• Hello, Guest! Welcome back, and be sure to check out this post for more info about the recent service interruption and migration.

Can this 575 logic board be saved?

jajan547

Well-known member
Yup!

As the chips are taking their time to arrive, I did some more chip-removal practice on the board with the last 3 dram chips (I think I did it much more cleanly this time).

I decided to test the board just in case it would work with just a SIMM (no on board memory). This is what I observed:
  • No SIMM or MB RAM: Chime followed by silence until manual power off.
  • 4MB SIMM, no MB RAM: Chime followed by ~ 3sec silence followed by death chime.
  • 16MB SIMM, no MB RAM: Chime followed by ~10sec silence followed by death chime.
So it doesn't seem likely that the board would booth without any on-board RAM, unless I have some more self-inflicted faults somewhere.
Board needs to have some on board RAM to boot, this further confirms an on board RAM failure. Try adding say two RAM chips and see.
 

Jagmn

Active member
Yup, I'm still waiting on the RAM chips. Though I had an idea and did a little experiment:

The logic board RAM chips have their own RAS line, separate from the SIMM. All of the other lines are shared. I have a functional 4MB SIMM I didn't mind temporarily mangling, so I disconnected its RAS0 and RAS2 pins (which are connected on the logic board, anyway) and temporarily tacked them together and to the logic board RAS line.

No dice, still the chimes of death. I might verify the DQ and A lines separately then see if there's a different CAS layout for the SIMM, but I might also just try and be patient and wait for the actual DRAM chips in the post.
 

jajan547

Well-known member
Yup, I'm still waiting on the RAM chips. Though I had an idea and did a little experiment:

The logic board RAM chips have their own RAS line, separate from the SIMM. All of the other lines are shared. I have a functional 4MB SIMM I didn't mind temporarily mangling, so I disconnected its RAS0 and RAS2 pins (which are connected on the logic board, anyway) and temporarily tacked them together and to the logic board RAS line.

No dice, still the chimes of death. I might verify the DQ and A lines separately then see if there's a different CAS layout for the SIMM, but I might also just try and be patient and wait for the actual DRAM chips in the post.
Just be patient better to wait then risk anything.
 

Jagmn

Active member
Right, I'm back to where I started; with the suspicious RAMDAC/CLUT/Antelope chip (the functioning boot but crazy video).

From probing the signals I was convinced that my SIMM hack should have worked. Funnily enough: I found a break in A4 between the on-board RAM and MEMCjr; in the same place and mode I've found other breaks: the connection between a track and a pad (they look fine under the microscope but do not conduct). I can only assume this was caused either by my initial hot-air mishaps or was already a potential failure mode caused by a previous electrolytic leak (same area, again).

Anyway: The board boots with either my hacked-up-SIMM-to-onboard-RAS or a regular 16MB SIMM with no hacking. So these boards do seem to be able to boot without onboard RAM.

I might not have time to progress the RAMDAC/CLUT investigation until next week but will report back with whatever progress I make.

Thanks for all the pointers and encouragement.
 

jajan547

Well-known member
Right, I'm back to where I started; with the suspicious RAMDAC/CLUT/Antelope chip (the functioning boot but crazy video).

From probing the signals I was convinced that my SIMM hack should have worked. Funnily enough: I found a break in A4 between the on-board RAM and MEMCjr; in the same place and mode I've found other breaks: the connection between a track and a pad (they look fine under the microscope but do not conduct). I can only assume this was caused either by my initial hot-air mishaps or was already a potential failure mode caused by a previous electrolytic leak (same area, again).

Anyway: The board boots with either my hacked-up-SIMM-to-onboard-RAS or a regular 16MB SIMM with no hacking. So these boards do seem to be able to boot without onboard RAM.

I might not have time to progress the RAMDAC/CLUT investigation until next week but will report back with whatever progress I make.

Thanks for all the pointers and encouragement.
Very interesting keep it up you can save it.
 

Jagmn

Active member
I actually had enough time to remove the RAMDAC/CLUT/U26 and give it a clean and even get it back onto the board.

Underneath it didn't actually look too bad, a bit of corrosion that I cleaned up but not much else.
IMG_6201.jpeg

Unfortunately no change to my symptoms. So, I'm guessing that the problem could be one of:

  1. Bad U26/RAMDAC
  2. Bad clock signal to U26 (I've traced what I believe to be the clock line going to U19, pin 1)
  3. Bad signals/setup between U26 and MEMCjr or PrimeTime
  4. Some badness on the signals to the VRAM SIMMS (though they all have continuity).
Given your previous nudges, do you suggest #1? If so, where would I find a replacement?

