from: midwayfair on January 31, 2013, 11:55:31 AM
Quote from: samhay on January 31, 2013, 11:31:52 AM
Looks like the recovery of the gate voltage can be tweaked by changing the value of the cap between the rectifying diodes.

Yes, in conjuction with the parallel resistor. Works out to uF * K = ms (according to PRR).

That rings a bell. The simulation bears it out too. Tweaking the resistor makes for a much more useful mod - 25k pot with 1 k fixed resistor perhaps - as the resistor does not change the attack time whereas the cap affects both.

Mike Burgundy · February 02, 2013, 06:24:00 AM

Samhay: I totally disregarded dB-ifying for the ratio, just looked at the rough voltages ;P
I think those diodes are there just for transients.
If you look at how this thing works, amplification for IC1A is 1 with the compression knob at 0. Turning it up it's just a gain setting, and the FET mucks with that afterwards. Volume is then used to change the level after compression - this means that everything in the first 5ms before compression kicks in gets amplified a LOT - there is a BIG transient in front. The LEDs aren't so much for limiting whatever signal is being put in, but for keeping that first transient from getting nasty.

Now, did anyone notice Vref is skewed? The only thing I can think of why this would be done is that the LED limiter is asymmetrical- but the question "why" immediately pops up again. It doesn't seem to influence the FET gate at all.

gritz · February 02, 2013, 09:30:20 AM

Quote from: Mike Burgundy on February 02, 2013, 06:24:00 AM
Now, did anyone notice Vref is skewed? The only thing I can think of why this would be done is that the LED limiter is asymmetrical- but the question "why" immediately pops up again. It doesn't seem to influence the FET gate at all.

Looking quickly at the schematic I'm assuming that the second opamp is going to be driven into rail to rail clip pretty easily. Shewing the reference voltage northward would allow more swing in the negative direction - which is the direction it needs to go to cut off the fet. Perhaps it's just a tweak to make the design more forgiving of fets with bigger cutoff voltages.

Regarding the schottky vs. silicon diode debate (forgive me if someone's mentioned this already as I've rather skimmed through this thread): compressors have an area of operation called the "knee". This is the transition between no compression and "full-on". Silicon diodes have a much sharper and more defined off ->on behaviour than schottkies, which are rather more "spongy" in their switch on. Hard knee compressors sound more aggressive and it's also possible that silicons cause more ripple voltage on the reservior cap, which would lend a bit of soft clip to the compressor's output on lower notes. [/conjecture mode]

Kipper4 · February 02, 2013, 06:38:55 PM

I finally got mine going after days of debugging and scratching my head.
I will say its very subtle.
Not unnoticable, it has a soft knee and is very nicely musical and understated. But then thats what i want in a compressor for electric guitar.

PRR · February 02, 2013, 08:27:10 PM

> When the pulse ends, there is no output (duh), but you can see the gate recover

Testing a limiter, what you want is a signal that goes big and then small (not off).

Traditionally 20dB below limiting level.

If your gate only gates, set up a second AC source at very low level (perhaps 2mV) and mix it with your gate.

Back in a real-world, we had one audio oscillator and a pulse-gate, added a "bleed" around the gate so it didn't cut all the way off. Now that audio sources are not physical boxes, it might be interesting to use a different frequency for the small tone.

midwayfair · February 16, 2013, 07:30:35 PM

Mike, you were right and I was wrong. I went back and measured extensively and absolutely: Lower Fv, bigger negative voltage.

But I'd still swear there is definitely something funky going on when using 1N4001 or other rectifier diodes that isn't the same as using general purpose diodes. In practice, it doesn't seem to matter a ton if they are silicon, but it would be a cool way of changing the ratio with a variable threashold.

Mike Burgundy · February 17, 2013, 10:35:54 AM

Thanks Jon, still didn't get a chance to do anything in the real world with this. I could not think of an explanation for that behavior, glad to hear it wasn't my mind going haywire ;P

Kipper4 · February 21, 2013, 07:16:50 PM

I built another today this time using BAT46 diode.
The compression (10k) knob doesnt seem to do much. Why is this? (am i just not hearing something?)
or is the range of compression so small its practicaly imperceptable.
I can tell when its off and overall i like it but, i just feel like its not working correctly still.

midwayfair · February 21, 2013, 09:48:47 PM

Quote from: Kipper4 on February 21, 2013, 07:16:50 PM
I built another today this time using BAT46 diode.
The compression (10k) knob doesnt seem to do much. Why is this? (am i just not hearing something?)
or is the range of compression so small its practicaly imperceptable.
I can tell when its off and overall i like it but, i just feel like its not working correctly still.

Does it get louder when you turn up the compression?

Measure the resistance to ground from the drain of the FET. Does it change when playing?

Do this: Tap the strings very hard (enough to hit the pickups, even) with the compression cranked. Do you hear a swell after that?

Kipper4 · February 22, 2013, 05:24:29 PM

it gets mildy louder when turned up and yes i can hear it. Especially with a OD/DS before the compressor.
Its just my stupid ears. I think i'm so used to coloured compressors that its hard to get my head round something that doesnt completely decimate the tone.
I cant wait to get the bits (chip and dip socket) so i can build the engineers thumb now.
Did you rate it as good as MerlinB did? pls Midwayfair The Engineers Thumb that is?
I know its off topic sorry guys.

News:

Lovesqueeze comp - howzit work?