Keith's "Blue Peter" distortion trim method (pics)

radiance · Mar 5, 2007

In his "New 'ultimate' SSL buss comp clone ;-)" thread Keith describes a method to trim your VCA's without a distortion meter. It can be found here....scroll down a bit.... At first it seemed a bit over my head but after reading it a few more times it started to make sense, so I thought I might as well give it a try. In the following picture sequence you'll see how I trim my "normall" (it's not an ultimate gssl a la Keith) GSSL compressor with gold can DBX202 VCA's.

Here's the test setup. What you don't see is my DAW (Logic Audio) which is used for generating the 1 kHz sine wave...

The 1 kHz sine wave is generated by the test tone of Logic's EXS sampler. On top of that I put the Channel EQ with a very narrow 1kHz boost to get rid of any harmonic distortion that might have been in the sine wave itself.

The sine wave goes via the GSSL and a Behringer Ultra curve (DEQ2496), with three stacked parametric filters set to narrow bandwidth and full cut at 1kHz, to one input of my scope. The same sine wave goes also directly to input 2 of my scope.

The scope is set up to display both waveforms with one input inverted.
The (blurry) sine wave on the top is the one which goes through the GSSL (which is bypassed with a relay hard bypass BTW!! so the signal does not travel throught the GSSL).

When you combine both sine waves you get this.....not a perfectly straight line as you can see. This is due to the latency from the Behringer ultra curve (a digital device). I could compensate this a bid by adding Logic's sample delay plug in set to a 10 sample delay (on the channel which goes directly to the scope)....

When I put the GSSL from bypass and the scope back in dual mode I got this....

When combined it looks like this...

When I turn the distortion trimmer fully CW..

And CCW....

When trimmed correctly I got this almost straight line which I guess is fine...

And back to dual mode it looks like this...voila, a nice sine wave...

As you can see the distortion is clearly visible even when the scope is in dual mode. It leads me to belive that you can trim the VCA's with only one sine wave going to the GSSL and a EQ and withouth another one to cancel it out.

Here a pic of the guts from my GSSL. It's with Greg's filter board, separate PSU, DBX202's, relay bypass and 23 pos Elma rotaries for Threshold and Make-up for easy recalling settings (and because I could get them for cheap on evilbay...)

And the front.....done by Schaeffer in "mittel bronze" aluminium.

pucho812 · Mar 5, 2007

when I was growning up "Peter" was slang for a man's private parts.
So reading the first part of the thread title had me cracking up...

SSLtech · Mar 5, 2007

Looks good... but did you try X/Y?

That allows you to look at the distortion order by "counting the points on the crown" as it were.

None the less, you've certainly trimmed the distortion down; probably just about as well as you could do it with a $2000 test set!

Incidentally, the word 'Bypass' is misleading here: nothing is in fact bypassed, and the signal still passes throughthe entire chain; just the gain reduction and makeup voltages are disabled.

The slight blur on the trace is noise: either RF pickup VCA or quantisation noise. Since you're looking at distortion products about 60-80dB down (and lower, in fact) then you have that much gain, and so some noise creeps into the measurement. -None the less, you can see that the blur is noise, and the repeating 'pattern' is the distortion. That helps you to 'see' the adjustment.

The difference (if you listened to the filter output as well as viewing it) would be a resulting "buzz" sound in amongst the noise, at (in this case) one octave above the test frequency (i.e. 2kHz for a 1kHz test signal).

But anyhow, well done! It's time consuming to set up, but that setup is useful for testing, measuring and gaining clues about other things you build and measure.

Remember, in electronic pioneering days, they had to make their own test and measurement setups also. this was in the days before Audio Precisions and Neutrik A2's...

:thumb: :thumb:

Keith

radiance · Mar 5, 2007

[quote author="SSLtech"]Incidentally, the word 'Bypass' is misleading here: nothing is in fact bypassed, and the signal still passes throughthe entire chain; just the gain reduction and makeup voltages are disabled.

[/quote]

I was not very clear here....The bypass is done with relays so when in bypass mode the signal does not go through the GSSL.
The relay hard bypass board is the blue-yellow-orange birds nest attached to the in/output side of the main pcb. The relays are soldered on the under side of the prototype board because they where SMT ones...