Don't trust the POST code to always be correct. This is where the BIOS dies, but it might not be what is causing it. I found the SIS 486 chipsets to be quite picky about RAM - I assume your 4MiB is 4 sticks of 30-pin SIMMs? Did you try both banks? Different set of sticks? If that still doesn't help, maybe there's an issue with cache or battery corrosion (RAM slots are pretty close to the battery).
Do you get a different error if you remove the RAM completly? That at least would tell you if it's detected at all (but not if it's bad, or the cache is). And BTW, did you try a non-VLB video card? It's best to get the system running with pure ISA cards and only then put VLB ones in, that way you know what is causing the problems. Keep in mind VLB is directly tied to CPU bus, if something doesn't work properly it'll bring the whole system down.

Thank you Deunan. Both RAM slots and cache slots are on the opposite side of the board. I’ve tried reseating the four sticks, but only on their banks, I haven’t tried moving them to the other banks. I’ve already removed the battery, started leaking a bit so I nicked it, applied vinegar, washed off with running water and finally washed with isopropyl alcohol and dried with cue tips. The only components next to the battery space are a couple of resistors and the keyboard connector. Both I already verified continuity with a multimeter.
I don’t have any other VGA cards to test the motherboard with.
I’m waiting for an external battery pack to arrive, maybe this motherboard does not like not having a battery…

I finally got a posting system. Ironically enough, the original board I thought was damaged by acid leakage, works. The SIS 85C461 Lucky Star motherboard has some other issue I haven't been able to pinpoint.

What I found was that the Lucky Star motherboard does not post with the DiamondStar Pro VLB in any slot, but the Gemlight GMB486UNP did. Unfortunately, the external AA battery pack I got does not seem to work with the CMOS.
Also, this is a first for me but it seems the seconds of the bios clock seem... arrhythmic? Is that normal? Could the quartz crystals be failing, somehow?

lavadrop wrote on 2022-04-08, 04:45:

What I found was that the Lucky Star motherboard does not post with the DiamondStar Pro VLB in any slot, but the Gemlight GMB486UNP did. Unfortunately, the external AA battery pack I got does not seem to work with the CMOS.
Also, this is a first for me but it seems the seconds of the bios clock seem... arrhythmic? Is that normal? Could the quartz crystals be failing, somehow?

I did say VLB can be tricky, that's why it's best to test with ISA cards.

As for your questions, external battery pack sometimes require certain jumpers to be moved, and it's best not to assume the pins you need to connect to are always the edge ones. Some of these 4-pin headers are wired in non-standard ways. Last option is always to just connect the pack in place of the original battery, though it's a good idea to add an extra diode to prevent charging (and again, some mobos have jumpers to allow for non-rechargeable batteries to be used). See if adding a jumper on the middle pins of the 4-pin conenctor makes it work.

The clock issue, is that something you observed in BIOS or DOS? BIOS has it's own refresh rate that is not very fast and not synced to RTC, so that might look like the clock is not advancing smoothly. Nothing to worry about. If it's in DOS however, it can be an issue with mobo clock (DOS pulls the time/date from RTC just once on boot and then uses timer and interrupt to keep its own clock).

Deunan wrote on 2022-04-08, 13:31:

I did say VLB can be tricky, that's why it's best to test with ISA cards. […]
Show full quote

lavadrop wrote on 2022-04-08, 04:45:

What I found was that the Lucky Star motherboard does not post with the DiamondStar Pro VLB in any slot, but the Gemlight GMB486UNP did. Unfortunately, the external AA battery pack I got does not seem to work with the CMOS.
Also, this is a first for me but it seems the seconds of the bios clock seem... arrhythmic? Is that normal? Could the quartz crystals be failing, somehow?

I did say VLB can be tricky, that's why it's best to test with ISA cards.

