The A-147 is a triangle/square-wave oscillator built around the pairing of an integrator and comparator. They work in a loop with the one feeding the other, back and forth. You can see them in action here.
The result of integrating a square wave is a triangle. But if we independently change the rate at which the integrator charges and discharges, we can skew the waveform from a rising to falling sawtooth via a triangle. Because the integrator and comparator are related, the square wave becomes a variable width pulse.
The easiest way to achieve this is to hijack the square-wave on its way to the integrator, split its path with two diodes and then set the balance with a potentiometer. You can see this manual arrangement in Ray Wilson’s Variable Skew LFO or Ken Stone’s Utility LFO.
To get proper voltage control over this ‘re-balancing’ might take a considerable amount of extra active circuitry, more than I felt the effect is worth, especially given the digital options available today. Here’s my simple kludge:
The idea is to massively reduce the size of the square wave going into the integrator so that the injection of an external voltage is enough to shift the bias of the square and hence alter the shape coming out of the integrator.
Resistor R14 (15k), which is to the left of the quad opamp (TL084) forms a voltage divider with the 1k resistor which also marked in yellow in the picture above. The output of this divider goes to the CA3080 OTA for integration. I found that swapping the 15k for a 1M resistor was enough to allow the bias to be shifted with a +/-5V CV. For the CV input I ended up using two resistors in series to give me 670k rather than the 1M in the picture.
Reducing the amplitude of the square wave to the integrator also reduces the overall frequency of the LFO, so I also swapped the 0.1uF capacitor C2 for a smaller 15nF to bump up the speed.
Below is how it looks and here are some construction pictures.
I like simple mods - minimum change for maximum effect. Skewing the wave shape with this method also changes the frequency of the LFO, which you might not want. There is a way to compensate for this, but again, I decided it wasn’t worth the extra circuitry as I like the effect of the LFO slowing down as it changes direction, almost like a thru-zero oscillator. But you could try the following schemes. One is relatively simple, the other might need an expander module to provide space for the extra PCB. Neither is perfect.
The idea here is to inject a fraction of the skew CV to the A-147’s FCV input. For example, given a skew CV of 5V I found I needed about 5-600mV FCV to correct the frequency. You can use the attenuator on the module for this or send your skew CV via a ca. 560k resistor to the output pin of the 50k FCV potentiometer (ahead of the FCV buffer). The ‘gotcha’ with this method is that it only works for positive offsets as we always need to speed up the LFO when skewing.
What is needed is an active full-wave rectifier to ensure that the compensation voltage is always positive. This is a simple enough circuit to add but for the lack of space behind the panel. If you want to try this, have a look at Jim Patchell’s Ideal Diode Tutorial hosted at René Schmitz’ Schmitzbits site.
The usual DIY disclaimers apply - you do this at your own risk. Have fun but take care of your module and yourself!
Thank you Tim Stinchcombe for identifying the relevant parts and for your thoughts and suggestions.
No comments:
Post a Comment