4051 Plexquencer

[luminiferous]

I think that it should be possible to mod the Plexquencer fairly easily so that you can feed it whatever voltages you have on hand. If you replace the 4081 with 2X7 SIL headers soldered to your prototyping board of choice, you could do something like this:

This is discrete AND logic circuitry adapted from a CGS module:

CGS pulse divider and Boolean logic - Synth DIY Wiki

There’s also an interesting clock divider on that page, BTW.

NB: I haven’t built this mod, or the Plexquencer itself. If you were to attempt this mod, I would recommend leaving the 14 pin socket on the PCB, and pushing the SIL headers into it, rather than soldering directly to the PCB, that way it’s easily reversible.

[analogoutput]

Interesting… the tack I’ve been thinking of taking is a replacement for the jacks board:

(Also easily reversible.)

Added: On another look at @luminiferous’s idea… wouldn’t you also want to lower the thresholds for In A–C?

[luminiferous]

Also interesting! We’ve each missed things that the other covered. I forgot about the address inputs on the 4051, and you forgot (or were unconcerned) about the voltage input. What I mean is that if the voltage input is less than 8.4V, there won’t be any gates. I realise that it’s normalled to 12V, but I think that it would be much more useful with a wider range of voltages. With what I suggested, you could do something like put an ordinary ±5V LFO into the voltage input, and get scaled copies of it at the CV outputs, and gates when it is positive (and the gate input is positive).

A simpler way to deal with the address inputs would be to cut the negative feedback traces and turn those op amps into comparators (as with your mod), but that’s not quite as easily reversible.

[analogoutput]

I briefly thought about the trace cutting thing too, but you need more than that — you need to put a threshold voltage on the inverting inputs. Preferably a positive voltage, large enough not to trigger on noise. One reversible but kludgey way to do it would be to bend up TL074 pins 2 (on U3), 6, 9, and 13 (on U1) (to break the feedbacks) and solder wires from a voltage divider to the bent-up pins (to set the threshold voltage).

I’m not sure if I forgot about the voltage input thing or just decided not to worry about it. There again you could bend pins 2 and 5 on the 4081 up, jam wires into those DIP socket holes leading to comparators, and solder wires from the comparator outputs to the bent up 4081 pins.

Or… replace both the TL074 and 4081 with daughterboards, along the lines you suggested, that implement comparators for all four inputs and for the two SWOUTs. That’d be easier if all four inputs used the same TL074 for their voltage followers but they don’t. Still, could be done.

[luminiferous]

@analogoutput

I briefly thought about the trace cutting thing too, but you need more than that — you need to put a threshold voltage on the inverting inputs. Preferably a positive voltage, large enough not to trigger on noise.

You’re right, of course! If I were doing this, I would make the daughterboard that I suggested earlier, add a voltage divider to it, cut the traces on the op amps, and run wires from the voltage divider to the inverting inputs of the op amps.

[klausklaerwerk]

nice! can you feed this divider any audio-range oscillator, or does it have to be a square wave osc?

[EddyBergman]

You can feed it any type of signal. The opamp comparator will change it into a pulse signal and feed it to the CD4024. So you can use it as a sub-oscillator with -2 and -4 octaves (the other outputs are too low to be useful I think). Naturally the output waves will be pulse waves only.

[klausklaerwerk]

cool, thanks! one more question: why do you drive the CD4024 with 8V and don’t just use the 12V rail? cheers

[EddyBergman]

That’s because I don’t want 12V output pulses. The voltage you use to feed the CD4024 will also determin the voltage of the pulses it outputs. The original schematic had a 5V regulator in it but I wanted a bit higher voltage pulses so I used an 8V regulator.

[analogoutput]

I’m considering now a daughterboard or a pair of daughterboards that plug in place of U1 and U4. U1 is replaced by another TL074 wired as comparators for InA-C (and still a voltage follower for the voltage input). U4 is replaced by another CD4021 with an added TL074 as comparators for its three inputs (and one unit unused). I’ll have to wait and see the PCB before figuring out much further.

[SynthesiS]

Wouldn’t it be easier to add transistors on the input to let it trigger with a lower voltage, and on the outputs to provide higher volrage outputs?
Or am I thinking too simple here?

[analogoutput]

The three problems here are: (1) Input gates lower than 8.4 V will not be seen as high signals by the CD4021 parts and you won’t get gates; (2) they also will not be seen as high signals by the CD4051 if the input voltage is the normal 12 V or is anything more than 1/0.7 times the gate amplitude and you won’t get bits into the shift register; and (3) the CD4051 outputs will not be seen as high signals by the CD4021 if the input voltage is less than 8.4 V and you won’t get gates.

Yes, you can use transistors, or op amp comparators, on the inputs to boost them to the level required by the CMOS parts. That was the intention of my jacks board replacement idea. That addresses problems (1) and (2) but not problem (3).

@luminiferous’s idea was to replace the CD4021 with discrete AND logic that has a lower threshold. That fixes (1) and (3) but not (2).

My more recent daughterboard idea is intended to address all three problems.

There’s no problem with the output gate voltages (unless you think they’re too high).

