PHASE 90 QUESTION

[Tombaxter97]

Hi, ive just tested my phaze 90 from fuzzdog, biased fets to 4.5v and im getting a really subtle effect no matter what the speed is, can this be attributed to unmatched 2n5457's?

THANKS

[Mark Hammer]

1) Near as I can tell from your description, you are confusing two different sorts of FET biasing.  If one is using FETs as gain stages, then yes, biasing to half the supply voltage allows for maximum voltage swing for each half of the audio signal.  Biasing of FETs when used as variable resistors, however, is different.  In that case, one is biasing the FET such that applying current to the gate from an LFO yields a change in drain-source resistance as that current changes.

2) The importance of "matching" FETs has to do with what I like to call the "turnaround".  The most musical phaser sounds are obtained when all control-elements - be they FETs, LDRs, or anything else - all move/change together.  They all continue to produce changing phase shift from one peak of the LFO cycle to the trough and back again.  If the FETs are not reasonably matched, it can be the case that, during the sweep cycle, one or more of the FETs reaches the boundary of how much the drain-source resistance can change, and simply stops until the sweep cycle has gone to the end and then returned to "pick up" that FET along the way.  There will be pleasing phase-shift for some of the sweep cycle but not all of it, and typical gets perceptibly weaker at "the turnaround".

Sometimes, one can get lucky and even a fistful of unmatched FETs turn out to yield changing drain-source resistance without having to do anything fancy.  And other times, you need to be picky and hand-select (i.e., aim for matching).