Switched capacitor phasers?

[ElectricDruid]

Hi all,

Has anyone experimented with using any of the many switched-capacitor filter chips for making a phaser?

The chips seem fairly expensive, but some of them offer 8 stages in a single chip, so they might be worth it. Another advantage is that using a VCO to sweep the phase shift stages is nice and simple.

The idea suddenly occurred to me today and I thought that someone around here is bound to have tried it at some point...

Thanks!

[Mark Hammer]

Mike Irwin spotted me a few MF6CN chips some time back, but I gather they are internally configured to be 6-pole lowpass, and not reconfigurable to be allpass.  Specs for the chip, however, indicate that the clocking rate, used to switch the caps, is generally high enough to not be  that audible on their own.  One wonders how that HF clock would interact with all the other clocks in other FX.  Strikes me serious efforts to protect against heterodyning would be needed.  I might be jumping the gun, though.

That makes me curious about the PWM designs that MXR went with for a little while.  I have no idea what the clock rate was on things like the Envelope Filter, the Analog DElay, or the PWM phaser.

The availability and low price of SSM2164 quad OTAs begs for a compact 4-stage design, or a pair for 8 stages.

[R.G.]

I dimly remember a switched-capacitor something, and I think it was an MXR product.  Not a world burner exactly, or I'd remember it better. I think it was CMOS switches and oscillators.  IIRC you can get switched resistance and switched capacitor both by different arrangements .

[duck_arse]

I've got an LMF100 here, pulled from a crappy Silicon Chip project, waiting to do something good. waiting, waiting, waiting,

[Fancy Lime]

I dimly remember a switched-capacitor something, and I think it was an MXR product.  Not a world burner exactly, or I'd remember it better. I think it was CMOS switches and oscillators.  IIRC you can get switched resistance and switched capacitor both by different arrangements .

Do you mean the MXR envelope filter? That uses a 4069 for an oscillator and a 4066 quad switch to modulate the effective resistance. But it is based on PWM, not switched caps. Frederick Lyxzén, who goes by Freppo in these parts, has a phaser based on PWM on his Parasit Studio site:
https://www.parasitstudio.se/building-blog/parasite-phaser

But yeah, PWM, not switched caps. I had pondered the idea of switched caps in a phaser in the past but ultimately decided that PWM was easier and less trouble with the clock, long before I even took it to the breadboard. And then I decided that a Phase 45 is plenty good enough for me, si the whole idea went the way of the buffalo. So I would certainly be interested in experiences with a switched cap phaser.

[ElectricDruid]

Yeah, I'm really thinking of things that use the dedicated switched-capacitor chips, not stuff using PWM.

There have been several examples of PWM phasers using 4066 or 4016 switches, mostly to control the resistance. I don't think I've seen one that uses them to control the caps, but it would be equally possible.

The major difference is that the the PWM phasers use the pulse *width* to control the notches, whereas the switched cap filter chips use the clock *frequency* to control the cutoff. Exactly how that works, I don't know, I haven't looked into it, but it's a significant difference. One has a fixed clock with a swept pulse width, the other has a variable clock.

Mark's right too - heterodyning is definitely to be avoided in anything like this. But I don't think that's a absolute reason not to try it.

[Fancy Lime]

Yeah, heterodyning is going to be so much more difficult to avoid with a sweeping clock frequency compared to PWM on a static clock. That's what I meant by "trouble" and I suspect this to be the main reason why switched caps are so rare (I've never seen them used) in diy audio. As a resistor on a chip, sure, as long as the clock is also part of the same chip. But discreet or semi-discreet? Difficult but certainly worth a try.

[ElectricDruid]

The thing is many switched-cap filters use a clock thats x100 of the cutoff. So say we sweep our notches between 500Hz and 5KHz, we'd need a clock that sweeps from 50KHz to 500KHz. Even at the low end, that's far from impossible to filter. In fact, it's easier than a lot of old BBD delay pedals, and don't forget they had variable clock frequencies too - just under control of a delay knob!

