I designed a basic S&H and built four of them.

[pugix]

An addition to my 5U MOTM system. But the circuit is good for any system/power supply.

https://pugix.com/synth/diy-quad-sample-hold/

[KSS]

Nice!
In reading your descriptiuon at the link I just had a moment. Why don't we put a switch to select different output resistors -or none- on modules which might be expected to preserve Voct?

We do it for AC-DC inputs. Why not outputs?
It's a game-changer thought for me. I'm going to have a re-look at all my 50SQ module designs.

Two things I'd add to your module. Two trigger level knobs so that the input trigger level for 2 of 4 here can be adjusted. It adds so much utility and variety to the results and it looks like you've got the room.

The second relates to something mentioned in the description. Selection of the holding cap. Small for fast aquisition and big for long holding time.

Why not both? Either using a switch to choose, or possibly simpler by normalling them in pairs for fast response *and* long hold. You end up with two channels when using this selectable configuration, but you also gain near permanent hold times and don't suffer in the sampling side!

The final thing I like to see in any S/H is a T/H option. It's again such a simple thing to add and brings so much to the result potential.

With 4 S/Hs here these additions could be parceled out amongst them.

Edit: I'd move all four LEDs up with an on-off-on between for L/R T-S/H and then normal vertical pairs. Knobs up top and toggles for short long below. Two of four could then be T-S chosen and all four remain as original with knobs CCW and toggles set to S.

Here's how that ends up. I'd add a third toggle betweenthe two at bottom to selectively short the output resistor for 1of2 Voct optimized output.

[pugix]

Nice!
In reading your descriptiuon at the link I just had a moment. Why don't we put a switch to select different output resistors -or none- on modules which might be expected to preserve Voct?

We do it for AC-DC inputs. Why not outputs?
It's a game-changer thought for me. I'm going to have a re-look at all my 50SQ module designs.

Two things I'd add to your module. Two trigger level knobs so that the input trigger level for 2 of 4 here can be adjusted. It adds so much utility and variety to the results and it looks like you've got the room.

The second relates to something mentioned in the description. Selection of the holding cap. Small for fast aquisition and big for long holding time.

Why not both? Either using a switch to choose, or possibly simpler by normalling them in pairs for fast response *and* long hold. You end up with two channels when using this selectable configuration, but you also gain near permanent hold times and don't suffer in the sampling side!

The final thing I like to see in any S/H is a T/H option. It's again such a simple thing to add and brings so much to the result potential.

With 4 S/Hs here these additions could be parceled out amongst them.

Edit: I'd move all four LEDs up with an on-off-on between for L/R T-S/H and then normal vertical pairs. Knobs up top and toggles for short long below. Two of four could then be T-S chosen and all four remain as original with knobs CCW and toggles set to S.

Feature creep! Ha ha.

As for trigger level, since the trigger circuit can take a sine or triangle, if those are biased up or down you get triggering at different points on the wave. I don't know of any S&H that have adjustable trigger level. Wait, I do! Bastl Tromso has a comparator with a CV input for threshold. The comparator then triggers the S&H.

Yes, fast acquisition time vs hold time is a tradeoff. Someone here did a nice design that uses two cascaded LF398 chips, the first for fast acquisition and the second for long hold. All of the S&H I have in any format just use one LF398. I don't worry much about droop. To me it's a feature!

I don't use Track & Hold very much, but if I want it I have three other S&H in MOTM format that do it.

Thanks for the ideas!

[cartertools]

Love the use of the inexpensive 2LFOSH Barton boards, I have two of them in my system but was thinking of building something just using the LF398 as it works so well and I often bypass the LFOs for my own trigger and voltage inputs anyway.

An addition to my 5U MOTM system. But the circuit is good for any system/power supply.
https://pugix.com/synth/diy-quad-sample-hold/

[KSS]

Added kludged panel pics above.

[KSS]

Bastl Tromso has a comparator with a CV input for threshold. The comparator then triggers the S&H.

