mki x es.EDU Noise/S&H

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!
Post Reply
User avatar
Girts23
Wiggling with Experience
Posts: 399
Joined: Mon May 20, 2013 2:22 am

mki x es.EDU Noise/S&H

Post by Girts23 »

Hi, guys!
Here's a next move in mki x es.EDU series - Noise/S&H module!
A sample & hold-circuit can often be the secret sauce that takes your patch from decent to distinctive whether you use it for self-sequencing, for adding a dash of randomness to your filter movements or accents in VCA. It's super simple JFET-based implementation – and then upgraded it with a couple of extra features like an internal clock generator or a dedicated slew output. An integrated noise generator has pink and blue noise outputs in addition to the standard white noise.
A full kit available from Erica Synths shop and at best DIY retailers!
And the demo from one and only Synth DIY Guy!
Enjoy!
Girts

ciso
Common Wiggler
Posts: 58
Joined: Thu Jun 30, 2016 10:36 am

Re: mki x es.EDU Noise/S&H

Post by ciso »

Anyone build this yet? It's next on my list after I finish my 2nd mki x es.EDU sequencer.
ciso
Common Wiggler
Posts: 58
Joined: Thu Jun 30, 2016 10:36 am

Re: mki x es.EDU Noise/S&H

Post by ciso »

So I built this today, and I'm experiencing a strange issue.

When I apply a trigger to the Clock In, the S&H samples the noise, but instead of holding, or even drooping, the voltage slowly rises.

I've reflowed all the joints in the upper left (upper right from the jack side) part of the pcb, no solder bridges, and it's still happening.

I replaced R27 as I thought it might be the culprit due to the resistor sitting so close to the board, I thought maybe the solder didn't reach both sides of the hole in the pcb, but that didn't change anything either. I also replaced the TL074 with anther I had on hand, no change. I"m at a loss as to why the voltage would slowly increase at all

Here's a link to the manual/build guide, and the schematic is on page 60
https://www.ericasynths.lv/media/SH_MANUAL_v1.pdf

Additional detail, the S&H out seems to linearly rise to 10.9v

I'm measuring -9.7v at TP5 (E) and 2.8v at TP4 (D).
ciso
Common Wiggler
Posts: 58
Joined: Thu Jun 30, 2016 10:36 am

Re: mki x es.EDU Noise/S&H

Post by ciso »

Update, I figured this out.

My kit came with a TL072CN for the slew circuit op amps. I swapped it out for a TL072CP that I had, and the problem is resolved.

I do not know if that means the CN is inappropriate and the kit came with the wrong op amp chip, or if the TL072CN that came with my kit is just faulty. I will email Erica Synths to determine if it was the wrong chip, but looking at the coding it seems as though they are roughly equivalent.

Edit: it seems the N vs P designation refers to the material the housing is made of, so it's likely I just had a bad chip.
Whatthevolt
Learning to Wiggle
Posts: 32
Joined: Fri Feb 18, 2022 3:07 pm

Re: mki x es.EDU Noise/S&H

Post by Whatthevolt »

Having trouble with this one. There’s no board layout in the manual, and the resistors below the right TL074.. the screen print makes it hard to tell what goes where.!

For instance, top to bottom of the left column, is it 33k, 1m, 470km 10k? Leaving the right pair, top to bottom, to be 100k and 20k?

Edit: in fact a lot are mega hard to identify, hah.

Thanks..
User avatar
KSS
Super Deluxe Wiggler
Posts: 10885
Joined: Mon Jan 25, 2016 7:28 am

Re: mki x es.EDU Noise/S&H

Post by KSS »

Whatthevolt wrote: Sat Sep 03, 2022 2:15 pm There’s no board layout in the manual, and the resistors below the right TL074.. the screen print makes it hard to tell what goes where.!

For instance, top to bottom of the left column, is it 33k, 1m, 470k 10k? Leaving the right pair, top to bottom, to be 100k and 20k?

Edit: in fact a lot are mega hard to identify, hah.
It looks like they published a second version of the manual separate fab PDF. It does include 2 board layouts. One with refdes and the other with component values.

The Board is marked EDUSH2. <--Does that match yours?
Label is upper left on panel PCB and next to power input on the main PCB.