[analogoutput]

Normally I don’t put a git repo online until I’ve built and tested the project, but in this case it seemed to me a good idea to get it out as soon as possible. So here is the completely untested pair of daughterboards in their current version:

Richard Holmes / plexdaus · GitLab

Use at your own risk, but do feel free to study, criticize, use, improve, modify, adapt, whatever, and keep us posted on your results.

[analogoutput]

I got these knobs for this project:

https://www.taydaelectronics.com/control-knob-black-color-13x11mm-shaft-hole-6-4mm.html

[Not actual size]

Not a very close match to Boss style but close enough and only 13 mm diameter.

(I moved to Kosmo to get away from tiny knobs like these! But okay, for a compact sequencer I can see making the exception.)

I also discovered Tayda now has Davies 1900H clones in aluminum, in a few colors; I’m considering a red one for a non Kosmo project. Not only the color and finish but the diameter at the top is a little different from the plastic ones. (And the set screw is hex key driven.)

Added 1/17/2025

Not gonna try to add yet another post here, I’ll just tag this onto my last one. Got my PQ panel/PCBs today and did a revision of my mods based on that.

I noticed another issue. The 4051 has absolute maximum input voltage V_DD + 0.5 V, so with +12 V power the inputs must be below 12.5 V. Not really an problem; the inputs are coming from op amps powered by +12 V so they’ll be below that (about ~10.5 V max). But — if you plug in an input voltage, that becomes the power voltage for the 4051, and if it’s below 12 V then the maximum input voltage is correspondingly lower, but the op amp is still sending ~10.5 V to it. In that case you might be exceeding the maximum.

In practice this might not actually cause any trouble. But there’s no guarantee the chip won’t start misbehaving, or in extreme cases become damaged.

CMOS chips like these do have protection diodes from the input pins to V_DD. But:

1. They’re intended for protection against transient large voltages due to static discharge. I’m leery of relying on them against DC overvoltage. The maximum DC current they can handle is pretty small. I’d rather have more robust external diodes.
2. Internal or external, these diodes can’t do any good without a current limiting resistor on the input.

So I’ve added resistors and Schottky diodes to my design.

I think what I’m going to do is to build the module, but with pin sockets instead of DIP sockets for U1 and U4. Then I can breadboard the mods and verify they work before getting the boards fabbed. Assuming that goes well I’ll have some extras to sell (to US addresses only, though), DM me if interested.

Added 1/19/2025

I’ve built the Plexquencer. I did solder in pin sockets instead of DIP sockets for U1 and U4, in anticipation of installing the mod daughterboards. But I connected them with jumpers to a breadboard with a TL074 and a CD4081, so it was wired up as designed and… it works with 5 V gates. In fact, it works with 2 V gates. I get CV and gate outputs as expected. Moreover, I still get gate outputs with input voltage down to about 4.5 V.

I’m surprised.

It may be that the behavior of these chips is officially undefined between the low and high threshold, but that they unofficially do respond to voltages in that range.

It seems the above modifications aren’t needed after all.

I do see one flaky behavior, but it’s unrelated to the gate amplitudes. If one of the gate outputs is being turned on and off by only a single stage, and then you switch that stage’s gate switch off while that gate is on, the gate output stays on. The reason is that the AND inputs are floating when there is no stage turned on. The fix for that is to add 10k pulldowns to ground at 4081 pins 2 and 5 (or from the cathodes of one of the diodes connected to each of those pins). Not too hard a bodge right on the PCB, I would think.

[klausklaerwerk]

so i built this and it works, but the a/b/c inputs are really picky and don’t work with everything.

• the lmnc LFOs don’t work unless you boost the signal.
• direct clock from the midimuso or from my arturia minifreak doesn’t work
• lmnc ADSRs do work if you loop them and set decay quite short and sustain to zero.
• trig out from the 2001 sequencer works.
• audio directly from VCOs works but is obviously really fast.

and i noticed the 2 gate outputs work the wrong way around. if you flick a switch to the left (i.e. to gate output 1) it puts out a gate on output 2 instead and vice versa. is this a mistakte in the design or did i do something wrong?

and finally some of the features are still quite confusing, like what does the gate input do? and the V in? @lookmumnocomputer it would be cool to have a proper video explaining all this

[lookmumnocomputer]

hello! interesting it works from the LFO and midimuso for me. anything above about 3v it triggers. i will double check.

VCO is good if you use 3 vco’s then multiplex them and then output of the sequencer into the filter

[lookmumnocomputer]

very strange. i havent had issues with the inputs. will chcek today

[lookmumnocomputer]

@klausklaerwerk

confusing, like what does the gate input do? and the V in? @lookmumnocomputer it would be cool to have a proper video explaining all this

yes i have i t on the list and then it fell off. i have half a video on it. ill get to it in the next coupele of days asmatter of urgency. fell off the list. bloomin list. haha

[lookmumnocomputer]

very odd. i had this in the case at the museum for the past few months and never plugged anything in and the gates hadn’t triggered. looking tonight. the shed that my synths are in has been sub zero last few weeks and deeply unpleasent to be in. its warming up now so can do none glovey work ha. will check these and see what i find.

[klausklaerwerk]

ok that is strange. do you send it clock from the midimuso or gate from a midi channel?

edit: also any word on the gate ouputs being the wrong way around?