How does the step mode on the Boss PH-3 work?

[freddd]

I love this sound! Could you mod a phase 90 to do this? I just don't understand how it is stepping through 'stages', I thought if you had a 4 stage phase it was stuck at 4 stages!

[Mark Hammer]

The rise, fall, and step modes on the PH-3 do not alter the number of stages.  Rather, they are sweep modes with a fixed number of stages (not sure how many).  "Step" mode generally involves a sweep cycle that proceeds in "stairstep" ather than a continuous upward and downward sweep.  In the case of the PH-3, the "step" would appear to be a random one, similar to what is used in the Maestro FSH-1, or in the Line 6 Liqui-Flange (though this is done digitally).

This type of sweep has been used for decades in synthesizers, and can generally be implemented using counter chips and a clock, or in the casse of random steps a noise source and a clock.  It is not the sort of thing you could easily mod a Phase 90 to do, although that has more to do with the existing circuitry than with the fact that it is a 4-stager.

[SeanCostello]

Step mode uses a "stairstep-triangle" waveform, not random. Think of an overly quantized triangle wave - maybe 16 steps? I haven't counted.

I'm pretty sure that the rise and fall modes use 12 stages, and would bet serious money that they configure the allpass stages as a Hilbert network, and use complex multiplication to get the "barberpole" effect.

Sean Costello

[Mark Hammer]

I'm not even gonna touch the "Hilbert thing" , but barberpole effects demand more stages, period.  There always have to be "enough" stages at different points in the ramp sweep cycle for the entire spectrum to appear to be occupied with moving notches.  Here, the "barberpole" analogy comes in handy.  Just try to imagine what a rotating barberpole might look like if it had only two or 3 stripes visible at once from any angle.  Much much less of an "it just keeps moving upward" illusion.

Step mode uses a "stairstep-triangle" waveform, not random. Think of an overly quantized triangle wave - maybe 16 steps? I haven't counted.

I listened to the sample that Boss provides on the site, and it had some properties of randomness.  Listened to some other samples on modezero following your comments, and you're right, it's a stairstepped triangle.  You can adjust the stepping rate with the speed control, but the number of steps per sweep cycle would appear to be fixed.  This is different than the FSH-1 approach.

[SeanCostello]

The PH-3 rise/fall modes sound like using a 12th order Hilbert network, without any additional allpass stages outside of the network. 3 notches rising/falling, unless my ears deceive me.

The Bode patent specifies additional fixed allpass stages outside of the network, cascaded with identical coefficients, in order to increase the number of notches. I have tried this, and to my ears this seems to bunch up all of the perceived action in a fixed frequency area - i.e. more notches, but all midrange. Then again, I prefer 4-6 stage phasers over higher order anyway...until you get to 24 stages, where things get cool again.

Sean Costello

[freddd]

I was just thinking about whipping out the existing LFO in the Phase 90 and replacing it with an 8 step sequencer.

[SeanCostello]

This could accomplish similar results, especially if you could create an up/down mode for the sequencer. Smoothing the control signal would also be cool.

Sean Costello

[Processaurus]

I tried the Geofex walking ring counter LFO with my Phase 100 to get a stairstep (not S+H) phaser and loved it!  I wanted to figure out how to make less (and simultaneously bigger) steps easily, like with a rotary switch, but  didn't get it enough at the time.  Might be worth another peek...