Your listed values and locations match what this version of the build manual says.

| 074
|________
33k
1M 100k
470k 20k
10k

Edit: attached EDUSH2-Fab.PDF and corrected reply text.
Attachments
Erica-EDU-SH2-Fab.pdf
(264.04 KiB) Downloaded 32 times
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.
depthbuffer
Learning to Wiggle
Posts: 37
Joined: Thu Jan 06, 2022 11:54 am

Re: mki x es.EDU Noise/S&H

Post by depthbuffer »

ciso wrote: Sat Aug 27, 2022 9:32 pm So I built this today, and I'm experiencing a strange issue.

When I apply a trigger to the Clock In, the S&H samples the noise, but instead of holding, or even drooping, the voltage slowly rises.

I've reflowed all the joints in the upper left (upper right from the jack side) part of the pcb, no solder bridges, and it's still happening.

I replaced R27 as I thought it might be the culprit due to the resistor sitting so close to the board, I thought maybe the solder didn't reach both sides of the hole in the pcb, but that didn't change anything either. I also replaced the TL074 with anther I had on hand, no change. I"m at a loss as to why the voltage would slowly increase at all

Here's a link to the manual/build guide, and the schematic is on page 60
https://www.ericasynths.lv/media/SH_MANUAL_v1.pdf

Additional detail, the S&H out seems to linearly rise to 10.9v

I'm measuring -9.7v at TP5 (E) and 2.8v at TP4 (D).
Did you buy your kit from Thonk by any chance? Just finished building mine, and it has the same (or at least similar) problem; I haven't measured the voltage, but there's an audible pitch bend upwards instead of a constant held voltage. Going to try the same solution once some TL072CPs arrive from CPC.
User avatar
KSS
Super Deluxe Wiggler
Posts: 10885
Joined: Mon Jan 25, 2016 7:28 am

Re: mki x es.EDU Noise/S&H

Post by KSS »

depthbuffer wrote: Tue Jan 24, 2023 1:10 pm
ciso wrote: Sat Aug 27, 2022 9:32 pm So I built this today, and I'm experiencing a strange issue.
When I apply a trigger to the Clock In, the S&H samples the noise, but instead of holding, or even drooping, the voltage slowly rises.
I've reflowed all the joints in the upper left (upper right from the jack side) part of the pcb, no solder bridges, and it's still happening.

I replaced R27 as I thought it might be the culprit due to the resistor sitting so close to the board, I thought maybe the solder didn't reach both sides of the hole in the pcb, but that didn't change anything either. I also replaced the TL074 with anther I had on hand, no change. I"m at a loss as to why the voltage would slowly increase at all

Here's a link to the manual/build guide, and the schematic is on page 60
https://www.ericasynths.lv/media/SH_MANUAL_v1.pdf

Additional detail, the S&H out seems to linearly rise to 10.9v

I'm measuring -9.7v at TP5 (E) and 2.8v at TP4 (D).
Did you buy your kit from Thonk by any chance? Just finished building mine, and it has the same (or at least similar) problem; I haven't measured the voltage, but there's an audible pitch bend upwards instead of a constant held voltage. Going to try the same solution once some TL072CPs arrive from CPC.
On page 9 of the manual you see -in the 2nd drawing from top- a voltage rising result.

This is during the charging of the hold cap.

If the sampling time is short enough, and if the capacitor is able to fully charge in this time, then you shouldn't be hearing this.

That you are, suggest that there may be misplaced or mis-valued caps in the build. For example if the capacitor in the sampling -JFET Gate- part of the circuit is too big then the time that the JFET is on will be longer than it needs to be and this means you have a track and hold instead of a sample/hold. <--Also a clue for a useful mod of the module.

Page 14,15 of the manual get into this a little.

You'd also want to use a 'timing' style capacitor. PolyStyrene is the classic choice. Large silvery plastic film looking thing. Other types of cap will have different soakage and this could relate to the results seen. A mylar or polyester cap is not a good choice for an S/H. A disc cap even worse.

Check that the caps used are placed correctly and verify their values in each position.

One last obvious thing is if the slew-lag is engaged. Ensire that's not happening.
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.
depthbuffer
Learning to Wiggle
Posts: 37
Joined: Thu Jan 06, 2022 11:54 am

Re: mki x es.EDU Noise/S&H

Post by depthbuffer »

Thanks for the quick & detailed response!

I'm taking from the direct S&H out, not the slew out. FWIW the slew out seems fine - from the S&H there's an audible jump at each clock pulse, then a "stabilization period" which is the problem I'm referring to, whereas the slew out gives a much more smoothed-sounding output and does respond to the slew pot. With the slew pot fully CCW it behaves identically to the S&H out. The three noise outs sound OK and the clock also seems fine.

