Hello, I’m new in the DIY scene and I’m having bit of a problem building VCO from Moritz Klein’s video Adding voltage controlled PWM to my VCO - YouTube (13:50 for schematic) which is based around 4069 VCO by René Schmitz from what I understand. My build is a bit different than that presented in the video because I also added triangle and sine output based on yusynth VCO and wanted to add a befaco-like/1222 VCO rotary switch for selecting octaves. The problem that I’m having is with the CV summing network and I would really like to hear some advice from someone more knowledgeable.
Trying to describe it in more detail wouldn’t be easy because it’s hard for me to comprehend it and describe it in English which is not my native language. Long story short I need to add voltage follower between rotary switch and CV summing network in order to avoid loading effect (or whatever it’s called) and with different voltage follower’s and summing methods I’m getting different averages and I can’t wrap my head around choosing which is right for my application. I made 4 “attempts” in falstad circuit simulator and you can check them either at the end of this post in URL format or as a file (Don’t know how to upload a text file yet).
1- is what I had first in mind, simple voltage follower going straight to summing network, the problem occurs when I try to add fine-tune pot, then for example if a rotary switch is at -2V I can swing between -1.8V to -2.8V. Technically fine-tune works correctly but I don’t know if it will work as a whole.
2 - This time instead of fighting with summing network resistors values I decided to make reference voltage (1v) for fine-tune pot and just sum them on the same resistor values, now I get 500mV response from fine-tuning and 0-6volts on a rotary switch which isn’t really what I’m looking for (probably) (got this idea from Sam’s 1222 VCO)
3 - This is probably the best I could think of but there is a major drawback, I don’t have space for another op-amp on the protoboard (I need to fit 6 of those VCO’s on one board and it’s a bit of a nightmare). Other than that, it’s perfect, 0 to -11V on the rotary switch and nearly exactly 1V of range on fine-tuning.
4 - this is just a test for resistors value for plan B in which I give up rotary switches for normal pots.
I had only seen 2(3 but I can’t remember the site with the schematic) VCO schematics that are using a rotary switch and I can’t choose which is right for my application. My only two options right now are to wait for some good samaritan who can help me or to solder-test-desolder-repeat that part of VCO and hope for the best, but I can’t even know if any one of my tests will be reliable because I would need to retune the whole thing for every one of them, test range of it and somehow measure mV on transistors for the full picture, even thou my multimeter isn’t that accurate.
And on top of that my other concern would be the probability that attenuverter for FM amount would “weight” down the whole CV summing network.
I will try to redraw the whole schematic with all the additions to eagle today and upload it in the comment (If it’s possible that is)
TLDR: Someone please help me to choose the right option to deal with a rotary switch which will select octaves in Moritz Klein’s VCO design.
An octave switch is really nothing but a discretized coarse tuning pot. Instead of variably selecting a division between the resistance above and below the wiper you select the division at specific points.
So the coarse tuning pot could be replaced by (or it could be added in parallel) for instance ten 10k resistors, or five 20k (or five 10k, it wouldn’t matter), or whatever, connected between terminals of a rotary switch, with ground and a voltage reference connected at the two ends:
For correct tuning the resistors should be matched at the 0.1% level or better. -Vref should be a precision voltage reference equal to 1 V per resistor, so in this case -5 V. You could use for instance an LM4040A50I (configured for a negative voltage). Since there’s a trimmer you can adjust the V/oct scaling, but the voltage reference gives you a stable voltage.
Unfortunately, I don’t have any precision voltage regulator, just matched 10k resistors but the problem wasn’t with precision but rather minimizing the loading effect and method to do that. But I see that maybe the loading effect could be removed with precision reference voltage, but I don’t know since in the circuit simulator that I’m using every voltage source is precise and doesn’t change while loaded (I think at least) and thus I don’t think it can be eliminated with just regulator.
It’s really the same thing as the coarse tuning pot. If the octave switch doesn’t work, then to some degree neither will the coarse tuning. I’m not sure what you’re referring to with “loading effect” but if you mean one CV affecting the sensitivity to another then really the circuit’s CV summing node needs redesign. Typically one sees an op amp in an inverting configuration, then the summing node is a virtual ground and the CVs will not affect one another.
edit if you look at the resistor ladder on the left you will see that all of them are not really perfect edit2 those inperfections are more noticable in real life unfortunately. edit3 worst part is that when I change fine tune knob, the voltages on the ladder are changing as well
this is the best solution I came up with but i can’t fit it on the board, switch is set to output 2V then its buffered and then inverted, fine tune is set at max with -1V to 0V range
Yeah, it’s just a passive summing node I guess, so for instance if you change the CV it’ll change the octave switch sensitivity to be more or less than 1 octave per step. In your second solution you’ve buffered the octave switch from the -12 V branch so the steps are 1 V at the switch — but my expectation is you’d still get non octave frequency steps when you change the octave switch, because the op amp still feeds a passive summing node. Whereas if the various CVs go through resistors to the inverting input of an op amp then you have an active summing node which ought to do better. In other words an active mixer for all the frequency controlling voltages, rather than just a buffer for the octave switch.
I think if you have no room for an op amp then you’re kind of stuck. This passive summing isn’t going to be very precise.
edit summing is pretty good, maybe max 7mV +/- at worts in terms of inaccuracy. edit only way is to sacrifice both attenuverters which is not out of a question and is better than not having reliable VCO edit3 Or I could desolder like 50 small components, some connectors, and a couple of IC, move everything around and pray that it fits which will take like a whole day to do it right and not destroy the board… so yeah I hope it will be worth it
I’m not sure why you have to put everything on one pcb, but couldn’t you mount a small extra pcb/stripboard on top of the other containing the extra circuitry (you don’t seem to have room for on the current pcb)?
I’m building bank of 6 of those on one bigger protoboard and I don’t have enough smaller size boards for adding those needed op amps, but I have figured about a layout that might fit everything
