Author |
Message |
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
Posted: Tue Sep 19, 2006 6:25 am Post subject:
|
|
|
i´m getting started on this and it drives me cracy!
got the dac from ad now, but all it gives me is not a stair but lots of jumps. It does cover quite a range, so maybe it doesn´t matter as long as it´s 1/12v steps. Don´t know what´s wrong.
The schematic from the first posting got pin 2,3 of the dac connected to ground. With that i don´t get any signal. Leaving them open gave me that nonstair.
Now i´m trying to get the ad781 s&h to do the job and that too doesn´t work.
It´s open all the time. All i have to do is add an inverter to it´s controll input, right?
Man, this gets frustrating... _________________ http://www.myspace.com/lorolocoacousticpop
http://www.myspace.com/petrolvendor
music and transcribed jazz basslines |
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Thu Sep 21, 2006 8:25 am Post subject:
|
|
|
Hey Zipster,
I dig what you're saying - I love a good mad scientist sound as much as anyone, but you're right - it doesn't take a $$ DAC and S&H chip to do it
The CD4024 isn't the best choice here, being a ripple counter. It's outputs are going to kind of tumble out like a row of dominos rather than just 'snap' out all at the same time with each clock pulse. This means, while its tumbling, it may excite your comparator, but finish its tumble before the S&H grabs it - this throws you off a step or two.
Not that I think that's what's going on here - to me, it sounds like the S&H pulse is too wide, or it's continually getting retriggered for some reason or another. What you might try doing first is getting a better pulse conditioner on there - for one thing, IMO, those two caps there just don't cut it for a nice, crispy pulse.
I've attached a schemo - if you have a CD40106 handy, you might try the top portion of it to see if it improves things any. It's a pretty quick pulse - adjusting the cap value will widen or narrow it. If you need to play around with delaying it, check out the Klee pulse delay used for the clock delay in Episode 3 of the Klee thread. You can do a similar thing on the first CD40106 stage, right after the comparator.
The bottom part of the circuit shows how you can use two CD4516s to replace the CD4024. The CD4516 is a synchronous counter, and I find two of them to be much better behaved for this app than the single CD4024.
Good luck and take care,
Scott
EDIT IMPORTANT
YIKES! Forgot to put a diode on the input of the first CD40106 stage - this is needed to keep the negative output of the comparator from plastering your CD40106 but good!!!!!!! Put it in, cathode to comparator, anode to CD40106. Put a 100K resistor to ground on the input pin of the CD40106 as well - this will pull it low and keep it from floating around while your comparator is not high.
EDIT 2 - I notice on my quantizer foray, I used values of 560 pF and 47K for the pulse width values shown on the second CD40106 on the schemo.
Description: |
|
Filesize: |
40.03 KB |
Viewed: |
447 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
|
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
Posted: Fri Sep 22, 2006 2:28 am Post subject:
|
|
|
Hi Scott
Thats cool, thank you. So you think my stair not being a stair comes from the 4024? You talked about it´s problems on your homepage, i just didn´t expect it to be such a problem...
Maybe it also has to do with the dac on the breadboard with no blocking cabs and no ground seperation. I´ll check that out. To me diy-synth always has that "no part is critical, it might just sound funny" feel.
Maybe the more i get into stuff like this a different aditude is needed.
The sample i posted was generated from the dacs output, not that of the s&h. I´ll get those other chips today and start over with this. See what comes out. _________________ http://www.myspace.com/lorolocoacousticpop
http://www.myspace.com/petrolvendor
music and transcribed jazz basslines |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Fri Sep 22, 2006 6:32 am Post subject:
|
|
|
Quote: | The sample i posted was generated from the dacs output, not that of the s&h. |
That would explain why the S&H sounded like it was being triggered more than once
I'm not sure it's the CD4024's fault - I do know things were a bit unruly til I got rid of it and went with two synchronous counters. Hope it works out!
