128k MoBo, garbled video and chime with Sad Mac

04E925 signifies a RAM error. Are the last four digits the same every time? They are supposed to be a bitmask that points you to the physical location of the bad memory chips on the logic board.

E925 in binary is 1110100100100101-- a lot of bits are set. I don't think that many memory chips would go bad on one logic board, so the problem is most likely elsewhere.

There's an interesting repeating pattern of 010 in there, which makes me suspect an address decoding chip may be out. I know the 74LS138 was a common chip for this task, are there any of those on your logic board? I don't have a board pulled right now to check. What are the full part numbers of the 74LSxxx chips near the RAM chips?

I'll pull it out tomorrow and check. I'm to tired right now and have got a a lot to do tomorrow!

ive started to notice more and more ram failures on the 128K and early 512K models.

Wonder if the early production RAM had contamination issues on the substrate?

The garbled video same as mine leads me to believe its not a ram issue. I did the piggy back ram and changed roms with no change in garbled video. I would guess that maybe the video chip is the issue.

I guess now would be the time to learn how to solder. It shouldn't be to hard to figure out.

And if it doesn't workout, I'll just fame it in a shadow box and hang it on my wall. For less then $70, it was worth the gamble.

I did all I could and am letting McD have a shot at it.

Here are the 47LS chips and their locations in the G and F rows (where the RAM is):

G-1: SN74LS393NDS

G-2: 74LS257ANB

F-1: SN74LS393NDS

F-2: SN74LS257ANDS

F-13: SN74LS166ND

F-14: SN74LS166ND

F-15: SN74LS151AN

I'm thinking that the "E" in the code could mean row E, so here are all the chips in that row (also has some 74LS chips and the processor):

E-1: HAL 16R8CN 8417

E-2: HAL16R4CN 8417

E-3: 10X-1-1222 B 8411

E-4 to E-10: MC68000G8 GN78416 (processor)

E-12: SN74LS244NDS

E-13: SN74LS244NDS

E-14: HAL16R8CN 8412

E-15: SN74LS161AN

I ordered new ram chips from my thread and it made no difference, that much ram going bad at once doesn't make a whole lot of sense. I would think its another issue, garbled video doesn't scream bad ram. Hopefully when McD gets time maybe he will have better insight.

See my thread here a couple of years ago documenting the repair of my 128K with bad RAM, aka "bit rot". It too displayed video corruption with the Sad Mac displayed.

viewtopic.php?t=14580

Well that would be great. I hope he replaces the ram chips and it fixes it. That would be great.

I'm thinking that the "E" in the code could mean row E

Click to expand...

Going by the "Single Chip Identification" table on the Apple page I linked in my previous reply, 1110100100100101 (E925 in binary) means that the RAM chips at locations G12, G11, G10, G8, G5, F10, F7, and F5 are bad.

that much ram going bad at once doesn't make a whole lot of sense. I would think its another issue, garbled video doesn't scream bad ram.

Click to expand...

That much RAM going bad at once doesn't make a lot of sense, you're absolutely right. This is why the address decoding circuit needs to be considered as well. But garbled video can have everything to do with RAM -- RAM is where the video buffer is stored.

Maybe someone more experienced than me can jump in and help troubleshoot the 74xx logic that implements the RAM chip select signals. I've reached the end of my technical knowledge. :?:

The video circuitry shares the system RAM to hold the bitmap of the display, so bad memory or the circuitry supporting it can definitely corrupt the video.

I would first check carefully for bent/shorting pins on the bottom of the board, then start looking at the memory address decoding circuitry. A logic analyzer would be ideal here, but a scope or logic probe can tell you a lot.

with such random address errors, I vote the RAM timing controller chips being bad. the RAS/CAS read/write/select logic. all it takes is a stuck or noisy 74 series gate, and your dead in the water.

Well, on macdrone's board, I replaced one of the chips is said was bad. That made no difference in the error code so I made the leap to think the DRAM wasn't the issue. So I went and one-by-one replaced every 74LS chip on the board - 13 of them. None made any differences. It's either a different chip or maybe there is some board rot in the traces?

That would suck, as that machine I don't think got any use whatsoever. I think it was a present never used.

i would like to take a guess and say the rom is corrupt.

You might be right uniserver. Should I try dropping in the ROM from the 512kE and see if it makes a difference? There shouldn't be any problem put 512kE (Plus) ROMs onto a 128k board, right?

You can put plus ROMs on a 128k and itll work. All 3 boards follow the same architecture and addressing scheme. Boards dont rot without being exposed to external forces, such as cap goo, Or, exposed to high moisture and/or corrosive air. (salt air by seaside destroys PCBs)

I put good roms on the board and it didn't change anything, but that was before the chips were changed.

Ok, I'll have to try that out either tomorrow or over the weekend. I should have time this weekend so I can do some diagnosing.

I made a video of the problem which can be viewed

.
When I had the 128k ROM in, it threw an error of 04E925. I slapped the 512kE (Plus) ROMs in, and the error code changed; 03E925. The second digit changed, the others stayed the same.

04=failed Mod3

03=failed byte write

And all of the times I rebooted it, the last 4 digits always stayed the same, E925. Could this mean there is a problem with the ROM and the RAM (or RAM decoder chip)?

i guess is the ram its self.

I had the same outcome you did with swapping roms. That's why I thought some other chip bit the dust. Interesting, maybe we are getting somewhere with this idea. if its not the roms, and not the ram what else is at boot? VLSI chip? Ram decoder?

McD any way to swap one of those off a plus board to try?

I have replaced half the chips on the board with no change. Coincidentally, I did not swap the ROMs. I have another 128 K/512 K ROM set I could stick in there for giggles.

really would be nice to find a common denominator!

plus roms from a plus board I gave you should work, or were they all SE boards? When I did it just the number changed like the other person with the 128 issue. It changed the 4 to a 3. 3 still listed as a memory issue. So whatever other chip is a memory checker or video ram deal is what I am suggesting. Whatever those two chips are.

Or one of the custom ICs is bad. Or, eroded trace.

So I was looking at uniserver's thread about un-upgrading a Plus upgraded 128k when I noticed something different between his board and mine:

[attachment=0]IMG_0284.jpg[/attachment]

Click to expand...

On his board, spot D-14 is empty, mine however does with the markings of:

TRW

8413

810-1-222G

Could this mean anything? I know this board was a NOS service part, maybe something was supposed to be removed before usage?

On all my 128 boards have that bridged piece at D14 like your talking about. Only the original board that McD is attempting to do something with has the same problem as you. My other boards are "upgraded" to 512 but still have that TRW part.

I still have a hunch its the video chip, as it accesses ram and probably wont/cant boot if its rotted. That would explain the video and the fact it reports ram errors. Just a hunch, its beyond me that's why I gave it to someone wayyyyyyy better than me.