It's a powerful addition. You're doing with an external offset-bias.
As with any mod, it has to work for the way you use your setup! The knob or knobs could be replaced with jack which brings its own utility in that now triggers will have to 'match' -be slightly above- this input. It lets the jack punch above its weight class. Booleans and other interactions become inherent. For your style use it could be a better alternative than knobs. Or do one of each. That's the beauty of your module, there are four S/H instances!
Same for the lower half. One toggle could choose output impedance and the other for short long hold. I don't find three toggles in 1U hard to use or look at, but two could still provide a positive expansion

Thanks for the ideas!

[Huba-Swift]

Have never used an LM398. Is changing it to T/H as simple as holding pin 8 high?

Might do a board layout and make a version for myself as I could use some Sample and Holds.

[KSS]

Have never used an LM398. Is changing it to T/H as simple as holding pin 8 high?

Yes. But I'd choose to call that a hold function.

For T-S/H it's often easier to short the cap in that part of the circuit.

Might do a board layout and make a version for myself as I could use some Sample and Holds.

398 is a really good chip. I had plans to use one in some work recently. You want to be careful about the speed of the pin 8 transitions, it's not always kindly to slow moves there. See also Rene's Schmitzbits YASH.

[Huba-Swift]

Have never used an LM398. Is changing it to T/H as simple as holding pin 8 high?

Yes. But I'd choose to call that a hold function.

For T-S/H it's often easier to short the cap in that part of the circuit.

Might do a board layout and make a version for myself as I could use some Sample and Holds.

398 is a really good chip. I had plans to use one in some work recently. You want to be careful about the speed of the pin 8 transitions, it's not always kindly to slow moves there. See also Rene's Schmitzbits YASH.

Thanks for the information KSS. If I do make a version I'll be sure to post PCB and panel info for anyone wanting to build one.

[Huba-Swift]

Here's my mockup panel for what I'm thinking so far. Would be built with two LF398 per to get the best of both slow and fast signals. Switches and jacks might be a bit tight for my own tastes, but should be workable.

-Four identical S/H circuits on one panel
-Noise normalled to all inputs
-When switch set to int (short for internal osc), the knob is used as a rate control for an internal oscillator normalled to Trig In
-When switch set to ext, knob functions as a threshold value for the external trigger signal
-Trigger input also functions as a square wave LFO output when set to Internal oscillator, meaning the module can also be used as a quad square LFO

[nigel]

Must be something in the air, I built a dual S/H just the other day. (4 HP Eurorack.) Just a bare LF398 with a two transistor gate conditioner circuit (as used in the Kassutronics envelope/slope generator).

[Huba-Swift]

Here's what I've come up with so far. I've managed to get a decent enough single Op-Amp VCO working which I'm happy with. Should be able to get the quad module going with 2x LM324 and 4x LF398, which should all be able to fit nicely behind the module without having to mount perpendicular.

I went with just one LF398 per S/H and just use a pin header on the PCB to select between caps for fast/slow acquisition. There'll be 4 of these on one PCB, so I figured best to keep parts count low.

Cap values for LF398 are placeholders for now, I'm not sure what exact values would be most useful.

Tomorrow I make the noise circuit and test the oscillator alongside the comparator. I've never made a comparator circuit with only single supply rails, so I wouldn't be surprised if I have to make changes to get that working properly. I Don't have any LF398 on hand yet so testing that will have to wait.

[atomd]

Why don't you put capacitive load compensation output to keep zero output resistance and protection at once?

[Huba-Swift]

Why don't you put capacitive load compensation output to keep zero output resistance and protection at once?

First time hearing of this. What would that look like? A quick search for capacitive load compensation doesn't seem to yield what you appear to be talking about.

[atomd]

https://www.analog.com/en/analog-dialog ... er-25.html

Check the in loop compensation

[Huba-Swift]

https://www.analog.com/en/analog-dialog ... er-25.html

Check the in loop compensation

Ohh, That makes sense. Here I thought you were just saying to put a cap in series with the output instead of the 1k resistor. Whcih didn't make sense to me.

I'm pretty sure I've seen CGS do what's shown here in the article with some of his designs. I'll include this as well.