• Please review our updated Terms and Rules here

C64 Memory Issue

robert_sissco

robert_sissco

Experienced Member
Joined
Oct 12, 2021
Messages
342
Location
Midwest U.S.A.
I am repairing a dead C64 I got a while back. I got it to the point it will boot (VIC-II, RF Modulator, and PLA were all shot). One those were replaced, it boots just fine...but with one issue. It never boots with the full 38911 bytes free, it is always a random number, and if I follow the decree of Saint Roy Trenneman and turn it off and back on again, the number is lower. If I keep turning it off and on again, it always loads with about 200 bytes less than last time. If I let it set for a while and cool off, I get back to about 36xxx free.

Colors are fine and it boots/runs from cartridges just fine (my Epyx Fastload and Star Trek carts seem to work fine) but when I load into memory from a floppy, then it fails.

I am thinking this is a back RAM chip. I do not have a dead test/diag cart yet (this is making me order one so I can get this to work when I get paid this Friday), but I wanted to try the heat test, so I booted the system and let is sit for about 10 minutes, all 8 RAM chips felt the same.

Anyone have anything else I can try while I await the diag cart? I also lack additional RAM chips otherwise I would try piggybacking them one at a time until I find a chip that gives me the full 38911 when it is piggy backed.

Good news is the SID works, so yay there.

Just more curiosity on my part on what I can try here since I think I know the current issue I am facing.
 
You can always use BASIC to poke and peek the RAM beyond the number found by the C64 itself. Maybe the difference in bits give you an idea which DRAM IC is faulty.
 
Well... those ram chips are 64k by 1 bit. So that's why there's eight of them... eight bit computer.

If one was dead completely, then the whole computer wouldn't function. So I would say the chip is flaky, and likely to die soon. The fact that it runs off of a cartridge isn't super surprising, considering the cartridge itself takes up a slot in the memory map in place of ram. Yep, trying to load floppy data into a flaky memory is going to fail.

What brand are the ram chips? Do you have a lot of MOS branded logic chips on the board? Could you snap a good quality photo of the board, and post it here?
 
Last edited:
I know, I just used the cart as part of the troubleshooting to see if it would still work since the disk loading was giving odd results. I didn't expected it to fail, and just confirming that it didn't.

All 8 of the RAM chips are NEC branded. I will double check when I can open the case tonight, but I think the only MOS branded are the CIA's, ROM's, Processor, SID and VIC-II. The PLA is a modern replacement. The memory controllers are TI I think, but I could be wrong on that, again, I can confirm tonight.
 
robert_sissco said:
I know, I just used the cart as part of the troubleshooting to see if it would still work since the disk loading was giving odd results. I didn't expected it to fail, and just confirming that it didn't.

All 8 of the RAM chips are NEC branded. I will double check when I can open the case tonight, but I think the only MOS branded are the CIA's, ROM's, Processor, SID and VIC-II. The PLA is a modern replacement. The memory controllers are TI I think, but I could be wrong on that, again, I can confirm tonight.
Click to expand...
NEC is good quality, so it's likely just one chip that's gone loopy.

Use a multi meter on continuity check to verify that all ram chips are connected properly. Make sure all the address lines are connected to all the other address lines on each chip. Din and Dout should connect to the MPU individually, and not to the other ram chips. Make sure each chip is getting the appropriate 5 volts and ground when operating.
 
Isn't there something about MOS 74xxx series chips that manage the RAM? It could be something like that flaking out which would give you random RAM amounts. The Dead Test is a good step. Look on YouTube for Adrian's Digital Basement (?) I think C64 repair 2/12 (#2 of 12 devices).
 
RetroNinja said:
Isn't there something about MOS 74xxx series chips that manage the RAM? It could be something like that flaking out which would give you random RAM amounts. The Dead Test is a good step. Look on YouTube for Adrian's Digital Basement (?) I think C64 repair 2/12 (#2 of 12 devices).
Click to expand...
I am subscribed to Adrian's Digital Basement and tend to binge his videos, he's actually while I took on this project to fix this board, and it was some info I got from watching his videos that got me to the point it boots with video, not I just need to resolve this memory issue.
 
robert_sissco said:
I am subscribed to Adrian's Digital Basement and tend to binge his videos, he's actually while I took on this project to fix this board, and it was some info I got from watching his videos that got me to the point it boots with video, not I just need to resolve this memory issue.
Click to expand...
Oh so you probably already knew about the shoddy MT ram, or the shoddy MOS logic chips. COOL!
 
