I’ve managed to get it basically working except the range on the 4.7k cutoff pot seems really limited, like getting no more than half of the travel on the knob before the sound cuts out at the high end. Anyone know of what might be causing this?
Yes it’s linear. I followed the layout as closely as I could, the only major difference that I can tell is using a surface mount LM13700 with a through hole adapter.
The high end? Normally a low pass filter would cut off at the low end, because the cutoff frequency there is so low it blocks everything in the input signal. Unless the pot’s wired the other way around, of course.
If that’s what’s happening then it’ll cut off at a higher frequency when the input signal is higher frequency, so… are you testing this with a low note or a high note going in?
Does it self oscillate at high resonance? If so you can verify which direction the frequency is going, and measure (or estimate by ear) the frequency range. If indeed the range is too low, so the cutoff frequency goes below even bass notes halfway through the pot range, my first guess would be a problem with the CV exponential converter. That’s the 100k, 1.8k, 470R, and 4.7k resistors, transistors, 4.7 nF capacitor, and op amp pins 12–14.
So what’s happening is I start turning the cutoff from minimum and the higher frequencies start coming through as expected, then about half way up there’s a burst of distortion then just above that it cuts out completely. Yes it is self-oscillating. I’ve tried to check the components pretty thoroughly but I guess I’ll have to have another go and see if I can come up with anything.
The same behaviour actually happens with no input signal - just the self-oscillating from the resonance knob will cut out when the cutoff is turned up.
Which layout have you used, and what is your input resistor value?
Does it happen more with higher resonance? I had a similar issue which was resolved by replacing that 10k input resistor with an 18k resistor.
My understanding from reading around when I had that problem is that it’s a common issue with the layout caused by the signal ‘hitting the rails’ of the op-amp if the input signal is too hot. I used a potentiometer wired as a variable resistor to reduce the input signal just enough to still let it get distorted and nasty with the resonance up, but without it cutting out. Measured the resistance at about 18k so put that in and have been happy since.
Well like I said above it happens without any input signal. Just turning the resonance up and getting the self-oscillation gives the same result - turn up the cutoff and about half way up the sound cuts out with a little burst of distortion. I have tried adding a trimpot to the input and it had no effect on the cutoff issue.
I’m again wondering if it’s the exponential converter… There’s an upper limit to how much control current you can send to the LM13700 before it, literally, burns up; I don’t know what happens if you exceed that limit for a short time, but I can imagine it could be this. The control current should be below 2 mA and preferably more like 500 µA. That goes into the two stages of the LM13700 via 10k resistors, the ones connected to one of the transistors. Then the voltage drop across each resistor should be under 2 mA x 10k = 20 V, so there should be no problem. However if those resistors were in fact 1k or something then there would be trouble. So I’d say triple check those, make sure they’re really marked as 10k.
Ok I solved it - by replacing the LM13700. So everything else was fine in the end. I suspect I must have wired the power in wrong at some point and damaged it, I’ve lost track of the steps I’ve taken on this thing now haha. Still, really appreciate everyone’s input, thanks