I've recently received a bundle of old hardware parts and among it were two AST 486 motherboards (AST Premmia - with EISA slots and Bravo
LP), equipped with integrated SX-33 CPUs. Unfortunately, both motherboards had all electrolytic capacitors, some wires connecting ICS on
PCB (AST Premmia M/B) and jumpers stripped down. As far as i know, both M/Bs were working before the stripping. I've been searching for
high ress photos of these motherboards everywhere, but couldn't find any, even bad quality ones, therefore, I have no other choice, but to ask
for help on this retro computing forum.
At this moment i need to figure out:
* capacitor values (I think both motherboards should be using 22uF caps, but prove me wrong);
* missing wiring on Premmia motherboard.
I've been working on these motherboards on and off for the past six years, but managed to restore (partially) just the Bravo LP motherboard. It does POST, everything works as intended, except for the video, which almost always outputs wrong colors. I suspect a bad RAMDAC, because other than colors being off, integrated video card causes no other issues.
The Premmia board is still in non-operational condition, though, thanks to one helpful guy on Ebay, I was able to figure out where the missing patch wires were supposed to go. He was selling a whole system about 2 years ago and I was lucky enough to contact him before the listing had ended (stumbled upon it by an accident), it was THE only time I had seen a full AST Premmia system for sale. I had explained my situation to him and offered cash for the BIOS ROM dump, but he was not tech savvy with the old computers and had absolutely no idea how to make system run and dump BIOS (PC had a leaky battery and needed BIOS reconfigured to boot). That was my closest chance of getting BIOS dump for the AST Premmia motherboard so far.
The situation: Premmia board came with BIOS chip missing and the best part is that AST BIOS ROM update files for the 486 boards (and probably the rest) did not contain the boot block update portion (BIOS update files can be found here). The most important part of the BIOS, in the BIOS update file, is missing. It is responsible for the chipset's initialisation and BIOS recovery when the main block gets corrupted. It contains a program which loads data from BIOS update diskette and flashes BIOS ROM from E000:0000 to E000:BFFF and leaves FC00:000 - FC00:3FFF untouched (boot block starts at FE00:0000). I know this because I had dissected AST Bravo LP's BIOS and disassembled its boot block. Bravo's boot block is not compatible with Premmia's Ti TACT84500 chipset at all.
I know that Premmia board is alive, I had written and flashed a small program into ROM chip, which was supposed to init 8254 timer and make PC speaker beep, it worked 😀
If there are any Vogoners or just about anyone reading this post who owns an AST Premmia 486 PC or just a mainboard with BIOS chip still in place, please dump its BIOS using P. Trauner's "DUMPAT" program and post it here/contact me. I had tested this program with Bravo LP motherboard and the generated file's content matches BIOS dump created using an external programmer 100%. This program will generate a 1MB file, BIOS should be at the very bottom.
I'd recomend to backup BIOS for every AST motherboard out there using this program, because once the ROM chip or boot block gets damaged, you'll end up in my shoes.
I'd be thankful If you could upload high-res pictures of LP motherboard from both sides. There is one suspicious area with kapton tape stuck under that motherboard, which might have had a patch wire once.
I just unearthed a Premmia 4/33 the other day and came upon your post while I was looking for more info. I've pulled the BIOS chip to make a copy, but unfortunately the lettering has been rubbed off; from what little I can still read, I'm assuming it's a PT28C020. Attached is the dump with this assumption.
Also attached are some pictures that might be helpful. Sorry, I'm not very good at taking pictures.
Please feel free to reach out if you need anything else.
I cannot BELIEVE my eyes... After six years of searching and numerous attempts to reverse engineer and rebuild a Bravo LP BIOS boot block, an original AST Premmia BIOS finally came to light! The BIOS image you have uploaded here is the only sample for the Ti TACT84500 chipset on the net so far!
I've just checked your BIOS image dump and it looks like a real deal (except for size, it should be 128KB), the main BIOS part is of a 2.02 version and the boot block seems to be intact. The original BIOS chip should be an intel 28F001BX-T I think, at least such chip was used on an AST BRAVO MB.
I'm going to set up and test Premmia board this evening as soon as possible.
By the way, I've noticed that some wire patches are wired differently on your board, what mfg. date code is stamped on it?
There are 4 patch wires on my board, two on the back and two on the front. Only one is in the video area. I've attached close-ups of those, and can provide alternate angles/details if needed.
At first I was stumped, because once board got populated with the newly provided BIOS image, it acted as if BIOS were not there. Motherboard did not emit any beeps and POST card showed absolutely no BIOS activity, though, a customized Bravo boot block still made POST card output at least something, so I knew that BIOS chip and its socket were not to blame, some more serious HW fault was causing BIOS not to init properly...
