MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

cv controlled sequencer/programable switch/analog wavetable
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2  Next [all]
Author cv controlled sequencer/programable switch/analog wavetable
meridic
Hope you like hand drawn photographed schematics!

http://holes.asdflkajsgoihwoijscvlxmkcvznzslhgfqweroityulkxergkljfbbda o.org/cv_controlled_seq.gif

Here we go, no clocks controlling this sequencer. We got ourselves a home made 3 bit Flash ADC made up of some resistors, 8 opamps and a 8 to 3 priority encoder. The encoder switches the 4051. Any signal can run this sequrencer.

The CV input is a percision rectifier, any negative signal will just output 000 from the encoder so might as well fold anything from under up. Signal is dumped into the comparator chain , each comparator triggers at a different level determined by the resistor chain. So we get pulse outputs as the comparators fire. The encoder reads the outputs of the comparators and picks the one of highest priority, this is then a decimal number. If the first 4 comparators go off the encoder only pays attention to the 4th and interprets that as 00010000, the numbers bellow 4 are just ignored. That is pretty useful for us since we can make that into a 3 bit binary number 100. Hey the 4532 does that for us as well. This three bit number controls the 4051 and gives us some pulse outs by way of some transistor switches that I stole from Mr Stone. The rest is pretty standard, the 4051s input is tied to +V and the outputs go through some buffers and are mixed down (Thanks again Mr Stone) to give us a nice CV output.

As of now the circuit stalls if a CV of more then 5Volts is applied, the encoder just outputs 111 for anything over 5Volts and 000 for anything below 0Volts. The CV input will eventually be reworked to keep things like this from happening. Along with adding a depth control and manual CV for more fun. Maybe tomorrow or something.

Edit, If you do decide to make this in the next few days don't build the percision rectifier, just leave space for a quad opamp and some parts. That are is going to get a good overhaul and be made much more fun.

Edit 2: opamps can be just about anything. Diodes are standard 4148. Make sure to decouple at all ICs and all that good stuff. Would not hurt to give the digital and analog their own grounds, the comparators would be digital, or maybe give them their own as well.

Edit 3: Along with sequencing this can also be used as a 8 stage wavetable gen, it even keeps the same pitch as the input wave, or twice the pitch if the input signal is bipolar. You will want to use a good fast opamp or better yet, proper comparators for the 8 in the chain if you want to run this at audio rate. As is you can do low audio rate and modulation, with something like the morphlag you will get nice wavetable modulation action synced to your source LFO.

Edit 4: Actually it may not hold pitch, it depends on your settings of the pots. What it will do is cycle through all 8 steps for every single cycle of the wave form, so you could do a 4 step wave table and double pitch, or do a 2 step wave and quadruple pitch.

Pulse/square waves are of no use on the CV input with this module. well not no use, they will just cause it to bounce back and fourth from 1 to 8 and skip everything in the middle. A sequencer would be a great input though, program the step order.
DGTom
Love the hand drawn photo!! For some reason I'm much better at graph paper + graphite than CAD but seriously, i just don't get it

Really nice! I'll chuck some 4532s in with my next f'lec order, they look really usefull paired up with the 4051.

Could you flip the 4051 around & negate the need for the 8 NPNs there?

Take +5V thru a pot & use the mux as 8 in, 1 out & then whack an op-amp on the output to buffer it, or maybe 2 to get an inverted version of the sequence.

Interested to see the re-vamped input section.
meridic
Quote:
Could you flip the 4051 around & negate the need for the 8 NPNs there?

Take +5V thru a pot & use the mux as 8 in, 1 out & then whack an op-amp on the output to buffer it, or maybe 2 to get an inverted version of the sequence.


Sure, the reason I did it this way mostly for the sake of having a left to right signal flow and nice straight lines on the schematic! There are some other reasons though. It came out to be a pretty much 50/50 trade off as to which way was more useful, for me this way won out for when used as a programable switch, with the addition of individual outputs that is. That is why the transistors are really there. But as you can see I ran out of paper.
meridic
Got the new CV input circuit done while waiting on the laundry. It will be a day or two before I can get into the computer, camera battery died and I have no charger. Looks like page two will be CAD. I will probably also update the schematic to show the programable switch feature.

