LFO modulating itself?

[birt]

i was wondering..

if you have an lfo that drives a led/ldr. and you use that LDR to control the depth or rate of that same lfo. what would happen.... melting universe? or something a bit more boring?

[jacobyjd]

[frequencycentral]

This is may favourite thread EVER. I'm chuckling my tits off - and that picture is awesome............

[dano12]

From a purely technical perspective, I'd imagine this:



From the off-hand perspective, I'd have to imagine that:

a) the analog drift of a typical LFO being driven by itself would lead to an interesting psuedo-random approach
b) You should use 5 watt carbon comp resistors in the LFO to ensure maximum randomness.

[frequencycentral]

Serious answer now:

In the commonest opamp LFO, a high resistance fom the speed pot means slow, a low resistance means fast. Lets say the LFO is driving an LED that gets bright on the upsweep and dim on the downsweep. The LDR is in parallel with the speed pot. As the LED gets brighter, the resistance of the LDR reduces and therefore the speed also increases. So the LFO gets through it's upsweep faster, and the downsweep lasts for longer. So what's really happening is that the duty cycle is being affected, rather than a 50/50 triangle you'd get a 20/80 or 30/70 or whatever, depending on how you set the 'depth' of the LED to affect the LDR. It might sound cool, it might be a hyper-triangular. Or it might be a straight line with the occasional rhythmic blip. Now I have to go try it............

[zyxwyvu]

For an LFO that is self-modulating its rate, we can get a good idea of the solution qualitatively.

Either the instantaneous rate will increase or decrease with increasing output voltage. The are pretty much just inverted versions of each other, so let's assume that as output voltage increases, rate increases (or, resistance decreases). This is the same direction an LED/LDR will work.



First, lets consider a simple square wave oscillator:



Assume that R1 is dependant on the output (the output of the opamp).

Since there are only two possible output states, the rate will have only 2 values. When the output is high, the rate will be high, and vice versa. The high and low times are now different. In other words, we've changed the pulse width. The depth of self-modulation determines the amount of width change, and the polarity of the modulation determines whether the high or low state is shortened (Higher output->Higher rate means that the high state will be shortened). Keep in mind this is not pulse width modulation - given a constant self-modulation depth, the pulse width is also constant.



Next, a triangle wave oscillator, such as this one:



If we use the square wave output (output of the left opamp) is used for modulation, we get a result similar to the above - a pulse width change. If we look at the triangle wave output, it will start looking more like a sawtooth:





If we instead use the triangle wave output (output of the right opamp), the output gets a bit more interesting. Now the rate is being continuously varied. As the output gets higher, the rate gets higher, so the high parts of the output get 'squished' in time, and the low parts get 'stretched'. The result is a kind of hypertriangular wave:



I didn't try out a self-modulating sine wave, but that would be quite a bit more difficult to build or simulate, and the result is (presumably) basically the same as the triangle.

[Processaurus]

Interesting, good point, it would just distort the waveform.  Could be an easy way to get a coveted hypertriangular LFO.

[cloudscapes]

is the LDR is slow-ish, you'll get pulse width adjustment