Anonymous Coward wrote on 2024-05-29, 01:11:

At least we know it belongs to another VOGONS member. I just can't believe you didn't dump the ROM yourself.

I presume you must have played with it before you re-sold it? Did you try the DX4 on it? And if so, did the BIOS report it correctly?

The board was completely non-functional, so he was giving it away, and I am trying to revive it currently.

I have tried several things so far, reflowed some bent legs on the back of the cache and BIOS sockets that looked to have separated from the solder pad, tried running with different sets of cache chips and tried swapping the keyboard controller (had this revive a different 486 board a while back) among other things.

So far, my POST card has shown no codes at all with the original BIOS installed, however, if I drop in some other unrelated 486 AMI BIOS chip I am able to get a POST code of 02 (hard to say what it means... possibly BIOS checksum or CPU related). I am trying it with an Intel DX2 66, which may not be compatible with whatever old 486 BIOS I've been using for testing. So there's a chance the original BIOS is the problem... which is actually unfortunate since that seems to be very hard to find.

I notice that the listing you posted above had two boards sold and the board I have now (that was pshipkov's) is definitely the exact board in the picture, so is there any chance another VOGONS user ended up with the second one? If so, it'd be awesome if they could dump the BIOS and post it here.

Anyway, I was able to dump the BIOS using my TL866A and it does have some readable data in it (similar to the older BIOS posted in this thread), but I have no idea if there is any corruption. If you want to try it on your board, please let me know if it works. If it does, then my board may have some other issue.

In the meantime, I am trying to find another BIOS chip I can flash the older version to just to see if I can get this thing to POST.

EDIT: Tried flashing the Dolch BIOS posted earlier to a different chip and using that in my board but it also does nothing. No POST codes, no beeps.

The POST card is showing that all voltages are present, including -5v, so that's a good sign... but this is a tough one to track down, for sure.

Have also tried a DX2 50Mhz, different RAM and 3 different keyboard controllers.

It's too bad, there doesn't seem to be any listed jumper setting to entirely disable cache on the board (other than switching turbo off), so I can't be 100% certain that it isn't an issue with a cache chip or socket.

Oh, and I did install an unused RTC for testing.

Wow. Thanks for the BIOS dump. I’ll have a look this evening.

By chance are you trying to run your board without the Dallas ramified clock chip? That might explain why it doesn’t light up. I think it should be a DS1387 that goes in the empty socket. Maybe a DS1488

Anonymous Coward wrote on 2024-05-29, 04:53:

Wow. Thanks for the BIOS dump. I’ll have a look this evening.

By chance are you trying to run your board without the Dallas ramified clock chip? That might explain why it doesn’t light up. I think it should be a DS1387 that goes in the empty socket. Maybe a DS1488

I have a DS12887+ in it which should still be good (the date is pretty old but it was brand new out of a tube today).

Please let me know if you are able to boot with that BIOS in your board. That'll rule out a pretty huge possible source of issues for me. 😀

Not all Dallas RTC modules are compatible with each other.
Some boards are super picky. The whole RTC stuff is such a PITA.
With that said - i tried with the entire spectrum of different models, including DS12887+.

Also the entire spectrum of 486 CPU classes.

Btw, what is so special about this particular motherboard (if anything) ?
Early Ozzuneoj indicated something like that.

pshipkov wrote on 2024-05-29, 05:58:

Not all Dallas RTC modules are compatible with each other. Some boards are super picky. The whole RTC stuff is such a PITA. With […]
Show full quote

Not all Dallas RTC modules are compatible with each other.
Some boards are super picky. The whole RTC stuff is such a PITA.
With that said - i tried with the entire spectrum of different models, including DS12887+.

Also the entire spectrum of 486 CPU classes.

Btw, what is so special about this particular motherboard (if anything) ?
Early Ozzuneoj indicated something like that.

