Weird clock lock state on certain range selections

[psioniclabs]

Hello! I'm troubleshooting a strange Klee problem and hoping for a little guidance. This is a new (many times delayed) build and I'm going through the Bring-Up Procedure. I had a couple of issues I've resolved so far, but am kind of stumped on this one. I discovered it on the Testing Voltages section when selecting a different range value.

Firstly, particular to my build I'm running off 12V and have analog and digital grounds tied together. I don't think this could account for what I'm seeing, but who knows.

The Problem:
On Range switches 2, 4, 5 and 6 I'm getting a weird locked state when I reach pattern step 11. It does not respond to clock pulses from either an LFO or a different clock source I've been using. A manual step will get it past the locked step until it reaches the locked state again. This problem does not occur with range switches 1,3,7,8. Clock steps gallop along just fine on those.

For range steps 4 and 6 I can set only pattern step 12, load the pattern and it will run all the way back to 11 and lock. If I set both 12 and 13, once the initial pattern switch 13 reaches step 11 it locks.

Range steps 2 and 5 exhibit a little different behavior. With a single pattern switch enabled, it's the same problem as 4 and 6. But if more than one pattern switch is enabled, it will lock after the first clock pulse once the pattern loads.

The Gates, Triggers and CV Outs all have stuck voltages associated with the locked states.

For Range Switch 2 and 5:
-Master Gate and Bus 2 Gate are 0V
-Bus 1 and 3 Gate are -5V
-All Triggers are -5V
-CV A and B are 1V
-CV A+B are 2V

For Range Switch 4 and 6:
-Master Gate and Bus 2 Gate are 2.5V
-Bus 1 and 3 Gate are -2.8V
-All Triggers are -2.8V
-CV A and B are 7.8V
-CV A+B are 8.7V

I've gone over the analogue board a couple of times so far and haven't found any solder bridges. I've hit solder joints for all the diodes to make sure I didn't have a cold joint. And I've done a sanity check on the switch cable. Oh and I've scoured over all panel mount components associated with step 11 and all continuity checks are fine. It just seems very bizarre that it locks particularly on step 11 only on those range switches.

Any clues as to what to look for next? Thanks in advance!

[psioniclabs]

Just an update on my testing. I've checked for shorts around U3 and U7 and haven't found any. I replaced U7 4051 and U3 4066. Still having the same problem, so I'm still open to suggestions.

[psioniclabs]

Aha!

While attempting to diagnose this with my scope I discovered something. When I ground the scope to the analog board this condition disappears! Looks like I need to reexamine my grounding. I'm not sure what needs to happen exactly because I've never had a problem before.

[EdisonRex]

Hi,

Out of curiosity how are you tying analog and digital ground? It should be via W1. Otherwise it sounds like you're on the right track.

[psioniclabs]

Yes I'm tying analog and digital ground via W1. Further testing while coupling the scopes ground shows that I'm not getting any CV from pattern pots 9-16. I suspect I may have some backwards panel wiring on those pattern steps creating this condition. I'll be back at it later today so I'll find out then.

[psioniclabs]

ha. So I'm a little bit closer, but not quite there...

Poking around I found that the pot for pattern 11 was shorting between all 3 pins. I figured it was bad, but after removing it everything checked fine. Must have been a short trace of some sort. Once I get it working, I hope to show off why it's pc mount

Anyway, with that fixed I no longer have the locked clock state on pattern switch 4 and 6! But wait a freaking minute, pattern switch 2 and 5 still lock immediately when more than 1 pattern switch is active Problem still goes away when I couple my scope's ground... I believe this coupling just circumvents the problem and there's no actual issue with ground. So I'm now back on the hunt for the remainder of this bizarre issue. It's gotta be some short somewhere.

[EdisonRex]

Panel wiring, no doubt. Beware of opens as much as shorts, as much of the wiring ties high or low, and hates to float.

[psioniclabs]

Yep! Honestly I'm surprised this worked at all. I was finding problems at each step. All due to my errors of course.

I missed connecting all my pattern switches and pots from the other grounds on my front panel PCB design. A quick kludge fixed that. Then I discovered some more issues with pinched wires on that tricky pot 11, grounding it to the panel. And then I discovered I got a little too heavy with the heat gun shrink tubing my patten switch connections.. I had melted through the shielding and fused together 3 and 6, causing some weirdness. And of course I got the glide pots backwards.

Persistence pays! Now to play around and get some video of it in action.

[EdisonRex]

That all doesn't sound too unusual. Although I'm intrigued by the "front panel PCB" remark - Maybe you could post a picture of it. In any case, congrats and happy Klee-ing!

[psioniclabs]

Sure thing. The clear shaft potentiometers I experimented with necessitated a PCB to mount the LEDs. Worked out very well I think. I squished it all into a fairly compact panel with this approach. My only regret is that I didn't leave space for a external transpose jack. But maybe..

[EdisonRex]

Impressive! I like the panel too. Very nice work.

[State Machine]

Nice, a Steam Punked Klee !!!!

Nice work.

Bill

[psioniclabs]

Thanks! I'm not sure if steam punk is entirely accurate though. Is there such a thing as space punk?..

[State Machine]