Step-limited Sample & Hold

Anything modular synth related that is not format specific.

Moderators: Kent, Joe., luketeaford, lisa

Post Reply
User avatar
pugix
Super Deluxe Wiggler
Posts: 3572
Joined: Sat Oct 31, 2009 7:23 am
Location: Asheville, NC
Contact:

Step-limited Sample & Hold

Post by pugix » Mon Nov 04, 2019 5:38 pm

I'll first describe what a Step-limited Sample & Hold is and then ask a question about it.

What I'm calling step-limited is the case where the output of a S&H is mixed together with the signal being sampled. Since the present output voltage is incorporated with the sampled signal, the result tends to stay closer to the present output, resulting a series of steps that wander more gradually, most noticeable when the input is noise. The only S&H module I know of that has this built in is the Oakley Modular Sample & Hold (of which I have two).

http://www.oakleysound.com/sample.htm

The Restriction pot on the Oakley internally fades between the input and the output. This produces more limited steps as more of the output is fed back. It's easy to patch up with a simple DC mixer. One mixer input takes the signal to sample, and another mixer input feeds back the S&H output. One way to think of this is a S&H with feedback.

Now my question: I suppose that different S&H implementations will have different acquisition times. The LF398 spec sheet states acquisition time as 25 microseconds max, when the holding capacitor is .01uf and delta V out is 10 volts. I'm wondering about what happens during the transition in the restriction scenario. Let's say that we're mixing the output so that its proportion to the input is 50%. Example: currently the output is at +4 volts and the input is +6 volts. The chip input sees +5 volts. Now the sample trigger happens. Ideally, we'd expect the resulting output to be +5V, after acquisition. But, during the acquisition the output is slewing from +4 to +5. The input is changing while it's being sampled. I can't wrap my head around what's going on. It probably doesn't matter too much in practical terms. But any effect would probably be amplified by a S&H with a much slower acquisition time. And I have a few S&H with quite slow acquisition times. I think will try patching up restriction using a Serge SSG, which has a voltage-controlled slew rate (corresponding to acquisition time?).

I'm curious if 1) people use this restriction or step-limiting technique, and 2) any thoughts about sampling part of the output each time.


Thanks,
Richard
https://www.pugix.com

"Everything in our world is actually always modulated by everything else." - Peter B

User avatar
Pelsea
Ultra Wiggler
Posts: 977
Joined: Thu Apr 19, 2018 11:46 am
Location: Santa Cruz CA
Contact:

Post by Pelsea » Mon Nov 04, 2019 6:30 pm

I use this technique a lot, but in software-- it's a drunkard's walk.

As far as the timing is concerned, yes, you would have a race condition during the transition if the output were connected directly to the input. However, in the Oakley module it is not. There is a second sample and hold which grabs the output of the first when the trigger falls. You could probably fake this up with two standard S&H, triggering the second (in the feedback path) on the inverse of the trigger.
Books and tutorials on modular synthesis at http://peterelsea.com
Patch responsibly.
pqe

User avatar
cptnal
Super Deluxe Wiggler
Posts: 3992
Joined: Wed Jun 14, 2017 2:48 am
Location: People's Republic of Scotland

Post by cptnal » Tue Nov 05, 2019 3:25 am

As far as the question goes I've no idea, sorry. But I see huge fun in store experimenting with this technique, so thanks for that. :party:
Is it finished?
Latest Tune:
Sounds: SoundCloud , Freesound
Racks: Big Case, Top Row, Funboat, Tinicase

User avatar
EATyourGUITAR
has no life
Posts: 5798
Joined: Tue Aug 31, 2010 12:24 am
Location: Providence, RI, USA

Post by EATyourGUITAR » Tue Nov 05, 2019 3:09 pm

On the LF398 you usually have a non-inverting buffer from a TL072. This is a high impedance jfet input. The only thing to worry about is cap leakage. The cap is only source or sink while sampling. That sample time can be way different with a 10uF poly film box cap. This is as big as I can go without the size being impractical. Most other people stop at 100nF. Even with the output connected to input, it is connected through a non-inverting buffer. The charge time does not change. The value at the input will change but it doesn't break anything.
WWW.EATYOURGUITAR.COM <---- MY DIY STUFF

User avatar
Synthbuilder
Super Deluxe Wiggler
Posts: 3058
Joined: Tue Mar 24, 2009 9:12 am
Location: Cumbria, UK

Post by Synthbuilder » Wed Nov 06, 2019 11:35 am

As Pelsea correctly states the current Oakley Sample/Slew module (and it's Eurorack equivalent, AJH Synth's Sample, Hold and Slew module) uses two sample and hold circuits to achieve this. However the original version of the module used only one sample and hold circuit. In this version the restriction pot acted as a blend between the input and a slewed version of the sample and hold output. The slewing behaviour of a simple RC network was used to mask the affects of any big changes during the acquisition period. It wasn't perfect but it worked reasonably well. The two S/H version is better.

Tony

User avatar
pugix
Super Deluxe Wiggler
Posts: 3572
Joined: Sat Oct 31, 2009 7:23 am
Location: Asheville, NC
Contact:

Post by pugix » Thu Nov 07, 2019 9:13 am

Synthbuilder wrote:As Pelsea correctly states the current Oakley Sample/Slew module (and it's Eurorack equivalent, AJH Synth's Sample, Hold and Slew module) uses two sample and hold circuits to achieve this. However the original version of the module used only one sample and hold circuit. In this version the restriction pot acted as a blend between the input and a slewed version of the sample and hold output. The slewing behaviour of a simple RC network was used to mask the affects of any big changes during the acquisition period. It wasn't perfect but it worked reasonably well. The two S/H version is better.

Tony
That explains my confusion. I have two of the Issue 1 Oakley Sample/Slews, in 5U.

In any case, I'm going to experiment with simple mixing (without limiting slew) of the input and output of several makes of S&H.

Thanks Tony,
Richard
https://www.pugix.com

"Everything in our world is actually always modulated by everything else." - Peter B

Post Reply

Return to “Modular Synth General Discussion”