VOGONS


First post, by CelGen

User metadata
Rank Member
Rank
Member

So I have this really bizarre Matrox E-VDP machine. It's basically a laserdisc player strapped to the top of a massive case with a hard drive, two floppy drives, an amplified front speaker and a passive ISA backplane. The computer itself consists of a 286 SBC board and two specialty ISA cards for video and audio. The computer controls the player using RS-232 and video feeds into the video card either for digitizing or to be overlaid on the normal video signal using some software package. The front of the machine has ports for a small keypad (or a regular keyboard) and a lightpen.
100_2367.jpg
IMG_7553.jpg
IMG_7555.jpg
IMG_7556.jpg

(click the images above for better photos of each card)

I spent the last few years working on and off with it to bring it back to life and eventually yielded a successful POST (with just the SBC and a random ISA video card) after dealing with a number of shorted tantalum capacitors.
The issue however that I am running into now is that the American Megatrends BIOS gives the two long, eight short "Video Failure or Video Card Not Installed" beep error when the extremely specialty VGA card is installed (and sits at code 1A while beeping which indicates that it is returning from the video ROM). It's essential as it does various things such as control overlays and data from the lightpen and you can't just replace it with some other video card.
I'm apparently the owner of the only one still known to exist (and the only person to ever photograph it inside and out) so there are no replacement parts or any support from Matrox. I can't dump the video EPROM as I do not own anything capable of dumping the extremely oddball Intel D27513 EPROM properly (I've still tried anyways but I'm not posting a link unless a moderator allows it) so I have to figure out what might be the fault, starting with how the 286 BIOS detects and initialized the video hardware, or how it rejects any installed hardware for use by the system at POST. Is there anyone here skilled enough to help answer a few questions?

Last edited by CelGen on 2014-12-22, 15:53. Edited 1 time in total.

emot-science.gif "It's science. I ain't gotta explain sh*t" emot-girl.gif

Reply 3 of 31, by CelGen

User metadata
Rank Member
Rank
Member

What's a bit more amusing is that for some reason I already owned the digitizer option for this machine. Matrox's MVP-AT. A full length two slot, three board monster.

IMG_7583.jpg
IMG_7575.jpg

You can see why I REALLY want this machine to live again. It's no 386 but my god it's a multimedia beast.

emot-science.gif "It's science. I ain't gotta explain sh*t" emot-girl.gif

Reply 4 of 31, by NJRoadfan

User metadata
Rank Oldbie
Rank
Oldbie

Jeez, and here I thought the Video Toaster card was a massive two slot affair....

If its a "standard" style VGA BIOS, it'll reside at C0000-C7FFFh and contain Int 10h services. That EPROM is kinda weird with the whole bank switching thing. Does the BIOS have any options for shadowing the video ROM? Maybe its trying to do that and the Int 10h services are being bank switched out by the card for some reason. I would hope that Matrox put the VGA BIOS in the "fail safe" bank 0, but who knows. I would check the bank switching logic in the card for faults if possible.

Misc. questions that may be related:
What does the memory map of this machine look like with all but the original video card installed?

Reply 5 of 31, by CelGen

User metadata
Rank Member
Rank
Member

Does the BIOS have any options for shadowing the video ROM?

I'm still fighting to enter the BIOS. So far all the usual keypresses have not worked (DEL, CTRL+DEL, F1, CTRL+ F1, ESC, ESC on reboot, slamming head into keyboard....)

I would hope that Matrox put the VGA BIOS in the "fail safe" bank 0, but who knows.

Because it's not useable for cloning, here's a partial dump of the video ROM which could not be verified because of the funky bank switching made it impossible to do so. Repeatedly it calls itself an EGA BIOS (but an HD15 connector???) but there's readable data in the first bank.

I could not tell you what the current memory map is, nor can I say what is where in memory as Matrox has zero technical information available. Is there an application somewhere I can use to list it more easily?

emot-science.gif "It's science. I ain't gotta explain sh*t" emot-girl.gif

Reply 6 of 31, by NJRoadfan

User metadata
Rank Oldbie
Rank
Oldbie

Many DOS memory managers included a program that can view the upper memory area and point out which blocks are mapped to ROM or are free for use.

The only thing I can think of is something is wrong with the bank select logic on the card. IE: The machine likely boots into bank 0 and something in the BIOS toggles another bank and it can't return to bank 0 when requested. VGA ROMs are usually larger than 16kb, so the ROM will be bank switched no matter what. If its an EGA ROM, then it likely fits in 16k, that would actually make sense. Given that Real Mode has plenty of ROM space, the card's logic likely switches to the appropriate bank based on what address the CPU is trying to read.

My best guess is as follows:

Reads from C0000-C3FFF trigger bank 0- Video ROM
Reads from C4000-C7FFF trigger bank 1
Reads from C8000-CBFFF trigger bank 2
Reads from CC000-CFFFF trigger bank 3

Banks 1-3 could theoretically be anywhere in Upper Memory outside of the BIOS area. Looking at your ROM dump, it appears to read the 16k Video BIOS properly from 0000-3FFF, after that it is than duplicated 3 times(the 55 AA option ROM signature at 4000,8000,C000 with the copyright string following is a dead giveaway).

Reply 7 of 31, by CelGen

User metadata
Rank Member
Rank
Member

All the memory managers I got on hand need at least a 386. What do you recommend I go find? It also occurs to me the CMOS battery is flat so give me a day or two to source a replacement and try to get into the CMOS again. I currently can't boot from the hard disk as the parameters can't be retained.