[Mark Hammer]

I don't think eliminating heterodyning is impossible in this context.  You just have to remember to DO it.

[Rob Strand]

IIRC, the MXR unit was a switch conductance type.   PWM'ing the filter resistor makes it look like a larger value.    (The MXR envelope filter uses the same idea.)

The switch capacitor filters tend to lead into the problems in this thread,
https://www.diystompboxes.com/smfforum/index.php?topic=125584.0

Near the end of the thread I summarized the Penfold phaser design.
https://www.diystompboxes.com/smfforum/index.php?topic=125584.msg1196783#msg1196783

[StephenGiles]

Wasn't there a switched capacitor filter/phaser project in Practical Electronics way back? - I search!

[PRR]

https://www.worldradiohistory.com/Practical_Electronics.htm

[Rob Strand]

The Penfold phaser *is* the Practical Electronics phaser using the switch capacitor filter.

The direct link is in the thread I referred to earlier,
https://www.diystompboxes.com/smfforum/index.php?topic=125584.msg1196606#msg1196606

[StephenGiles]

The Penfold phaser *is* the Practical Electronics phaser using the switch capacitor filter.

The direct link is in the thread I referred to earlier,
https://www.diystompboxes.com/smfforum/index.php?topic=125584.msg1196606#msg1196606

So it is!

[ElectricDruid]

I've ordered a couple of LTC1059 chips to play with. At €11 a chip they're not cheap though. These are a "slightly faster" (says the datasheet) version of the LTC1060. The 1060 has a 2-pole allpass filter as one of the examples in the datasheet. Well, it's Allpass, Bandpass, and Lowpass, so that covers some options!



Since each chip is a dual filter, that's 4-pole allpass on a chip. The clock can be set to 50:1 or 100:1 for the cutoff control. For starters, I shall use the MULTIFLANGE chip that I've already done for the Flangelicous, since it generates output frequencies from 25KHz to 500KHz and has lots of LFO waveforms. That'll do for testing.

We'll see where I get to with it. I really didn't need to start another project when I have ten others unfinished on my desk already...

[cordobes]

Ray experimented with various effects using 5 pole low pass filter type integrateds. Taking the envelope follower concept (here) that he developed would be a good starting point.

[ElectricDruid]

Ray experimented with various effects using 5 pole low pass filter type integrateds. Taking the envelope follower concept (here) that he developed would be a good starting point.

Yeah, that's the kind of idea that I'm coming from, but I want to do it with allpass filters for a phaser instead. I've already done an autowah thing so I'm not so interested in that, but if I could get Lowpass and Bandpass modes out of this as well as a Allpass/Phaser mode, that'd be cool. It's possible - it's just a question of how *complicated* that is, and whether that level of complication is worth it for the result.
The other question I suppose is "envelope follower driven or LFO driven?"- Phasers are mostly LFO driven. Autowahs are traditionally envelope driven. Both things can be either. Doing a single pedal that does all those options *well* is difficult. Most do one thing better than the other, if they even do more than one option at all.
This is just early days and a bit of an experiment at this point. I've got no particular end point in mind.

[amptramp]

There was a Paul Nelson phaser from a while back that used CD4066 switches to simulate the variable resistance to ground for each stage of a phaser.  I used to have the schematic but it seems to have fallen into the bit bucket.  I should go and find a link to it as I may have had it on a different laptop.

[PRR]

Paul Nelson phaser from a while back that used CD4066 switches

https://www.diystompboxes.com/smfforum/index.php?msg=993202

Nelson visits downthread. His website is expired.

[R.G.]

Wayback Machine.  The link below probably doesn't work directly but searching the Wayback Machine for parasitstudio.se will get you there eventually.

https://www.parasitstudio.se/uploads/2/4/4/9/2449159/parasite-phaser-final.pdf

I 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.