Matched fets?

[col]

I noticed that a lot of threads relate to matching jfets for phasers etc. When I was reading one of the Penfold books (Practical Electronic Music Effects units) I noticed that he uses the two n-channel devices in a 4007UBE for a phaser and wondered if this could be used more widely as it would be cheaper than two J201s and presumably are ready matched,

Col

[e178453]

Don't know the practicality of using other than specified.  Economically, not sure it's worth the effort.  Checked out 2N5952 fets fro a phase 90.  Mouser sells them for $.12 each if you buy 25.  So for $6 you could buy 50. Geofex has a very simple matching circuit, check it out, guy there says he got 4 or 5  good matched sets and a couple OK sets of 4 out of a lot of 75.

[R.G.]

Both right.

The CD4007 is a "do it yourself" CMOS gate that has matched MOSFETs inside. I've used this for matched MOSFETs before (See the MOS-Doubler at GEO). However, the MOSFETs in a CMOS gate share some problems with all MOSFETs - the MOSFET substrate/body diode. This looks like a reverse biased diode in parallel with the drain-to-source. That means that you can't put more than about 300mV in the reverse direction across them before that diode starts causing distortion.

EH used a CMOS gate in a phaser once. I think others have too. If it were really, really good, other people would have too, and the famous phasers would not be th P90 and Small Stone, but instead those CMOS phasers. That didn't happen, ergo other designers didn't like them either.

There's not much unplowed ground in analog effects.

[Mark Hammer]

Mike Irwin built the ETI phaser using a 4049, and reported that it was particularly noisy.  Those familiar with that project will note that it lacks any regeneration.  The reason?  Too damn noisy.  Can be improved upon with some design changes, apparently, but in the end nowhere near as quiet and well-behaved as an LDR-based phaser.