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

Track & Hold in MU
MUFF WIGGLER Forum Index -> 5U Format Modules Goto page 1, 2  Next [all]
Author Track & Hold in MU
defutura
What T&H modules are there in MU/5U format?

The only one I've found is the Oakley Sample & Slew, which has a Track mode. Are there any others?

Are there any suitable DIY candidates?

Can a regular S&H (say, Arrick's) be modded to exhibit a T&H behavior, rather than S&H?

I'm looking to use it with a ribbon controller to make it 'remember' the last held pitch.
ranix
I'm not clear on the difference between track and hold and sample and hold. Does the dotcom sample & hold behave like a "track and hold"?
Sideshow
You can make the track and hold yourself out of a regular S&H and a voltage controlled switch:
defutura
ranix wrote:
I'm not clear on the difference between track and hold and sample and hold.


"S&H = samples the input signal. Every time the gate is high a new sample will be available at the output.

T&H = follows the input signal as long as the gate is high. If the gate is low the last signal is sampled and presented at the output."

I've borrowed this explanation from user e-grad over at the general forum. Credit where credit is due.

ranix wrote:
Does the dotcom sample & hold behave like a "track and hold"?


I don't think so, but if it does I'd be happy to know.
defutura
Sideshow wrote:
You can make the track and hold yourself out of a regular S&H and a voltage controlled switch


Thanks! I'm trying to wrap my head around this, and I don't have a switch module, so I can't test it.

But just from looking at it, I don't understand what it is in this patch that keeps the S&H from going on sampling when the gate goes low (and the ribbon controller drops to zero CV).

What is the clock source? Is the clock source very slow? If so, how can I tell the S&H to sample at every new gate ON?
ranix
ah, I see

It doesn't, I just checked and it does have the "sample and hold" behavior. I was able to create "track and hold" behavior using a q117 and q962 together to switch between the held voltage and the dry signal. Needs an x2 clock though, a q128 switch could probably do it better.
Sideshow
defutura wrote:
But just from looking at it, I don't understand what it is in this patch that keeps the S&H from going on sampling when the gate goes low (and the ribbon controller drops to zero CV).

The S&H will continue to sample, but the VC Switch will make so that you get the free running signal when the gate is low (or vice versa if you like...)

defutura wrote:
What is the clock source? Is the clock source very slow? If so, how can I tell the S&H to sample at every new gate ON?

The clock is whatever you like to use as clock but if you would like the S&H to sample at every new gate, then just connect the gate into clock inout.
defutura
Sideshow wrote:
The S&H will continue to sample, but the VC Switch will make so that you get the free running signal when the gate is low (or vice versa if you like...)


But the thing with T&H is that when the gate is low, you don't want the free running signal, you want the last voltage held until the next gate.

From the description of the Track mode on the Oakley module (emphasis mine):

"The second mode is 'track'. [---] With a clock input above 1V then the output of the module will follow in the input signal. When the clock input falls below 1V the output will then be held and not change again until the clock input rises again above 1V."
Sideshow
defutura wrote:
But the thing with T&H is that when the gate is low, you don't want the free running signal, you want the last voltage held until the next gate.

Shouldn't that just be to switch the two inputs on the VC Switch?
Just me
I have also wanted a Track and Hold. Looks like the only easy way would be to convert the Doepfer module to MU. http://www.doepfer.de/a148.htm There were other circuits for it out there and in the DIY forum, but all connected to much larger builds. (appendage comes to mind)
There was a large complicated and very cool project a few years back that gave all sorts of control over ribbons. T&H was a very small part of it.
defutura
Just me wrote:
I have also wanted a Track and Hold. Looks like the only easy way would be to convert the Doepfer module to MU.


What about the Oakley module, then?

Just me wrote:
There were other circuits for it out there and in the DIY forum, but all connected to much larger builds.


In DIY, there's also this module from Synthrotek. I've emailed them to see if it runs on 15v.

Just me wrote:
There was a large complicated and very cool project a few years back that gave all sorts of control over ribbons. T&H was a very small part of it.


Interesting. Here or on Electro-music? Do you have a link?
johny_gtr
This is quite interesting S/H from Frequency Central:
http://www.frequencycentral.co.uk/?page_id=1363

I hope they will have MU panel for DIY or ready-for-use module.
defutura
Sideshow wrote:
defutura wrote:
But the thing with T&H is that when the gate is low, you don't want the free running signal, you want the last voltage held until the next gate.

Shouldn't that just be to switch the two inputs on the VC Switch?


