Bravo! I'm glad to see it live again. What do you plan to do with it? Fair warning, 9x isn't really what you want on this platform, and if you're going NT4, I'll suggest keeping the hardware roster simple. There might be a Plug'n'Play service, but don't let it fool you, it can't configure resource allocation at all, and since these boards' BIOS has very weak resource configuration, I ended up having to ditch it for 2000. I'm still in the process of setting up, but it's shaping up to be fun. I'm going to have to trim it down quite a lot. Maybe run litestep instead of Windows Explorer, maybe dig into the services... A PPro @ 233 and 128MB of RAM aren't exactly a lot for 2000.

NT4 or 2k will be good choices for that board. and also, you should add (Solved) to the end for the post's name.

Congratulations! I knew it would live all along 🤣

feipoa wrote:

I should get some of that green 32 AWG wire that I see some manufactures use on motherboard for fixes. It is really inconspicuous. At least I think it is 32 AWG. The white kynar cladded wire shown in the photos is 30 AWG and the wire on MB fixes I've seen appears smaller.

That ripped trace/pad can still be restored to the point, that no one would ever tell it was damaged. For that you'd have to desolder the chip again and solder a trace, borrowed from any PCB (the point is, it must be flat), on top of the existing one.

Should look something like this:

Off topic: This whole BIOS boot block recovery business, gave me an idea of how it would be possible to test a motherboard (for basic functionality) w/o it's original BIOS. I was succesfull in testing one motherboard, for which I was unable to get its original boot block since 2014! At least, now I know it works (though I think MB doesn't see RAM at all) 😀

Sedrosken wrote:

Bravo! I'm glad to see it live again. What do you plan to do with it? Fair warning, 9x isn't really what you want on this platform, and if you're going NT4, I'll suggest keeping the hardware roster simple. There might be a Plug'n'Play service, but don't let it fool you, it can't configure resource allocation at all, and since these boards' BIOS has very weak resource configuration, I ended up having to ditch it for 2000. I'm still in the process of setting up, but it's shaping up to be fun. I'm going to have to trim it down quite a lot. Maybe run litestep instead of Windows Explorer, maybe dig into the services... A PPro @ 233 and 128MB of RAM aren't exactly a lot for 2000.

I'm probabbly going to set it up to run Win98SE (or 95c), NT4, and W2K. I have four AAR-2400A RAID cards sitting in a box (they were free) that I'll probably use with it, just to diversify the types of controller cards are in my systems. https://www.ebay.com/itm/123470463883 . The case has space for 4 hard drives, so I figured I might setup a noisy IDE RAID system. User Kixs mentioned that his motherboard works with the 1 M Pentium Pro chips at 233 MHz, so I'll shoot for that. Going for 256 MB ECC EDO RAM (you use parity and the logic turns it into ECC). If its too slow in ECC mode, then might just use parity.

SSTV2 wrote:

Congratulations! I knew it would live all along :lol:

I'm tempted to try updating the Batman's Revenge board now.

SSTV2 wrote:

That ripped trace/pad can still be restored to the point, that no one would ever tell it was damaged. For that you'd have to desolder the chip again and solder a trace, borrowed from any PCB (the point is, it must be flat), on top of the existing one.

I had considered using a strip of copper tape (though cutting off a donar trace might be easier), but was hesitant because of the added complication and possibility that the EEPROM hack would not work. If I was 100% confident that the EEPROM would have worked, I may have put more effort into doing a proper trace repair. These traces are really tiny and tightly packed. Perhaps for someone with steadier hands. Also, isn't there the risk that solder point under the EEPROM would reflow away from its point of contact, toward the heat source which is touching the EEPROM pin during reflow? Is it kosher to place an un-cladded conductor across other traces? Would have to ensure there weren't any scratches. I should mention also that the EEPOM solder pad just to the right of the damaged pad was also damaged and I had to straighten it, so the spacing between those two pads is a lot closer than it is normally (its somewhat arched). I'd also be concerned that the thickness of the trace would be enough to make the EEPROM pins not be in contact with the pads for the reflow to work well on much of that side of the chip. I weighed all this in and went with the an external approach. Its not a rare high-value motherboard by any stretch. I got it for $30.

Speaking of boot block, I wonder what Intel changed in the .18 boot block code for this board? And does the recovery mode replace the previous boot block? I don't recall now which of the 3 boot blocks I put on the EEPROM.

EDIT: Here's some official track patch work done by Acer on a V20T motherboard.

feipoa wrote:

Also, isn't there the risk that solder point under the EEPROM would reflow away from its point of contact, toward the heat source which is touching the EEPROM pin during reflow? Is it kosher to place an un-cladded conductor across other traces?

No, if soldering point would be as far as I've marked it, then the trace would not desolder and drift away, while soldering IC, you can also grease it with epoxy and glue in place. The trace can be uninsulated, it won't short with other traces underneath it, the solder mask prevents it.

feipoa wrote:

Speaking of boot block, I wonder what Intel changed in the .18 boot block code for this board? And does the recovery mode replace the previous boot block? I don't recall now which of the 3 boot blocks I put on the EEPROM.

There probably is an option somewhere in the BIOS update utility, that lets you select whether you wish to update the boot block, otherwise it remains untouched, but when MB is in recovery mode, maybe it gets rewritten?

feipoa wrote:

I'm probabbly going to set it up to run Win98SE (or 95c), NT4, and W2K. I have four AAR-2400A RAID cards sitting in a box (they were free) that I'll probably use with it, just to diversify the types of controller cards are in my systems. https://www.ebay.com/itm/123470463883 . The case has space for 4 hard drives, so I figured I might setup a noisy IDE RAID system. User Kixs mentioned that his motherboard works with the 1 M Pentium Pro chips at 233 MHz, so I'll shoot for that. Going for 256 MB ECC EDO RAM (you use parity and the logic turns it into ECC). If its too slow in ECC mode, then might just use parity.

