Switched capacitor phasers?

Started by ElectricDruid, June 12, 2024, 04:55:08 AM

Previous topic - Next topic

ElectricDruid

Quote from: R.G. on June 16, 2024, 02:51:23 PMI will need to ponder this a bit, but it might be that a DDS NCO in a PIC would drive the PWM inputs directly and have the pulse widths fall out naturally; triangle or nearly-true sine modulation for free.
It'd be nice, but I doubt we'll be so lucky. What are you suggesting, exactly?

If you use the PIC's NCO, the problem is jitter. I'm going to see how bad that is for a phaser when I try a prototype with the MULTIFLANGE chip that uses that technique.
You can avoid the NCO jitter if you use the PWM modules, but then you only get frequency division, so the frequencies get sparser and more steppy as you go up. With an NCO, that's the other way up - the jumps are largest for the smallest increments and you have good resolution for large increments. But for division, each frequency is *fixed* with no jitter, so no noise is introduced like that. There are benefits and costs to both methods, as usual...

More specifically for this project, the frequencies are going to need to be relatively high. For best filtering of the filter clock, we should probably aim for the 100:1 ratio of clock:cutoff. So for a phaser sweep from 500Hz to 5KHz, we'd be looking at 50KHz to 500KHz. That's roughly the range the Flangelicious NCO+LFO chips use, which is why I'd thought to use one of those for testing, but we'll see how that goes!



R.G.

It's not a fully baked idea, of course. Yeah, NCO will have jitter, but I was speculating that since the thing that matters with switched caps and resistors is the general duty cycle more than evenly timed pwm events, as long as the frequency is high enough, individual jitter might not be a problem. I haven't been through the math and background enough to be certain of that; it just seems right.
R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.

amptramp


Mark Hammer

Not to take anything away from Paul's work and diligence, but it is a switched-resistor design, similar to the MXR unit, not switched capacitor.  Many thanks to amptramp for posting the schematic and clearing that up.