1V/oct Keyboard PCBs : leaky S&H

Discussion & Documentation for Ray Wilson's synths, modules and noise boxes.

Moderators: Kent, dslocum, synthcube

Post Reply
johnstilton
Common Wiggler
Posts: 141
Joined: Tue Oct 09, 2018 10:31 am
Location: West Chicago

1V/oct Keyboard PCBs : leaky S&H

Post by johnstilton » Sun Dec 01, 2019 9:01 pm

Tonight I (partially) successfully completed my first build: added series resistor networks to the cleaned switches of this Lowery keybed and two MFOS 1V/oct keyboard PCBs.

The left side is 20 keys and the right side is 37 keys.

The sample and hold circuits are very leaky though. The output voltage starts to drift away immediately after a keystroke. Worst on the right hand circuit.

This voltage is supposed to be stored on C19 a .01uF polystyrene capacitor. I wonder if they are leaky? Also I used J310s in place of the 2N5457s for the sample switch FETs Q2 & Q3 and I wonder about this substitution now.

I made sure to rinse off the flux thoroughly around the S&H components as Ray wrote “any leakage path to the capacitor will cause it to (droop).”

I’m going to walk away feeling OK about this progress for now but I wonder what other builders of this circuit have to advise ?Image

User avatar
Bodo1967
Veteran Wiggler
Posts: 678
Joined: Wed Aug 10, 2016 2:27 pm
Location: Germany

Re: 1V/oct Keyboard PCBs : leaky S&H

Post by Bodo1967 » Tue Dec 03, 2019 3:06 am

johnstilton wrote:This voltage is supposed to be stored on C19 a .01uF polystyrene capacitor. I wonder if they are leaky?
I doubt that, tbh. These caps usually don't cause any problems and don't leak (unlike electrolytic caps). I have them in several circuits and never encountered any "strange" behavior.
johnstilton wrote:Also I used J310s in place of the 2N5457s for the sample switch FETs Q2 & Q3 and I wonder about this substitution now.
Using another FET type might at least be worth a try.

What about those red mats on the picture? Is it possible that these are not 100 % non-conducting? Personally, I'd place the PCBs freely for testing so they don't have any contact with any surface whatsoever. You could use M3 screws and nuts in each corner to give the PCB "legs" (like a table). Or just have 'em dangling in the wind from some string or wire ;).
... why buy it for $100 when you can build it yourself for $150?

johnstilton
Common Wiggler
Posts: 141
Joined: Tue Oct 09, 2018 10:31 am
Location: West Chicago

Post by johnstilton » Tue Dec 03, 2019 10:16 am

I found a thread for the same issue :

viewtopic.php?t=177166&highlight=

It doesn't look like the original poster found any resolution but the last writer did resolve their issue by installing new FETs.

I'm going to order both a polypropylene caps and the specified obsolete FETs and see if either improves the operation.

I did try it with the pcbs resiting on a pair of plastic pens.

johnstilton
Common Wiggler
Posts: 141
Joined: Tue Oct 09, 2018 10:31 am
Location: West Chicago

Re: 1V/oct Keyboard PCBs : leaky S&H

Post by johnstilton » Mon Dec 23, 2019 9:50 am

Update.

On the leakier circuit changing the FETs to 2n5457 and the cap to a wima FKP4 and installing a ground ring resulted in the same behavior. So I removed all of the S&H components and using new components I breadboarded the circuit and wired it to the PCB. On the breadboard the S&H works great.

I wonder if there is something wrong with the PCB itself. Or there is flux resin reside I cannot see or wash away with water? What do you all think?

On the less leaky board I removed the FETs and the cap and washed the board well with alcohol this time before replacing the FETs and cap with the new components and washed again with alcohol. Now this circuit has something different wrong with it. The voltage sampled crashes and runs to -vcc. I’ll have to look into that

johnstilton
Common Wiggler
Posts: 141
Joined: Tue Oct 09, 2018 10:31 am
Location: West Chicago

Re: 1V/oct Keyboard PCBs : leaky S&H

Post by johnstilton » Tue Jan 21, 2020 9:39 am

Bit of an update on this for the record:

Using a solderless breadboard, I was able to make two S&H/Buffer circuits that did not leak at all. I used new Wima polypropylene caps and new FETs.

