Fet "matching" with trimmers

[ildar]

I was wondering if it would be possible to use trimmers to "match" FETs for phaser circuits. If possible, this could make trying to find matching FETs unnecessary. Has this been covered before?

[R.G.]

I was wondering if it would be possible to use trimmers to "match" FETs for phaser circuits. If possible, this could make trying to find matching FETs unnecessary. Has this been covered before?

It is possible, but quite tedious.

You need at least two trimmers per FET.

One trimmer sets the DC back bias on the gate relative to the source; this sets the cut-in point where the JFET begins to not be completely cut off.

The other sets the range of LFO voltages. Some JFETs need bigger or smaller LFO voltages so they stay in the good resistance range (not necessarily tracking, just in range) while all the JFETs sweep.

My take on it is that since I can get a JFET cheaper than a trimpot, I'd prefer to buy three times as many JFETs as I need and match them than to buy the right number of JFETs plus twice as many pots and then tweak two pots per JFET.

But it can be done.

[Rob Strand]

In practice, you only really need one pot per JFET.  It's more of a problem stuffing around trying to adjust all the pots by ear, but it can be done.

In the context of single pot adjustment:  The main problem with JFET matching is when the voltage is near the pinch-off voltage.  If the JFETs are mismatched in this region the resistance errors are *enormous*.  That's why JFET matchers that match the pinch-off voltage produce good phasers.

To a large extent an adjustable bias alone for each JFET solves the main problem with JFETS.  

Matching at other resistance points isn't very critical at all.   For example you rarely hear people complaining about opto-coupler base phasers being out of match yet opto couplers have pretty poor tolerances.  The key thing about optos is they don't have the *massive* mismatch like you get with mismatched JFETS at one end.  If you look at the univibe the time constants are mismatched from the start, by a large factor, yet in practice the sound is only slightly different - nothing like a phaser with mismatched JFETs.

[Paul Perry (Frostwave)]

If money is a serious problem, and you don't mind wasting a lot of your life, you can always use trimpots, then replace the trimpots by selected resistors.Using one of those handy calculators that show which two 1% resistors in parallel give the value you want.
It sure makes one's day when you try servicing something where this has been done  :?: :x  :cry:  :cry:  :cry: