modding the attack/release times on compressors

[themagicmanmdt]

random question.

so, I was digging deep in Recording The Beatles, and had a thought hit me reading the Altec mods section.

is there a particular component that causes attack time to be what it is?

a small line in RTB alludes that attack time was altered by simply switching in different values of capacitors.

is this true for all compression circuits? Does attack time depend on a simple component?

if so, is it possible to take your favorite compressor and with a quick substitution, have a different set of attack times?

[GooberNumber9]

I'm not an expert.

From what I know, mostly "yes" is the answer to your questions, although technically the attack time is not based solely on one single component in any circuit. Most likely attack time is related to an RC circuit somewhere in the detector. There's a formula somewhere (but not in my memory) for calctualting the time constant of an RC circuit. The important thing in terms of your questions is that the R and the C work together, so changing the C will definitely change the time constant, but the R is not something to be ignored either.

Now I will get out of the way for someone who really knows this stuff to help.

Todd Wilcox

[lacquer_monkey]

Attack and release times are set by resistance/capacitance (RC) networks. The time constant is the resistance times the capacitance. In the old tube compressors this is often a fixed value. By changing the capacitor in the network both the attack and release times will be changed. By changing the attack or release resistor the atack or release can be changed without effecting the other. Using a pot allows variable timing.

[widdly]

Of course it probably wont work so easily for an optical compressor.

[earl parameter]

in most cases yeah just adding in a cap in parallel or even increasing a pot value will help you get different ranges but there are a lot of other factors depending on the design. its often easier to gets slower rates then it is to speed them up.

did you have something specific in mind? if you can post a scheme i can make suggestions


steven

[themagicmanmdt]

just thinking of speeding up an older altec unit to do something faster than half a second.

[earl parameter]

if you have a scheme i can point you there


steven

[brianroth]

Of course it probably wont work so easily for an optical compressor.

The attack/release time constants for an opto compressor are almost always due to the response of the photoresistors. Example: in the LA-2/3, the audio signal is used to directly drive the electroluminescent panel, so the "time constants" are due solely to the specs of the photoresistors.

Bri

[earl parameter]

right and there are ways to slow it down but unless there is already something there slowing it down there isn't anyway to speed it up short of replacing the vactrol with a different model but that can cause other oddities that might not be wanted or force you to change other components and in the end, change the sound, which i'd assume is unwanted

so are you guys saying the altec he's talking about is opto? i have no idea myself


steven

[djimbe]

Sometimes you get lucky with a particular box and the design and documentation are such that the mods jump right out at you. I'm thinking of the CBS Audimax 444 right now. That's NOT an opto unit. It is mostly a broadcast limiter which means it has a pretty fast attack and a LOOOONG release, all hardwired in the unit. But the circuit is designed such that you can change that release time just by swapping out a resistor that was conveniently landed on a set of stand off pins on the board. Just like it was made to be altered. 'Cause it was.

"So I can change the release time by using a different resistor, huh? So if I build a m-b-b rotary switch with several resistors I can have several release times, I betcha".

See how that works? Mind you, I'm not smart enough to have figured out which is the right resistor on my own, but having the correct documents for the box sure made it easy. Yes, my 444's do NOT have the stock release system...

[widdly]

Of course it probably wont work so easily for an optical compressor.

The attack/release time constants for an opto compressor are almost always due to the response of the photoresistors. Example: in the LA-2/3, the audio signal is used to directly drive the electroluminescent panel, so the "time constants" are due solely to the specs of the photoresistors.

Bri

There are opto compressors that do have attack and release controls though. Like the Avalon VT-747s (http://www.avalondesign.com/vt747sp.html)

I've never been able to figure out how that works.

[A.David.MacKinnon]

What I'd really love to know is how to implement the "Hold" switch on EMI's Altec compressor. From the discription in Recording The Beatles it sounds like the equivelent of setting the release to infinity.

[nclayton]

The 436's were vari-mu, and I'm pretty sure the CV was negative, applied at the input transformer's center tap and from there to the grids. Usually this type of compressor can be pretty fast on attack. I can't remember if that model had a resistor to slow it down, but if you feel like it's pretty slow, then it probably does. If it's really a half second, then it definitely does.

There's going to be a small rectifier tube, maybe a 6al5, and it will have the time constant network attached to its anodes (plates). If you can figure out which pins are the plates (they should be wired together) then see if there's a resistor coming straight from them to a capacitor whose other side is grounded. If you can find something like that then you can speed it up by making the resistor smaller or wiring a second resistor in parallel, or reducing the size of the capacitor.

This is important: a lot of old compressors drive the rectifier directly from the output tube without an intermediate buffer. I don't know about 436. If there's a slow down resistor going into the cap, then reducing it will also reduce the impedance the output tube sees and could increase distortion. So you don't want to reduce the slow down resistor size TOO much unless you're sure the output amp can handle it. Better to put in a smaller cap (although this will also shorten the release time).

Since the compression control voltage just drives the grids of the input tubes, it needs a bleeder resistor to provide a release time constant. There should also be a second resistor going from either the rectifier's plates or the RC junction to ground. It will most likely be a fairly large resistor. If you want to speed up the release, make that resistor smaller (but not too small or you could make audio distortion as above) or if you want to slow down the release make it bigger.

So if you want to speed up attack, but keep the release, you could reduce the cap and increase the bleeder by the same ratio.

If you just remove the bleeder resistor than there's no official path for the control voltage to discharge through. The rectifier can pump more current in to the cap to charge it to a bigger negative voltage, but then it will just get bigger and bigger. The only place for the charge to flow out is through leakages, and that will take a long time. I can only imagine that the emi mod mentioned for holding the CV just amounts to removing 1 resistor. If you look at the platees of the rectifier and find the large value bleeder resistor that goes to ground, all you'd have to do is remove it and there you go.

[earl parameter]

There are opto compressors that do have attack and release controls though. Like the Avalon VT-747s (http://www.avalondesign.com/vt747sp.html)

we are not talking about having a knob to vary the attack, we are talking about changing the overall range that the attack control normally has

good write up nclayton.

i have not grabbed a copy of the RTB yet so i've not even heard of the hold idea but nclayton's idea should work. there might be some unforeseen side effects depending on the design so play with it a bit before you settle


steven

[emrr]

The hold position is indeed one which disconnects the bleeder, so the CV never bleeds down.

I've read elsewhere that the 436/438 family work as fast as they are able and still sound good and I'm inclined to believe it. Very much faster and you have distortion thumps with transients. Getting seriously faster requires an involved re-design, which really means you need to move to a different limiter. It's much easier to take a limiter like RCA BA-6A or Gates SA-39 and make them act like the slow compressor families (Altec 436, Gates Stalevel) than it is to go the other way. They are more complex for that very reason.