The capacitor you're referring to - is that C21, 100nF? The one that came with my kit looks suspiciously like it might be polyester - I don't have it in front of me to take a pic but this is the closest visual match I could find: https://www.rapidonline.com/kemet-r82ec ... or-10-5826

I think it's the "correct" one in the sense that it's clearly the right physical dimensions for the layout, and matches the BoM, but reading your post and the manual side by side, it does sound like you're on to something! I'll double check my soldering in that area. If I do end up wanting to replace it, I'm struggling to find 100nF polystyrene caps online; the closest I can find is a tenth of the capacity! Would something like this be worth trying?

https://www.soundtronics.co.uk/10nf-50v ... citor.html
User avatar
KSS
Super Deluxe Wiggler
Posts: 10885
Joined: Mon Jan 25, 2016 7:28 am

Re: mki x es.EDU Noise/S&H

Post by KSS »

C12 -and R22- set the length of the pulse which turns on and off the JFET. Page 22 of the manual talks about this possible reason to have rising voltage. C12 cap type is not critical. <--shorting C12 with a switch gives you a T/H S/H option.

C21 is the hold cap, and the box type used there can be the right or the wrong 'film' type. It is best to look up the actual part number to see what type of film is in this kind of 'box' cap. Yes, usually if they're white they are mylar-polyester. 'Better' films are often red or yellow. <--But you can't just use the color , you need to look it up. A good modern box cap of a PS or PPS type will be fine here too.
Possibly a PP type too. <--These are often much larger values than used here.

With all this said, you have to respect that these are not intended to be master modules! They are meant to be 'good enough' functionally and inexpensive learning tools. The manual discusses this too when it talks about sampling accuracy in bold type on page 17.

The cap you linked *is* the preferred PS PolyStyrene type. But you also see it's big and may not fit well between the JFET and the '074. That's one of the downsides of this type cap. At least when you're trying to fit them into an existing PCB layout. But sometimes you can simply put a part on the other side. Not ICs! But diodes, resistors, caps and transistors all can be mounted to the other side of the PCB to make space <--as long as you make sure to keep polarity and which leg goes in which hole aligned properly. So if you put the JFET on the opposite side you'd have room for that PS cap. You might not even need to do that, but it's good to have a backup plan!

In *any* S/H you are dealing with possible leakage. You *must* be absolutely *sure* the PCB and traces which connect the S/H core are *clean*. Not just 'looks good' clean but *really* clean.

One of the biggest differences between how a beginner troubleshoots compared to a more seasoned builder is that they will start trying new parts. Not taking enough time to *really* check the placement and values of inserted and soldered components. Not cleaning he PCB thoroughly.

Don't be too quick to want to change parts. We can assume that the ones which came with the kit will work at least well enough to meet the stated goal for the module. Even that polyester cap, if it is polyester.

One part which *does* have built-in variability -but more importantly- heat and static sensitivity is the JFET. It's not hard to mess one of these up in soldering. Since it's at the core of the module's primary function here, it should get some special attention. First by cleaning the PCB well. Both sides.

Then by trying to sample something besides wildly variable noise. Choose a slow moving ramp or triangle. Even DC. <--Feed the same CV you're using on a VCO to the S/H input. And feed the S/H output back to the same or another VCO. If the same, then the original CV and the sampled CV will add together. If the original CV is 1V, then the pair -original and S/H output should raise the VCO one octave. If you use a second VCO -and they are both calibrated- then they should both be able to sound the same pitch. Line up the starting pitch of both first by zero-beating.
depthbuffer
Learning to Wiggle
Posts: 37
Joined: Thu Jan 06, 2022 11:54 am

Re: mki x es.EDU Noise/S&H

Post by depthbuffer »

Thanks again, all this advice is absolute gold. I haven't spent much time on testing yet as I finished assembly late yesterday evening, then came straight to this thread and immediately found someone talking about what seemed like the same problem - and at the moment I'm buying a lot of components anyway, trying to build up to having a small supply of useful things on hand for my own repairs/experimentation, so yes I have been very quick to jump to the idea of replacement opamps & capacitors!

Definitely need to test with external inputs, I like the idea of comparing two oscillators. Will set that up when I next have time - and then my first port of call needs to be inspecting the area around C21 and the JFET, as it is almost 100% going to end up being something I've done, either a poor joint or heat damage. Will also check the clock pulse is coming through clean to the gate. As an electronics noob I don't think I have (for example) any J113s lying around, and it makes sense to stuff baskets with other useful things at the same time, so if I don't fix it by poking with a soldering iron I'll probably buy multiples of both replacement caps and JFETs. (And if replacing parts does fix it, it was probably because I broke them in the first place.)

