Can you try using the BIOS ROM from the other board into this one? Wonder if the BIOS is corrupted or the ROM chip somehow damaged. Have you checked if the traces from the battery to the BIOS are good?

bregolin wrote on 2023-05-23, 13:26:

Can you try using the BIOS ROM from the other board into this one? Wonder if the BIOS is corrupted or the ROM chip somehow damaged. Have you checked if the traces from the battery to the BIOS are good?

Thanks, I also thought about the BIOS. I'll check out it and battery trades in a couple of hours when I get home. I also found the DRAM Parity switch in the manual, should I pay attention to it?

Huh, I had that 386/486 mess of characters thing when I was playing about with feeding the CPU voltage from an external regulator. I thought I'd messed up the computer but clearing the CMOS got it back to working normally.

How do you have the RTC battery connected. Is it going to the pins where the original rechargeable battery was, or into the 4 pin external battery header?

Thermalwrong wrote on 2023-05-23, 15:10:

Huh, I had that 386/486 mess of characters thing when I was playing about with feeding the CPU voltage from an external regulator. I thought I'd messed up the computer but clearing the CMOS got it back to working normally.

How do you have the RTC battery connected. Is it going to the pins where the original rechargeable battery was, or into the 4 pin external battery header?

Thanks for your reply. It is connected to a 4 pin connector.Sould i connect it in place of the onboard battery? I'll also try another power supply

Thermalwrong wrote on 2023-05-23, 15:10:

Huh, I had that 386/486 mess of characters thing when I was playing about with feeding the CPU voltage from an external regulator. I thought I'd messed up the computer but clearing the CMOS got it back to working normally.

How do you have the RTC battery connected. Is it going to the pins where the original rechargeable battery was, or into the 4 pin external battery header?

So I tried ECC memory (with DRAM parity enabled), different power supply, different bios chip and checked battery tracks. Nothing helped, but: the other bios chip turned out to be PCI 1.0 version (instead of 1.15 without PCI) for the version with a PCI riser, and with it the meaningless characters disappeared, but the board just hung before detecting the drives. As a result, I returned the chip back, put the jumpers correctly and got not 386/486 and a lot of characters, but "DX2-66 (2 random characters)MHz" I also sometimes get a Real Time Clock error, which is quite important in this case, I think. When I have a rest, I'll try to measure the crystal oscillators on the board with an oscilloscope.

I also tried the second board: it sees the lack of memory (1-3-4-1 beep code), but when it is installed, it does not boot. And I just can't find the ISA post card. And, by the way, a few years ago it was flooded with dirty water, which conducts electricity a bit, but six months ago I washed it pretty well. I found some places that weren't fully cleaned, but there's no way to clean them right now.

I forgot to attach photos

crapasanya wrote on 2023-05-23, 15:46:

Thermalwrong wrote on 2023-05-23, 15:10:

Huh, I had that 386/486 mess of characters thing when I was playing about with feeding the CPU voltage from an external regulator. I thought I'd messed up the computer but clearing the CMOS got it back to working normally.

How do you have the RTC battery connected. Is it going to the pins where the original rechargeable battery was, or into the 4 pin external battery header?

Thanks for your reply. It is connected to a 4 pin connector.Sould i connect it in place of the onboard battery? I'll also try another power supply

Nah that looks good, I use the external battery header too. The only reason I asked was because the on-board battery might get charged up while the external won't, so something like a CR2032 is best on the external battery connector.

Thanks for adding the pictures - I've never seen those type of ZIP memory expansions before for the VGA. Have you tested with/without those or do they otherwise work?

Yeah those are non-standard ZIP memory upgrades for PB AFAIK, doubt that would effect boot up but never know.

Thermalwrong wrote on 2023-05-24, 00:52:

crapasanya wrote on 2023-05-23, 15:46:

Thermalwrong wrote on 2023-05-23, 15:10:

Huh, I had that 386/486 mess of characters thing when I was playing about with feeding the CPU voltage from an external regulator. I thought I'd messed up the computer but clearing the CMOS got it back to working normally.

