Fall, Rise and Staircase up/down waveform generator?

[strungout]

Tom: Yeah, you're right, the symmetry control won't be useful. It's supposed to be a trim to make sure the wave outputs have 50% duty cycle. Thought it might vary the step (as opposed to the rise) length and have long and short steps...

Now that I have a working square wave generator, I'll try the schem you posted. See how that acts.


EDIT: The circuit you posted works, but I'm getting... not a staircase, clearly. It's not random, but kinda looks like it.

[ElectricDruid]

EDIT: The circuit you posted works, but I'm getting... not a staircase, clearly. It's not random, but kinda looks like it.

No, the point of that one is that you can set the height of each step individually. That lets you do stairs up, stairs down, stairs up/down, whatever. Given a random set of pot values, what you'll get will sound like a repeating sample-and-hold waveform.

Tweak the ten pot values a bit and you'll get some staircases. What would be nice would be to have slider pots for a thing like this, so you had a clear visual representation of what it's doing. That's what some of the Zvex stuff that uses a sequencer does, and it makes a lot of sense.

[anotherjim]

FWIW, 4017 clock is the positive edge and the duty cycle is of no consequence. There is a Schmitt trigger in the input which is helpful. It can be clocked negative edge if you tie the clock pin high and drive the CE pin instead but that bypasses the Schmitt trigger. This trick is useful when cascading chips for longer sequences as you want the next stage to clock when the preceding last Q output falls from high to low - so a negative edge clock is necessary.


If you built the above don't make that common gate output at the right - as drawn it shorts the outputs although I think the thick vertical line is meant to represent a diode OR gate. There's always a note-on with this sequencer and it would make more sense to use the clock as a note-on trigger.

This one...

Has an error. Don't tie Q9 to 0v. The chip won't like it. If the selector switch fails open circuit it will run on to Q9 which would be shorted to 0v. Unused outputs don't need connecting to anything. The final step should be Q7, no need for Q8 in the sequence except as the last step reset and you need to use a contact of the selector switch for Q8 to get a proper 8 step sequence. The selector switch should be a break-before-make and a pull-down resistor on the reset pin (100k is fine) to stop it from floating when the switch is moving between contacts.

The fixed resistors in the ladder could be 10k pots as variable resistors (yay! now you only need x4 pots!). I would try to employ slide pots for this. The output should be buffered by an opamp unity gain buffer. Another opamp wired as a x1 inverting amplifier could follow to give an inverted sequence.

[strungout]

Tom: Damn, I didn't consider that each pot's value varies. Thanks for refreshing my memory! I get stairs, now. And I was misunderstanding how the pots work. I thought the middle would give me a staircase but, actually each pot is a voltage divider and must be adjusted going up, down or up then down or vice-versa. THis also allows for random steps, obviously.
Slider pots would be ideal, yeah, but: space. I'm definately putting in an LED to indicator to visualize the increasing/decreasing steps with along with its brightness.


Jim: I didn't connect the gate out. Figured the same.
What do you mean by "note-on trigger'?

And the RESET pin, it resets the count, I'm assuming. Do I really need to use it? Hmm four slide pots would be more manageable for the enclosure space...


I'll try the final circuit, the complex one, with the 40106/4516/4051/LM358. But if it does pretty much the same as Tom's circuit, I'll use the latter.   

I think I should stick with 4/8/16 steps, no? So it can be timed to the beats.


EDIT: it occurs to me that both do exactly the same thing, one with more parts and only 8 steps. I'll just remove the 9th and 10th step from Tom's schematic.

[anotherjim]

Jim: I didn't connect the gate out. Figured the same.
What do you mean by "note-on trigger'?

And the RESET pin, it resets the count, I'm assuming. Do I really need to use it? Hmm four slide pots would be more manageable for the enclosure space...

A note-on trigger is more a synth thing. If you want to sequence an envelope generator it will want a start pulse (trigger) as each new CV step occurs or whatever, but this really calls for being able to switch steps off for rests then a gate output only sends a trigger if a step is enabled. The same basic circuit idea can do all that with additions.