I'll be honest, I don't have a lot of experience with Dallas RTC modules aside from just plunking in a new DS12887+ when a board needs one. I got a pile of them new old stock a while ago for a fraction of the normal per-unit price of brand new ones. So far, every board I have put one in has worked, but it is pretty lame that there are different types and they aren't cross compatible. It is certainly possible that this one doesn't work for this board, and maybe the other stuff I did to the board fixed other problems (lots of bad solder joints), so if anyone knows exactly what type of RTC this should use I will try to get some.

As for the board itself, others here can probably offer a lot more insight than I can, but I think what makes it special is that it is EISA + VLB, and the later revision uses 72pin SIMMs and supports 3.3v CPUs which should allow it to use any 486 (except maybe 40Mhz FSB? I don't see that listed...) all the way up to the fastest 5x86 processors. From the standpoint of someone who just wants an early 90s gaming PC, this doesn't make much of a difference. But if you like to swap out hardware and tinker with a variety of things, this is a great combination of features... for example, testing or comparing ISA, EISA and VLB video cards or IDE\SCSI cards in one system, all running with the same CPUs on the same chipset.

There aren't exactly a ton of EISA cards out there, but personally I would like to have a board like this up and running to tinker with them.

Thanks for the notes.
I guess this hardware is entirely for the EISA fans.

A DS1287 or compatible isn’t going to cut it. It must be a DS1387 with the extra ram for the eisa config data.

I had a look at the BIOS in a hex editor. I think it’s probably not corrupt. Unfortunately it looks like the same 1992 core BIOS that usually comes with the SIS EISA boards like my SuperEISA. I didn’t see anything that would indicate support for the DX4. However I saw something about power management, so that might be interesting.

I just did a quick test with your BIOS. It is not corrupted. It works properly. So I guess you're happy, right?
Sadly I'm a little disappointed. It seems like A.I.R. did little to nothing with this BIOS over the standard 6/6/92 release.
The CPU is identified as a 486DX or 487SX. Though, the CPU speed was reported at 100MHz...I think my board just assumes 66MHz if there is clock multiplying on a 33MHz bus.
Compared to my board the advanced settings in the BIOS are extremely limited. It's almost like an OEM BIOS.
There are no power management options that I could find.
The BIOS still doesn't like my Cyrix 5x86.
Fortunately 1024kb of cache is still detected on my board despite the 486EI only supporting up to 512kb.

Probably the only remaining hope for a BIOS upgrade is if the 486EI had Award BIOS as an option, but I suspect if such a thing were to exist it would be similar to the case with the AMI.

Anonymous Coward wrote on 2024-05-29, 14:48:

I just did a quick test with your BIOS. It is not corrupted. It works properly. So I guess you're happy, right? Sadly I'm a litt […]
Show full quote

I just did a quick test with your BIOS. It is not corrupted. It works properly. So I guess you're happy, right?
Sadly I'm a little disappointed. It seems like A.I.R. did little to nothing with this BIOS over the standard 6/6/92 release.
The CPU is identified as a 486DX or 487SX. Though, the CPU speed was reported at 100MHz...I think my board just assumes 66MHz if there is clock multiplying on a 33MHz bus.
Compared to my board the advanced settings in the BIOS are extremely limited. It's almost like an OEM BIOS.
There are no power management options that I could find.
The BIOS still doesn't like my Cyrix 5x86.
Fortunately 1024kb of cache is still detected on my board despite the 486EI only supporting up to 512kb.

Probably the only remaining hope for a BIOS upgrade is if the 486EI had Award BIOS as an option, but I suspect if such a thing were to exist it would be similar to the case with the AMI.

Well, that is mostly good news. I did fix a lot of probably-bad solder joints, and I did get at least some POST codes when using a non-EISA BIOS, so perhaps it's working now and just needs the proper DS1387. I will see if I can order some today. Thank you!!

Also, I'm not sure how much of this applies to your board revision, but do you think there's any chance that changing the CPU type jumpers (as mentioned in the 2.21 manual here) would help get your CPUs running? Maybe the different board revision + BIOS combo interprets the setting differently. I do wonder if there's anything else going on with the little socketed dip20 chip right next to the CPU socket that has a paper label marked EI4120 on mine. Whenever I see something socketed I assume that there's a reason for it (updates, etc.), but I have no idea what this chip would be for since I am unfamiliar with EISA boards in general.

