Anyone seen this weird fault.. Seq 1 Step 3 pot not working properly, but hardware looks ok

The Moog Forum Moog Semi-Modular Analog Synthesizers
1

I have an issue with a Moog Subharmonicon, FW version 1.1.1

Sequencer 1, Step 3 pitch is not working correctly, but investigation seems to show neither the pot, the multiplexer or mux output buffer are the issue.

When step 3 pot is at any position other than fully CCW, the pitch output at step 3 is the lowest note pitch (i.e. same pitch as for other steps when their pot is fully CCW)

When step 3 pot is full CCW a higher note plays, but it is not always the same note. Each time the pot is moved fully CCW a new note is selected, which continues playing at step 3 until pot is moved again.

The issue is consistently with step 3. Steps 1,2 and 4 are working normally.

The potentiometer is working correctly, with a steadily increasing voltage between 0 and ~3.5V output at the wiper terminal as pot is turned through full range

When checking output pin 1 of multiplexer IC U43, the voltages of all 4 steps of sequencer 1 can be seen to be passing correctly through the multiplexer. The step 3 voltage can be seen changing as the pot is turned, in the same way as other steps.

There are some “spike” artifacts visible on the mux output trace and also on test point TP23 on buffer amp output. However similar artifacts are also visible on the trace for TP32 (buffer output for sequencer 2). Since sequencer 2 is working normally I assume these artifacts are normal and not the issue?

When the mux output was isolated from the buffer op-amp (TLV9064 on top side - no U designator but adjacent to R116) and an steady externally-supplied 1V level is applied to the buffer input, the buffer output is steady with no spike artifacts yet step 3 still sounds lowest pitch note though steps 1,2,4 pitches follow the applied voltage as expected.

Firmware 1.1.1 was re-applied to rule out a corrupted firmware issue. All other functions of the Subharmonicon appear to work just fine.

Is this a known issue? what can be done to resolve this?

video

seq1 mux output

2

That is unusually well isolated already - pot sweeps cleanly, mux passes the step 3 voltage, buffer output looks steady with an injected 1V, and reflashing 1.1.1 changed nothing. Given that, the analogue path looks fine and the odd behaviour is downstream in how the step value is being read/quantised, which points at something after the buffer rather than the pot. Two cheap tests that would narrow it further: swap the step 3 and step 4 pot connections (if the fault stays on step 3, it’s not the pot at all), and try a different quantise/scale setting plus a sequencer reset/init to see if step 3 has a stuck stored value. The “each full-CCW move gives a new random note” part smells like an out-of-range or noisy reading being clamped, so it may be worth checking the voltage right at the ADC/MCU pin rather than at the buffer output. Also try it with nothing patched and with only the CV out patched - I’ve seen phantom steps caused by a jack normalling contact. If it survives all that, I’d send the readings you’ve already gathered to Moog support; that write-up is far more than they usually get and should shortcut the diagnosis.

3

Hi Helen, thanks for your reply and insight! Moog tech support have been responsive so far and the ticket has been passed to level 2 support. They suggested a couple of things, such as recalibrating the oscillators, but made no difference, so waiting for next response.

I think I’ll wait to hear from them before doing more to the PCB (e.g. isolating the step pots) for risk of damaging it, but swapping them round is a good idea, although I’m pretty sure the hardware upstream of the mux/buffer is ok.

I hadn’t thought to check the sequencer 1 output via the patchbay, but that looks as expected (i.e. step 3 CV jumps from zero to “random” high values when pot is turned full CCW) and pitch is consistent with this when patched to OSC2. All other steps work just fine.

My guess is that somewhere in nonvolatile memory there might be some factory-set calibration data on the voltage range of each sequencer pot (to making them act consistently regardless of manufacturing tolerances on the pots). My guess is that this data has got corrupted somehow and I hope that there will be some undocumented button combo or sysex command to recalibrate or clear it (I can hope!!)

Do you know if there is a full schematic available for the Subharmonicon? I was unable to find one when I last looked, and I wasn’t too keen to go poking blindly around the MCU in case I broke something:)

4

Your NVM calibration theory fits the symptom set better than anything mechanical, and it would explain why recalibrating the oscillators did nothing: if the stored range for that one pot is bad, the firmware is mapping the whole sweep to the bottom of the range and then producing garbage at the extreme. Two things that would help level 2 confirm it without you opening anything further. First, log the actual numbers: pot position at a few marked points against the CV you measure at the patchbay for step 3, then the same table for step 4. A side-by-side showing step 3 flat at minimum while step 4 tracks linearly is much harder to argue with than a description. Second, try a full factory reset or init if the Subharmonicon has one documented, and note the firmware version before and after, because if a reset clears it that is direct evidence the fault lives in stored data rather than silicon. On the schematic, I have not seen a full official one published for the Subharmonicon either, which is another reason to let support drive rather than probe the MCU blind. Given the CV out at the patchbay already shows the fault, you have proved it is upstream of anything you could damage, so I would hold off on the pot swap until they answer.