Hypertriangle LFO questions.

[ThePastRecedes]

I've LFO here that outputs 3 wave shapes, square, triangle, and hypertriangle from ~ 0.2Hz to ~ 20Hz.
The ratio of R1 & R2 set the amplitude, it's currently set fairly high. I'm sure that they'll be adjusted when it comes to applying the circuit.

Currently I have two future applications for this LFO:
1) 6 stage phaser
2) PT2399 chorus with two delay lines sweeping 180 degrees out of phase.

Once they're complete full schematics will be going up here at diystompboxes.com since I've learnt so much from you guys.

(Click for larger Images)



at around 0.2Hz, triangle and hyper triangle outputs:



at around 20Hz, triangle and hyper triangle outputs:



The issue I'm having with ltspice, maybe it's my circuit?



This is my first ltspice simulation so bare with my stupid questions  When the LFO is set at its lowest frequency and the simulation starts from t=0 there seems to be a 50s period before it starts to oscillate. Any ways to stop that lag? I've yet to breadboard this, will this exist in reality?

I'll be adapting the circuit values for when I use it for my phaser & chorus projects but is there anything here that is poorly designed that I should reconsider?

Also anything useful I can do with that left over opamp?

The the amplitude of the square wave amplitude is always twice the triangle and hypertrianlge, what's the best way to reduce it?

Thanks in advance!

[armdnrdy]

Considering the source...birthofasynth I doubt if the time lag is there in the physical build.

The circuit is so small...I would just bread board it to see what you get.

On the drawing you wrote "yet to calculate best pot taper" for the rate control. Reverse audio or "C" pots generally work well for oscillator controls.

[Digital Larry]

The amplitude of the square wave amplitude is always twice the triangle and hypertrianlge, what's the best way to reduce it?

Also anything useful I can do with that left over opamp?

I would run the IC1_A pin 1 to a dual 10k (or whatever. within reason) voltage divider.  Voila.  If you need low impedance output, oh look, you have an extra op amp stage!

[anotherjim]

The start up time?
Operating, the Integrator is swinging +/-4v ; before it can do that, the timing cap has no charge on it, so it has to pre-charge first. I'm guessing that's all it is, because at the lowest rate, and with the negative feedback the cap places on the op-amp, the initial charging is very sloooow.

[duck_arse]

this is something similar:
http://www.diystompboxes.com/smfforum/index.php?topic=98974.msg886084#msg886084

much messier and some bad stuff I can see now .....

[jonnyeye]

In LTSpice, it's often helpful when working with oscillators to check the "Start external DC supply voltages at 0V" box in the transient simulation menu.  It simulates power being applied to the circuit which often gives the kick needed to start oscillators, well, oscillating.

[mth5044]

In LTSpice, it's often helpful when working with oscillators to check the "Start external DC supply voltages at 0V" box in the transient simulation menu.  It simulates power being applied to the circuit which often gives the kick needed to start oscillators, well, oscillating.

Excellent tip, thank you.

[Mark Hammer]

There is certainly value in being able to select the waveform one wants.  However, where the character of square and triangle waves seems to remain constant, regardless of speed, the utility of more complex waves, like the hypertriangular, declines as speed increases.  At very slow speeds, the sinusoidal sweep at the low end accentuates the effects, while the triangular part at the top end of the sweep gets in and out quickly.  Once the sweep rate goes above 1hz (and maybe even a little slower than that), a hypertriangular wave starts to sound more pulse-like, almost like a 30$ duty-cycle, and can be a little "unmusical".

EHX had an interesting way of addressing this.  In the earliest-issue Small Stone - the 6 x 3094 model - the LFO fed the Iabc pins of the 3094 chips via a 10k/50uf output network on the LFO's output.  In addition to providing current-limiting, so the 3094 chips don't fry, this forms a lowpass filter, rolling off variations in the "AC" from the LFO at 0.3hz.  As the speed gets turned up, the filtering action of that 50uf cap becomes more pronounced and the "spike" of the top part of the sweep gets rounded off, such that the sweep becomes more symmetrical in the upward and downward directions...but only for fast sweeps.  Clever.

I've implemented this on the "Ropez"/Ross, that essentially mimics the 6-OTA Small Stone, but with LM13600s.  Works like a charm.

[ThePastRecedes]

There is certainly value in being able to select the waveform one wants.  However, where the character of square and triangle waves seems to remain constant, regardless of speed, the utility of more complex waves, like the hypertriangular, declines as speed increases.  At very slow speeds, the sinusoidal sweep at the low end accentuates the effects, while the triangular part at the top end of the sweep gets in and out quickly.  Once the sweep rate goes above 1hz (and maybe even a little slower than that), a hypertriangular wave starts to sound more pulse-like, almost like a 30$ duty-cycle, and can be a little "unmusical".

EHX had an interesting way of addressing this.  In the earliest-issue Small Stone - the 6 x 3094 model - the LFO fed the Iabc pins of the 3094 chips via a 10k/50uf output network on the LFO's output.  In addition to providing current-limiting, so the 3094 chips don't fry, this forms a lowpass filter, rolling off variations in the "AC" from the LFO at 0.3hz.  As the speed gets turned up, the filtering action of that 50uf cap becomes more pronounced and the "spike" of the top part of the sweep gets rounded off, such that the sweep becomes more symmetrical in the upward and downward directions...but only for fast sweeps.  Clever.

I've implemented this on the "Ropez"/Ross, that essentially mimics the 6-OTA Small Stone, but with LM13600s.  Works like a charm.

Thanks for the reply! I didn't realize that this would be a problem at higher speeds. I googled "Ropez Mark Hammer site:diystompboxes.com" and I found a few of your posts about this circuit which was a good read so far.

Another questions which is kinda related is the sweep function on the moog mf103 phaser. It sounds like it controls which frequencies are accented in the sweep in the demos I've seen. With a guess I assume it moves the position of the "x-axis" of the LFO. That is just an assumption since I'm not too sure what that portion of the circuit does. Could you explain what is exactly going on there? Thanks in advance!

http://i.imgur.com/Ct4aMQR.jpg

[Mark Hammer]

Maybe somebody can, but I'm afraid that one is beyond me.