Memorymoog "DEAD OSC" on all voice cards OSC3

In a Moog Mood? Here's a forum for discussion of general Moog topics.
whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Thu Nov 26, 2020 6:39 pm

I'm looking at the common analog. Here are the bullet points to the problem just so I don't ramble:

-OSC3 low on all VCs; turning up OSC3 mixer section produces a squeal which leads into the note. Sometimes the OSC3 sounds like it's dying a slow death. Autotune will read the hex data from OSC1 and 2 but all OSC3 show dead osc.
-OSC3 on the remaining cards will work if S12 is disconnected from Voice card #5. Mixer control will work, but autotune will not read any of the oscillators (all dhow dead osc)
-With S12 disconnected, OSC3 will sound at full volume if 2/3 KB TRACK is enabled, but mixer control will not work.
-Last night I replaced a bunch of 4017s and 4051 chips in the DMUX board for the hell of it which didn't change or affect anything.

*Keep in mind I've done the autotune upgrade on this so I have changed things and re-routed connections per the instructions. I'm pretty sure the problem only surfaced after I did the upgrade, not when I re-capped it. But I'm not 100% on that, unfortunately. I did check every board back in the synth after re-capping but I was mainly looking for 'is it broken or working'.

So Marky, you think this is at the common analog board? It's so weird! It was suggested to me in a FB group to check the voltages at the S&H on the DMUX board and I scoped a bunch of places that all gave me exactly what the schematic showed.
memorymoog #1257 | minimoog model d reissue

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Thu Nov 26, 2020 9:08 pm

Ok, I pulled out U13 3360 from voice card E instead of the ribbon and I get volume from the others. So I'm now thinking I have either a bad 3360 or something around it.
memorymoog #1257 | minimoog model d reissue

Markyboard
Posts: 354
Joined: Wed Jul 30, 2003 6:46 pm
Location: Springfield, VA

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by Markyboard » Fri Nov 27, 2020 11:54 am

Trying to avoid some more misunderstandings so please clarify the following:

whinylittlerunt wrote: Thu Nov 26, 2020 6:39 pm
-OSC3 low on all VCs (I assume you mean low output level?) .; turning up OSC3 mixer section produces a squeal which leads into the note. Sometimes the OSC3 sounds like it's dying a slow death. Autotune will read the hex data from OSC1 and 2 but all OSC3 show dead osc.

I assume you mean the C7 routine. Autotune is a separate routine that reads those values and provide correction voltages. I’m guessing running the actual autotune shows “ 0 tuned” when complete?

-OSC3 on the remaining cards will work if S12 is disconnected from Voice card #5. Mixer control will work, but autotune will not read any of the oscillators (all dhow dead osc)

Does Osc 1 and 2 also show dead Osc for all voice cards with S12 disconnected? Or did you just mean all Osc 3s?


-With S12 disconnected, OSC3 will sound at full volume if 2/3 KB TRACK is enabled, but mixer control will not work.

Check your +15Vand -7.5V with S12 connected and disconnected


-Last night I replaced a bunch of 4017s and 4051 chips in the DMUX board for the hell of it which didn't change or affect anything.

*Keep in mind I've done the autotune upgrade on this so I have changed things and re-routed connections per the instructions. I'm pretty sure the problem only surfaced after I did the upgrade, not when I re-capped it. But I'm not 100% on that, unfortunately. I did check every board back in the synth after re-capping but I was mainly looking for 'is it broken or working'.

Yeah, not familiar with the specifics of the upgrade as I had mine done very early on and then my whole unit was replaced in 1983 or 84.

So Marky, you think this is at the common analog board? It's so weird! It was suggested to me in a FB group to check the voltages at the S&H on the DMUX board and I scoped a bunch of places that all gave me exactly what the schematic showed

Well the DMUX s&h was my original thought when it sounded like an output level issue. But I think there’s more going on. Need more info before I can draw further conclusions/ suggestions.

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Fri Nov 27, 2020 4:50 pm

Image

My luck keeps getting worse and I'm not rushing, so I'm not sure what I keep doing wrong here...