The idea for this circuit actually came from the other sequencer thread of wetterbergs. Someone mentioned that sequencers must have a program input for step select to be useful, to which wetterberg said there were no simple circuits with this feature out there. This was as simple as I could come up with and still be useful. at least as simple of a circuit that I could come up with while making my bread. I actually did have an easier one, made the ADC with a diode string, XORs for a priority checker and diodes for the 8 to 3 encoder. It was fun but but did not work so well, the linear response was kind of nifty though.

Edit: Actually I am going to dump the new CV input circuit maybe, while it would allow for some interesting effect, I think I might just make the ADC resolution adaptive to the input CV. That way there will be no dead spots regardless of the input, make it swallow both bipolar and unipolar waveforms as well. The fun stuff the input circuit could do can be done externally anyways.
meridic
Ok, the proof of concept on the adaptive circuit was a success it seems. Slightly more complex circuit but it no longer needs the +5 volts. Now swallows bipolar signals ease with so no need for the rectifier on the input. Next will be the redesign of the output section, get rid of Kens buffer/summer and see if I can make the output adapt and spit out a bipolar wave if the input is bipolar. Not sure I can pull that off without a drastic increase in complexity. A switch or two outputs is more likely. I will also add in the the input so it can be used as a programable switch. But perhaps a 8in 1out switch would be more useful then a 1 in 8 out. To bad the complexity of making it bidirectional is prohibative. Or perhaps just dump the programable switch and work on the other features more.

Laundry is done, bread is done and it is bedtime for now.
fonik
nice one. i like your use of the CMOS.

what about additional pulse outputs ANDed with a clock? i know the tempting idea of your circuit is NOT needing an external pulse. maybe sometimes one would like to sync it. however this could be done by feeding the module with a CV from a S/H anyways...
zthee
Why not feed the output buffer -V and just have one inverting output instead of 2?

I like the idea. But to be honest it seems like there is something more needed... I don't know what though.. smile
frijitz
meridic wrote:
... We got ourselves a home made 3 bit Flash ADC made up of some resistors, 8 opamps and a 8 to 3 priority encoder.

Have you thought about using the comparator chain built into the LM3914 bar graph chip?

grin

Ian
wetterberg
Wonderful layout!!! *THIS* is what we need!

Ian, I've heard that the lm3914 does a short crossfade between states, making it difficult to use in this manner - but perhaps when connected to something like an 8-to-3 decoder this would be a non-issue?
meridic
Quote:
what about additional pulse outputs ANDed with a clock? i know the tempting idea of your circuit is NOT needing an external pulse. maybe sometimes one would like to sync it.


Thats what the binary outputs are for. They are going to be improved, edge detector on each one and a 3 input OR gate. Will be capable of some rather odd timmings with the right input. I have thought about a clock input, but decided against it.
Quote:

Why not feed the output buffer -V and just have one inverting output instead of 2?


As I said the output buffer is not mine, I just swiped it since it happened to be sitting right in front of me at the times and would work. and there is no -V.

Quote:
I like the idea. But to be honest it seems like there is something more needed... I don't know what though..


sigh... "more" .....

http://www.youtube.com/watch?v=qDuW3NvjqJY
Quote:

Have you thought about using the comparator chain built into the LM3914 bar graph chip?


Only for a moment right after I read your post. I will need to look into those chips abit. My conclusion reached during that moment was that for 3Bits there is not much savings in cost, layout or complexity when switching to the bargraph chip. So unless it brings something new and fun in I doubt I will switch over. I suspect I will be digging through the data sheet within the next 24 hours though.

Quote:
Wonderful layout!!! *THIS* is what we need!


Enjoy! It should be getting some improvements in the next fe days as well.

I think I am going to follow Tom and flip the 4051, decided I do not care about the programable switch feature so might as wiell simplify. I will try and get the new circuit up in the next few days.