Take care,
Scott |
|
Back to top
|
|
|
bugbrand
Joined: Nov 27, 2005 Posts: 846 Location: Bristol, UK
Audio files: 1
G2 patch files: 1
|
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Fri Sep 22, 2006 8:24 am Post subject:
|
|
|
I'm not sure that bleedthrough is going to be function of pulsewidth, but rather a function of whether your sample pulse is high or low and how well the IC functions when it's supposed to be holding. It can also be a function of how quickly the S&H reacts to a high or low, but I imagine that point is moot, owing for the fact that S&H ICs are designed to react very, very quickly.
The S&H IC, be it an AD781 or LM398, allow, when the pulse is high, anything to pass through and when the pulse is low, it blocks all input and the output's high impedance prevents the cap from discharging. The LM398 can be configured to do the opposite, of course (block when high and pass when low). I imagine the AD781 can be configured likewise?
Anyway, if theres a way to short across the cap that differentiates the incoming high into a short pulse, it converts the function to a 'track and hold' - as long as the input is high, the input passes through unimpeded and, when it's low, the output 'freezes' (a very fun function, BTW - imagine an LFO sweeping, freezing at a certain point, begining its sweep from a different point, freezing, ad nauseum). It's a simple mod for a lot of S&H's out there.
Back to the issue of pulsewidth - if the input is wiggling around a lot and the S&H pulse is wide enough so that the ear can detect that wiggle, you'll hear that on the output. The Buchla 266 SOU S&H does just that - it's got a pretty wide S&H pulse width - somewhere around a ms IIRC. If one is feeding it a very fast signal, such as a VCO or anything moving at an audio rate, one or more cycles can pass through during the time the sample pulse is high. This translates more to a percussive attack at each sample rather than as a discrete tone - not enough of the signal gets through for the ear to decode it as pitch.
Of course, one can also make the input pulse too narrow for the cap its supposed to charge - in this case, its output shuts off before the cap has had a chance to charge/discharge fully to the S&H output.
I use the LM398 mainly because they're easy to get hold of here and cheaper than the AD781. They seem to do just fine. If you go into mass production with the crusher, you might see if it fits the bill - it would lower BOM cost if anything.
Cheerio,
Scott |
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
Posted: Fri Sep 22, 2006 11:58 am Post subject:
|
|
|
For modulation sources that behavior of the s&h ics can be very usefull.
Can be used like a kind of "switch and hold"
A slow lfo is passing. Each time the controll goes low the lfo seems to be stopped, when the controll goes high again the lfo seems to continue from a different point.
Ok, i get multiple triggers because my "stair" is jumping across the cv all the time. Seems logic. I thought it wouldn´t matter if it´s actually a stair or jumping around, as long as all semitones are present. Didn´t think enough there... All understood.
AD was so kind to send me two of the dacs, since they both act the same it has to be either the 4024 or my double checking the wireing still missing something.
Scott, i saw on your page that you sent pin3, "ref out" to some opamp circuitry. What´s it for? _________________ http://www.myspace.com/lorolocoacousticpop
http://www.myspace.com/petrolvendor
music and transcribed jazz basslines |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Fri Sep 22, 2006 12:53 pm Post subject:
|
|
|
Hi ZipZap,
Pin 3 is the +5V reference output of the DAC. I buffer and invert it to provide a 5V offset to center the staircase at 0V - this is for bipolar operation.
Cheerio,
Scott |
|
Back to top
|
|
|
Coriolis
Joined: Apr 11, 2005 Posts: 616 Location: Stilling, Denmark
|
Posted: Sat Sep 23, 2006 12:49 pm Post subject:
|
|
|
Don't mean to hijack the thread, but I have a general question about quantizers: I've been looking at the quantizer designed by Scott Stites http://mypeoplepc.com/members/scottnoanh/birthofasynth/id15.html
and was wondering about the basic operation of this. Is it correct, that if I have, say, a 4 ch sequencer, then I can take the summed output of each and put them through separate channels of the quantizer, and get 4 different melodic lines for instance?