I'll start with: "I don't know when that blob of solder dropped on the traces"; it could have been earlier on, it could have been today when I was tinning a new ground wire. I just don't know and don't have any pictures to reference from before I pulled the board. At any rate, I removed that blob and my problem didn't change so. But I did find it before putting the board back in at any rate, so I may have just avoided a huge catastrophe...

But I came to this when I was measuring voltages on the DMUX again earlier and found that at pin 10 of U3 (7417) which by the schematic looks to go back toward the OSC3 amount, I was reading 5v with a drop anytime keys were played, and it's supposed to read 15v. I realized that the 7417 was in that spot right next to the huge gob of solder for the ground wire, so I decided to pull the board and clean that up, replace the socket and chip and see where I'm left off. After I did this, I noticed the blob of solder I dropped. So, it may have been when I was tinning the ground. Either way that didn't change anything and I replaced the surrounding 7417s just to be sure I didn't short anything... no change.

To answer the questions:
I assume you mean low output level?. Yes, with S12 connector connected on Voice5, all OSC3s are extremely low. Pressing KB TRACK 2/3 brings the volume back to life but mixer control does not work.

I assume you mean the C7 routine. Autotune is a separate routine that reads those values and provide correction voltages. I’m guessing running the actual autotune shows “ 0 tuned” when complete? Yes, sorry, when I run C7 and check the hex data for each voice/osc, if S12 is disconnected from Voice 5, all oscillators will read DEAD OSC. If I plug in S12 on Voice 5, and run C7 on any given voice/osc, it will read the hex data on OSC1 and OSC2 but all OSC3s will show DEAD OSC.

Check your +15V and -7.5V with S12 connected and disconnected I did and it didn't seem to make any difference. Everywhere so far I've checked on the DMUX is giving me good, correct voltages.

As a test I plugged in 2 extension sockets to voices 5 and 6 so I could get my scope on it. I tested the good voice card, voice 6, on pin 3 of S12 which is the 2/3 KB TRACK. When S12 is disconnected from Voice 5, pin 3 on S12 voice card 6 is high, 10v. When S12 is connected, pin 3 shows no activity on the scope and reads about 1.2V, and will only go high when I physically turn on 2/3 KB TRACK.

That's where I'm at currently...
memorymoog #1257 | minimoog model d reissue

Markyboard
Posts: 354
Joined: Wed Jul 30, 2003 6:46 pm
Location: Springfield, VA

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by Markyboard » Fri Nov 27, 2020 7:01 pm

My blobs will put your’s to shame. Good you spotted it but I wouldn’t worry about it anymore-bleep happens.

whinylittlerunt wrote: Fri Nov 27, 2020 4:50 pm
But I came to this when I was measuring voltages on the DMUX again earlier and found that at pin 10 of U3 (7417) which by the schematic looks to go back toward the OSC3 amount, I was reading 5v with a drop anytime keys were played, and it's supposed to read 15v.

You lost me here. U3 is the DAC. Tried to figure out which part you may have meant instead but no luck. Unless our schematics are different?


To answer the questions:
I assume you mean low output level?. Yes, with S12 connector connected on Voice5, all OSC3s are extremely low. Pressing KB TRACK 2/3 brings the volume back to life but mixer control does not work.

I assume you mean the C7 routine. Autotune is a separate routine that reads those values and provide correction voltages. I’m guessing running the actual autotune shows “ 0 tuned” when complete? Yes, sorry, when I run C7 and check the hex data for each voice/osc, if S12 is disconnected from Voice 5, all oscillators will read DEAD OSC. If I plug in S12 on Voice 5, and run C7 on any given voice/osc, it will read the hex data on OSC1 and OSC2 but all OSC3s will show DEAD OSC.

Looking at the DMUX S54 carries all the voice mod signals, the Osc 1,2,3 amounts and the keyboard tracking amounts. All of these seem to be some how involved in what you’re observing. My first thought when I listen to your videos was it sounded like voice modulation. I can’t remember if the signals go from the DMUX to VC6 first or VC1 first. But when you remove the S12 connector it’s affecting/interrupting the Osc 1 and 2 and 3 amounts. That’s why Dead Osc also appears on Osc 1 and 2.

