8 channel vactrol-based video sequencer

smrl · Post by **smrl** » Wed Nov 14, 2012 9:18 am

hi, here's something i've been playing with for a few days. Just got the 2nd prototype working. It's an 8-step sequencer which can operate at full video bandwidth. It's based on vactrols that I got from here

So there are eight steps, each with an input, which are advanced with a clock pulse. There's also a reset jack included to reset the sequence with a pulse. The clock/reset inputs are designed to work with 5v logic. The other inputs can be pretty much anything.

It works great! Of course, there is the typical vactrol decay time...

attached is a very quick sketch of the schematic.

interested to hear thoughts/suggestions/etc. photos & a little demo video forthcoming...

numan7 · Post by **numan7** » Wed Nov 14, 2012 9:30 am

cool - that sounds very interesting, smrj - i look forward to seeing what your vactrol video sequencer is capable of! are you patching it with LZX modules, btw?

cheers

smrl · Post by **smrl** » Wed Nov 14, 2012 9:56 am

yeah, i am. But I really only just tested them for functionality - just grabbed eight ramps and multed them to both units, and multiplied the two outputs with a tvfkg... Haven't really tried anything more than that yet. Perhaps when I'm done working today...

But these should work with any signal levels (anything that the op-amps can handle, anyway)

Arturo00 · Post by **Arturo00** » Wed Nov 14, 2012 10:38 am

smrl
This is awesome! Thank you so much for sharing this! I think I'm gonna try giving this a go in the near future

I assume the components in the blue boxes are the vactrols. Will you be willing to provide a more detailed list of components? there are a few with missing values.

smrl · Post by **smrl** » Wed Nov 14, 2012 11:57 am

of course, that's why I uploaded a schematic

As I said, it was just a quick draft. The blue boxes are the vactrols, which is why there's no value given there - it's variable.

The transistors I used were BC546's but I'm sure 3904's or any common NPN would work here - they just have to handle the current you're driving the vactrols with (I played around with 4-30mA, vactrol response didn't seem to change much - but you probably just want to pick a current based on your LEDs and calculate R25 based on that. I used 150 ohms which gave me ~30mA with a 1.7v LED).

The schottky diodes used are 1N5819. Those are there to protect the 4017 inputs, and I didn't try sending it voltages outside of the range it expects, but i'm pretty confident this is fine.

Also not shown: bypass caps, power supply (the 4017 is run off of 5v & gnd)

If you wanted you could easily extend this design into a 10-step sequencer.

lizlarsen · Post by **lizlarsen** » Wed Nov 14, 2012 4:37 pm

AWESOME WORK SMRL!!!

It is important to note, just because this has not been brought up yet, that the schematic as-is won't work with 75 ohm terminated sync-inserted, standard composite or monochrome video signals. This circuit is for use with sync-stripped sources and oscillators inside of a video synthesizer system.

It could be modified to work with 75 ohm standard video signals, but they'd all need to be genlocked, and with DC restoration, sync separation / insertion, etc.

Also, LM319 Dual Comparator would be a great cheap addition to this if you wanted 1V LZX clock/reset logic level inputs with a 0.5V comparator threshold.

fletch · Post by **fletch** » Wed Nov 14, 2012 8:55 pm

cool!

smrl · Post by **smrl** » Thu Nov 15, 2012 12:29 am

Thanks for the notes, Lars. I'll update my schematics and throw in the comparators for reference, in case anyone is interested in doing it that way.

And yeah, DC restoration, sync insertion, etc... all this extra effort would take this project from something cheap and easy to something quite complicated. Though it could be interesting, someone would still need that many genlocked sources for it to be worthwhile...

[video][/video]

Veqtor · Post by **Veqtor** » Thu Nov 15, 2012 4:16 am

That looks really cool, I like the vactrol based "ease in" effect

Matos · Post by **Matos** » Thu Nov 15, 2012 4:47 am

Flapping sweet brother! Looks great! You should whip up a batch for your gimp handed brethren.

samn · Post by **samn** » Thu Nov 15, 2012 10:19 am

Agree with veqtor that ease-in is really cool

smrl · Post by **smrl** » Thu Nov 15, 2012 12:24 pm

You'll notice in that video - both sequencers are running, and one was built with faster and one with slower vactrols... you can definitely see the difference. If you changed the impedance of the op-amp circuit and/or picked a vactrol with a lower off-resistance you could make it take a bit longer.

I hand-selected the vactrols for this. I was looking for something with ~150 ohms on resistance, with dark resistance in the tens-of-megaohm range. The lower the on-resistance, the more closely matched the amplitudes will be - as any any additional resistance beyond the 5.1k impedance of the mixer circuit provides attenuation. what this means is, with an ideal 'on' resistance of 0 ohms you'd have a gain of 1, as the resistance decreases to 5.1k you'd have a gain of 1/2 and by the time your resistance is at 1M, you've got a gain of .005. The curve is exponential, and with the vactrols i've used it's significantly off within a short period of time. But if you picked a vactrol with a lower dark resistance, you'd likely get slower fades at the expense of decreased offness.

so 150 ohms is a decent compromise which leads to about a 3% amplitude reduction - and this could have easily been compensated for by making R26 slightly larger. But the variation from vactrol to vactrol is significant, especially with these surplus devices. My impression is that these are factory rejects, which are out-of-spec. If you decide to build one of these with the electronic goldmine optocouplers, get a couple of packages of these - they vary widely.

wcfields · Post by **wcfields** » Thu Nov 15, 2012 1:02 pm

An estimate of how much all the parts cost from Mouser/Digikey?

smrl · Post by **smrl** » Thu Nov 15, 2012 1:29 pm

I've got all of this stuff just sitting around. Just looking at it though, a really conservative estimate would be $30. Jacks would be the most expensive thing, the op-amp is a couple dollars, cd4017 is $.50, transistors might be $.10/pc and the resistors are pennies. The vactrols you can't get at mouser/digikey. If you find them at boutique audio places, they can be $5-10/pc. This project really only makes sense because electronic goldmine happens to be blowing out these reject vactrols.