My bar for functionality is not "perfection", but currently, my module's hold performs obviously, audibly worse than the example in the overview video earlier in the thread.
depthbuffer
Learning to Wiggle
Posts: 37
Joined: Thu Jan 06, 2022 11:54 am

Re: mki x es.EDU Noise/S&H

Post by depthbuffer »

Well, I'll have to stop for the night, but I'm a litle further along. Only a little, though.
  • I noticed the tuning wasn't stable on my EDU VCO module, found & fixed a loose capacitor on it, and recalibrated it. Unrelated, but a win, so I'll take it.
  • It's not DA2 (the left-hand TL074CN when looking at the component side of the board). I removed it so as not to damage it whilst resoldering a few things, and replaced it with a new one as I have a few spare, and it made no difference.
  • It's definitely a pitch bend up. If I run the same voltage through two oscillators and tune them, when I patch in the S&H into one of them instead, it begins each interval at the correct pitch, then slowly & steadily rises a quarter tone or so, and goes back in tune on the next clock pulse. (In fact, if I turn the "rate" knob fully CW, running the clock so fast that it should basically act as a passthrough, the tuning is correct and stable.)
  • If I plug in an external clock (5V pulse), the pitch just keeps rising; it drops slightly on each clock pulse, but not all the way to the input voltage.
  • Attempting to measure the voltage at test point E, I've noticed that if I short it to ground, the pitch again just keeps rising, regardless of internal or external clock pulses.
I'm beginning to wonder if it's nothing to do with the S&H part of the circuit itself, and more to do with the clock generator. The trouble is, there are several things I'm not sure of. I might be able to figure it out if I stare hard enough at the diagram and rack my brain, but I'm not sure exactly what I *should* be seeing at each of the test points under correct operation, and I'm also not sure whether the module is *expected* to work with a 0V low/5V high clock pulse, or only something more closely resembling its built-in -12V/+12V. (I think it *should* work fine, because looking more closely, it looks like the -12V/+12V signals are generated *after* the clock input jack, so I guess it's designed to generate those voltages in response to smaller input pulses? Need to study more.)
depthbuffer
Learning to Wiggle
Posts: 37
Joined: Thu Jan 06, 2022 11:54 am

Re: mki x es.EDU Noise/S&H

Post by depthbuffer »

The problem WAS exactly the same as ciso's! Replacing the included TL072CN with a brand-new TL072CP fixed it.

I probed various points on my board, and figured the clock generator was good, the gate-to-trigger converter (rectifier & high-pass filter - manual page 22) was good, and the comparator was good; I was getting a nice strong ~22Vpp impulse, with highs & lows around +/-11V, probing the stripe end of VD5. I also probed pins 8 & 14 of IC DA2, i.e. the outputs of the opamp buffers just before and after the S&H core itself; the input matched the CV input exactly, but the output rose & fell matching the audible pitch bending. And it was definitely rising to *higher* than the input voltage, meaning the hold capacitor was getting extra charge coming in from somewhere - and the only place it can come from is up from the slew circuit.

However, note that when using an external clock, it seems it really needs a good strong pulse - a simple 0-5V clock doesn't seem to suffice. What happens is (I think) the JFET doesn't get held open long enough during the sampling phase when the voltage coming into the trigger generator portion isn't sufficiently high; it results in the output voltage taking multiple clock pulses to converge on the input value. This is despite the fact that the clock pulse - whether internal or external - is going into a gate-to-trigger converter & comparator, which should mean any external clock with highs above 3V should work. I can get the same effect with the internal clock, though, when creating large enough swings in input voltage - it is clearly audible at 3 or 4 octave jumps. So at this point I think we're getting into the territory KSS was talking about, with these not being "master module" designs; this is just how the circuit behaves.

I *might* experiment with swapping the hold capacitor for one of those polystyrene caps, but at this point I'm pretty happy; the module is working as designed, I'm just butting up against design limitations - limitations you likely won't notice when sampling random voltages, or quantising the output, both things I'd likely do if using this for generative sequencing.
User avatar
KSS
Super Deluxe Wiggler
Posts: 10885
Joined: Mon Jan 25, 2016 7:28 am

Re: mki x es.EDU Noise/S&H

Post by KSS »

:tu:
I hope both of you let the supplier know about the potentially* faulty OPAs being sent out with kits!

* :hihi:
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.
Post Reply

Return to “Music Tech DIY”