Ross extra stages

[jimbob]

Has anyone built the Tonepad Ross Phaser and added extra stages. I have never (as far as I can remember) listened to or used a Phaser w extra stages. Is there a big differece between 4 and 16? What about if I added 32 more? Does more stages make the phaser better sounding or more full? I dont get it.

[Mark Hammer]

In a flanger, as the notches are swept downwards (i.e., as delay time is increased), the notches get closer together such that you can have a lot of notches.  As it sweeps upward, the notches space out.  In a phaser, the notches are fixed in number and spacing.  As you sweep upward, you don't really "lose" notches, but more of them start to be above the functional bandwidth of your rig so you just don't hear them.  I've heard phasers with as many as 24 stages, and while they can sound great, there are noise and current consumption issues stemming from the large number of chips involved, and you start to get diminishing returns on anything more than 12 stages unless you have a very harmonically rich  and sonically complex signal and speakers with wide bandwidth.  For guitars, you start to run out of real benefit after 12 stages or so.  But 12 stages sounds one whole helluva lot more interesting than 4, I'll tell you.

[RaceDriver205]

Mark, what would be the best way of implementing 12 to 24 stages?
I.e which would have the least noise, parts, board space, or cost?
I was thinking JFET style with 6 TL074s, or 12 NE5532 if noise is a problem.

Thanks, Mr Oracle, sir.

[Mark Hammer]

One thing you probably ought to do in making a FET-based unit would be to select the chip based on the resistances used.  In some instances, you will see a 10k feedback resistor and 10k inverting input resistor.  That strikes me as a good spot to use a 5532.  If the same components were now 100k, a BiFet like a TL072/74 might be more useful. These chips show different noise characteristics under different use conditions.

A second strategy to adopt is to stick a small feedback cap in every 4 or 6 stages.  That is, in tandem with the feedback resistor in that stage, stick in a small cap to roll back the high end.  If the feedback resistor was 100k, go with something like 100-120pf.  You'll lose a bit of the "zip" in notches way up high, but you'll also lose some of the accumulated hiss in the process.

T'wer I, though, matching FETs for that many stages will prove to be one big headache.  Far better to go with OTAs for anything that ambitious.

[RaceDriver205]

Mmm. But my concern with OTAs is that they require an opto per stage, and the only real supplier is smallbear (their NJM13600D is 0.75c).
On the other hand, I can get a bucket of bf245s or j201s or something for a fair price from futurlec, and use quad ops.
But then I dont know of a way of applying a hypertriangular clock to a 'jfet-style' phaser.
Argh!

Would it be possible to graft a hypertriangular clock from an OTA-based phaser to a JFET-based phaser?
Otherwise I may just get a huge bulb with LDRs around it or something.