Or is it the other way round - every step of, say, a single 8 step sequencer, needs it's own quantizer channel?
A bit confused as you can see! Sorry again for hijacking, but I feel I'm missing the finer points of this excellent thread!
C |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sat Sep 23, 2006 6:33 pm Post subject:
|
|
|
Hi Coriolis,
Each channel of the quantizer will quantize its input signal individually. Say you put a sequencer into channel 1, and LFO into channel 2 and a sample and hold into channel 3.
Channel 1 will output the quantized sequencer voltage, channel 2 will output the quantized LFO voltage and channel 3 will output the quantized S&H voltage.
I've been meaning to redesign that quantizer with less range - 10 octaves puts very narrow decision points in there, almost too narrow.
Cheers,
Scott |
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
Posted: Sun Sep 24, 2006 2:53 am Post subject:
|
|
|
Hey guys, know what was wrong? I do (someone in a german electronics forum did). In the shematic all outs of the 4024 have to be connected in opposite direction (and use pin11, not 10 on the 4024).
So q1 goes to db0, not db11... q2 to db1 and so on.
Now i´ve got a nice stair. The lowest step is a bit longer than the others, don´t know if that´s a problem though.
Disconnecting some of the steps makes new scales.
Sometimes it´s so easy.
Now i move on to the comp section. _________________ http://www.myspace.com/lorolocoacousticpop
http://www.myspace.com/petrolvendor
music and transcribed jazz basslines |
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
|
Back to top
|
|
|
Coriolis
Joined: Apr 11, 2005 Posts: 616 Location: Stilling, Denmark
|
Posted: Sun Sep 24, 2006 6:32 am Post subject:
|
|
|
Thanks for the explanation Scott! May I suggest about 5-6 octaves then?
C |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Mon Sep 25, 2006 7:31 am Post subject:
|
|
|
Yup - that's the range I was thinking.
Cheers,
Scott |
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
|
Back to top
|
|
|
Coriolis
Joined: Apr 11, 2005 Posts: 616 Location: Stilling, Denmark
|
Posted: Tue Sep 26, 2006 7:54 am Post subject:
|
|
|
Yeah, is it just a matter of plopping a resistor on either end of the trimmer to limit range, or...
Cos then we could do it ourselves, and you could concentrate on the Super Klee...
C |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Tue Sep 26, 2006 9:33 am Post subject:
|
|
|
I was afraid someone would ask that question Last year I could have answered it right off the top of my head, but now I have to remember what the idea was....
I'll chew on it for a bit here.
Cheerio,
Scott |
|
Back to top
|
|
|
fonik
Joined: Jun 07, 2006 Posts: 3950 Location: Germany
Audio files: 23
|
Posted: Thu Sep 28, 2006 6:16 am Post subject:
|
|
|
hi zip zap,
i want to try that circuit, too. (including the improvements scott and you suggested). what OPA you're using for the oscillator/clock? or do you use the schmitt triggers scott suggested?
cheers,
matthias |
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
|
Back to top
|
|
|
fonik
Joined: Jun 07, 2006 Posts: 3950 Location: Germany
Audio files: 23
|
Posted: Sat Sep 30, 2006 9:51 am Post subject:
|
|
|
4584? a hex schmitt trigger or a genlock (pll)? |
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
|
Back to top
|
|
|
zipzap
Joined: Nov 22, 2005 Posts: 559 Location: germany
Audio files: 24
|
Posted: Sat Oct 14, 2006 2:27 am Post subject:
|
|
|
Just came back to this, been away for a while.
It´s not working. Overtested chaining those 4516s. On breadboard, on pcb. The clock is ticking, but the 4516s outs are all high. What could be wrong? somebody got an idea?
Thanks _________________ http://www.myspace.com/lorolocoacousticpop
http://www.myspace.com/petrolvendor
music and transcribed jazz basslines |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Oct 14, 2006 3:59 am Post subject:
|
|
|
Have you got the "Carry In" input held low? If it isn't it won't clock. _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
|