jakethompson1 wrote on 2020-10-09, 02:16:

There should be three extra jumpers to set for a WB CPU. I believe these connect the WB/WT#, HITM#, and CACHE# pins. You can read a bit about what changes were made to the 486 pinout to accommodate WB it here. http://datasheets.chipdb.org/SGS%20Thomson/x86/486/4193.PDF

Tks Jake!

(I kinda deserve the newbie badge!)

I have found these pages on the manual which would point differences on JP28, 29 and 30 from a standard DX4-100:

Tryed that P24D pin-out (green jumpers)

But still ended up with this:

Any suggestions on wherelse I should poke on this board?

I've also done the BIOS update for the AMI Dec 1995 version (from an Award Bios) but still no juice!

TIA

Weird, and you definitely have a V1.7 board?
I'm using the same jumpers for my Am5x86, except for JP31 which needs the 2x setting which it interprets as 4x.

jakethompson1 wrote on 2020-10-09, 03:39:

Weird, and you definitely have a V1.7 board?
I'm using the same jumpers for my Am5x86, except for JP31 which needs the 2x setting which it interprets as 4x.

Yep! V1.7 here! Just like the one in the 2nd post on this thread!

Maybe a BIOS thing?

TIA,

Ricardo

I'm using the 12/01/1995 BIOS I uploaded earlier in this thread. I think the "X" version out there works too. Even if it were possible to misconfigure the BIOS somehow, I don't think it's possible to do it so badly that it would change the CPUID.
You said the CPU is definitely an &EW right? No missing pins?
Out of ideas otherwise, maybe try an Am5x86 if you can get one. They're faster too!

I did the same for the bios - just used the one that was linked here. As for the &EW, it does says that on the IC. I'd love to try one Am5x86, although they are pretty rare to find here in Brazil.

I'll sleep over this and see if someone also throws in new ideas!

Thanks again!

Wait you have the original manual ? Can you scan it please 😁
The only ressources we have are HTML pages and MTL pages but no original manual scans.

Deksor wrote on 2020-10-09, 08:37:

Wait you have the original manual ? Can you scan it please 😁
The only ressources we have are HTML pages and MTL pages but no original manual scans.

I have the manual for the DX-6900 - which I understand is one of the versions of the M912. This manual was from my board when I was a kid - and I no longer have it. It is NOT the one that came with my current board (ie my board came with an Award BIOS and the Manual has comments on AMI).

I'll scan it through the weekend and post it.

Best,

RC

I should probably scan my manual as well. It was originally for the v1.4 board but there are some inserts and a bunch of notes for the 1.7 board.

The AMI BIOS is better as it actually supports larger HDDs.

I can't get either working reliably with cache enabled with the POD83 or the Evergreen AMD 5x86 133 upgrade.

Waiting on some new EEPROM chips to arrive so I can try the latest BIOS. The AMI one I currently have seems like a slightly older compared to the latest BIOS and I don't have any EEPROM chips that I was able to get to POST with the newest AMI BIOS.

I did find I actually have a second v1.7 board in one of my "spare" systems when I was going through stuff yesterday so I plan on taking that one out and seeing if it has a different BIOS.
That system came to me with a Cyrix DX4-100 installed so I'll be testing with that first.

Another thing.. I haven't been able to get the DUAL BIOS thing working either. It only seems to work with the AWARD BIOS.

Alright fellas,

I've scanned the manual, but the file is larger than the allowed 5Mb here. I was not able to login on the vogonsdriver.com to upload there (I guess it is a different access allowed). I have uploaded it to the "vintage PC Enthusiasts" Facebook group. Still under moderation by the time I'm writing this, BUT it may have been released by the time you read this.

Anywhere else I should upload that would work with the forum?

https://www.facebook.com/groups/VintagePCEnthusiasts/files

As for the scan: I tried my best for having it organized, but I just don't know how to fix the huge 1st page. Tried a couple of tricks but couldn't make it better and my scanner was not very friendly today.

