[ ? ] Continuously variable phase LFO?

[moosapotamus]

Well, even tho most of this thread is over my head, just gotta say... "way cool!"

~ Charlie

[Cliff Schecht]

I probably missed the post, but did anybody mention all-pass filters? You don't get the 360 degrees of freedom that some may want, but you can get nice quadrature phasing without amplitude variations.

Or could you not combine two cascaded single-pole all-pass filters (using a dual-ganged pot, of course) with an inverter to get your 360 degrees?

[TELEFUNKON]

I probably missed the post, but did anybody mention all-pass filters? You don't get the 360 degrees of freedom that some may want, but you can get nice quadrature phasing without amplitude variations.

Or could you not combine two cascaded single-pole all-pass filters (using a dual-ganged pot, of course) with an inverter to get your 360 degrees?

see reply #41.

[puretube]

... from sketching the pulses on paper, it seems that my idea in reply#131 is another no-go... 

...However it lead to a new waveform: the "Crownwave", which now is an essential part of my brandnew
"MindButcher" Fundamental-Centrifuge...     

[Cliff Schecht]

I probably missed the post, but did anybody mention all-pass filters? You don't get the 360 degrees of freedom that some may want, but you can get nice quadrature phasing without amplitude variations.

Or could you not combine two cascaded single-pole all-pass filters (using a dual-ganged pot, of course) with an inverter to get your 360 degrees?

see reply #41.

I was mistaken here, a single all-pass filter and signal inverter will give you your 360 degrees of freedom. If the RC section is designed in a high-pass configuration with the cutoff low enough (i.e. a 2.2 uF cap and 1 meg resistor), you can get your 180 degrees phase shift at low low frequencies (LFO frequencies at least). With good matching of the inverter resistors, you can get a good amplitude response as well (H(w)=1 throughout range of interest).

[R.G.]

All pass sections are good for specific phase shifts at one frequency, or for a variable phase shift as frequency changes. H(w) is indeed 1 at all frequencies, but phase changes slowly from not-quite-ever 0 degrees to not-quite-ever 180 degrees as frequency changes. For an LFO, you could indeed get almost 0-360 by cascading two allpasses, but if you turn the frequency pot on the master LFO, the relative phase changes too, unless you make the allpasses also track the change in LFO frequency. Ugly. Tracking. Nasty.

You could use allpasses overlapped in a chain to give a main and a quadrature output over a reasonable range of frequencies. The so-called dome or Hilbert filter does this OK if you take the time to tweak it in. From quadrature, you can get any phase shift between main and slave by combining the proper amount of I and Q. But that's a bagfull of opamps, tweaking, accuracy, etc. eetch.

[Cliff Schecht]

All pass sections are good for specific phase shifts at one frequency, or for a variable phase shift as frequency changes. H(w) is indeed 1 at all frequencies, but phase changes slowly from not-quite-ever 0 degrees to not-quite-ever 180 degrees as frequency changes. For an LFO, you could indeed get almost 0-360 by cascading two allpasses, but if you turn the frequency pot on the master LFO, the relative phase changes too, unless you make the allpasses also track the change in LFO frequency. Ugly. Tracking. Nasty.

You could use allpasses overlapped in a chain to give a main and a quadrature output over a reasonable range of frequencies. The so-called dome or Hilbert filter does this OK if you take the time to tweak it in. From quadrature, you can get any phase shift between main and slave by combining the proper amount of I and Q. But that's a bagfull of opamps, tweaking, accuracy, etc. eetch.

I guess it all depends on the application. I am using a variable phase LFO for a tremolo and for the small range of frequencies that I'm trying to offset, the all-pass filter works well. If you're looking for phase-tracking accuracy and need the gain to be consistent throughout a decent range then yeah, an all-pass filter can be a nightmare. For something like a trem, however, where small amplitude variations really aren't noticeable (I'm sure you know that the human ear is much more sensitive to relative pitch) and relative phase shift isn't ultra-important (it just needs to be offset-able for the stereo effect), they work well.

[R.G.]

Different horses for different courses. 

[puretube]

Creating a square-clock, shifting it from 0° to 180°, double the frequency (or quadruple it and flipflop it down to symmetrical "double"),
and use this as a generator-clock for an MF-10 should do the 360°-trick, shouldn`t it ? 

[R.G.]

?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ??
?? ?? ?? ?? ?? Are we still chasing this one?  ?? ?? ?? ?? ?? ??
? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ??

[johngreene]

I haven't been on here very much over the past few years and I see I've missed some discussions that I would have found interesting! I developed something similar to a few things here using an Attiny MCU. It has the following features:
1. 3 LFO outputs, 120 degrees apart.
2. Phase adjustment on one of the LFOs to vary it 0 - 360 degrees to LFO 1.
3. Offset adjustment.
4. Amplitude adjustment.
5. Frequency adjustment (.07Hz - 20Hz).
6. Symmetry adjustment
7. Ramp function (rate and width) for ramping between two speeds.

Switchable functions:
1. Tap Tempo input
2. Wave Selection (triangle, square, sine)
3. Range select (allows 3 'presets' for different 'modes', currently not used).

And it is this big:

[puretube]

While working on an improved stereo nano-wiggle-wobble solidstate pedal
loosely based on the"Modzky" (as mentioned in reply#87),
using a derivate of a low frequency voltage-controlled quadrature oscillator (VCQO)
like the one mentioned in reply#70, or the one mentioned in reply#68,

a nice noticeable observation on the scope-screen was made when trying to mod
the modulation a bit more lively...

An extra pot (P2) + diode (D2) and an inserted diode (D1) offers 2 great new opportunities:

1.) vary the waveform continuously from pure sine-wave via lopsided-sine ("seasick") to saw-tooth
at one output - and - (simultaneously ! )
2.) vary the waveform continuously from pure sine-wave via "soft-hyperbolic" to "fullwave-rectified-sine-like"
at the other output...

The (quadrature-output-) voltage-controlled voltage-controlled quadrature LFO (VCVCQO)
is made of above mentioned few components added to one of the Quad-LFOs/VCOs from further above like this:




Swapping P2 from one output to the other, swaps the ouputted (?) waveforms from one to the other
on either outputs...


(the theoretical construction of both waveforms can easily be computed by those
practised in the art of trigonometry and/or modulation-technique,
or by studying the works of Fourier, Bode, Sunstein, and numerous others -
a  practical simple implementation like above might inspire other DIY "breadboard-jockeyz"
to new beginnings...)