Unusual oscillator for Tremolo

[Fancy Lime]

Hi there,

since tremolos have come up again recently, I revived an old project that got stuck at some point: Envelop controlled tremolo. The point were I got stuck was, how to make the oscillator change the wave shape (and possibly amplitude) when the frequency changes. What I want is a square wave with high amplitude at high frequencies and a triangle wave with lower amplitude at low frequencies. The opposite is easy: Just run a square wave oscillator with variable frequency through an appropriately sized integrator. But I am unable to figure out if there is a way of designing an oscillator the way I want it to behave without lots of extra parts like a voltage controlled amplifier that inverts the natural amplitude behavior.

Any ideas?

Thanks,
Andy

[bool]

The only simplistic solution I can think of at this moment would be to incorporate a sort of frequency-dependant waveshaping: a non-inverting opamp with a "high-pass" in the nfb; into a clipper of some sort. At low freqs below the knee it would remain triangular (actually "shark-fin" shape) and gradually more square-ish at freq above the knee.

Could be done with inverting topology as well, but probably requires more tinkering.

[Rob Strand]

The simple solution is to use something like the MXR Phase 90 LFO  with a dual-ganged pot.  The second gang is at the Schmitt-trigger output for high frequencies and the cap end at low frequencies.   You can add dividers to set the levels at each end and perhaps a buffer on the cap side to make life easier.    You might need to tweak the DC levels at each end.

The topology is simple and flexible at the cost of a dual-ganged pot.

[duck_arse]

if you are using an integrator/schmitt type oppie osc, you have sq and tri already. if you have enveloped freq, you have envelope already. why not string an ldr type divider between sq and tri, al la shoot the mooon lfo, and sweep it with the freq env output?

[Fancy Lime]

The only simplistic solution I can think of at this moment would be to incorporate a sort of frequency-dependant waveshaping: a non-inverting opamp with a "high-pass" in the nfb; into a clipper of some sort. At low freqs below the knee it would remain triangular (actually "shark-fin" shape) and gradually more square-ish at freq above the knee.

Could be done with inverting topology as well, but probably requires more tinkering.

Hi bool,
I don't quite follow. Wouldn't a highpass in the nfb have the exact opposite effect: square at low, triangle at high freq?

Andy

[Fancy Lime]

The simple solution is to use something like the MXR Phase 90 LFO  with a dual-ganged pot.  The second gang is at the Schmitt-trigger output for high frequencies and the cap end at low frequencies.   You can add dividers to set the levels at each end and perhaps a buffer on the cap side to make life easier.    You might need to tweak the DC levels at each end.

The topology is simple and flexible at the cost of a dual-ganged pot.

Hi Rob,
sadly, dual gang pot won't do, since i want to control the frequency and the waveshape with an envelope follower.

Andy

[Fancy Lime]

if you are using an integrator/schmitt type oppie osc, you have sq and tri already. if you have enveloped freq, you have envelope already. why not string an ldr type divider between sq and tri, al la shoot the mooon lfo, and sweep it with the freq env output?

Hi Stephen,
yeah, that'll probably what I will end up doing in the end, or something similar following the same basic idea. I was just wondering if there is some kind of oscillator design out there that changes the waveform automatically in the way I want, when frequency changes, since a design that does the exact opposite does exist. I like it when things are neat, elegant and uncomplicated. Alas, that is rarely how things work out...

Andy