Best,

Ricardo Contieri

I made a improved version that's been ocr-ed and is now available right here http://www.win3x.org/uh19/motherboard/show/573 😀

Just to add to the thread: I couldn't make the board recognize the WB in that processor. And to just found out that I have a relabeled processor hiding the real specs! I got so in doubt of what the processor was, that I decided to remove it. Took the cooler out, the thermal paste with isopropyl alcohol and surprise: a remarked processor!

Already started a dispute with the seller (Mercado Livre, a sort of eBay from Brazil). Probably the seller sourced from China and didn't have a clue about what had in his hands...

So bare with me kids: if your processor looks too nice, you might have been fooled!

Best,

Ricardo

Interesting! I kind of thought of that when I started asking if it could be missing pins but also thought there was no way.
I wonder whether it was remarked recently or back in the 90s when it feels like that might have been more worthwhile...
Are there any telltale signs on the underside that would have told you it's a WT?
Go for an Am5x86 this time? 😁 Wonder why those aren't popular in .br?

jakethompson1 wrote on 2020-10-11, 02:04:

Interesting! I kind of thought of that when I started asking if it could be missing pins but also thought there was no way. I wo […]
Show full quote

Interesting! I kind of thought of that when I started asking if it could be missing pins but also thought there was no way.
I wonder whether it was remarked recently or back in the 90s when it feels like that might have been more worthwhile...
Are there any telltale signs on the underside that would have told you it's a WT?
Go for an Am5x86 this time? 😁 Wonder why those aren't popular in .br?

I have no clue when it was remarked. The looks was great before I cleaned it! And came quite neatly packed. It did look like a pulled IC though.

We were not big fans in AMDs back in the days here in Brazil - we thought they were just the "competitors/cloners" and not the "originals". I guess that Intel was better at branding their products around here than AMD.

And to top that: I'm trying to rebuild a computer I had when I was in my teens.

Cheers!

I have a bios release date 12/02/1995D
Ami date 07/25/1994

The IBM Cx86C 100 that I have shows as 66mhz. I can't set it to 100mhz, won't go past boot screen and freezes.
Speedsys gives a score of 31.30, and 39.22 for Intel Dx4 100.

But is the one that has fake cache. I'll try to save it.

Nexxen wrote on 2020-10-11, 14:55:

I have a bios release date 12/02/1995D Ami date 07/25/1994 […]
Show full quote

I have a bios release date 12/02/1995D
Ami date 07/25/1994

The IBM Cx86C 100 that I have shows as 66mhz. I can't set it to 100mhz, won't go past boot screen and freezes.
Speedsys gives a score of 31.30, and 39.22 for Intel Dx4 100.

But is the one that has fake cache. I'll try to save it.

I tried an ami flash utility (ver 895) but the chipset is not supported.
Wich utility should I use to have it work?
Thanks

Are you sure you have an EEPROM and not just an EPROM on the board? All of the ones I have came with EPROM.

I ordered a bunch of EEPROMs and just waiting for them to arrive from China.

cyclone3d wrote on 2020-10-11, 21:01:

Are you sure you have an EEPROM and not just an EPROM on the board? All of the ones I have came with EPROM.

I ordered a bunch of EEPROMs and just waiting for them to arrive from China.

I have no idea of what I have.

I'm short on other things and I'll get back to it in a few days. I hope that putting 256kb o real cache will actually work. After 25 years I don't sweat it and I'm happy things are sorting out well.

Thanks in advance.

Nexxen wrote on 2020-10-11, 14:55:

I have a bios release date 12/02/1995D Ami date 07/25/1994 […]
Show full quote

I have a bios release date 12/02/1995D
Ami date 07/25/1994

The IBM Cx86C 100 that I have shows as 66mhz. I can't set it to 100mhz, won't go past boot screen and freezes.
Speedsys gives a score of 31.30, and 39.22 for Intel Dx4 100.

But is the one that has fake cache. I'll try to save it.

Have you checked the multiplier jumpers? Looks like they are set for 2x and not 3x.

As for the BIOS, maybe you will need to use a recorder such as the TL866 and get new EEPROMs, as that generstion of machines still didn't take advantages from the flash memory technology back then.

Best!

Just browsing this thread but could not fully understand that can a board version 1.4 be BIOS updated to support 5x86?

Yes, I have done that and amd 5x86 works on my rev 1.4