We have the jumper settings here as well as some decent photos to help you seeing where the traces were originally:
http://www.win3x.org/uh19/motherboard/show/809

minimum, bios chip, CPU and speaker. You should get beeps telling there is no memory installed.

You can check the voltages on the power supply before connecting that to the board.
You can check the +12v lines on the board itself for short-circuits (tantalum caps sometimes short out).

Thanks for this, however I didn't mess with the jumpers and I also own the original manual.
Regarding the photos: there are not even ones from the back side, are you joking? 😉

Power supply voltages with no load are 5,36V, -4,97V, 11.12V, -10.05V.

On the Motherboard I measure ~20MOhm between +-12V, -5V and GND, and ~1.8k between 5V and GND. Is that too low?

gedobbles wrote on 2021-04-28, 15:01:

Thanks for this, however I didn't mess with the jumpers and I also own the original manual. Regarding the photos: there are not […]
Show full quote

Thanks for this, however I didn't mess with the jumpers and I also own the original manual.
Regarding the photos: there are not even ones from the back side, are you joking? 😉

Power supply voltages with no load are 5,36V, -4,97V, 11.12V, -10.05V.

On the Motherboard I measure ~20MOhm between +-12V, -5V and GND, and ~1.8k between 5V and GND. Is that too low?

You will not get accurate readings from the power supply without load. Try to attach a hard drive or something that draws a bit of power, then measure voltages again.

1.8k+ is good between the power rails. It would be problematic if you had 10 ohms or lower.

computerguy08 wrote on 2021-04-28, 15:26:

You will not get accurate readings from the power supply without load. Try to attach a hard drive or something that draws a bit of power, then measure voltages again.

1.8k+ is good between the power rails. It would be problematic if you had 10 ohms or lower.

That's good, with the power supply connected, the voltage levels are fine.
I just measured without it being connected, because evasive said so, 1s of thinking reveals that there is no sense to do so.

evasive wrote on 2021-04-28, 14:58:

You can check the voltages on the power supply before connecting that to the board.
You can check the +12v lines on the board itself for short-circuits (tantalum caps sometimes short out).

What else could be the problem?

I would install a bios battery I have had plenty of motherboards that will not post or act weird if the battery is dead,low, or missing. Did the original battery leak and cause and damage to the system?

I will try doing that. Yes, the old battery leaked and lead to corrosion on some traces (see my first post).

When I attached the battery, I observed, that there were 5V between battery + and GND, turns out that battery - is not connected to GND, but it should I guess?
The acid probably completely removed the connection between the battery - holder and the GND pane inside the PCB.

Fixing that (connecting battery - to GND) powers the RTC chip as expected, but still it doesn't beep.
Any ideas?

My best advice would be to give the board a good cleaning with vinegar to remove the akili from the battery and then a good rince with 99% IPA and then see if traces need to be fixed and go from there.

Edit: fixed a spelling error.

chrismeyer6 wrote on 2021-04-28, 18:35:

My best advice would be to give the board a good cleaning with vinegar to remove the akili from the banter and then a good rince with 99% IPA and then see if any traces need to be fixed and go from there

What is "removing the akili from the banter"?

That was supposed to say battery sorry about that. Not sure how my phone did that. I usually use this site on my computer and apparently my phone's auto correct decided to have fun.

After reading trough the beginning of some BIOS dump, it seems that one of the first things to do is to read the status register of the 8042 keyboard controller (through port 0x64) and read/write some CMOS RAM in the RTC (through ports 0x70, 0x71).
Does anyone know, what happens, if a read to these ports fail, so does it just return whats left floating on the bus as other port reads do or does it somehow block on that read?
May a malfunction in the 8042 or the RTC be the cause the BIOS doesn't even get to emitting beep codes?

gedobbles wrote on 2021-04-29, 08:09:

After reading trough the beginning of some BIOS dump, it seems that one of the first things to do is to read the status register of the 8042 keyboard controller (through port 0x64) and read/write some CMOS RAM in the RTC (through ports 0x70, 0x71).
Does anyone know, what happens, if a read to these ports fail, so does it just return whats left floating on the bus as other port reads do or does it somehow block on that read?
May a malfunction in the 8042 or the RTC be the cause the BIOS doesn't even get to emitting beep codes?

