How to get optimal sweep on the Ross/Ropez?

[Mark Hammer]

I've built the Ross/Ropez phaser from the Tonepad layout about 4 or 5 times.  Sometimes the sweep is nice, and sometimes the sweep is awful, and I can't figure out what is responsible for it.

I was looking at the trace from the LFO on a scope tonight.  The output of pin 8 is supposed to be a square wave, if one goes by this diagram:

What I get is something that has a higher edge on the rising side, that looks sort of like this, rounded on one side and trapezoidal on the other, and I can't for the life of me figure out why, or what to change.  The asymmetrical square results in a hypertraingular sweep from pin 9 that is not particularly smooth.  It feels like a triangle comes up, disappears for a while, then comes back again, like cars zooming down the highway as you're trying to sleep beside it.  Not good.  So what would do that?

[Quackzed]

could it just be the caps filtering highs?

[Lurco]

Leaky cap? Try to replace with a smaller value in the nF range (which also makes it easier to watch on the scope), and poke around some more.

[Mark Hammer]

What is the role of R29 in the circuit shown?  Is it something to fine-tune?

[J. Luja]

from the drawing of your square wave output, it shouldn't really affect the shape of the triangle in a noticeable way.
the above schematic has a slight error in that pin 3 of the integrator should connect to the junction of r29 and c11, which seem to form a low pass filter to suppress "ticks". it should work roughly the same without it.  does the triangle output look ok? it sounds like you're describing the upper part of the sweep as going out of range and then coming back, like ota's are being overdriven by the lfo. maybe adjust the 10k feeding them?

[Mark Hammer]

Not quite.  The "hypertriangular" shape (triangle at the positive peak, sinusoidal at the lower end) is there, but it's a a long enough period of inactivity between successive upward sweeps that it sounds like pulses rather than a continuous up and down.