You must use reset for any count under 10. The 4017 resets back to Q0 (it doesn't have a condition where all outputs are 0 like binary or bcd counters do after reset. It resets when the reset pin goes from 0 to 1 so you use the next Q output after the last one in the sequence to reset.

[ElectricDruid]

+1 what Jim said. It's a 10-step counter, so for 8 steps, you have to tie the eighth output back to the reset pin. Then it goes 0-7,0-7,0-7 etc. Just dropping two pots will give you 0-7, blank, blank, 0-7, blank, blank. etc

For the slider pots, what I really want is something like this (this is only 30mm travel, so this is pretty small)



We could then use the LED as the one in the schematic, and it'd show you not only the position, but also the step you're on...KEWL!!

I almost don't care how much panel space that takes up. It's just got to be done, and I'm sure it has, somewhere!

[strungout]

Thanks for the explanation, guys.


And using that slider would be really cool, indeed!

[PRR]

.....almost don't care how much panel space that takes up. It's just got to be done, and I'm sure it has, somewhere! ...

For practical purpose: ARP 2500 synth, 1027 sequencer.
https://i1.wp.com/alanrpearlmanfoundation.org/wp-content/uploads/2020/09/2500_1047-2_10271050.jpg?ssl=1
https://alanrpearlmanfoundation.org/the-arp-module-1027-clocked-sequential-control/
http://www.guitarfool.com/ARP2500/1027.html
More 2500: http://www.guitarfool.com/ARP2500/

Differences: rotary pots (but three banks!!), lights to the side, and so old they were incandescent lamps (IIRC).

MFOS DIY-level plans http://musicfromouterspace.com/analogsynth_new/TENSTEPSEQUENCER/TENSTEPSEQUENCER.php

[strungout]

Paul: Well, you've pre-emptively answered the question I was coming here to ask! I wanted to be able to select at which count the 4017 reset. I think that's what the MFOS sequencer does with the rotary, right?

[ElectricDruid]

I think that's what the MFOS sequencer does with the rotary, right?

Yes, but it's pretty complicated in that circuit because it's tied up with a load of stop/start logic and all sorts of chaos on page 2.

For really simple, just take the switch common back to the 4017 reset pin. The rotary switch should be break-before-make to avoid shorting outputs together, and the reset pin should have a 10K pull-down resistor to ground to stop it floating while the switch changes.

[anotherjim]

Space-saving reset switching...

What may be making reset hard to understand, is that it looks wrong to use an unwanted step in this way. That steps Q output will go high, but only for the time it takes the internal logic to clear the count back to 0 when the resetting Q will snap back to the low state and Q0 will go high until the next clock pulse moves it to Q1.
It will cause a slight glitch in timing but that numbers in nanoseconds and of no consequence to us.

The logic sequence for a 4 step.
Start after reset Q0=1
1st clock Q1=1
2nd clock Q2=1
3rd clock Q3=1
4th clock Q4=1, Reset, Q4=0, Q0=1
5th clock Q1=1
etc...
There is a flaw as the first step from the first clock is Q1 and not the expected Q0! This is quite logical really as the chip works internally from a 4 bit BCD counter the output of which is decoded into 10 separate Q outputs. Reset of the BCD count will be binary 0000 which decodes to Q0=1. A musical sequencer would have to ensure the start of the sequence prevents the 4017 from being clocked for the 1st step and act on the setting for Q0 until the 2nd step is due.

[ElectricDruid]

Space-saving reset switching...

I like this, Jim. Very neat.

The same principle works for basically any three options. Choose your three favourites and put the two shortest ones on the switch, and the longest one hard-wired.

[strungout]

EDIT: Quick demo: https://soundcloud.com/user-165425177/sequqencedphaserdemo/s-BHLHfOlno4Z?si=53494fd259614d1ebfbb43340855a448

You can't miss the thumping and ticking!

Alright, progress!

Right now I have the sequencer hooked up to the ross phaser, through a 100k. A 3.9k in series with an intensity indicator LED. The sequence is down to 8. And it is awesome! It works as I thought. I'll try and record some samples.