How do you have the RTC battery connected. Is it going to the pins where the original rechargeable battery was, or into the 4 pin external battery header?

Thanks for your reply. It is connected to a 4 pin connector.Sould i connect it in place of the onboard battery? I'll also try another power supply

Nah that looks good, I use the external battery header too. The only reason I asked was because the on-board battery might get charged up while the external won't, so something like a CR2032 is best on the external battery connector.

Thanks for adding the pictures - I've never seen those type of ZIP memory expansions before for the VGA. Have you tested with/without those or do they otherwise work?

Yes, I checked without them, zero difference. I'm most suspicious of the OPTi chipset chip right now, but first I'll check the crystal oscillators with an oscilloscope. I also found a postcard - let's see which post code everything hangs on. By the way, can the cache be the cause of this behavior? I think this is the only thing I haven't tested yet. But I turned it off in BIOS and it didn't help.

bregolin wrote on 2023-05-23, 13:26:

Can you try using the BIOS ROM from the other board into this one? Wonder if the BIOS is corrupted or the ROM chip somehow damaged. Have you checked if the traces from the battery to the BIOS are good?

I've checked what you asked in one of upper messages

Thermalwrong wrote on 2023-05-24, 00:52:

crapasanya wrote on 2023-05-23, 15:46:

Thermalwrong wrote on 2023-05-23, 15:10:

Huh, I had that 386/486 mess of characters thing when I was playing about with feeding the CPU voltage from an external regulator. I thought I'd messed up the computer but clearing the CMOS got it back to working normally.

How do you have the RTC battery connected. Is it going to the pins where the original rechargeable battery was, or into the 4 pin external battery header?

Thanks for your reply. It is connected to a 4 pin connector.Sould i connect it in place of the onboard battery? I'll also try another power supply

Nah that looks good, I use the external battery header too. The only reason I asked was because the on-board battery might get charged up while the external won't, so something like a CR2032 is best on the external battery connector.

Thanks for adding the pictures - I've never seen those type of ZIP memory expansions before for the VGA. Have you tested with/without those or do they otherwise work?

I found a postcard and this is what it shows: during a long POST, it hangs at 38 (3A is the previous one), when it hangs after it before the disks are allocated, it hangs at the d2 code (68 is the previous one), then after 5- 10 seconds the CE postcode appears (d2 is the previous one), then the system freezes completely. By the way, I get strange behavior on the oscilloscope from the crystal oscillator near the IDE connector: random (at least at the first glance) smooth voltage surges. Maybe I'm just measuring wrong?

I tried a motherboard without cache, with a different cache, and all 4 variations of the cache and its settings (old cache + 128k setting, new cache + 128k setting, old cache + 512k setting, new cache + 512k setting) - no difference.

I also connected the postcard to the donor - it is in a cycle between several postcodes. I have no idea at all what the reason is, so I'm attaching a video of what is happening. Maybe we should switch to it? I don't care which board will work.

By the way, I found that the RTC is located right in the chipset. So far the most obvious option for me is to replace the chipset, but this is the most difficult element to replace on this board, so I don't want to rush into it.

I tried to clean the board more thoroughly from dust - no result.
By the way, I forgot to write what the post codes mean: d2 is unknown interrupt error (Calling the interrupt handling procedure from an unidentified source), CE is displaying messages on the screen and initializing the digitizer

Have you tried to write a rom for the ISA riser version in to a new/different ic? or try with the pci riser version bios chip, but with the ide controller mapped to ISA(from the bios)? At 'fresh/cold' start, if you enter the bios and leave the pc like that for ~30 minutes, does it hang? Is the clock working ok? if so then the 32.768 khz crystal is used for RTC in combination with the 82c602 http://www.bitsavers.org/components/opti/data … ices_199411.pdf and it should be ok and not be the cause of your problem.

bogdanpaulb wrote on 2023-05-24, 19:55:

Have you tried to write a rom for the ISA riser version in to a new/different ic? or try with the pci riser version bios chip, but with the ide controller mapped to ISA(from the bios)? At 'fresh/cold' start, if you enter the bios and leave the pc like that for ~30 minutes, does it hang? Is the clock working ok? if so then the 32.768 khz crystal is used for RTC in combination with the 82c602 http://www.bitsavers.org/components/opti/data … ices_199411.pdf and it should be ok and not be the cause of your problem.

