Found the motherboard manual so leaving the link here just in case its useful for anyone else - https://theretroweb.com/motherboard/manual/32947.pdf

CRG wrote on 2023-06-01, 10:21:

Found the motherboard manual so leaving the link here just in case its useful for anyone else - https://theretroweb.com/motherboard/manual/32947.pdf

As your dealing with a board with EISA slots you'll need a suitable EISA configuration utility (ECU) and the correct configuration file that the ECU can process for the board - is there any info on the screen re. model number when the system powers up (pics might help). Currently I can see at least two possible configuration files which seem to match your board layout but have so far been unable to track down the ECU that ICL would have used.

PC Hoarder Patrol wrote on 2023-06-01, 16:45:

CRG wrote on 2023-06-01, 10:21:

Found the motherboard manual so leaving the link here just in case its useful for anyone else - https://theretroweb.com/motherboard/manual/32947.pdf

As your dealing with a board with EISA slots you'll need a suitable EISA configuration utility (ECU) and the correct configuration file that the ECU can process for the board - is there any info on the screen re. model number when the system powers up (pics might help). Currently I can see at least two possible configuration files which seem to match your board layout but have so far been unable to track down the ECU that ICL would have used.

Sorry for the late reply. I've now got the board to post although if I try to boot to floppy it quickly crashes. Equally I haven't been able to get it to boot to HDD and to be honest I think that as you suggested I need the configuration disk.

I've attached a copy of the post screen although it doesn't provide much info.

Alternatively if it is of any interest my video on this machine has just went live, if you skip to around 21:50 you'll see the post screen as it comes up or if you skip to around 23:00 I take a look through the bios even if it is very limited.
https://youtu.be/aNQB2ME7hsM

The crash issue I'm seeing when trying to boot reminds me of a memory issue and I've been told that the configuration disk may let me set options such as memory timings. I suppose its also possible the ram module I have fitted isn't fully compatible with this board but my stock of parity 72pin simms is very limited.

This should be compatible EISA configuration utility

https://web.archive.org/web/20020917202721/ht … ml/7000_dl.html

CRG wrote on 2023-06-04, 13:41:

Sorry for the late reply. I've now got the board to post although if I try to boot to floppy it quickly crashes. Equally I hav […]
Show full quote

Sorry for the late reply. I've now got the board to post although if I try to boot to floppy it quickly crashes. Equally I haven't been able to get it to boot to HDD and to be honest I think that as you suggested I need the configuration disk.

I've attached a copy of the post screen although it doesn't provide much info.

Alternatively if it is of any interest my video on this machine has just went live, if you skip to around 21:50 you'll see the post screen as it comes up or if you skip to around 23:00 I take a look through the bios even if it is very limited.
https://youtu.be/aNQB2ME7hsM

The crash issue I'm seeing when trying to boot reminds me of a memory issue and I've been told that the configuration disk may let me set options such as memory timings. I suppose its also possible the ram module I have fitted isn't fully compatible with this board but my stock of parity 72pin simms is very limited.

Good video...watched it all with plenty of detail.

Hopefully @bitterlemon is right re. this Acer J3 motherboard https://theretroweb.com/motherboards/s/acer-j3 , as one of the ICL .cfg files does reference the ErgoPRO E5 mentioned in the aka list.

I've attached the ICL .cfg files which seem to match - !ICL00C1.CFG seems most likely, but !ICL0081.CFG and !ICL0101 do have a matching board config and all cover models from the TeamServer Ei series

bitterlemon wrote on 2023-06-05, 00:15:

This should be compatible EISA configuration utility

https://web.archive.org/web/20020917202721/ht … ml/7000_dl.html

Good shout on the possible ECU

Just saw your video as well, great stuff.

In addition to creating a boot disk with the ECU + .cfg files as mentioned above, I strongly suggest modding that DS1387 chip to add a 3V coin cell battery. Unlike a DS12887, its purpose is not only RTC + CMOS configuration data but it also stores the EISA configuration you set using the ECU. If the battery is low/depleted, each time you power up the machine the config data may be in an invalid state, which can cause all kinds of errors like the ones you're experiencing.