That read is to determine if the reset is a power-on event or a warm software induced reboot. This is pretty important as in the second case the BIOS will skip most of the system init code, like say RAM refresh, and do a quick boot. If certain CMOS area is set with certain values the reboot is almost transparent - this is a mechanism back from the days of 286 where you couldn't switch CPU back to real mode except by reset, and the switch between modes was used to access memory over 1MB (though HIMEM.SYS can use undocumented LOADALL instruction to work around that even faster and without reset - not sure which version introduced this).

So if the power-on is being detected as warm reboot the mobo will be pretty much uninitialized and will not work. You could maybe hack the BIOS to try to overcome that, but if you want to get past the early stage boot you also need to recalculate the checksum byte.

You would see this sort of stuff on a POST card, in the form of hex codes. If you don't have one already, you should get it, it's really useful.

computerguy08 wrote on 2021-04-29, 17:08:

You would see this sort of stuff on a POST card, in the form of hex codes. If you don't have one already, you should get it, it's really useful.

Yes, that's what I also think, I found some schematics on the internet to built one myself, but then I was wondering, if the standard CMOS logic circuits (74HC...) are suitable in the means of speed and voltage levels,
or if I needed TLS (74LS...) or 74HCT ones.
Has anyone experience with this?

Additionally it may be useful to have some EEPROMs to test the board with some custom code, like beep with the speaker directly on boot, then test reading/writing CMOS RAM etc.,
however I don't know, what the pinout of suitable EEPROMs is and where to get some (+ programmer...).

From the DYI cards I see, all of them seem to use ALS/HCT chips with TTL voltage levels. That would make sense as that is what the PC design was based on.

TL-866II programmer. As for the EEPROMs, check the postings here on what EEPROMs are being used.

evasive wrote on 2021-04-30, 08:17:

TL-866II programmer. As for the EEPROMs, check the postings here on what EEPROMs are being used.

I don't quite want to peel the sticker off to know the type of the EPROM, does anyone know a working and availeable EEPROM replacement?

What you can do is use uniflash, usually that will detect what eeprom is mounted.
https://www.rainbow-software.org/uniflash/

I see rom.by has restarted the development of uniflash:
https://soggi.org/motherboards/bios-update-fl … h-utilities.htm

If it's a single-chip BIOS on a 486, there's 99% chance it will be a 64KB part. Either a 27C512 or W27C512 (Winbond EEPROM) should work. The latter are pretty cheap and still widely available and won't need a UV source to erase and re-programme if needed.

evasive wrote on 2021-04-30, 12:53:

What you can do is use uniflash, usually that will detect what eeprom is mounted. https://www.rainbow-software.org/uniflash/ […]
Show full quote

What you can do is use uniflash, usually that will detect what eeprom is mounted.
https://www.rainbow-software.org/uniflash/

I see rom.by has restarted the development of uniflash:
https://soggi.org/motherboards/bios-update-fl … h-utilities.htm

haha, the original chip (almost for sure) is an UV-eraseable EPROM (I can feel the window through the sticker),
my plan is to eventually replace this by an EEPROM temporarily for testing/educational reasons.

However I don't have any idea about the pinout of this EPROM and whether compatible EEPROMs can be found.

(While I was writing this, megatron-uk was faster typing a reply):

megatron-uk wrote on 2021-04-30, 12:57:

If it's a single-chip BIOS on a 486, there's 99% chance it will be a 64KB part. Either a 27C512 or W27C512 (Winbond EEPROM) should work. The latter are pretty cheap and still widely available and won't need a UV source to erase and re-programme if needed.

Thanks for that, I will look into these.
EDIT: wow these EPROMs are even availeable at my favourite electronic parts shop!(the EEPROMs however are not)