And a big question to the hot-air de-solderers out there: I think I've worked out why so many of my pads are loose after removing a chip. My flux gets very hot and very liquid as I move the hot air around; I'm pretty sure it is melting the pad glue and absorbing it as I heat and remove the component. Any shout outs to what I should do differently? I had several loose pads after letting U26 fall off of the board via flux+hot air. I used MG Chemicals 8341 no-clean flux paste from a syringe.
 

jajan547

Well-known member
I’d check 3. And 4. I don’t think the chip is bad but I do think there may be some very fine breaks you otherwise can’t see test everything for continuity even if it looks solid and check the vias under the chip and make sure they connect to the back of the board.
 

jajan547

Well-known member
I’d check 3. And 4. I don’t think the chip is bad but I do think there may be some very fine breaks you otherwise can’t see test everything for continuity even if it looks solid and check the vias under the chip and make sure they connect to the back of the board.
How’d new chips turn out?
 

Jagmn

Active member
I haven't put the new chips on yet; having the board boot with the SIMM is enough for me to try to get to the bottom of the video issue.
Unfortunately I haven't had time to do much more investigation; there's what looks like sensible continuity between U26, the VRAM SIMMS, RAM (data), MEMCjr and PrimeTime.

Does anyone have a functional pinout for the CLUT? I'm building up a bit of a schematic of how its all bolted together in my notes but actually knowing if, for example, the 13th pin on the bottom of U26 (counting from the left) is its clock pin (wired up to U19 / 343s0161) would be helpful.

I'm considering researching and investing in a logic analyser to let me see what is going on between the various components; something like that would be useful anyway.
 

jajan547

Well-known member
I haven't put the new chips on yet; having the board boot with the SIMM is enough for me to try to get to the bottom of the video issue.
Unfortunately I haven't had time to do much more investigation; there's what looks like sensible continuity between U26, the VRAM SIMMS, RAM (data), MEMCjr and PrimeTime.

Does anyone have a functional pinout for the CLUT? I'm building up a bit of a schematic of how its all bolted together in my notes but actually knowing if, for example, the 13th pin on the bottom of U26 (counting from the left) is its clock pin (wired up to U19 / 343s0161) would be helpful.

I'm considering researching and investing in a logic analyser to let me see what is going on between the various components; something like that would be useful anyway.
I have an oscilloscope, can take a look again at my board when I get home.
 

jajan547

Well-known member
I haven't put the new chips on yet; having the board boot with the SIMM is enough for me to try to get to the bottom of the video issue.
Unfortunately I haven't had time to do much more investigation; there's what looks like sensible continuity between U26, the VRAM SIMMS, RAM (data), MEMCjr and PrimeTime.

Does anyone have a functional pinout for the CLUT? I'm building up a bit of a schematic of how its all bolted together in my notes but actually knowing if, for example, the 13th pin on the bottom of U26 (counting from the left) is its clock pin (wired up to U19 / 343s0161) would be helpful.

I'm considering researching and investing in a logic analyser to let me see what is going on between the various components; something like that would be useful anyway.
U26 13th pin from left on the bottom connects to no pins on U19. How's the update on this?
 

Jagmn

Active member
That's odd; on my board that pin connects to R100 (which measures 48Ohms) and then to the top-left pin of U19. Where 'top' is having the ports facing away from the viewer. I count the pins starting from '1'.

I haven't progressed this much further but I think I need to transfer my notes into a schematic as I've got a fairly detailed connection layout around U26 now. It seems to have an 8-bit data path connection (shared between VRAM, RAM and MEMCjr) and 4 pins connected to MEMCjr (at least 2 of which are also connected to PrimeTime/U12), with the rest connecting to either VRAM (S[1-2].U[1-4].SIOxx), gnd/5v, video out signals x4, and some analog connections that look like RC networks and/or decoupling. Oh, there was another connection to MEMCjr (pin 16 on the right, counted from the top) that seemed to go to U17 pin 9 through ~23Ohms, I haven't been able to trace that more thoroughly yet.
 

Jagmn

Active member
It's alive, Jim!
IMG_6254.jpeg
I managed to lay my hands on a second LC575 board, albeit a fully functional one, and compared the CLUT connections to the ones I had already mapped. It turns out that there was indeed a dead-pin, the one I investigated back in post #35 (bottom-right but one on the CLUT). It should have been connected to R88 (48Ohm) which links it through to pin 8 of MEMCjr/U17 but it showed as unconnected to anything I could probe.

So I bodged the other side of a VIA to R88 back to that pin and… the display now functions. Everything else seems to check out, so far. I'll have to see how it handles when I re-solder the motherboard RAM (though, now it's a bit of a franken-mac board, I'm wondering about using those signals for a second SIMM).
 

jajan547

Well-known member
It's alive, Jim!
I managed to lay my hands on a second LC575 board, albeit a fully functional one, and compared the CLUT connections to the ones I had already mapped. It turns out that there was indeed a dead-pin, the one I investigated back in post #35 (bottom-right but one on the CLUT). It should have been connected to R88 (48Ohm) which links it through to pin 8 of MEMCjr/U17 but it showed as unconnected to anything I could probe.

So I bodged the other side of a VIA to R88 back to that pin and… the display now functions. Everything else seems to check out, so far. I'll have to see how it handles when I re-solder the motherboard RAM (though, now it's a bit of a franken-mac board, I'm wondering about using those signals for a second SIMM).
Hey that’s awesome!
 
Top