VOGONS


First post, by Lazer42

User metadata
Rank Newbie
Rank
Newbie

I have an SC-55mkII with a dead MIDI in port on the rear. The front port works.

I ran the MIDI test as described in the service manual and get an NG for the 422 readout. The manual points to either the MIDI connector itself or one of two ICs to look at.

Is there a resource which could help me understand how to test the ICs so I can confirm whether one of the them is in fact bad? Or, is there something else I ought to consider looking at? The solder looks okay to me on the connector, and other obvious things like the capacitors look okay.

I am able to use a multi-meter and do repairs on this sort of stuff to a point. At this stage I am just trying to establish whether whatever is wrong with it is beyond my skill or not.

Reply 1 of 13, by keropi

User metadata
Rank l33t++
Rank
l33t++

I would start by reflowing the connector anyways - there can be hairline cracks that are hard to spot

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 2 of 13, by Lazer42

User metadata
Rank Newbie
Rank
Newbie
keropi wrote on 2021-05-18, 20:13:

I would start by reflowing the connector anyways - there can be hairline cracks that are hard to spot

Thanks for the reply. I have more information: although I haven't yet had a chance to try what you suggested (something I plan to try unless this information yields different advice), I did previously check for continuity between the pins on the connector and the solder on the bottom of the board and it is there in all cases. If I recall correctly I was also able to establish continuity between the connector pins and the solder connections for the coils that the connector runs into. I have not been able to establish continuity between the connector and the resistors or the optocouplers that follow in the circuit path, but I think that is more a matter of either my not having the right probe or my not knowing how to do it with the particular sorts of mounts these components have on the board. (I couldn't establish continuity between the pins and these components for the working connector, either.)

However, I remembered this morning that when I did the self test, the unit reported a result of "OK" for both IN1 and IN2. The NG (bad) result was for the third option, "422," which I think is the serial port? I'm not sure I understand this correctly, though. In any case, the two ICs the service manual points to for this result are the subprocessor, into which all the inputs (including the working one) go and the IC for the serial port. This might suggest the unit is set to read the serial input rather than the MIDI in, and I have come across one case online where a person had a similar problem to mine and the problem was the switch which swaps between the serial and DIN port. In that case, the switch was bad and so it was "stuck" on the serial in.

In my case, though, I've tested the switch with the meter and when I move the switch, continuity does change between its pins. In other words, the switch seems to be working correctly.

Reply 3 of 13, by Cloudschatze

User metadata
Rank Oldbie
Rank
Oldbie

The behavior does seem to suggest some sort of issue with the switch. If that's in any position other than "MIDI," any input going into MIDI IN 1 will not route to the internal tone generator.

Reply 4 of 13, by Lazer42

User metadata
Rank Newbie
Rank
Newbie
Cloudschatze wrote on 2021-05-20, 16:15:

The behavior does seem to suggest some sort of issue with the switch. If that's in any position other than "MIDI," any input going into MIDI IN 1 will not route to the internal tone generator.

Agreed. Like I said, though, the switch itself, as an isolated component, seems to be working correctly, so I'm not sure what to check from here.

Reply 6 of 13, by Lazer42

User metadata
Rank Newbie
Rank
Newbie
Cloudschatze wrote on 2021-05-20, 16:43:

In Test 4, does the LCD reflect the current setting of the switch as its position changes?

Wow, thanks for that: I didn't even notice that part of the test existed. I'll run it again when I get home and report back!

Reply 7 of 13, by Lazer42

User metadata
Rank Newbie
Rank
Newbie
Cloudschatze wrote on 2021-05-20, 16:43:

In Test 4, does the LCD reflect the current setting of the switch as its position changes?

Progress. The indicator does not change. It is stuck on 232-2 even as I change the switch. Now I need to try to figure out why. I had previously checked the switch and found that continuity changes between its pins when moving the switch.

Reply 8 of 13, by Lazer42

User metadata
Rank Newbie
Rank
Newbie

I finally got it working. The switch is connected directly into an IC via a trace on the board and it operates by a change in voltage. Moving the switch basically adds resistors in serial and the IC reads the voltage to choose the input.

What had happened is that a capacitor (C6) which the trace runs right next to had leaked and corroded the trace - though because there is also some screen printing directly on top I had a very hard time seeing exactly in what way, how much, or exactly where. I ran a bodge wire to the IC and it works correctly now. https://imgur.com/a/ePMScKG

Now I have to ultimately replace the capacitor, though I am not entirely sure I understand why the unit is still working. It's a 2200 uf capacitor reading at 0.4 uf, but other than the incidental damage to the trace everything else is working fine.

