Script Phase 90 (GGG build) with unmatched JFETS question

[gbkd80]

About a week or so ago, I built up the script Phase 90 from scratch, etched the board etc... while I was part-ordering, I neglected to focus on the "matched" JFETs needed for this circuit, so I just bought a handful of 2N5952s.  It wasn't until I was putting the board together that I re-read the BOM and realized "oops" - I used them anyway after some quick n' dirty research on the subject. 

The circuit works.  Well, i think it works... It phases.  It sounds sorta "right" but I'm only able to get phasing with the trimpot in a very narrow sweet spot.  There's not much room for adjustment.  I just want to be sure that the reason this is happening is because I used unmatched 5952s and had I matched them, it would have had more room for play?  I am only guessing this by looking at where it lies in the schematic.  Anyway, just curious more than anything.. like I said, it seems to work fine in that sweet spot.  Unfortunately I don't have a real Phase 90 to even compare it to other than video demos..

[R.G.]

It's good that you get phasing. It is possible that the phasing would be more intense with matching. Or you could have lucked into a set that's OK-ish.

If you have a sweet spot and like the sound, it's fine. Come back to it when you have the time and inclination - and more JFETs! - to tinker and find out whether it's better with matched.

[Mark Hammer]

The matching is not required to obtain phasing.  Where the matching helps out is in the "turnaround"; i.e., at the upper and lower extremes of the sweep.

The matching increases the likelihood that all four JFETs (or however many the circuit employs) will show a change in resistance in response to the LFO.  Or stated another way, it decreases the likelihood that one or more JFETs will say "Nope, my shift's over.  This is as far as I go." as the sweep goes out to the lowest or highest points of the sweep.  When they are matched, you have greater assurance that whatever bias setting you put it to is optimal for all 4 JFETs, and not just optimal for 2, pretty close for 1, and not quite there for the 4th.

[gbkd80]

Great, thanks for the explanation RG and MH - makes it a lot clearer.  I was thinking about getting a matched set and socketing them to see what works best before boxing this all up.  My OCD almost won't let me finish it unless I know for sure I tried it both ways. 

And really, what in this case constitutes "matching" anyway?  I only really know about matching germaniums for high and low gain...

[Mark Hammer]

My personal view, only, but I think if you have the Vgs matched within 100mv or so you're pretty good to go.

Steve Daniels sent me an assortment of 2N5458s last year, that had been selected to cover a range of values; each one labelled with its individual Vgs.  I found I was able to get decent "musical" phasing from just about any 4 consecutive values I used.

[nickbungus]

Hi

Did you socket the transistors?  I've got hundreds of J112's already matched, I can send you some FOC.  Just PM me.

[rightbeforeimpact]

I also built this GGG script phase 90 but I did use a "matched" set. My friend and I actually built 2 of them together and mine seems to "phase more" than his and it only did so in a narrow band of biasing the trimpot. It sounds like you're fine, but I wouldn't turn down some free JFETs from nickbungus.