Good luck with that board, I've been tinkering a lot with EISA lately as well so this is really interesting!

Thanks for the info. That is 100% the board I have so fingers crossed I can now get it going.

As for that dallas chip, I didn't know it it stored configuration data like that so yeah its probably best to just modify it.

In terms of the boot disk, its it just a simply MS-DOS boot disk with the utilities on it or do I need something in particular?

CRG wrote on 2023-06-05, 09:09:

Thanks for the info. That is 100% the board I have so fingers crossed I can now get it going.

As for that dallas chip, I didn't know it it stored configuration data like that so yeah its probably best to just modify it.

In terms of the boot disk, its it just a simply MS-DOS boot disk with the utilities on it or do I need something in particular?

You don't really want the full Acer Altos boot disk, just the ECU app (CF.EXE), its help file (CF_HELP.HLP) and your 3 ICL .cfg files on a bare MS-DOS floppy boot disk, so I've extracted the 2 Cf files, attached below.

Problem remains that you still need to properly boot the system with this disk to use the ECU which atm you can't (the flat Dallas may have some part in that)

I have nothing to contribute but to say it was a great video!

PC Hoarder Patrol wrote on 2023-06-05, 03:54:

bitterlemon wrote on 2023-06-05, 00:15:

This should be compatible EISA configuration utility

https://web.archive.org/web/20020917202721/ht … ml/7000_dl.html

Good shout on the possible ECU

@PC Hoarder Patrol Thank you, I hope we helped @CRG enough so we Can watch Part 2 🙂.
Recently, I had some experience with Acer mobos which spiraled into various dead ends. That's how I learned how and where to find stuff. As far as I understand this motherboard is not made by Acer, instead it's made by Atlas Systems - company later bought by Acer. Acer offered it to various computer manufacturers, including ICL. I could be wrong here, so take it with a bag of salt.

I just watched the video. In my opinion, replacing the battery in the Dallas and running the ECU (though both of these things will ultimately need to be done), will not correct the crashing observed. EISA ECU boot disks always boot into some flavour of DOS before running the configuration utility, so there is no reason it should not be able to do this prior to configuration.

Based on the observed behaviour, I'd wager the problem is one or more bad cache chips. I would suggest physically removing them and trying again.

So quick update and I'm now fairly sure there is indeed bad cache, I've been able to get it to boot but I had to replace the bios first with the ACER one from the link provided above (https://theretroweb.com/motherboards/s/acer-j3). This bios has far more options and allowed me to disable the cache, interestingly though so far it'll only boot with BOTH the internal and external cache disabled so it is rather slow but at least I now am at a command prompt and can run the EISA software.

That itself raises another problem though as the config file kindly provided above now no longer matches the vendor ID which I assume is being read from the bios, the software now thinks the machine is an acer but at least its progress!

One good bit of news is that the dallas chip seems to be charging. It initially only held the clock data for a few minutes but when I powered it up last night it must have been 24 hours and it was still good. I didn't think those chips had rechargeable batteries in them?

CRG wrote on 2023-06-06, 08:22:

So quick update and I'm now fairly sure there is indeed bad cache, I've been able to get it to boot but I had to replace the bios first with the ACER one from the link provided above (https://theretroweb.com/motherboards/s/acer-j3). This bios has far more options and allowed me to disable the cache, interestingly though so far it'll only boot with BOTH the internal and external cache disabled so it is rather slow but at least I now am at a command prompt and can run the EISA software.

That itself raises another problem though as the config file kindly provided above now no longer matches the vendor ID which I assume is being read from the bios, the software now thinks the machine is an acer but at least its progress!

One good bit of news is that the dallas chip seems to be charging. It initially only held the clock data for a few minutes but when I powered it up last night it must have been 24 hours and it was still good. I didn't think those chips had rechargeable batteries in them?

Can't you restore the ICL bios now its booting (please tell me you backed up the ICL one!)

In the EISA world the three letter vendor EISA ID is embedded to SLOT 0 (ie the mainboard itself) which the ECU uses to scan for a suitable .cfg file. After the bios update it now thinks its an Acer system, but which one as the Retroweb link doesn't say.

