duhvoodooman wrote:
Tried the same output tranny bias as Morgan, hooked up the same way. In my case, this made only a slight difference. Then again, I have never noticed a distortion problem with my Phase Royal, nor have I had any problem getting a good, deep phase cycle on it. So maybe the normal variation inherent in component tolerances can result in these apparently significant differences between builds?? I dunno...just a theory.
I believe your theory to be correct. It seems like just about all DIY Phase 90 clones have had this inherent "one works great, the next one doesn't" issue going back to the beginning of the century. You can see it in build reports of the Tonepad version, and it used to show up on DIYSB for the GGG project. Some builds would have this ~ 6 dB boost and the typical 'fix' would be to change the 150k feedback resistor on the output amp to a 120k. My phase royal had this when I first built it, I swapped it for a 120k and it got rid of the boost, but it still clipped too much and the phasing wasn't very deep. I ripped it out of the enclosure and experimented on it for about a year, tweaking different parts of the circuit, measuring & replacing fets, until it was this f-ed up POS. I finally recently returned it back to dead stock, and curiously the boost was gone but the clipping remained. The 68k mod patched it up to the best it ever has sounded.
Also, I built 5 or 6 of the previous BYOC phase 90 clones and I got exactly one build that sounded like a really great vintage script Phase 90. The others were good, but lacking in depth. All of them had matched fets. I think Tark is onto something with the vref adjustments and the output stage bias fix.