Looking at the picture it appears that there is still some more corrosion to clean off. I've had alot of old and modern motherboards play dead or just act strange and randomly with a low/dead,missing bios battery. I'd give the while board another thorough cleaning and try to get a bios battery hooked up to it and try it again.

I used isopropyl and toothbrush - is there a better chemical/technique? Photo of the original state attached.

You'll want to try white vinegar to clean the battery corrosion off since their alkaline-based batteries the vinegar neutralize it. Then give it a good rince with distilled water and then isopropyl and let it dry in a warm sunny place.

There's a lot of green still on that board. The general rule is that if you can see it, there's probably just as much that you can't see.

Unless you get it all, and that includes neutralising with white vinegar and then cleaning that back off, AND any repair work for traces and replacement ICs, you're just fighting a losing battle.

There's definitely corrosion on those dram modules, on the IC at RP2, in the (fpu?) Socket at the foreground of the image, and likely more.

Done the cleaning as suggested - vinegar, de ionosed water wnd then isopropyl alcohol - see the photos for the end result. Also, attached battery (4.8v, original is three serial of AA).

The behaviour is still the same - beep code 2-1-1 with a slight twist that it now appears dead without the battery attached. Just be sure I have checked the PSU with and without load and it is all fine (5.05 and 11.7).

I have tested many tracks that are visually affected by corrosion and they're all fine. However, I didn't test all.

My understanding is that the board has four 256K DRAM chips and that's where the problem is. Do these easily/ commonly fail? They are surface mounted and I have no tools to check if they work.

What are my options now?

Two more close ups of memory chips

Enness wrote on 2023-07-25, 09:52:

My understanding is that the board has four 256K DRAM chips and that's where the problem is. Do these easily/ commonly fail? They are surface mounted and I have no tools to check if they work.

DRAMs can fail but I doubt the on-board 1M is your problem. Look at the RP2 - a yellow chip-like resistor pack near the empty DIP40 socket. It appears to have pretty much all of the traces going to it affected by corrosion, some could be actually open. There might be left-over corrosion under the socket as well, this can easily cause some signals to short (the residue is conductive).

You'll have to go over this whole area and inspect all the traces, as well as any nooks and crannies that could retain any corrosion (and do make sure to properly dry the board after washing, leave it in the open for a few days at room temperature). Some traces might need removing the solder mask and upper layer of corrosion to tell if there is still enough copper under it or not - a pen with glass fiber tip is great for that (but it leaves behind tiny bits of broken fibers, do your best to avoid getting those into your eyes or lungs). Exposed copper should be protected afterwards - tinned, and perhaps painted over as well, but tinning is usually enough.

In general always suspect broken connections and/or shorts after corrosion like that. Chips and other parts are only to be replaced if visibly damaged or you have a solid clue they are the culprit.

You cant clean yourself out of corrosion damage. Cleaning is just the first step of the process to find broken traces/connections/corroded vias. If you cant visually find any next step is blind reflowing/taking off components to look under them.

Spent some of time testing connection between chipset and memory modules and have found one bad trace. It does look visually fine, see the photo, yet it is definitely bad. The red line shows no connectivity between leg and pad, the blue line shows connectivity between same pad and another leg. Two more chips are also connected to same trace / pad on same leg. There is also connection to the chipset above the cpu via a resistor.

Could not find pin outs for such DRAM modules (but they look similar to ones on SIMM modules - any help here? It would help quite a bit).

What is the best way to repair this? DRAM is surface mounted and have no experience of soldering of SM components. How sensitive is DRAMto heat?

Also, any ideas why it has developed this fault?

Added a jump wire and nothing, still same 2-1-1.