4mspedals · Post by **4mspedals** » Sun Nov 18, 2012 1:34 am

Nice! Thanks for the tip on goldmine-elec, depending on how many good ones you find, you're getting a pretty cheap video switch out of that deal....

I might build me one...

krz · Post by **krz** » Sun Nov 18, 2012 6:33 am

MmmmmmAWESOME smrl!

mono-poly · Post by **mono-poly** » Sun Nov 18, 2012 10:55 am

aren't there potentiometers on each step?

smrl · Post by **smrl** » Sun Nov 18, 2012 11:09 am

dann-

I looked at my stock of those vactrols... I started with 60, at least 50% of the ones I got were usable for this (i've successfully used 36 of them in the last week!). I found only 3 dead ones in the whole lot. The rest just had either very low dark resistance, (1-500k), or an on-resistance greater than 200 ohms. Surely they're still useful for other stuff. Also, it might be interesting to experiment with some of the former and try a kind of PWM-driven video mixer. You might have some experience with that kind of thing? Easy high-bandwidth VCA Matrix?!?

Also, I'm in the middle of debugging another similar project, a 'scanner' rather than a 'sequencer' - rather than taking a clock pulse, it scans a bank of 10 inputs based on an input joystick. It's based on the LM3914 bargraph display IC. More details when I sort out the few bugs I've got at the moment. IMO, though, the sequencer seems to make more sense. A scanner is kind of odd with the crossfading/decay, and with the slow modulation limitation. Odd but also awesome.

4mspedals · Post by **4mspedals** » Sun Nov 18, 2012 3:10 pm

Hmm... For PWM fading, I bet the vactrol LEDs roll-off in the low audio range, but maybe a FET switch would do it? Or there's the pricey but very nice LT1251 for a real VCA. I still have a backburner project to make a 3x3 Video-rate VCA Matrix with these chips

The scanner would be fun too. If you added 8 gate inputs to the design, then you could patch up an A-151 for a sequential switch, or A-152 (or that TGS version of the RCD code) for a scanner, or different clock divisions to get various effects...

Also I'm sure you recall that the Zarlink MT8816 crosspoint chip works in the video range

... More overhead in the programming, but maybe a simple 8-gate-input AVR chip to interface would be versatile.

smrl · Post by **smrl** » Sat Nov 24, 2012 12:32 pm

Hey y'all - i decided I'd make a kit out of this. I don't know how many I'll be able to offer, but I figured it might just be best if I make up a set of PCBs, and buy out Electronic Goldmine of their vactrol stock. I called them recently and it didn't sound like they had too many left. So it probably makes the most sense for me to get all of them, then I can characterize them and redistribute along with the other parts needed for the kit. That will allow me to make the most consistency between vactrols in a given kit. Also it will maximize the number of kits I can offer.

I'm currently routing the boards. There are 3 PCBs, one faceplate board, one jack/LED board and everything else on a third. It's got a connection to the normals on the jack board so you could get several and make a normal buss between them that could be replaced by signals at the input jacks on the front. It's all through-hole, and I fit it all in 6hp! (hope I'm not speaking too soon, still have a bunch of nets to connect!!!)

Let me know what you think. PCB color? (I was thinking perhaps white soldermask with black silk screen or black with white, to more closely match the LZX scheme) LED color?

So don't buy any vactrols from Electronic Goldmine. And if you need any because you want to build something else or something along these lines, get in touch and I will try to help you out. I've also got a ton of VTL5c3's.

zapp550 · Post by **zapp550** » Sat Nov 24, 2012 2:00 pm

I'd be interested in a kit if you ended up making them

smrl · Post by **smrl** » Sat Nov 24, 2012 4:48 pm

mono-poly wrote:aren't there potentiometers on each step?

perhaps it would be more clear if i called it a 'sequential switch'? Eight full-bandwidth inputs rather than eight CV's set by pots. Of course you could normal 8 pots so that you'd have a more ordinary sequencer. But you would still have a fixed slew rate / crossfade determined by the vactrol response.

smrl · Post by **smrl** » Sat Nov 24, 2012 4:54 pm

Also-- I've added two comparators at the input so you can select logic transition voltage -- ~.5v for LZX interfacing or ~2.5v for just about anything else.

and reset currently works in a gate-like fashion -- it will hold at step 1 while high. thinking I should leave it this way, but I could make it trigger instead...

laserpalace · Post by **laserpalace** » Mon Jan 28, 2013 5:51 pm

smrl
I was just thinking about how great a "sequential scanner" interface would be for a video system after watching this little documentary on the SYNKIE

[video][/video]

Glad I found this thread! Any updates on your process or possibly providing kits for this?

mono-poly · Post by **mono-poly** » Wed Feb 06, 2013 7:22 pm

smrl i'd be up for a pcb for this!

MOD WIGGLER

MOD WIGGLER

8 channel vactrol-based video sequencer

8 channel vactrol-based video sequencer