Once working, I swapped in the components from the previous attempts on the PCBs and "most" of them worked just fine in the breadboard. One FET and one styrene cap caused a little bit of voltage droop, but nothing like when installed in the PCB.

Alarmingly, I discovered though when I installed an IC socket into the breadboard, the voltage leaked through the socket (!).

Thinking I was on to something, I re-installed new components into one of the PCBs without a socket....and it still leaked.

So, to date, the only way I can get these two builds to work properly is if I use a solderless breadboard to hold the S&H circuits. Which I have done successfully twice.

Quite the kludge.

Makes me apprehensive for the rest of my MFOS builds, specifically the S&H in my Ultimate PCB, but in general as well.

:despair:

User avatar
Bodo1967
Veteran Wiggler
Posts: 678
Joined: Wed Aug 10, 2016 2:27 pm
Location: Germany

Re: 1V/oct Keyboard PCBs : leaky S&H

Post by Bodo1967 » Fri Jan 24, 2020 4:35 am

johnstilton wrote:
Tue Jan 21, 2020 9:39 am
Alarmingly, I discovered though when I installed an IC socket into the breadboard, the voltage leaked through the socket (!).
[...]
I re-installed new components into one of the PCBs without a socket....and it still leaked.
Just a brainfart: I know it sounds weird - but perhaps try measuring the resistance between two random points on the sockets and the PCB? However, this might of course take some rather sophisticated (lab grade) measuring equipment - I doubt a common $9.99 DMM would be able to measure in such ranges...

Anything less than "infinite" might indicate a problem with non-100%-isolating plastic material or such :despair: .
... why buy it for $100 when you can build it yourself for $150?

johnstilton
Common Wiggler
Posts: 141
Joined: Tue Oct 09, 2018 10:31 am
Location: West Chicago

Re: 1V/oct Keyboard PCBs : leaky S&H

Post by johnstilton » Sat May 09, 2020 9:32 pm

I let this topic go for awhile to concentrate on building the main boards for my project. Coming back to the drooping S&H problem, I wondered if the voltage could be leaking through other sockets, like the op amps of the CV creation path. I desoldered all of the sockets in the PCB and soldered all of the ICs directly.

After cleaning the PCB thoroughly and letting it dry over a weekend I found that the voltage droop was within acceptable specs. I lost .008V in 60 seconds where Ray’s documentation was a bit more than that.

Tomorrow I am going to do the same for the other PCB and hopefully this will also work within spec and I will not have to write any further on the subject other than DONT USE SOCKETS ON IN THIS DESIGN

User avatar
synthcube
you will be assimilated
Posts: 2174
Joined: Wed Dec 21, 2011 6:42 pm
Location: Boston
Contact:

Re: 1V/oct Keyboard PCBs : leaky S&H

Post by synthcube » Sat May 23, 2020 2:36 pm

Thank you for the updates, we look forward to your results.
facebook: https://www.facebook.com/2SynthCube

synthcube webstore www.synthcube.com/cart
music from outer space: www.musicfromouterspace.com
blacet research: www.blacet.com
MOTM DIY Analog Synthesizers: www.motmsynthesizers.com
music from outer space euro smt: www.mfoseuro.com

johnstilton
Common Wiggler
Posts: 141
Joined: Tue Oct 09, 2018 10:31 am
Location: West Chicago

Re: 1V/oct Keyboard PCBs : leaky S&H

Post by johnstilton » Wed May 27, 2020 8:22 am

I removed the sockets from the second one and the leakage is now also within acceptable levels. I used the good Mil-Max sockets that Ray specified in the Ultimate's BOM with the Mouser link. I have used these before as well when I did PCB repair work at a previous job. My lesson learned is not to use sockets in this design, and maybe not in other similar S&H circuits.

Post Reply

Return to “Music From Outer Space”