br44 said:
Oh so you probably already knew about the shoddy MT ram, or the shoddy MOS logic chips. COOL!
Click to expand...
Yep, first thing I checked when I had it narrowed down to the RAM was checking the brand on those chips, and non of them were sus so I figured it was likely a bad RAM chip that wasn't always reporting back good since it was failing.
 
robert_sissco said:
Yep, first thing I checked when I had it narrowed down to the RAM was checking the brand on those chips, and non of them were sus so I figured it was likely a bad RAM chip that wasn't always reporting back good since it was failing.
Click to expand...
Yep, I agree. That's likely what the issue is. It gives a logical avenue to pursue at least. Do the multimeter tests, and then do the diag cart when it comes in. I'd bet money that'll fix it.
 
Diag cart came in today. Says U21 is BAD in RED, and then the test freezes and all RAM reports BAD, but in blue.

RAM has been ordered and once I get it in I will piggy back it on U21 and see if that works, and if it does, time for the desoldering gun and a swap.
 
robert_sissco said:
Diag cart came in today. Says U21 is BAD in RED, and then the test freezes and all RAM reports BAD, but in blue.

RAM has been ordered and once I get it in I will piggy back it on U21 and see if that works, and if it does, time for the desoldering gun and a swap.
Click to expand...
That sounds like a logic chip rather than ram. I find it highly unlikely all the ram would be bad all at the same time, unless either something dramatic happened to the board, or it was from a known bad manufacturer.

Oh I should probably ask what revision of the board you have.
 
U21 is the 4164 for bit 0. I know that people use this piggy back trick but I advise against it strongly. Main reason: you don't know what is wrong with this DRAM. If its output has been blown and it keeps outputting a hard zero, it could blow up the new IC as well. OK, this is just an example, if it did output a hard zero, I think the whole board would not work. Just be warned!

Just desolder the U21, place a socket and then insert the new IC. If things go wrong again, you at least have much less work in placing a new part.
 
br44 said:
That sounds like a logic chip rather than ram. I find it highly unlikely all the ram would be bad all at the same time, unless either something dramatic happened to the board, or it was from a known bad manufacturer.

Oh I should probably ask what revision of the board you have.
Click to expand...

It could be, but it could also be bad RAM. I have seen on Adrian's where a bad RAM chip can throw others off in the diag. If that fails, I will try other things, but I do think that the U21 is the culprit, but I have been wrong multiple times before on this.

Also, it is an assembly 250407 board.
 
robert_sissco said:
It could be, but it could also be bad RAM. I have seen on Adrian's where a bad RAM chip can throw others off in the diag. If that fails, I will try other things, but I do think that the U21 is the culprit, but I have been wrong multiple times before on this.

Also, it is an assembly 250407 board.
Click to expand...
Oh goodie, you and I have the same board revision. That makes things easier.
 
It was actually 4 bad RAM chips, U21, U22, U23, and U9. Once they were removed and replaced with new chips in sockets, I got the 38911 bytes free every time. To make sure that it was not one bad chip causing the Diag to report bad chips every time, I swapped the old chips back in one at a time and each time I was getting the incorrect amount of free RAM, and when I replaced them all back with the new good ones, I am getting 38911 bytes free every time I boot.
 
robert_sissco said:
It was actually 4 bad RAM chips, U21, U22, U23, and U9. Once they were removed and replaced with new chips in sockets, I got the 38911 bytes free every time. To make sure that it was not one bad chip causing the Diag to report bad chips every time, I swapped the old chips back in one at a time and each time I was getting the incorrect amount of free RAM, and when I replaced them all back with the new good ones, I am getting 38911 bytes free every time I boot.
Click to expand...
Well dad gum.... four whole bad chips! All at once!

Suffice to say, you should probably swap out the other four, just to be sure. If the first four squirted the bed, then the other four are probably not far behind.
 
Yea, I was shocked as well. And I plan on replacing the others as well. U24 fails the diag but I have left the system running for 10 minutes and then restarted it and it still came up as 38911 free. SO it is working for now, but more are starting to give up the ghost.
 
Just a final update, got the other RAM chips in, so now in this board all 8 RAM chips are socketed and new, and Diagnostics are passing everything except the things I need a harness to test (which makes sense to me as I lack a harness (for now))
 
You must log in or register to reply here.
Back
Top