This really smells of a bad connection on that ribbon cable. Maybe even a short. Not sure if you can reverse the cable as it probably won’t reach all the boards. But maybe try swapping board 1 and 5. See if that reveals anything new.


Check your +15V and -7.5V with S12 connected and disconnected I did and it didn't seem to make any difference. Everywhere so far I've checked on the DMUX is giving me good, correct voltages.

As a test I plugged in 2 extension sockets to voices 5 and 6 so I could get my scope on it. I tested the good voice card, voice 6, on pin 3 of S12 which is the 2/3 KB TRACK. When S12 is disconnected from Voice 5, pin 3 on S12 voice card 6 is high, 10v. When S12 is connected, pin 3 shows no activity on the scope and reads about 1.2V, and will only go high when I physically turn on 2/3 KB TRACK.

I know what you’re saying but again your pin outs don’t seem to match my schematics - at least the ones I have here in my iPad.


That's where I'm at currently...
Last edited by Markyboard on Sat Nov 28, 2020 8:25 am, edited 4 times in total.

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Fri Nov 27, 2020 7:14 pm

This really smells of a bad connection on that ribbon cable. Maybe even a short. Not sure if you can reverse the cable as it probably won’t reach all the boards. But maybe try swapping board 1 and 5. See if that reveals anything new.
Yup, already ahead of you - I just pulled all the cards again so I could re-inspect voice card 5 and I put voice 5 in the voice 1 spot (meaning using all voice 1 or A cables) to rule out a connection issue down the line, and the problem does follow the card. I did the C4 routine to disable the cards other than voice 1 so I can toy with the bad card. About to do some more checking under the magnifier....

And sorry I meant to say U7; I don't know why I keep dyslex-ing this stuff....

Yesterday when I pulled the ribbon cable off I did a point-to-point continuity check on it from each pin on the first connector all the way down the line and they all checked out...
memorymoog #1257 | minimoog model d reissue

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Fri Nov 27, 2020 10:26 pm

Ha.. I removed and re-did my ribbon connector sockets on that voice board with the good gold-plated machined ones and I *think* I'm in good shape now. I'm getting OSC3 now. I stared at that voice card for at least a half hour.

Now - I'm not sure I'm out of the water yet; I'm not getting noise (noise mixer dial not working yet), the pulse-width adjustment sounds disgusting, and the octave of OSC3 is very high compared to the rest (even at 16') but that may just be a calibration now that it's operating correctly and I probably calibrated it wayyyy out originally not realizing it wasn't working. If I remember correctly the noise is digital pink noise, not analog white noise.

Ok, going to get this thing buttoned up with all 6 voice cards and see what's what. I'm praying!
memorymoog #1257 | minimoog model d reissue

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Sat Nov 28, 2020 12:27 pm

Okay, things are better, but not totally fixed yet...

I do have OSC3 now, and I can calibrate the voice cards with all oscillators in the correct footage/range (by the way someone did at some point reverse the trimpots on these cards so, it may be time to replace; they are a little wonky). So everything calibrates out, but when I hit autotune I still get 0 TUNED.

I ran C7 and checked the hex data for voice card 1, and all oscillators are coming back in the 86 86 86ish range, so they're being read. Then I changed the patch to something with 2/3 KB TRACK and it immediately dropped the oscillator range on OSC3.... so I'm still having a problem in the KB TRACK area.
memorymoog #1257 | minimoog model d reissue

Markyboard
Posts: 354
Joined: Wed Jul 30, 2003 6:46 pm
Location: Springfield, VA

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by Markyboard » Sat Nov 28, 2020 6:44 pm

Good progress! But it still seems like the S12/S54 ribbon cable signals are getting screwed up. Since you’re doing a refurb It’s definitely worth making new ribbon cables. Always nice to have a spare set so even if it doesn’t fix this problem its not wasted effort.

