I am no expert at this, but if you have a access to a scope (or maybe a decent logic probe), you could start comparing signals between a working and a non working card . This might give you a hint as to what is wrong .

That Goldstar chip appears to be SRAM , maybe it's gone bad?

darry wrote on 2021-11-16, 03:23:

I am no expert at this, but if you have a access to a scope (or maybe a decent logic probe), you could start comparing signals between a working and a non working card . This might give you a hint as to what is wrong .

I do have a scope. Unfortunately, I don't have any experience using it for these kind of diagnostics.

AppleSauce wrote on 2021-11-16, 03:30:

That Goldstar chip appears to be SRAM , maybe it's gone bad?

I've been working my way across both boards, socketing chips as I go. So far I have found that the replacement 74HC74 was good but 74HC10 below it was also bad. This produced the same error as the bad board when swapping only this chip but there is still some other problem that is still causing it to fail even after replacing this chip as well.

I tasted the SRAM and that checked out fine. Still a couple more to go...

The 74HC688 was also bad. I wonder what kind of fault had to occur to cause so many chips to die. It seems to be isolated to the left side of the board.

I've run out of sockets and I don't want to solder any of these back directly onto the board so I can't confirm if this is the last faulty chip. The 74HC573 and 74HC245 also tested fine. The only IC I haven't tested specifically is the 74HC139 and the PC900V DAC. I am also assuming that the crystal is working. The rest are all passive components.

most likely some power supply failure caused this
PC900V is an optocoupler for the MIDI-IN signal, it will not cause this failure if it's bad
you can test the card without it just to make sure

I've ordered all of the replacement parts. I'll update this thread in a month or so. Most of the parts are coming from China so who knows.

Still waiting on one more chip for the repair. I ordered these in November and the ETA says February. Maybe it's stuck on a boat off the coast of California.

To pass the time I dumped the ROM. I found that it was identical to the one here.

I'm attaching it here for archival in case anyone ever needs a copy. I've also noted the kind of chip used and included a photo as well.

By "ETA" February apparently meant "will arrive tomorrow". I got the last set of chips in the mail the other day and I'm looking at them now. There is at least one casualty but hopefully the others are alright. This is why I like to order extras when buying old chips. You never know what condition they will arrive in and how many will actually work.

I found that three of the chips were damaged. One had its pins mangled, another had a pin break off while trying to straighten it, and a third got dented on a pin somehow.

I managed to straighten the pins on the other two and thought about what to do with the broken leg on the third. The chip could still be salvaged by running a bodge wire as necessary, but I felt it was a waste to not try and reattach it. This wouldn't be entirely useful if the chip is to be soldered as the leg would fall off on its own when heated again. But a bit of reinforcement with solder should hold it on well enough for use in a socket.

I didn't have any way to hold the leg and the chip in position with my soldering iron and solder wire in-hand so I propped it up and stuck it in place with some mounting putty and tacky flux. This was somewhat regrettable afterwards as it melted onto the back of the leg and had to be scraped off but I still managed to flow some solder well enough to keep it from moving. This wasn't the end of it though as it was only secured on one side with the break still through to the other.

After cleaning off the melted putty I held the chip down against some Kapton tape and fed in a bit more solder. This made a small blob which I then spread along the surface while holding the leg firmly with ceramic tweezers.

The result was excellent and the leg held firm, now secure on both sides of the break. A bit more solder on the outside surface would have been ideal but this was an extremely precise repair and any further manipulation could have easily made it worse. This was sufficient and I confirmed that it was a strong connection by inserting and removing it from a socket a few times.

Of course, there's no way I would remember that I had repaired this; which chip in a pile of chips and which leg on the thing. So as I always try to do, I added a helpful little note to myself. This is to remind me that if I use this chip to put it in a socket and not to solder it directly as this would cause the leg I just repaired to fall off. Which leg? The one with the arrow pointing to it, of course. 😀

Since I was already adding sockets to a few of the chips I ended up just socketing everything on the board. This also will provide me with a good reference in case I ever need to swap chips again in the future for testing.

When socketing everything onto the board I had a moment of panic as I was immediately greeted by the same "Could not detect MPU-401" error -- which was a problem because I had already confirmed that the board was working before adding the rest of the sockets. But this was a false alarm! I had accidentally flipped the dip switch on position 2. After setting it back to the ON position everything worked perfectly. All of the chips I had ordered also tested fine with no defects in functionality.

Success! My Midiman MM-401 is repaired.

nice rescue!!! enjoy!

Bravo!

tl;dr was the address decoder '688 the only thing wrong with this card?

maxtherabbit wrote on 2021-12-22, 15:13:

tl;dr was the address decoder '688 the only thing wrong with this card?

It was three chips on the left side which failed: 74HC74, 74HC10, and 74HC688. Somehow 74HC125 survived whatever catastrophe occurred.

I knew something was wrong initially because not only did the card not detect but 74HC74 got very hot to the touch. This chip did not even get warm on my other working cards. Replacing this one chip did not fix the issue so I started swapping chips one by one. After socketing and swapping three chips down the left side and having it still not work I decided to socket the entire board. I found out later that there was only one more dead chip but I didn't know it at the time.

Here is a photo with most the chips labeled for reference. This is the second board I was using to swap chips with. Labeling them helped me keep track of which chip was what at a glance.

Note the hand-written sticker on the EPROM chip. I bought several cards from someone who used to work at M-Audio so this would have been written by either him or another engineer. The SRAM was already socketed on this board before I touched it so it might have been used for testing. I guess I finished the job by socketing everything else.