Reply 9 of 13, by yawetaG

User metadata
Rank Oldbie
Rank
Oldbie
Lazer42 wrote on 2021-05-21, 05:01:

Now I have to ultimately replace the capacitor, though I am not entirely sure I understand why the unit is still working. It's a 2200 uf capacitor reading at 0.4 uf, but other than the incidental damage to the trace everything else is working fine.

Failing capacitors do not immediately lead to a non-working electronic device. It can take years before the capacitors are failed enough to make something non-functional (although you'll likely notice instability or other problems beforehand).

Reply 10 of 13, by Lazer42

User metadata
Rank Newbie
Rank
Newbie
yawetaG wrote on 2021-05-21, 05:52:
Lazer42 wrote on 2021-05-21, 05:01:

Now I have to ultimately replace the capacitor, though I am not entirely sure I understand why the unit is still working. It's a 2200 uf capacitor reading at 0.4 uf, but other than the incidental damage to the trace everything else is working fine.

Failing capacitors do not immediately lead to a non-working electronic device. It can take years before the capacitors are failed enough to make something non-functional (although you'll likely notice instability or other problems beforehand).

Sure, but like I said, it appears to be failed pretty completely - at least based on the reading I took.

In any event, does anyone have a good source for replacing these?

Reply 11 of 13, by auron

User metadata
Rank Oldbie
Rank
Oldbie

that's interesting, are you really sure that cap leaked? it's kind of hard to tell much from the pictures you've posted. what device did you use for the capacitor readout? in-circuit measurements aren't that reliable - both that measurement and the idea that a cap of that size would fail to that extent are strange imho, not to mention the mess that would follow such leakage. thing is, people have already mistaken the glue on the crescent-shaped marking for leakage, see here: Roland SC-55 MKII Mini-Headphone Jack Help now whether that glue could have actually become corrosive is a different matter...

if you do want to replace that cap, these are very likely just general purpose caps so about everything should work. if you want the new cap to look the same, these are SU-series which i believe isn't made anymore, however roland used the very similar looking M-series in the 88st pro, which i'm guessing are the successor of the SU caps. part number for that cap on the M-series is ECA1CM222.

Reply 12 of 13, by Lazer42

User metadata
Rank Newbie
Rank
Newbie
auron wrote on 2021-05-21, 09:46:

that's interesting, are you really sure that cap leaked? it's kind of hard to tell much from the pictures you've posted. what device did you use for the capacitor readout? in-circuit measurements aren't that reliable - both that measurement and the idea that a cap of that size would fail to that extent are strange imho, not to mention the mess that would follow such leakage. thing is, people have already mistaken the glue on the crescent-shaped marking for leakage, see here: Roland SC-55 MKII Mini-Headphone Jack Help now whether that glue could have actually become corrosive is a different matter...

if you do want to replace that cap, these are very likely just general purpose caps so about everything should work. if you want the new cap to look the same, these are SU-series which i believe isn't made anymore, however roland used the very similar looking M-series in the 88st pro, which i'm guessing are the successor of the SU caps. part number for that cap on the M-series is ECA1CM222.

There may have been some glue in that area, but there was also corrosive discharge. I can't post a photo as I already cleaned it up in an attempt to get the circuit working again, but for one thing there was a lot of stuff that was similar to what you get on a leaking and corroded battery.

I admit I am not an expert at testing these things. I took the readings off of the leads using my multimeter which has a capacitor tester.

Why would you say it's strange to think it would fail, though? 30 year old electrolytic capacitors fail all the time.

Reply 13 of 13, by auron

User metadata
Rank Oldbie
Rank
Oldbie

i didn't say 30 year old caps can't fail, i was rather specifically referring to that 0.4uf reading which, if legit, points at a catastrophical failure of some description rather than regular aging. from my own experience even certain bloated and leaking caps can measure with fully intact capacitance and just have high ESR. did this capacitor you pulled show any kind of bloating or visible leaking from the bottom bung? was this corrosion on the board you are talking about in the striped crescent area where the glue was or at the cap vias?

i in fact did a (spoiler: in hindsight pointless) preventive mt-32 recap about 6 years ago, so that device would have been roughly as old as your sc-55mkii is now. looking back through my notes, measured with my atlas ESR60, all of those caps (mostly sanyo branded) i pulled had fully in-tolerance capacitance and ESR that was in line with general purpose caps you can buy today. that's just one example of course - in someone else's mt-32 they may very well be in much worse shape if that specific unit hasn't been powered on for 20 years.

did you try putting that multimeter to some other capacitors on the board, or any capacitor really, to check if it gives expected capacitance readings otherwise? sometimes dirt on the leads can also interfere with measurements - my ESR60 usually cancels the measurement in those cases.