But the trims are a bit hard to set, though the effect is rewarding. I imagine knobs might make it easier. Those LED sliders would've been ideal.

A big problem is ticking, but it's not the clock. It's not regular. It depends on where the trims are set. Low, the risk thumping as it switches to the next Q. High risks ticking.

That's why I have a 100k res in series to the LFO in of the phaser. Higher values get rid of any ticking but, of course, it also lower the depth of the effect. Lower values = opposite effect. Where do you guys think the ticking is coming from and what can be done about it?

[iainpunk]

isnt it easier to make a triangle oscillator with independent leading and trailing edge controls. this then followed by a sample and hold with its own frequency control.

its a bit way more chaotic, but its just my 2 cents, haha

cheers

[ElectricDruid]

A big problem is ticking, but it's not the clock. It's not regular. It depends on where the trims are set. Low, the risk thumping as it switches to the next Q. High risks ticking.

That's why I have a 100k res in series to the LFO in of the phaser. Higher values get rid of any ticking but, of course, it also lower the depth of the effect. Lower values = opposite effect. Where do you guys think the ticking is coming from and what can be done about it?

I'd guess it's ticking because the jump from one step to another is basically a square wave - an instantaneous change in value implies harmonics out to infinity! The further apart the two values are, the bigger the jump (=higher volume of square wave) and the worse the problem.
Adding some smoothing to the edges should help, but it will "blur" the waveform a bit. That could be a simple lowpass, or some kind of integrator/slew-rate limiting.

I'd also be very careful to make sure that the sequencer doesn't output values that are too low or too high for the Ross phaser circuit (the original LFO is within quite defined limits) since I know from doing similar experiments with a LM13700-based phaser that going too low caused awful thumping.

[anotherjim]

Surely the best way to match any mod source to any control is the vactrol?

Iain, I saw in the LM3900 applications a staircase generator built into the sq/tri osc.

[ElectricDruid]

Surely the best way to match any mod source to any control is the vactrol?

Definitely not *any* source to *any* control!

Imagine what would happen if you tried to use vactrols to control voltage-controlled oscillators with a voltage coming from a keyboard! You'd press a key and get a portamento-slew from one note to another whether you wanted it or not, and then the note would drift towards the correct pitch for about a minute without ever actually getting there and being *just a bit* out of tune the whole time! Argghh!

[strungout]

Iain: would that have allowed as much control?


Tom: The output I get from the sequencer is 1.31V-7.61 or so. The phaser seems to respond well to the changes in steps. but if I needed to reduce the range, how would I do it?

And what should be my target frequency for the lowpass?

I added a simple buffer at the output of the sequencer. That helped a lot with the ticking and the response of the filter. It'S still there but more subdued.


Jim: That circuit you posted with a diode OR gate, I assume it could just be repeated for every step? Using a rotary 1P8T.

[Rob Strand]

The output I get from the sequencer is 1.31V-7.61 or so. The phaser seems to respond well to the changes in steps. but if I needed to reduce the range, how would I do it?

The most generic way is to run the output through a pot/trimpot to adjust the range.  However you might want to connect one end of the pot/trimpot back to an adjustable "Vref" so you can tune the DC offset/centre of the sweep.

You can see the idea here,
https://www.hobby-hour.com/electronics/s/boss-bf2-flanger.php

In its simplest form the pot/trimpot becomes a resistive divider and the Vref another resistive divider on the supply rail.   The trick is setting up the range and DC offset to suit what you want in the first place - that's where pots help.

And what should be my target frequency for the lowpass?

Try about 10Hz to 20Hz.  Best done by ear.

A second order (2-pole) filter can work better but you want to keep things simple if you can.

Iain, I saw in the LM3900 applications a staircase generator built into the sq/tri osc.

That's a really cool idea.

[iainpunk]

Surely the best way to match any mod source to any control is the vactrol?

Iain, I saw in the LM3900 applications a staircase generator built into the sq/tri osc.

wow, thats a really really cool chip, shame it only comes in large 4 amp packages, i'd love a dual amp package, i really want to build a distortion with this kind of chip now, just for the fun of it.

cheers