Thanks for the answer. No, I just swapped the chips around. The donor board I swapped chip from was once working with a PCI bios and an ISA riser (so it should work), but then it flooded and was left like that for several years. And, although there are no noticeable oxides, this conductive rubbish still remains on it in inconspicuous places. I plan to buy myself an ultrasonic bath or ask to use one in the service

crapasanya wrote on 2023-05-26, 15:08:

bogdanpaulb wrote on 2023-05-24, 19:55:

Have you tried to write a rom for the ISA riser version in to a new/different ic? or try with the pci riser version bios chip, but with the ide controller mapped to ISA(from the bios)? At 'fresh/cold' start, if you enter the bios and leave the pc like that for ~30 minutes, does it hang? Is the clock working ok? if so then the 32.768 khz crystal is used for RTC in combination with the 82c602 http://www.bitsavers.org/components/opti/data … ices_199411.pdf and it should be ok and not be the cause of your problem.

Thanks for the answer. No, I just swapped the chips around. The donor board I swapped chip from was once working with a PCI bios and an ISA riser (so it should work), but then it flooded and was left like that for several years. And, although there are no noticeable oxides, this conductive rubbish still remains on it in inconspicuous places. I plan to buy myself an ultrasonic bath or ask to use one in the service

This is what i meant: you said that with the other bios chip, it hangs at ide detection, some motherboards have in bios the option to change the irq or map the integrated ide controller to the PCI bus or the ISA bus, so you can enter the bios before ide detection and try to change the irq or map the ide controller to the ISA bus if you have the options.

bogdanpaulb wrote on 2023-05-26, 17:12:

crapasanya wrote on 2023-05-26, 15:08:

bogdanpaulb wrote on 2023-05-24, 19:55:

Have you tried to write a rom for the ISA riser version in to a new/different ic? or try with the pci riser version bios chip, but with the ide controller mapped to ISA(from the bios)? At 'fresh/cold' start, if you enter the bios and leave the pc like that for ~30 minutes, does it hang? Is the clock working ok? if so then the 32.768 khz crystal is used for RTC in combination with the 82c602 http://www.bitsavers.org/components/opti/data … ices_199411.pdf and it should be ok and not be the cause of your problem.

Thanks for the answer. No, I just swapped the chips around. The donor board I swapped chip from was once working with a PCI bios and an ISA riser (so it should work), but then it flooded and was left like that for several years. And, although there are no noticeable oxides, this conductive rubbish still remains on it in inconspicuous places. I plan to buy myself an ultrasonic bath or ask to use one in the service

This is what i meant: you said that with the other bios chip, it hangs at ide detection, some motherboards have in bios the option to change the irq or map the integrated ide controller to the PCI bus or the ISA bus, so you can enter the bios before ide detection and try to change the irq or map the ide controller to the ISA bus if you have the options.

Strange. I don't know what happened, but now with the PCI BIOS it hangs on the first reboot with the battery connected at post code 68 (66 is the previous one) with a black screen, and after a time I get 2 meaningless characters and memory errors. Also, I found only the ability to disable the ide and floppy controller, but they are not saved without a battery, and with battery I can not get into the bios at all. By the way, there may have been a misunderstanding: both BIOSes hung before the IDE detection, but on the postcode D2 (unknown interrupt) or 68 (previous before D2, which means the enabling of external and CPU caches).

Thank you all very much for your help. Here's the happy ending: it turned out that the board needs a special battery with a special voltage (around 3.7v), and plus has to be connected to the plus of the built-in battery and minus to the ground (NOT the negative battery contact), in my case i just connected it to case. A regular 18650 li-ion battery works perfectly here.

Finally this PC was completely disassebled and cleaned (even the power supply, CD drive and floppy). Now the case went to a retrobrite (it lay near the window for a long time, so it was not even yellow, but orange), the PC is mostly assembled and I already play on it from time to time. I am happy, another PC of my childhood has been restored.