Coolio! -I may have forgotten to mention point 'A', but I think I did mention 'C' in my first list of points to link; I think I may have described it as something like "detector output after the diode" or something. I don't think that I did link the ±12Volts between the two boards, but I did have them linked directly to the slave's board, which should be the same thing.
-I reckon you don't need those ±12V links, in fact having them means that if either regulator has a slightly differenct notion of what is precisely 12 volts, there could be a noise issue... try removing them (assuming that you have the slave ratio board fed it's ±12V from the main slave board, of course!) and seeing if it shill works... it should, and since you won't have the two regulator outputs connected directly to each other, there should be no chance of a "fight"...
Rob, one of us should probably do a summary of all the modifications, to save the next person from having to 'trawl and compile'...
Mine has been out on loan for the last 12 weeks, at a local SSL studio where they do a LOT of rock music. I can't get it back because they're in the middle of a 3rd album mix, and it's been used on EVERY SINGLE tune in those 3 albums. -The chief engineer gave me a pack of mic stands brand-new-in-the-box as a thankyou present the other day, just because I mentioned that I needed some mic stands, and they'd been trying to think of some way of thanking me... if that tells you anyghing!
Radiance, there's one way to trim distortion with a scope and a notch filter (or a few narrow-bandwidth maximum-cut EQ bands)... it's tricky, but it CAN work...
Switch the scope to X/Y, connect a pure 1kHz input tone to the input signal to the scope horizontal input and the input of the compressor.
Take the output of the compressor and feed it through the notch filter, or "gang-up" at least 4 bands of EQ all set to match the input frequency, maximum-cut, narrowest bandwidth.
listen to the output signal, and also connect it to the scope's vertical deflection. Wind the gain up MASSIVELY, so that it is very sensitive indeed.
Tweak and tune the bands' center frequencies so that the signal you hear and the "diagonal' deflection that you see on the scope is minimized.
Now, if you start to see 'waves' on what should (in a zero-noise, zero-distortion case) be a perfectly flat horizontal line, you're looking at harmonic multiples of the input frequency. 2 waves is second harmonic. 3 waves is 3rd harmonic, and so on. Noise will be a blurry fuzz.
Tune the symmetry/distortion trim deliberately far from center. You should see some waves and hear a 'sharp' sounding note relating to the tone start to pull through the noise and background tone. Back off the adjustment you made and it shoudl start to fade back out, and at some point it will be replaced by another -usually slightly different sounding- sharp tone.
Somewhere in the middle of this, you should find that there is a sort of null point, perhaps where neither 'sharpness' is perfectly vanished, but perhaps where they are at an agreeable minimum. -Congratulations... that's your lowest distortion point!
A few notes: You'll need an oscillator which doesn't drift up & down in frequency, or you'll never get it nulled out by the filters/EQs long enough to adjust. Likewise you'll need an EQ (or EQs) with similar center-frequency stability.
If your oscillator has 1%THD or something (many do) then you'll hear this distortion LONG before you hear the compressor's distortion, so eliminate it by doing the OPPOSITE EQ "trick", by filtering off everything above and below 1kHz, and perhaps boosting 1kHz on a very narrow 'Q' while you adjust the level of the oscillator down to avoiod overload.
Sure, it's convoluted, but it's resourceful and you CAN do it. Basically it's how an analog distortion analyser works inside the box, and you'll learn a lot about how distortion "sounds" and start to think about the composition of that distortion, and what it sounds like.
Top work Rob, -Now, Radiance: -prove me right... -Build one and trim it up using the "Blue Peter" method I described!
:wink: :thumb:
Keith
