understanding phaser and resonance...

[swt]

I know....i read geofex tech years ago...but i want to build a phaser with a compilation of ideas from different ones. And this is something not brought as a subject, or at least couldn't find it using search functions.
Taking a look at most of the diy phasers, seems like most of them have the feedback path entering the second stage of the neg feedback opamp. with the exception of the small stone, but as mark said, using more stages that way, regeneration sounds worst than the ross for example. But case is, that the classical ones, like mutron phasor II, biphaser, boss ph2, moog 12 stage, all have the feedback coming to a differential opamp.
What's the difference sonically between the two?. How can i use positive and negative feedback with the same circuit?. can i use the differential stage and use switching to feed the signal there, or at different stages?. should i lift the voltage divider to the noninv input if i do that?. Or can i have that resistor and cap to vr on that stage, and use it as an inv input with a set input imp?.
Sorry for the thousands of qs, and my bad english...Thanks!!

[DougH]

My best advice would be to breadboard a phaser circuit, like a phase 90 for example. Then start trying out your ideas and experimenting with injecting feedback to different areas of the circuit and noting how the changes affect the sound, and so forth. Your questions are fairly broad, and the best way to find out IMO is to roll up your sleeves and get your hands dirty. I have a phase 90 breadboarded right now and I tried various ideas out on it. (This time I'm building it stock, but anyway...) For me, it's about the most complex complete circuit I would want to breadboard, but it actually wasn't too bad. It's fundamentally 3 dual op amps, 4 matched JFETs and yes, after one minor mis-wire correction it fired up and works fine.

In any case, I think breadboarding is a great way to learn how circuits behave and you can let your ears do the work. A scope and sig generator would be ideal too, for *really* understanding the electrical behavior. Another idea would be to try ideas using a circuit simulator but in the end it's always going to come down to your ears anyway.

[Mark Hammer]

I have a DOD FX20 phaser.  This is essentially very similar to the Ross LM13600-based phaser, except that it has 6 stages and provides a sum output and difference output.  For one of them, adding resonance produces a very audible and pleasing effect.  For the other, adding feedback does nothing.

Where yo take your feedback path from seems to matter almost as much as where you feed it back to.  Typically, feedback is taken from the last phase-shift stage and returned to a point where it passes again through an odd number of stages (1, 3, 5, etc).  But, you can also take the feedback signal from the output of the mixing stage and feed it back to the beginning.  In other words, the feedback is not "internal" to the allpass sections, but feedback of the notched output product.

[mac]

Before building my P90 I breadboarded it and experimented with feedback. I took the signal from the final stages and sent it back to the inverting and non-inverting inputs of stages 1, 2 and 3.
IIRC if you send the signal back to stage 1, 2 or 3 you get more trebly or bassy sound. The (-) and (+) where different? I do not remember, maybe. Be careful to lower the amp or fx volume if you use a very low resistor, <15k in a P90, because it may howl like crazy.

There is a sound sample of my Fet Phaser v1 and v2 with regeneration paths back to different stages:

http://www.diystompboxes.com/smfforum/index.php?topic=67167.0

mac

[nordine]

yeah, breadboard all the way

i remember i was really pleased when started experimenting with the Easyvibe on breadboard... first started changing the lfo (for the ross phaser lfo), then fiddling with lfo offset and feedback path, oh and matching the caps,

got mega huge phasing, real thick and wobbly, something i didn't hear neither from p90 nor small stone

[mac]

It is always suggested to match fets for a P90, but the four 473 caps and 10k resistors should be as close as possible too. Although the 5% tolerance of caps and resistors is fair enough, I used almost identical metal film resistors and poly caps. A little improvement.
Also the two 150k resistor mixing the dry and wet signal should be replaced with a trimpot because opamp losses can deviate the center point a little. That's why a used a pot in my own design.

mac