Phase 45 FET matching

[idlefaction]

hi all

i built up a phase 45 - it's nowhere near as noticeable as the sound sample on runoffgroove.  i was wondering how closely people match their FETs, and what method is recommended?

i've read up some very good threads on FET matching, esp. one between rob strand and mark hammer where they compared their FET matching circuits and did some random spice FET generation etc.

the FETs i ended up with were some 2N5458s, and i matched them in Mark's opamp bridge circuit, with a Rds of 5k6, 10k and 100k.  the FETs were closer or furthur away in each case.  i went with some that were matched at Vgs of -1.264V and -1.260V at Rds of 10k, as this seemed about the right Rds for the circuit.  i thought i had it sussed...  until it put them in the circuit and it sounded really tame!!

i have 20 FETs from the same batch - i'm tempted to just stick in some sockets and start trying combinations until i find something i like    the 20*19 permutations scare me though.  any advice?

maybe i just need a six-stage phaser  

[brett]

This is a surprising result....My phase 45 works well with any JFETs matched to within about 0.02V (0.05V sounds ok too).

Are you sure that you're testing them the right way around, etc?  I'm not sure what I did (maybe legs in wrong sockets?), but my early tests on JFETs were all wrong, and I thought I had JFETs with low, matched values, when they were actually all over the place.  It might be worthwhile checking that you measure a Vgs @ 0.5mA (10k@4.5V) in a J201 of about -0.5V (ie quite high) and in an MPF102/2N5486 about -3 or -4 or -5V (ie quite low).

You should get a fairly strong whoosh out of it.

Good luck!

[idlefaction]

i should make some sound samples - maybe my idea of 'tame' is your idea of strong!  

but yeah, i'm sure they went in the test circuit the right way.  i actually tried some the wrong way around, and got a reading of 3.31V for every single one of them which was kinda interesting.

funnily enough, i did test a bunch of 8x J201s i had, two of which were matched very closely and most of which were 0.3V-0.4V, and two 2N5486s i had which were at 2.91V and 3.13V at that 0.5mA level.  the 2N5486s couldn't bias in this circuit, getting the trimpot in the right place pulled the voltages all over the show, and the J201s were really touchy on the trimpot and didn't sound very noticeable either.

i'm thinking maybe i have a cap in the wrong place or read a resistor wrongly or something!!  time to go check all the component values again.  

[bwanasonic]

I take it you have a bias trim pot and you've twiddled that already? I have a couple of DOD 201s, and tweaking the bias pot really helps dial in the effect. Once you get it working the *proper* way,  be sure to try the *univibe* cap values as per JC Maillet's (sp?) thread. If you click on *links* at the top of the page there is a link to the thread.
My local music store is nice enough to let me come in and A/B my homebrew stuff against their stock when I need a frame of reference. Maybe you could try a Dunlop reissue for comparison?


Kerry M

[idlefaction]

yep, bias pot twiddled.  i only get effect over a very short travel of the pot.  need more tiiiime