Maybe it should. But for this to work you need to tell the S&H to sample at GATE OFF, whereas it normally samples at GATE ON (from the clock/trigger signal).

For the ribbon controller use case, you want the free running signal while the gate is ON (you're playing), and when you let go and the gate goes OFF/LOW you want the last held pitch to stay on until the next GATE ON.
defutura
johny_gtr wrote:
This is quite interesting S/H from Frequency Central


Yeah, I noticed that module, too, thanks. But it doesn't seem to do T&H.
Just me
Oakley sags over time and the cost is beyond what I'm willing to pay. I will probably just make my own at some point.
defutura
Don't all S/H circuits exhibit voltage drop, due to the electric charge being stored in a capacitor?

Are you saying the Oakley module has a worse hold time than the others, making it musically less useful, compared to say Doepfer, Dotcom, MOTM or similar circuits used by ARP or Moog in the past?
sduck
MOTM 101 does T&H. Don't know if it's available anywhere anymore.
defutura
sduck wrote:
MOTM 101 does T&H. Don't know if it's available anywhere anymore.


Thanks, sduck. Yeah, it seems Synthcube's selling it.
defutura
I don't know if I'm reading the spec sheets correctly, but it seems so me that the Oakley and MOTM T/H circuits work in opposite ways:

MOTM wrote:
...in the TRACK position, the output follows (tracks) the input when the clock is low (less than 1.5 volts) and holds the output when the clock is above 1.5 volts.


Oakley wrote:
The second mode is 'track'. [---] With a clock input above 1V then the output of the module will follow in the input signal. When the clock input falls below 1V the output will then be held and not change again until the clock input rises again above 1V.
sduck
It can be implemented either way. Just use an inverter if it doesn't work the way you want.
ranix
defutura wrote:
Sideshow wrote:
The S&H will continue to sample, but the VC Switch will make so that you get the free running signal when the gate is low (or vice versa if you like...)


But the thing with T&H is that when the gate is low, you don't want the free running signal, you want the last voltage held until the next gate.


that's the exact behavior you get out of q117 + q962. The output switches between tracking and holding on the rising edge of the clock. The falling edge does nothing. The q128 would provide switching from tracked to held on the rising edge, and switching from held to tracked on the falling edge (or vice-versa)
defutura
Hmmm, intriguing... Many thanks for the information.
ranix
don't rush out and buy a q962 for this, there seems to be some strange interactions using it for this patch and I am trying to figure them out now. It's not as easy as it should be for some reason, my 962 is doing weird stuff
ranix
I think I found the problem - both of my q962s freak out if the Shift CV drops below ~-100mV

This probably means the q962 + q117 won't work well for you, since you probably can't clock it off whatever you feel like without a signal processor. I have one other sequential switch, the Moon 564, which I don't use in the audio path due to switching noise. The q962 doesn't have that problem.

This is not a problem for me because I have many workarounds.
Synthbuilder
All analogue sample and holds will drift. The key question is by how much. The Oakley Sample/Slew is not optimised for slow droop although it should be similar to the MOTM one since both circuits use the same core sample/hold circuit. The Oakley has two S/H circuits internally to allow for the 'restriction' control which is my implementation of the correlation control on the Buchla 265. The Oakley Sample/Slew module was devised in the main to generate random and semi-random CV and audio rate events. It samples relatively quickly but with no great emphasis on accuracy or hold time.

As with all engineering problems it's a compromise scenario. You can build a sample and hold capable of fast sampling with a certain amount of droop, and you can build a sample and hold that allows for less droop but requires a longer time to sample the input. It's not a trivial issue to have an analogue sample and hold that will sample signals across the audio band and slow moving CVs and still have an accurate low droop output across all of that.

For accurate sampling and infinite hold times I'd chose a digital solution. No such product exits from me although one, and its bigger brother a quantiser, have been on my list for quite some time.

For a ribbon controller the sample time can be quite long, several mS, and therefore it is possible to design an analogue sample and hold circuit with small enough droop time to be usable over a minute or so. However, you need to be able to tell the circuit to hold the sample before you take finger off the ribbon. My first thoughts are that this is not straightforward as the CV output from the controller may become unstable the moment before the finger is removed and the gate signal resets to zero.

Either way, I don't think the Oakley Sample/Slew module is the one for this job.

Tony
MUFF WIGGLER Forum Index -> 5U Format Modules Goto page 1, 2  Next [all]
Page 1 of 2
Powered by phpBB © phpBB Group