Autotune giving 0 tuned and yet C7 shows good values? Maybe a timing problem with sampling each VC. The VCF envelopes are used to enable the Osc to be read by the processor. The VCs are “Or-ed” into the auto tune signal going back to the DB for a timing measurement. Maybe the envelopes are not activating on the right VC. For example when you run C7 and select VC 1 Osc 2 maybe it’s really enabling VC 6 Osc 2.

Just a thought.

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Sat Nov 28, 2020 8:12 pm

Well it's also possible that the sockets I used suck; I had to replace the sockets on the 'bad' voice card since I had gone and replaced them all anyway preemptively. Apparently something there went wrong so maybe I need to re-address the cards individually again, knowing that all it took was one card in the chain to completely eff the synth...

I'm about to pull the analog board and check those too. The noise source also isn't working.

I did check that ribbon cable and I tried swapping it with the other one from S53 earlier on as a test and the same problem carried. I do think you're right in that I should replace the ribbon cable at this point though.

EDIT: and actually what the 2/3 KB TRACK is doing is dropping the transpose of OSC3; not by an octave, but when it's switched on only OSC3 goes from normal to a lower note. And just as I sort of expected, when I booted up with P28 cable removed from the analog board, the 2/3 KB TRACK on OSC3 remains fine, but the pitch is lower than the rest of the oscillators.
I tested on only voice card 5 also with the ribbon cable and the problem with the KB TRACK is still there so at this point I think i'm ruling out the cable itself being bad.
memorymoog #1257 | minimoog model d reissue

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Sat Nov 28, 2020 10:26 pm

I'm embarrassed to say I don't think I quite understood the sample and hold circuit....

I'm rethinking my DMUX board again. In doing the autotune upgrade, the first instruction was to replace 13 capacitors on the DMUX board. This is where I failed to understand what I was replacing, but replaced per the instructions. I'm now realizing by studying the schematic that those caps, coupled with the corresponding resistor and opamp, are making up the S&H circuit. 13 of those caps were changed out, which have a large ground plane on the top side of the board, almost impossible to solder through to and you can't get to the bottom of the cap from the top side, they sit flush with the board. However, I questioned this early on when I had forgotten to install 2 of those 13 caps and 2 of those 18 resistors like a dope, which was creating almost a fast beating LFO sound on OSC3 everywhere, until I went back and fixed that.

Notice how everything always revolved around OSC3?

I did do a connectivity test on the caps I replaced on the DMUX last time it was out on the bench. I tested connectivity from the solder point on the bottom of the board of that particular cap to the ground plane it sits on and they were all making connection. I used .047uf film caps rated at like 630v because that's all I could find in the right package and fit...

Let's put it this way; before I pulled it apart and starting doing this autotune upgrade, it always read and tuned voices. Used to tune 6, started to tune 4-5 so I decided it was time for the overhaul. When I originally got it I got it for a song because the power supply was cooked, and I fixed that.

I guess I'm trying to understand the best way to check that the S&H circuit is functioning properly as a whole.
memorymoog #1257 | minimoog model d reissue

Markyboard
Posts: 354
Joined: Wed Jul 30, 2003 6:46 pm
Location: Springfield, VA

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by Markyboard » Sun Nov 29, 2020 9:26 am

Let me start by saying there’s nothing to be embarrassed about. Your Memorymoog is totally mind f’ing with you and you need to kick it in the balls. Fear not, it has plenty to spare :D .

The service manual gives a pretty good explanation of the s&h function. What’s you’ve done is a fairly sizable and invasive change. Plenty of places where you could have messed something up.

But...if everything looks good from an inspection point of view then I would start taking a more analytic approach to the problem, Or much better yet each specific problem. That dmux is really a set of individual s&h circuits fed by a common DAC and timing circuit. Here are some examples of individual issues to tackle

Look at the specific s&h signals for the tracking 1/3 and 2/3. Those are just on/off outputs as opposed to continuously variable. With your scope on the DMUX signal (or at the VC)you should see the level toggle with each button push. Nothing else should affect these signals but do turns the Osc levels and play with the voice modulation controls to confirm. If anything else is affecting these tracking levels you need to figure out why by tracing back with the scope. Try not to perturb anything else and focus on just this problem.