EDIT: Whelp... definitely looks like the DS1387 has been out of production for a while. I can get a couple of them from Aliexpress for under $10 shipped and they'll be here in a couple weeks. Otherwise there is a seller on ebay asking $7.50 each for them but the stupid shipping is almost $6 for me, and the date code on them is from 1998. My understanding is that the internals don't start "running" (using the battery) until it has been used, but that's quite a bit of money to spend on something that old that cannot be recharged.

I just need something that will definitely work for now so I can try to get the board working.

Has anyone tried building one of these to make a DS1387 replacement? I can't really tell if the components exist to make one if I don't already have a dead DS1387.

Cyrix 5x86 is just simply too new and confuses the '92 core AMIBIOS. I've already tried all possible jumper combinations. I confirmed it's a BIOS firmware issue on another board using a 92 AMIBIOS, which I swapped for a '94 MR-BIOS. '92 AMIBIOS only knows about the classic 486s, 286, 386 and Cyrix SLC/DLC. I think I need at least Cyrix DX2 support for the 5x86 to boot without issue. The good news is that am5x86 work in our boards without issue.

It's hard to say what that white labeled chip next to the CPU socket does. It's probably something rather mundane. It's almost certainly a GAL/PAL, and they're usually put in sockets as they tend to be less reliable than standard TTL logic.

The easiest root to DS1387 would be to buy a dead one and do the coincell mod. Amazingly mine is 30 years old and still holds a charge. The kit you linked to is also pretty nice. That would be the best option if you wanted to put in the effort.

Anonymous Coward wrote on 2024-05-30, 01:39:

Cyrix 5x86 is just simply too new and confuses the '92 core AMIBIOS. I've already tried all possible jumper combinations. I confirmed it's a BIOS firmware issue on another board using a 92 AMIBIOS, which I swapped for a '94 MR-BIOS. '92 AMIBIOS only knows about the classic 486s, 286, 386 and Cyrix SLC/DLC. I think I need at least Cyrix DX2 support for the 5x86 to boot without issue. The good news is that am5x86 work in our boards without issue.

It's hard to say what that white labeled chip next to the CPU socket does. It's probably something rather mundane. It's almost certainly a GAL/PAL, and they're usually put in sockets as they tend to be less reliable than standard TTL logic.

The easiest root to DS1387 would be to buy a dead one and do the coincell mod. Amazingly mine is 30 years old and still holds a charge. The kit you linked to is also pretty nice. That would be the best option if you wanted to put in the effort.

I just in case, I will give some newer CPUs a try once I get this thing up and running. It probably won't work, but it's worth a shot with the 2.21 revision board, eh? 😀

One of the newest\fastest I have is a Gainbery 5x86 120Mhz which I think is just a 100Mhz Cyrix 5x86 on an interposer (like this). I also have a couple of Pentium Overdrive 83Mhz chips... not sure if those work with the jumpers set for an "overdrive" CPU, as the manual describes it.

Regarding the DS1387, I just ordered four on Aliexpress for about $12 shipped, which is a fraction of what it would cost to get them domestically, and there is at least some chance that I will end up needing more eventually. I will be very busy the next couple of weeks anyway, so it's not a huge deal that I have to wait for them.

If they ever die, I will definitely do the coin cell mod to them. 😀

POD5V83 works in this board. Well, it POSTs, but I didn't get to fully test to see if both caches were working.

Cyrix 5x86 most likely won't work with the AMI BIOS. I could not get 3x/4x chips working at all. The 2x/3x chips I can get "working" but it involves some hackery. You have to set the interposer multiplier jumper to "2.5x" which is not supported by Cyrix. However, this allows the chip to boot up in 2x mode. Then using a software utility, you can set it into 1X mode, manually change the interposer multiplier jumper to 3X, and then finally set it to 3X in software. You other option is to use 2x50 which requires only using the "2.5x" mode.