Good choices! I run an SL25A 1M PPro 200 chip at 233 myself. I poked around with 2000 on there and it ran surprisingly well, but I did end up reverting to NT4 as games like Quake III Arena run significantly worse on 2k probably due to the lackluster Banshee drivers I had for it. I've got it dualbooted with MS-DOS 7.1 on a 2GB FAT16 partition for booting with a FAT32 data partition taking up the remainder of the drive. With the FastFat32 driver it Just Works (TM). I was a bit annoyed with how it played musical drive letters once I re-enabled the IDE channel that had my zip and DVD drives after getting it up to date (to prevent the zip drive stealing C:) but at least it settled where it's at... Having my programs drive be E: on NT but D: on DOS is annoying but I'll get over it. It's a shame the disk utility in NT 4 is more of a suggestion than anything.

As for the I/O conflicts I was getting -- my SCSI card is ISA, non-PnP, no boot ROM. The Adaptec AVA-1502E, also known as the EZSCSI apparently, and it works with the NT driver sparrow.sys. I set its resources via jumpers, under 98 and 2000 I could get it to play ball with everything else, but without a device manager my ability to do the same on NT is at the least very hampered and totally untenable at worst. I could buy another ISA SCSI controller, forget SCSI entirely, or try to figure this mess out. My current situation is that I can have SCSI or my video driver properly loaded -- one or the other -- not both. It's a shame the PnP configuration in the BIOS is so weak or I'd try to solve it there.

Edit: It would figure that the one setting I hadn't tried would be the one to work. The SCSI host adapter needed to be on IRQ11. I got hinted this by checking the configuration for the driver under the SCSI adapters control panel. Then, setting it so, and moving the reserved IRQ for ISA from 10 to 11 for the card, it magically worked. And with that I'm done ironing out major issues with this install -- now just to finish dialing in exactly how I want this to run and what I want it to run. I'm actually editing this post from the PPro machine, running NT4, through Opera 10.63.

Do you use the SCSI AVA-1502E just for the CD-ROM drive? I typically use 1522 or 1542 ISA SCSI cards. I'm probably going to use the onboard IDE for the CD-ROM.

Drive letter shuffling with multi-boot systems are a common annoyance.

I have found that certain cards in certain system seem to prefer one IRQ over the other, even if all IRQs tried are available.

I remember when I tried upgrading my dual PII-400 to W2K from NT4 in the year 2000 and thinking "oh dear, this is a lot slower." I ultimately downgraded back to NT4 until around 2003.

I'll considering using a Diamond Stealth III S540 (S3 Savage4 PRO) graphics card in the system, again just to diversify the systems. What would be even more exotic might be a 3Dlabs oxygen VX1. Seems like 3DLabs card were more workstation oriented, just as the PPRO was. Might make for a nice union, though game support will suffer.

I use the SCSI card I have just for external SCSI peripherals, it has no internal header. Namely I have a Jaz drive I use it with. I ultimately got it working. Now I'm playing around with games and stuff on there.

Ah yeah, I recently did the Intel recovery procedure for a different 440. Took a lot of tries. I had to make a Win98 boot disk, and extract the files from the (very hard to find) BIOS download. It wasn't at all obvious that the system expected the BIOS (I think an EXE) to be extracted and placed on the boot disk.

Failure = drive light remains on, nothing happens
Success = one beep (initialization), then two beeps (completion).

Yeah, need to extract the download*.EXE , then extract the BIOS.EXE - a double extraction.

I don't recall hearing any beeps, but it worked.

It looked pretty bad... but you managed to fix it. Great work 😁

kixs wrote:

It looked pretty bad...

SSTV2 was by far the most optimistic in the thread. Whether it worked or not, at least I'll have learned something new that interests me.

I think I'm going to try it out on the Batman's Revenge board.

Well done!

- Brief update -

I was able to update my Batman's Revenge board, which had the Gateway BIOS of 1.00.10.AF2T, to the Intel BIOS 1.00.13.AF2. The BIOS update utility, FMUP or BIOSUP, would not flash this when using the program from within DOS or a DOS boot disk. It would complain that the flash was corrupt or incompatible. I first had to downgrade the BIOS using an official Intel boot recovery disk, which had version 1.00.08.AF2 on it (file name is BATREVRD). You must set the recovery mode jumper when you do this or you'll garble your BIOS otherwise. Once you've downgraded your BIOS to an official Intel BIOS, you can next upgrade to 1.00.13.AF2 using FMUP.EXE, which is supplied on the Intel BIOS update download.

Sorry for kicking this old topic up but I have a similar yet slightly different issue with the same motherboard.

In my case the board shows no post codes at all (just four times a minus sign) and the BIOS does not feel warm at all.

I did check if the board provides a pwrgood signal and it does so all should be alright. I also checked voltages on the socket and they seem fine.

Is there any way to check if the BIOS starts? I only have basic tools and no oscilloscope.

Hello.

I have the same board. with the same issue..

No boot, no beeps, no codes in the diagnostic card.. basically, a dead board.

Could you burn a new bios chip for us ?

I'd have to re-read the whole thread to figure out what I did. Couldn't you follow the procedure laid out as well? EEPROM programmers aren't very expensive.

Iirc this board uses a soldered TSSOP flash so using it in a programmer is quite non-trivial

It is trivial. download/file.php?id=66865&mode=view The adaptors come in a kit with the programmer.

The hardest part would be desoldering and resoldering, but they'd have to do this anyway.

Oh nice. I got some PLCC and SOP adapters with mine but no TSSOP