Thanks for the comments.
wetterberg
SlayerBadger! SlayerBadger! SlayerBadger!
zthee
meridic wrote:
Quote:
I like the idea. But to be honest it seems like there is something more needed... I don't know what though..


sigh... "more" .....

http://www.youtube.com/watch?v=qDuW3NvjqJY
[/quote]

applause w00t lol
frijitz
wetterberg wrote:
...I've heard that the lm3914 does a short crossfade between states, making it difficult to use in this manner - but perhaps when connected to something like an 8-to-3 decoder this would be a non-issue?

In "dot" mode there is an overlap between the stage outputs. (Wish it *was* a crossfade!). But for the present appplication "bar" mode would be used. It seems to me this should work fine. The overlap region might slightly delay the transitions, but this could be a good effect, as it might prevent chattering at the x-over points. I believe chattering at low ramp rates could be a problem using comparators without hysteresis.

I've used the LM3914 for a number of aplications, but I never thought of using it with a priority encoder. d'oh! d'oh!

It's going to be very interesting to see how this develops.

grin

Ian
wetterberg
frijitz wrote:
I've used the LM3914 for a number of aplications, but I never thought of using it with a priority encoder. d'oh! d'oh!
actually, just a simple cv->3-bit gate out type converter would be super cool to have. Talk about simple, it would probably be simpler than a Baby10.

right, wherever this ends up, can't we print some simple PCBs for it? That way we can all play along, no matter how lousy our breadboard chops are help
meridic
The 3914 certainly seems like it could have some benifets. The biggest is for increased number of steps without the massive comparator increase and can be chained for more steps. But it has 10 outputs which just does not work well in this case. If it was not for the overlap in dot mode it would be the perfect solution, just dump the encoder and use the outputs to switch on 4066s, but no luck there. a 16 to 4 encoder and a 16 channel mux will work, but I sure hate the idea of all those unused i/os.

I have some 3916s kicking about, I think I will see what I can get them to do.

Chances are this is going two break into to seperate projects, the original idea deffinetly wins out for audio manipulation and modulation. But the bargraph chip is going to win out for sequencing. At least that is my guess.

Wait, I gave up on the whole switch idea, The mux is not even needed, which means the binary is not needed. Just need to get decimal output from the ADC, run it through pots and sum it all up.

I had the epiphany of not needing the mux and also had second epiphany that with the mux/old circuit we can make a really nifty wave shaper.
meridic
Quote:
right, wherever this ends up, can't we print some simple PCBs for it? That way we can all play along, no matter how lousy our breadboard chops are


PCBs suck.

But something can probably be managed.
fonik
bargraph drivers came to my mind too, but i did not mention them because i thought one would need to add inverters. aren't the LM3914/15 stages current drains actually? i am not sure if this is the correct term. i mean the stages are normaly high and the active stage would switch to GND. so one would need inverters, right?
alas, i never did get around the massive logic, but this reminds me of tom gambles ADC flasher:
http://www.modular.fonik.de/files/EFMforum/orgEFMfiles/adcflash.pdf
wetterberg
delegate, my brotha! smile No need to worry about that at this time.

re. the 10 outs of the lm3914; isn't this just a matter of tying outputs 8-9-10 together? That way when the voltage gets that high, it just turns on the last step, as you would want it.

I don't know, LM3914->4532->4051... seems lovely and simple... one kind of wants it to be more complicated, hehe.
fluxmonkey
grant richter's Analog Tracking Generator might be relevant, it uses the 3914: http://www.musicsynthesizer.com/DIY/Grant/CVtwister.html

b
meridic
Quote:
i mean the stages are normaly high and the active stage would switch to GND. so one would need inverters, right?


Well, I need to make a priority checker for it anyways, I can just go with lowest priority instead of highest.
Quote:

re. the 10 outs of the lm3914; isn't this just a matter of tying outputs 8-9-10 together? That way when the voltage gets that high, it just turns on the last step, as you would want it.


That would work, but it would create a sort of hold feature every time the input of 9 or 10 triggered.