Anonymous Coward wrote on 2024-05-30, 03:59:

POD5V83 works in this board. Well, it POSTs, but I didn't get to fully test to see if both caches were working.

Cyrix 5x86 most likely won't work with the AMI BIOS. I could not get 3x/4x chips working at all. The 2x/3x chips I can get "working" but it involves some hackery. You have to set the interposer multiplier jumper to "2.5x" which is not supported by Cyrix. However, this allows the chip to boot up in 2x mode. Then using a software utility, you can set it into 1X mode, manually change the interposer multiplier jumper to 3X, and then finally set it to 3X in software. You other option is to use 2x50 which requires only using the "2.5x" mode.

Yikes... that's quite a process. Definitely sounds like a good candidate for a PODP5V83 then, at least for those times when "maximum performance" is needed. 😅

Any updates on those DS1387s you ordered?

Anonymous Coward wrote on 2024-06-05, 16:44:

Any updates on those DS1387s you ordered?

It will be a while yet before I get them. Unsurprisingly, the June 11th-20th expected delivery date with an on-time guarantee immediately showed June 20th once the order was placed. 🤣

But it's okay, I won't have much time to tinker with things for almost a week after today.

Anonymous Coward wrote on 2021-09-04, 08:45:

Whatever the reason your system can't detect more than 64MB, the BIOS is not the issue. I tested your PAC60 BIOS on my SuperEISA […]
Show full quote

Whatever the reason your system can't detect more than 64MB, the BIOS is not the issue. I tested your PAC60 BIOS on my SuperEISA board and it still counted 128MB.

Unfortunately, the PAC60 BIOS didn't solve my Cyrix 5x86 problem. Even though ELHVB and TH99 claim V1.2 of the 486EI supports 3.3V CPUs, I think it must be a mistake:

https://www.elhvb.com/webhq/models/486vlb3/486ei12.html
http://www.uncreativelabs.de/th99/m/A-B/33191.htm

I think V2.0 is what I need. It's kind of hard to believe it hasn't popped up by now. The 486EI was a pretty popular board.

If I take a look at any of the known revisions below 2.x they lack the JP12. Like with so many other TH99/Microshouse entries, the revision given is more than likely the revision of the manual and not the board itself. Without JP12 on your board (between the CPU and the front edge), there is no way to set 3.3V in the first place.

Since I happen to own elhvb.com I will try and dig in a bit further into this and correct the page where necessary.

Just as a side note:
I once had a AIR 486IEI running with 256 MB. But Himem.sys would only detect 64 MB by default.
It seems that there was a certain bit not set by default to interpret the memory size correctly.
Somebody wrote a small program (just a few byte) to set this bit and the memory was fully available. I tested it with a big ramdisk and also with windows and it worked.
I only remember having graphics corruption when I used that strange ELSA Winner card (EISA on one edge, ISA on the other, each side with a VGA connector). As I had no other, I used an ISA card just fine.
Adaptec EISA SCSI controller worked fine as well.

Unfortunately I don't have the hardware anymore as somebody stole it from our basement.

Anyhow, if there is interest, I can check my old notes I have floating somewhere on my NAS.
If I am not mistaken, the small program should be there as well.

Yay! My order of DS1387 modules from Aliexpress arrived early, I installed one in the board and it's up and running! Since pshipkov said he had tried running it with a DS1387 and it was dead, most likely what fixed the board was reflowing all the broken solder joints on the back. I would say there were at least a dozen legs on cache chips that were bent over to the point that the joint had been compromised. I reflowed probably 40+ joints on cache and BIOS chips and other components.

Skorbin wrote on 2024-06-10, 09:04:

Anyhow, if there is interest, I can check my old notes I have floating somewhere on my NAS.
If I am not mistaken, the small program should be there as well.

If you have a chance to do this, I would definitely appreciate having any other info about the board that you are able to dig up.

Awesome. Glad to see the board up and running. Good Joe.