As for your questions, external battery pack sometimes require certain jumpers to be moved, and it's best not to assume the pins you need to connect to are always the edge ones. Some of these 4-pin headers are wired in non-standard ways. Last option is always to just connect the pack in place of the original battery, though it's a good idea to add an extra diode to prevent charging (and again, some mobos have jumpers to allow for non-rechargeable batteries to be used). See if adding a jumper on the middle pins of the 4-pin conenctor makes it work.

The clock issue, is that something you observed in BIOS or DOS? BIOS has it's own refresh rate that is not very fast and not synced to RTC, so that might look like the clock is not advancing smoothly. Nothing to worry about. If it's in DOS however, it can be an issue with mobo clock (DOS pulls the time/date from RTC just once on boot and then uses timer and interrupt to keep its own clock).

Yes, I agree, the only problem is that now a simple 16 bit VGA card is as expensive as a VLB card. I guess these motherboard's VLB slots were never tested and thus maybe be damaged somehow.
The good thing now is the other motherboard has better documentation available. Unfortunately, there doesn't seem to be any jumper to configure the board for an external battery other than the pins to connect it.
By adding a jumper in the middle of the 4 pin connector, what do you mean. exactly? it's not physically compatible to add anything between the 1 and 4 pin. I did try flipping the connectorbut nothing changed.

The manual states pin 1 is the VDD (6v).
Is it possible the motherboard needs even more voltage when using an external battery? 3 AA batteries just provide 4.5v. I can see on original photos that the board came with a 3.6v battery.

Yeah, it was inside BIOS, I haven't tried installing DOS or any other software adventures yet.
Oh and also one of the RAM sticks appears to be dead, I can only tally up to 3072 Mb RAM during the memory count.

lavadrop wrote on 2022-04-08, 16:34:

By adding a jumper in the middle of the 4 pin connector, what do you mean. exactly? it's not physically compatible to add anything between the 1 and 4 pin. I did try flipping the connectorbut nothing changed.

Don't do that. Figure out which pin is GND using a multimeter and that goes to battery pack (-). There should be a diode there but if not you risk frying some mobo chips.
I assume your pack has a 4-pin connector, with just the 2 side ones populated/wired? And the two middle ones are empty? If so you can't put a jumper on the middle pins, yes. I assumed your pack has wires with 1-pin connector on each one.

Speaking of multimeter, your mobo is old enough to use 146818A RTC/NVRAM chip. Pin 12 is connected to GND and pin 24 should have power when the mobo is powered or the battery is present. So check if you do have some 4V or so (assuming 3xAA pack) on pin 24, and if not try a "jumper" made of piece of wire. Take some thin copper wire, make a loop or an 8 with that on the middle pins on the mobo, and then stick the battery pack connector on top. See if that brings power to pin 24 but DO NOT power the mobo with that wire in place. If nothing changes, remove the wire. Only when it does provide the battery power you can try powering up via PSU.

lavadrop wrote on 2022-04-08, 16:34:

The manual states pin 1 is the VDD (6v).
Is it possible the motherboard needs even more voltage when using an external battery? 3 AA batteries just provide 4.5v. I can see on original photos that the board came with a 3.6v battery.

6V is usually what external pack back in the day used. Frankly the 146818A (especially older non-A variants) can be somewhat power-hungry for an RTC chip. But surely it's not any worse than my 286 and this one works just fine with 3xAA (I was using 3xAAA before but that drains too quickly if the mobo is mostly unused). 3xAA should also be safe to use instead of original 3.6V NiCD, especially if an extra diode is present. The RTC chip will work OK-ish down to some 3.3V on power but might start slowing down clock more than usual below that. TL;DR: A pack with 3 cells is pretty much perfect for this mobo no matter where you connect it. 6V can only be safely used on the external battery connector, and 3V is just barely enough and will cause problems once the batteries start loosing even a bit of their voltage.

Thanks for that, now I know how to find polarity with a multimeter. 😁
Somehow, now the battery pack is working with the CMOS.

Just have to figure out where to put it inside the case, away from the components.