mulletchuck said:
ah, thanks Amaziad.
mnats, no room for the Stereo Link boards from Hairball, I see...
I've just completed two stereo g1176 clones. One based on OEP (with R84 mod) and the other on the original Lundahls.
They are both super silent (thanks a lot to an off board PSU) and I'm delighted.
Thanks to Mnats for the boards, Purusha for the case, and Gyraf for the design.
I bought and sorted through 100 BF245A's during the build using a matching procedure like this one http://www.nrgrecording.de/html/fetmachting.html
It only cost something like 30 euros for 100. And it took me maybe 2 hours to do the measurements on a breadboard.
I ended up with two sets of a quad of 4 closely matched FET's = 2 boxes * (1 FET for GR compression + 1 FET for GR metering) * 2 channels /box matched at 3 points on the Vgs v. resistance curve (i.e. at -1 -3 -6 dB GR to approx ±0.02V). I used the closest matched pair for the actual GR compression FETs on channel 1 & 2, and then the next nearest match to them as a pair for the GR metering. Same for the other box. Once I'd set them up it was pretty cool to watch and hear.
In the end, these things tracked in the stereo pair to within 1dB over the whole compression scale on virtually any setting. OK the internal metering was sometimes a tad off (±1dB) compared to my more accurate external metering, but the cross-channel tracking is so tight in fact that you couldn't really see when the link switch is open or closed when the same signal is applied to both channels and you pre-align the control settings.
For stereo you really can just connect the two pad 7's together. The UREI link board was to compensate for mismatched FETs with different DC biases (and not any tracking by the way).
So who needs stereo link boards if you take the time to match FETs?