If you can't restore the ICL bios all I can suggest is make a full copy of the Acer Altos ECU boot disk (which contains a load of Acer .cfg files) and boot with that and re-run the ECU, though even that may not work if it doesn't have the correct .cfg file.

Also, pretty sure the Dallas is not rechargeable

PC Hoarder Patrol wrote on 2023-06-06, 09:54:

Can't you restore the ICL bios now its booting (please tell me you backed up the ICL one!) […]
Show full quote

CRG wrote on 2023-06-06, 08:22:

So quick update and I'm now fairly sure there is indeed bad cache, I've been able to get it to boot but I had to replace the bios first with the ACER one from the link provided above (https://theretroweb.com/motherboards/s/acer-j3). This bios has far more options and allowed me to disable the cache, interestingly though so far it'll only boot with BOTH the internal and external cache disabled so it is rather slow but at least I now am at a command prompt and can run the EISA software.

That itself raises another problem though as the config file kindly provided above now no longer matches the vendor ID which I assume is being read from the bios, the software now thinks the machine is an acer but at least its progress!

One good bit of news is that the dallas chip seems to be charging. It initially only held the clock data for a few minutes but when I powered it up last night it must have been 24 hours and it was still good. I didn't think those chips had rechargeable batteries in them?

Can't you restore the ICL bios now its booting (please tell me you backed up the ICL one!)

In the EISA world the three letter vendor EISA ID is embedded to SLOT 0 (ie the mainboard itself) which the ECU uses to scan for a suitable .cfg file. After the bios update it now thinks its an Acer system, but which one as the Retroweb link doesn't say.

If you can't restore the ICL bios all I can suggest is make a full copy of the Acer Altos ECU boot disk (which contains a load of Acer .cfg files) and boot with that and re-run the ECU, though even that may not work if it doesn't have the correct .cfg file.

Also, pretty sure the Dallas is not rechargeable

Back... up... errr...

Yes I backed up the ICL one and I tried going back to that after its booting to no avail. I get the impression there is a cache fault, the ICL bios has the cache enabled with no option to disable, the ACER one does let me turn it off. I made a mistake in the first video, the datasheet for the motherboard and cpu card is incorrect in terms of jumper settings. On closer inspection there is a table printed on the card and I don't think its possible to turn off the cache using jumpers only. Even just removing it crashes the ICL bios, not tried this with the acer bios yet.

It did give me the file name of the cfg file it's looking for, I just can't remember it off the top of my head but if there is a an Acer boot disk that may resolve the issue.

And yes I too was sure those chips weren't rechargeable but for whatever reason this one is now retaining its clock information.

There is a slight chance its the power supply, running with cache disabled is much slower thus smaller power demand. How is 5V ripple?

rasz_pl wrote on 2023-06-06, 11:21:

There is a slight chance its the power supply, running with cache disabled is much slower thus smaller power demand. How is 5V ripple?

I'll check it although I've used the same PSU to run other machines with no problems. It does make me wonder though, the board has a secondary power connector on which I've only connected 1 of the molex leads, maybe I should connect the other as well which is delivering 5v. Its all connected on the board though, all the same 5V rail.

I dont know the purpose for second power connector, those only started being necessary around Pentium 4 😀
Another idea - CPU card might rely on external ground? Maybe inside the case its screwed into a solid bar going across, and here just plugged in the air missing that ground makes interferences around CPU too big?

rasz_pl wrote on 2023-06-06, 13:11:

I dont know the purpose for second power connector, those only started being necessary around Pentium 4 😀
Another idea - CPU card might rely on external ground? Maybe inside the case its screwed into a solid bar going across, and here just plugged in the air missing that ground makes interferences around CPU too big?

The CPU card is held in place by a rail but crossing the case but there is a plastic mould on it with a slot that positions the CPU card so there is no connectivity from the CPU card to the case.

I'm going to try an alternative set of cache chips that I have to see if that resolves the issue.

It's almost there, does now boot and I've discovered a cache error but I'm waiting delivery of parts to try and resolve that.

Latest video on it is live, give it a look if you're interested.

https://youtu.be/BVxUnFwFdWk