Analog Barberpole Flanging...

[DrAlx]

Stephen Giles got me thinking in a recent post.  Has anyone actually put together a BBD based barberpole flanger?
In particular, like this...

Take waveforms from a quad LFO (and their inverse waveforms) and use them to drive 4 separate delay lines, with delays that go from 1 to 5 ms like this.  


Each of those delays lines gets combined with the clean signal, so there are four separate flanging sounds.
They need to be mixed together in a way that means you only hear the upward sweeping sounds (coming from the downward slopes of the LFO waves).
So the 4 flanging sounds need to be mixed together with variable gains that go something like this


Looking at just the first flanger's LFO and associated gain:



You can see the gain looks a bit like a deformed sine wave.  All I did to produce that gain profile was take one of the other LFO waveforms and exponentiate it.
A simpler option is to just take that other LFO waveform and half-wave rectify it, or amplify it in some way that causes it to clip on the downside.
It's then a case of using that gain profile to set the amplitude of the flanger sound using an OTA.

The part count is pretty big:  1 quad LFO,  4 BBDs and their associated VCOs,  4 OTAs for the variable gain + few more op-amps for mixing.
I was wondering if anyone has actually tried doing barberpole flanging that way, or if there is an easier non-digital way of doing it.

[Dave W]

i can safely say puretube has done this. he's the rosetta stone of flanging, all things bbd and a whole lot more! 

[Mark Hammer]

All those clocks, operating at different frequencies, in such a closed environment?  Not touching that one with a 10ft pole!

I also wonder if the barberpole aspect would command so much listener attention that audible differences between phasing and flanging (assuming we're talking about  more than 4-allpass stages) would simply be overlooked and underheard.

[DrAlx]

Not touching that one with a 10ft pole!

10ft (barber's) pole 

[StephenGiles]

The parts count would be pretty big - maybe 2 rack size PCBs stacked, or in my case vero(!!!!!), definitely big boy stuff. But then what would Robert Moog gave done back in the 1970s ??

Just a thought, consider one master clock running 4 times as high as necessary, which is divided down separately for each BBD. Is that a plan or absolute drivel?

[DrAlx]

Just a thought, consider one master clock running 4 times as high as necessary, which is divided down separately for each BBD. Is that a plan or absolute drivel?

Never tried dividing a clock down so don't know how it works or if you can get a smooth range of clock rates from it.  e.g. all the rates to give you a smooth set of delays going from 1ms to 5ms.

[ElectricDruid]

I was wondering if anyone has actually tried doing barberpole flanging that way, or if there is an easier non-digital way of doing it.

I've never tried it, and I don't know if it breaks your definition of "non digital" but I'd be looking for a PIC/AVR to generate all the control waveforms. Ideally, it could produce the modulated BBD clocks too, but I don't know of a chip with the right  mixture of peripherals for that.
This approach would have the advantage of getting the parts count down a bit, whilst keeping all the audio path analog.

This idea hit me at the right moment, since I've been working on PIC-based BBD LFO/Clock chips recently, and I'd been idly wondering about the possibility of doing a multiphase triple chorus - not a million miles from what you've got in mind, and with similar problems in terms of parts overload!

One final thought - a triple chorus uses LFO waves at 120 degree spacing. Perhaps with triangle wave modulation you could do the same. That saves a few more parts (one set of VCO, BBD and OTA).

T.

[DrAlx]

Yes, ElectricDruid I've been following your PIC thread with interest.
The barber pole thing will work with 3 phases.
However I can't see how you can pull off the same trick that I describe above with 4 phases   i.e. taking the LFO waveform for one delay line and then reusing it as the mix control for another delay line (after half-wave rectifying it with a diode).

[StephenGiles]

I say sod the parts overload!

[ElectricDruid]

The barber pole thing will work with 3 phases.
However I can't see how you can pull off the same trick that I describe above with 4 phases   i.e. taking the LFO waveform for one delay line and then reusing it as the mix control for another delay line (after half-wave rectifying it with a diode).

Ah, ok. I get it now. I was wondering what the motivation was for having four stages, but I agree that does make sense. If the clocks were digital with the LFO incorporated, then you'd (want to) have the PIC generating the necessary gain control signals and there starts to be a benefit from having 3 phases not 4, but otherwise I see it doesn't help much.

And yeah, to hell with parts overload. Some things are just big circuits.

T.

[puretube]

Beware of the VCA-warble...
(and of the latency, while you `re tryin`to get it below audible frequencies...)