SSL Clone compressor ratio measurements

Having completed a couple of compressors for a few folks round here lately, I've noticed that the effective ratios vary depending on what VCAs have been used for the sidechain and the main signal VCAs.

My suspicions were first roused when one that I measured (built as per Jakob's diagram, with 2181s in all three locations) actually produced 'past-infinity' results when set to 10:1 ratio... Obviously the two audio VCAs are turning the signal down more than the sidechain VCA...

I set a second one on the bench this evening which used 2001 VCAs in the audio path (Class 'A') and a 2151 or 2150 -I forget which- in the sidechain position. I noted that it went nowhere near 10:1 on the maximum ratio.

I tried a couple of different resistances in parallel with the (1k) control port series resistance, and I found one that produced a 1dB rise in output for every 10dB rise in input above the threshold, when the 10:1 ratio was selected. If I remember correctly, it took a 5.6K resistance in parallel with the 1k on each channel.

The 4:1 setting also matched reasonably well once above the (softer in that ratio setting) knee area, and also the 2:1 setting seemed to track fairly well... I think it was about 1.9:1, but close enough! -It should be noted that at this setting, the knee is very soft indeed! Only the 10:1 ratio has a sharp corner, both the others have an "over-easy" (dbx tm) soft knee behaviour, with the 2:1 setting being very soft indeed.

The chap with the 2181s is on vacation at the moment, but I think I'll get his back on the bench when he gets back if he wants the ratios calibrated... I think that the results will be useful to everyone. -If you want to vary the ratio but keep the knee very sharp (for whatever reason) you can simply increase or reduce the control port resistances on the two signal path VCAs... but bear in mind that you can go past infinity if you reduce it too much... a wierd effect, and one that most feed-back compression systems cannot achieve...

Keith

Is going past infinity like traveling faster than the speed of light?

what does that sound like?

dave

soundguy,

I think it sounds like you are hearing it before you hear it...

I'm going to make one of these, and this has most likely been asked before somewhere, BUT, what is the best and easiest VCA to get?

Nobody really has to answer this, but if they do it will be something like two men digging three holes in the ice somewhere in Alaska wondering about VCA's and...does my mother know I'm doing this sort of stuff as a hobby, I hope no one will tell her I'm suppose to be in law school. I think my cat is hungry, I think Grape Nuts cereal is DA BOMB!!!!!

Bluebird, did you fall off?
:shock:
Keith, how did you like the sound of the 2001 once you got it dialed?
:roll:
Shreaded Wheat with sliced bananas on top is daaa booombbb!
:grin:
Chew your cud, boy!
:guinness:

Fall off? Was he ever on?
3 months! Sheesh.... :shock:

chef

hmm....I´m using 2159 vca which I got from Frank.....curious how they measure ratiowise. but on the other hand the comp sound perfect on my drumbus so I think I´ll leave it as it is :thumb:

I´ll check it out when the current production is finished. thanx keith!

steff

Thanks, Keith!

I haven't thought about checking ratios with the recent developments in VCA's - I see there is some more experimentation needed here..!

Jakob E.

I'd like to check this in the GSSL I have. I'm using three DBX2151s.

For example, at 2:1, can I just set a threshold point then input a signal 4dB greater than the threshold, then measure the output, which should have increased by 2dB? Then do something similar for 4:1 and 10:1?

Greg,

stay away from the threshold area... specially at 2:1. There's no sudden corner, since the knee is soft.

A better method:

Set the threshold so that there's about 8dB of compression. Set the make-up gain to read a known figure... perhaps 0dB on whatever scale you choose. Then increase the input by a calibrated amount... I suggest 10dB. At 2:1 compresion, the output should increase by 5dB for a 10dB input increase. -starting with a goodly amount of GR at the baseline measurement before you increase the input signal will make sure that you're away from the 'knee'.

For each ratio measurement, re-check your baseline level and re-adjust the makeup gain to get you to a convenient point, if necessary.

Keith

[quote author="soundguy"]Is going past infinity like traveling faster than the speed of light?

what does that sound like?

dave[/quote]
Well, you've probably heard this before...

:wink:

keef

SSLTech - Thanks for the tip... simple enough. Should be able to get this done this evening. I'll post some results after.

I did a quick ratio check on my clone yesterday evening. I didn't have my scope or multimeter with me, so I used my console VUs. I did just what SSLTech suggested to stay away from the knee.

2:1 - Input increased 10dB, Output increased 4dB.
4:1 - Input increased 10dB, Output increased 2dB.
10:1 - Input increased 10dB, Output decreased 1/2dB.

Am I reaching the "infinity" point at 10:1?!?

Your actual ratios are:

in the 2:1 setting, you have 2½:1,
in the 4:1 setting, you have 5:1,
in the 10:1 setting, you have 20:1

-Sounds like your ratios are all just a bit on the high side, you might want to increase the 1K resistors a little... perhaps try 1.1k (stick 100Ω in series with it) and see what that does for you. (You only need to do one side and measure that side... when you find the right value, stick that value oin series with the other side and measure that to make sure there's a good correlation.)

You want to get to the following:

2:1 -Increase input 10dB, output increases by 5dB.
4:1 -Increase input 10dB, output increases by 2½dB.
10:1 -Increase input 10dB, output increases by 1dB.

If 1.1k is too far the other way, try 47Ωinstead of 100Ω in series... or if it's not enough, try 150Ω... you get the picture, I'm sure. -Good luck and let me know how it goes!!!

Keith

Hey Keith,

"I tried a couple of different resistances in parallel with the (1k) control port series resistance, and I found one that produced a 1dB rise in output for every 10dB rise in input above the threshold, when the 10:1 ratio was selected. If I remember correctly, it took a 5.6K resistance in parallel with the 1k on each channel. "

If I remember it correctly it was a 560R resistor.... I'll see if I can open it up and check again..
btw, thanks a bunch for all the help with the clone...

Gil

Gil, that was for the 2001 VCAs....

For the different VCAs like the 2151s, and the 2181s (like in Nick's one) it seems that there are slightly diffrent values.. yours was under compressing, Nick's is actually over compressing, so his resistances need to be increased...

Actually, now that I think about it... Greg also needs a reduction in resistance, but I don't think it will need to be quite as severe as yours... If Since I had yours on the bench it was easy to have the Neutrik plot the amplitude curves for me...

My head hurts...

Keith

I finally have the chance to get my GSSL out of my studio and I want to fix my ratios. Before I did this I wanted to make sure I adjust the correct resistors. Are the 1k's circled in red below the correct resistors to adjust?



Thanks.

Hey Greg, thanks for the picture. :grin: I can see what you are talking about. Now, my question is...from the design notes on THAT Corp's site, design note 137 in particular is R2 the culprit that is causing the extra R value that is making the 2181's clamp down harder than we(the design for 2150's) require. The VCA contains an internal 27 Ohm resistor that should eliminate a parallel resistor attatched to the postitive supply voltage. Is this what makes the 2181's hyperactive in the compression department? :shock: :shock: :shock:

Hey this will be my future bug to fix since I have the 2181's...which I can't wait to trim .

Anyone ?!? SSLTech, Gyraf? I'm itching to get this corrected. Just want to make sure those are correct resistors.

Greg,

Yes.

If you parallel these two resistors with other resistors, you increase the VCA's reactions to control voltages.

Jakob E.

I have an idea :grin:

if I would set the ratio switch to 10

and connect a pot with the 1k's in series then I would have full sweepable ratio control, from 1:1 - 10:1 :?:

mat