Look at the osc3 amount/ level coming from the DMUX (or again monitor at the VC). Now toggle the kB tracking 1/3, 2/3 buttons. Do you see that Osc 3 level signal changing? If you do trace it back to the specific 4051 output. If it’s not changing then figure out what else could affect that level.

You say the Kb track is causing Osc 3 to transpose? Then you should be able to seethe Osc 3 SUM level change when you engage/disengage the kB track buttons. Work your way back from here. Maybe that the transpose and kB track are shorted somewhere. But instead of guessing you can nail it with the scope.

These are just examples. Once you nail one or 2 problems it may fix the rest. My point is that you have a scope and the ability to logically figure out what’s going on with each of these problems as opposed to the blanket approach of pulling boards, swapping chips, sockets, ribbon cable etc. which only gets you so far.

You’ll learn a ton about how this stuff really works. But yeah it’s going to hurt your head.

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Sun Nov 29, 2020 2:03 pm

Thanks... I hear you, and I realize I’m messing around with a synth that most techs won’t even touch, but I’m not afraid of stuff like this even with limited knowledge. It’s all just parts... and you’re right, I did make a real large change to the system including a new OS that I programmed myself on 3 EPROMs, so I’m thinking if that wasn’t working, I’d probably know by now...

Latest findings: since I swapped out the connector sockets when I had the boards out originally, I decided to swap them again for the gold plated machined sockets now that I fixed my issue on that voice card. But, that didn’t make any difference. What I did notice though is when I pull P28 off the analog board (which the orange/white wire affects 2/3 on that oscillator), the note shifts up an octave. It only does this with OSC3. And as far as I can tell it’s doing it on all voice cards. So something between P28, the DMUx board and the voice cards is transposing the entire octave which also explains why autotune is kooking out on just OSC3...

So it’s like something as a whole is messing with the transpose somewhere. That’s what I need to nail down. And yes, I have tested a lot of this with my scope and have seen the hi/lo enabling on 2/3, so I don’t think it’s a switching issue.

If I keep P28 connected, and then range/scale up the oscillator to match another good one, autotune will read it as PITCH HI. So it’s definitely a transpose problem on just that oscillator and not the trim pots either. I’m so tired lol...
memorymoog #1257 | minimoog model d reissue

whinylittlerunt
Posts: 45
Joined: Thu Mar 16, 2006 4:30 pm

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by whinylittlerunt » Thu Dec 03, 2020 11:31 pm

Well, I've made some progress (yay?)

I fixed my noise problem (or, lack of noise control). Apparently I bent a fairly important pin on the 4051 in the sample/hold circuit that just happened to be the noise amount. Didn't notice it, again, until I pulled the DMUX board, again :roll:

I replaced all the ribbon sockets on the DMUX with the machined ones. I don't think that did anything different but it's worth doing anyway.

Replaced several trimpots on the common analog board and plan to do the rest when I get the correct ones.... (I also have all the trimmers for the voice cards on order)

OSC3 is still an issue but - interestingly, not sure if it was that messed up pin on the IC, but I'm now actually getting the autotune circuit to tune voices. I tried it just for giggles and it came back with 5 tuned! So, I'm well on the way.

I had originally thought that because I noticed CR3 diode was missing from the board, and just happened to fall right in line with OSC3 2/3 KB TRACK, that it could have been my problem all along, but in the FB group I was informed that it was a factory goof and instead used a flyover diode across 2 resistors to compensate.

So I'm going to move forward and put the new trimmers on the cards, get that sorted out, and see where OSC3 ends up.
memorymoog #1257 | minimoog model d reissue

Markyboard
Posts: 354
Joined: Wed Jul 30, 2003 6:46 pm
Location: Springfield, VA

Re: Memorymoog "DEAD OSC" on all voice cards OSC3

Post by Markyboard » Fri Dec 04, 2020 9:42 am

Cool - figured you were hard at it. Never give in, never surrender!

Post Reply