Quote:
I don't know, LM3914->4532->4051... seems lovely and simple... one kind of wants it to be more complicated, hehe.


Those three will not play well with each other, there would be a dead spot as I think I explained above, not sure how well I did. While the LM3914 seems simpler, I do not think it will be with a finished circuit, I suspect they would be about even.

The one real problem with the LM3914 that I see is it will be harder to extract a good clock from it. This is more fallout from dumping the encoder/mux then an issue with the IC. Although I bet I am overlooking a simple way of doing it.
meridic
Quote:
grant richter's Analog Tracking Generator might be relevant, it uses the 3914:


When Ian mentioned the 3914 I remembered this circuit and for a moment I thought I reinvented both the 3914 and Grants circuit. Thee are certainly some simularities and overlap in the capabilities of the two, but some differences.

It is a good circuit that I had forgotten about and now have been reminded of twice today. May have some goodies in it on how to elaborate on mine.
frijitz
fonik wrote:
bargraph drivers came to my mind too, but i did not mention them because i thought one would need to add inverters. aren't the LM3914/15 stages current drains actually? i am not sure if this is the correct term. i mean the stages are normaly high and the active stage would switch to GND. so one would need inverters, right?


Or ... drive the input with a down ramp?

grin

Ian
frijitz
wetterberg wrote:
... the 10 outs of the lm3914; isn't this just a matter of tying outputs 8-9-10 together? That way when the voltage gets that high, it just turns on the last step, as you would want it.


Much simpler than that. You can vary the number of output steps by adjusting the relative magnitudes of the input signal and the bias voltage at the top of the comparator ladder.

This is how my 5Pulser waveshaper works.
http://home.comcast.net/~ijfritz/sy_cir8.htm

grin

Ian
meridic
After some thought I decided to put off experimenting with the bargraph chips. I am not seeing anything good in the way of off the shelf parity checkers so it would require one rolled from gates. It also lacks a few areas of future exploitation that the original has. Mainly I got an idea of how to alter the ADCs response that I want to play with in the future. For now it will be just expo, but It may go linear, I am on the fence as to which is more useful, or if one is more useful then the other even. I also decided to to dump the adaptive resolution, and went back to my improved input circuit. New features are a clock output and a resolution CV input. The programable switch feature has been scrapped and the 4051 has been flipped for a simpler circuit, I think I tried to make it to useful and was losing some of the ability it had in other areas, and zthee wanted more?

The circuit is mostly drawn up in the computer land, but having to relearn xcircuit has made it slow. I am not going to have time to test it for a few weeks most likely, but I can help troubleshoot any problems for the brave souls who attempt it. For the most part it will work, the majority of the circuit has been tested in one form or another, it just may not work completely as I intended. Not sure if anyone other then myself will be able to notice that though.

Oh, perhaps I should add that hysteris as Ian reccomended as well.

I will get the schematic up in a day or two.
meridic
I have been trying to figure out a simple way to add hysteresis the to ADC but am having no luck coming up with a good simple way to add it to all of them and not individually. Anyone have any ideas?

Best Idea I have had is running the clock output into the resistor chain. I think it might work but not sure, it would not be perfect but perhaps good enough. The pulse should last long enough to get the comparator past the its indecicisive stage to were its state should be an obvious choice by the time the pulse goes low. Perhaps that clock should be inverted as well.

OK, I did it, I reinvented Grants circuit. Well not exactly but I have gotten pretty close to it. I am going to have to think on this abit to decide if mine is worth the time to sort out since there is already a working circuit pretty close to mine out there. Clock output would be nice though which would be tricky on his. I also have a few future Ideas to extend the usefulness of mine which could not be done on his, but that is for a later date when I have more time. But I can also see some tweaks that could be done to his circuit to make it more useful as well.

So what do you all think? Enough difference in my circuit to make it worthwhile to persue further?
MUFF WIGGLER Forum Index -> Music Tech DIY Goto page 1, 2  Next [all]
Page 1 of 2
Powered by phpBB © phpBB Group