typical phaser lfo waveforms?

[daverdave]

Hey all, I was just wondering what the typical lfo waveform was used in most phasers. Is it triangle or sine?
I've been looking at building one up from scratch, I was going to have an lfo which was switchable between square and triangle, or square and sine. I know that a few I've seen use a triangle, like craig andertons EPFM phaser and the vermona 16 stage rack phaser. I wanted to know what the more common phasers use, like mxr's and small stones, cheers.
Dave

[frequencycentral]

MXR's are triangle. The Ross is hyper-triangle, maybe the Small Stone is too? ROG's new Tri-Vibe has a very interesting LFO which does sine, triangle and hyper-triangle.

[daverdave]

ah cheers mate, never heard of hyper triangle though, what's that?

[frequencycentral]

Ross LFO with graphs: http://userdisk.webry.biglobe.ne.jp/000/024/65/1/200510_img_4.gif

The ROG Tri-Vibe, with some very useful LFO information and graphs: http://runoffgroove.com/tri-vibe.html

[Mark Hammer]

Both the Ross and the Small Stone use "hypertriangular" LFO waveforms for slow sweeps, and something different for faster sweeps.  The "Color" switch on the Small Stone simultaneously selects between two LFO waveforms/outputs and two resonance settings.  The Ross opts for one sweep setting, but offers variable Resonance.  Below is the schematic of the first issue SS, which has the identical LFO "engine" that the Ross does, only with CA3094 chips and a few component value differences.

Both the Small Stone and the Ross have a 270k/100k divider going between V+ and V- (ground), providing a reference/bias voltage to pin 3 (with a .05uf smoothing cap).  [Osamu's drawing shows pin 14, but the Tonepad drawing will show pin 3; same function, just the other side of the dual chip.]  The SS has 280k of series resistance (100k+180k) from pin 5 to V+, and the Ross has 270k from pin 1 of its chip.  When you flick the Color switch on the SS, 100k of the 280k series resistance gets bumped over so that it is now in series with the 270k from pin 3.  In other words, the 270k/100k divider feeding pin 3 is now 370k/100k, while the 280k resistance between pin 5 and V+ is now 180k.  This changes both the overall range of sweep speeds, as well as the shape of the waveform. 

I've installed the functional equivalent of what you see on the Small Stone into a Ross phaser, and it works fine.  In principle, one can play with both the 270k/100k bias voltage and the 270k resistance on pin 1 (Ross) independently, so as to alter the speed range and waveform independently.  I haven't done it yet, but look forward to doing so.  I recall seeing some experiments that Osamu did, varying the 270k and achieving waveforms that were different in their degree of hypertriangularity.

Generally speaking, hypertriangular waveforms are optimal for slower resonant sweeps of both phasers and flangers, triangular a bit better for medium speed, and sine-wave optimal for vibrato.  One of the reasons why the old Boss CE-1 Chorus ensemble is so cherished is because it changes between triangle wave for chorus and sine-wave for vibrato setting.

[daverdave]

Cheers for the replies, sounds interesting. I was using an opamp astable for the squarewave and then an intergrator to obtain a triangle wave from it, would it be possible to get a hyper triangle from this setup? I have both opamps biased at around 4.5v and the output of the squarewave through a voltage divider to get 2v pk-pk.

[Mark Hammer]

Check out this old article.  http://hammer.ampage.org/files/hypertriangleclock.gif

[daverdave]

that's great, I'll have a try with that when I get back to uni.

[daverdave]

Just out of interest, what should the output waveforms be biased at? Since they're roughly 2V pk-pk are they best being near 1V so that they swing bewteen 0 and 2V? I was going to use optocoupler to drive the stages.

[MoltenVoltage]

Whatever you do, consider using rail to rail op amps to get the full sweep range.  You will be surprised how much better it sounds.

We have a schematic posted for doing that with our MV-53 tap LFO chip on the Phase 90, but you can modify any LFO in a similar manner.

http://www.moltenvoltage.com/mv-53.html

[PRR]

> what should the output waveforms be biased at?

They come out of the LFO at whatever the LFO wants to give you.

They go into the controller "centered"(*) around whatever you consider "mid-scale".

If you use a 400V 12AX7 VFO (the lovely Fender tremolo) and a DBX VCA (20mV control), you will have to pad and shift a lot. Controlling JFETs, you may need a few volts but "centered" a little below the bias, which may be zero or half-supply depending on the audio design.

(*)Since pitch and amplitude are log scale, and many controllers are linear, the notion of "centered" "mid-scale" may be considerably skewed.

Much of the real "art" is finding VFOs and controllers which fit together sweet. The McConkey plan makes a linear triangle, but then semi-LOGs it, so that a linear controller (BBD?) gives LOG frequency-- smooth "pitch" instead of linear Hz.

[Processaurus]

Hey Karl, does your tap tempo chip's pulse width knob make the sine wave into a hypertriangular sine wave, or make it lopsided, approaching a sawtooth?  If the former is the case, I think it will work out well for a phaser project I'm getting around to.

[MoltenVoltage]

Hey Karl, does your tap tempo chip's pulse width knob make the sine wave into a hypertriangular sine wave, or make it lopsided, approaching a sawtooth?  If the former is the case, I think it will work out well for a phaser project I'm getting around to.

By "pulse width knob", you are talking about the Duty Cycle Control.

It doesn't make a "hyper triangular" waveform.

Instead, it adjusts the relative playback speed of the first and second halves of the selected waveform.  When you look at a duty cycle adjusted sine wave on the scope, it is skewed sort of like a sawtooth wave, but it is much more rounded.  The eleven (11) Duty Cycle ratios correspond to the most common musical subdivisions, and let you make the LFO swing. The relative length of the first half is indicated by the numerator, the overall length by the denominator: 1/8, 1/6, 1/4, 1/3, 3/8, 1/2, 5/8, 2/3, 3/4, 5/6, 7/8

You can always get a pure triangle wave by programming a sawtooth up > sawtooth down, 2 step sequence.

We can program a custom chip with an order of 100 units.  Mathematically, the "hyper triangular" waveform looks like its just the repeated bottom half of a sine wave.

[Lurco]

rectify a sine?