After a thorough search I've identified two main problems with the board, a damaged resistor pack RP17 had a chipped corner, which separated a +5V power suply to the rest of the resistors in the pack (not a fatal issue) and a loose pin of a U72 chip. Once those two issues were rectified, board came to life at the first power-on 😀
There are still some unanswered questions regarding wire patches, I'll try to compile all the info I have about them and post my findings here.
Hinoserm, I can't thank you enough, this board is pretty rare/obscure and frankly speaking, I wasn't expecting to see this BIOS dump any time soon, but now a 486 board with an unique EISA chipset can kick-butt once again 😀
There are 4 patch wires on my board, two on the back and two on the front. Only one is in the video area. I've attached close-ups of those, and can provide alternate angles/details if needed.
No sweat, photos were not mandatory, I have figured it all out 😉
Am sure you already looked into it but supposedly there is a v2.03 bios (05508313.EXE) but all mirrors and files found so far create an unusable disk AFAIK.
Many of the other disk images (inside the .EXE are .IMZ) when found are actually openable with WinImage. http://web.archive.org/web/19970719190151/htt … bls.htm#Premmia
Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun
The chassis of the computer is stamped 03-12-1993, but I couldn't see any markings on the PCB to indicate what date/revision the motherboard might be -- I'll try to upload some high-res photos a little later and maybe you can figure it out. Mine does not have the integrated 486SX CPU (there's the spot for it, but not populated) and instead has a CPU installed in the socket.
In the past when I've played with EISA I needed a special setup disk to configure it -- does this machine have that? Does anyone know where to find the disk? I'm assuming I'll also need Windows 3.11 drivers for the onboard video.
Since someone mentioned there's a BIOS update, if you manage to get that working, I would appreciate a new image of the BIOS chip so that I can update mine without overwriting the original chip -- it would be more convenient for me to burn a new one.
I opened the self-extracting archive in 7Zip, extracted the DISK.IMG, and used WinImage to write it to a diskette. After that, update was performed through the BIOS setup utility. There were no issues.
Am sure you already looked into it but supposedly there is a v2.03 bios (05508313.EXE) but all mirrors and files found so far create an unusable disk AFAIK.
Many of the other disk images (inside the .EXE are .IMZ) when found are actually openable with WinImage. http://web.archive.org/web/19970719190151/htt … bls.htm#Premmia
I keep forgetting about the power of archive.org 😀 There is a, mostly working, AST support site on it, which hosts pretty much everything for AST computers, EISA files included, though I'm not sure if those are neccessary for this MB.
I opened the self-extracting archive in 7Zip, extracted the DISK.IMG, and used WinImage to write it to a diskette. After that, update was performed through the BIOS setup utility. There were no issues.
-Hinoserm
According to software bulletin "SB0139.TXT", BIOS version 2.03 added support for "new chipset ATI LX HJ", which is supposed to be a "MACH32-LX" chip. I think they just changed the video BIOS with that release, video benchmarking results should tell if that upgrade is worth it.
Regarding wire patches - there are three main differences between our boards, your MB lacks a wire patch between U56 pin 76 and RP27 pin 6 and a patch that links U26 pin 2 with pin 10, also, your board has a link between U19 pin 9 and U39 pin 6, which my board lacks. Perhaps link U56->RP27 is formed on the bottom side of your board?
There is still something wrong with that board, it won't show any signs of life with an external CPU installed, CPU does heat up, but MB won't fetch even a single instruction from the BIOS chip, hmm...
Am sure you already looked into it but supposedly there is a v2.03 bios (05508313.EXE) but all mirrors and files found so far create an unusable disk AFAIK.
Many of the other disk images (inside the .EXE are .IMZ) when found are actually openable with WinImage. http://web.archive.org/web/19970719190151/htt … bls.htm#Premmia
I keep forgetting about the power of archive.org 😀 There is a, mostly working, AST support site on it, which hosts pretty much everything for AST computers, EISA files included, though I'm not sure if those are neccessary for this MB.
Thanks ! I looked at ari-service.com before but hit some road blocks. Your link at least lets me browse around. Bookmarking it in my AST archives. Thanks again !
Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun
Completely got rid of long patch wires. Wires that were connecting U56 with U57 and U75 got replaced with two 33 Ohm micro SMD resistors (added for series termination), resistors were soldered on existing traces, that were severed during PCB production. A trace between U56 and RP27 was also severed during PCB production and later re-patched with a wire, thus trace only needed to be reconnected.
This motherboard shows an unusual L1 cache and RAM throughput performance at data write and move operations, memory bandwidth remains unchanged, regardless of the CPU frequency. What could possibly limit memory throughput at such abysmal levels, perhaps chipset's register settings or maybe some remaining harware fault?