Last edited by CelGen on 2014-12-22, 06:03. Edited 1 time in total.

emot-science.gif "It's science. I ain't gotta explain sh*t" emot-girl.gif

Reply 8 of 31, by NJRoadfan

User metadata
Rank Oldbie
Rank
Oldbie

To add, I would try burning the data from 0000-3FFFh in your dump to a 27128 and putting it on the card to see if the machine will POST (assuming that part of the ROM dumped properly). Pins 1 and 27 have different definitions on the 27513 (mostly for the bank selecting), I don't know if its safe to a 27128 or the card to leave them connected or not. Even if the bank select logic is faulty, it should consistently read the Video BIOS at C0000-C3FFFh if a non-bank select ROM is on the card.

Overall this is a pretty silly design, they should have had a separate 27128 on the card just for the video BIOS, but went the cost saving route instead.

Does MSD show a memory map? Its been awhile since I have used it.

Reply 9 of 31, by CelGen

User metadata
Rank Member
Rank
Member

Everything up to location 0x008000 checks out but the other half fails with random bits. Again , it's probably because of the oddball chip and I was reading it back in 27512 mode but 'll try digging a 27128 up and see what happens.

emot-science.gif "It's science. I ain't gotta explain sh*t" emot-girl.gif

Reply 10 of 31, by NJRoadfan

User metadata
Rank Oldbie
Rank
Oldbie

Note the above would rule out a problem with the bank switching ROM. There could also be a fault in the video memory as well, which would be pretty damned annoying.

Reply 11 of 31, by JayCeeBee64

User metadata
Rank Retired
Rank
Retired
NJRoadfan wrote:

Does MSD show a memory map? Its been awhile since I have used it.

Yes, it does (I use it on occasion to do quick system memory layout checks). But have no idea if it will work with a customized machine like this.

@CelGen: I must say this is quite unique. I've never seen anything like this before and it's too bad the specialty video card doesn't work. It would be really something to see it in action again.

Ooohh, the pain......

Reply 12 of 31, by Stiletto

User metadata
Rank l33t
Rank
l33t

Ookay, comments coming in from Lord Nightmare:

Lord Nightmare wrote:
- 27513? Don't you mean 27512? […]
Show full quote

- 27513? Don't you mean 27512?

- Guessing video card has bad RAM, which isn't super hard to fix, assuming it doesn't use something obscure.
If it uses those Toshiba ZIPram DRAMs, those are a bitch to find because people use them to upgrade their Amigas, so the best place to get them is to steal from other video cards, which kinda sucks.

- there's a missing DIP40 socket on one of the cards, is it for a 287 coprocessor?
If so, is there a DIP switch on the main board to indicate "Coprocessor present"?

- If we had a dump of the Matrox video editor box BIOS and video BIOS we might be able to figure out the fault reason.

"I see a little silhouette-o of a man, Scaramouche, Scaramouche, will you
do the Fandango!" - Queen

Stiletto

Reply 13 of 31, by CelGen

User metadata
Rank Member
Rank
Member

27513? Don't you mean 27512?

Nope.
Intel D27513.
IMG_7549.jpg

Guessing video card has bad RAM, which isn't super hard to fix, assuming it doesn't use something obscure.

Closest thing I see to memory on the video board is eight LH2464-10's directly above the two CHIPS ASIC and Bt 471KPJ DAC (what a weird place to put it)

there's a missing DIP40 socket on one of the cards, is it for a 287 coprocessor?
If so, is there a DIP switch on the main board to indicate "Coprocessor present"?

Yes and no idea. None of the jumpers are at all marked (unlike the ribbon cables which are done beautifully).
In fact, aside from that unoccupied 20 pin DIP socket and three test points there aren't any jumper blocks on the SBC at all.

If we had a dump of the Matrox video editor box BIOS and video BIOS we might be able to figure out the fault reason.

I need a mod's approval before I'm posting dumps of the system BIOS. I know some forums get antsy about these things.

It would be really something to see it in action again.

IF we ever get the hardware problems sorted I've already discovered a second issue. The drive is a Connor with a SCSI interface. It runs fine and the gasket has not failed but that I can tell so far it's been formatted and data not related to the machine has been put on it. If an unformat can be performed, we're probably going to experience some form of permanent data loss. I need to find someone else with one of these machines or ANYONE who might have the software needed to control this extremely specialty hardware. 😢

emot-science.gif "It's science. I ain't gotta explain sh*t" emot-girl.gif

Reply 14 of 31, by Matth79

User metadata
Rank Oldbie
Rank
Oldbie

http://www.eetimes.com/document.asp?doc_id=1276912 - article referencing bank switching EPROMS - might be some clues

Datasheet here http://www.cpu-galaxy.at/CPU/Ram%20Rom%20Epro … intel_erpom.htm

Reply 15 of 31, by Stiletto

User metadata
Rank l33t
Rank
l33t
CelGen wrote:

If we had a dump of the Matrox video editor box BIOS and video BIOS we might be able to figure out the fault reason.

I need a mod's approval before I'm posting dumps of the system BIOS. I know some forums get antsy about these things.

Well, I'm a mod. I can't approve any and all things and I DO know what the disclaimer says at the bottom of every page, but I personally think that linking to firmware to obsolete PC hardware for the purposes of repair and diagnostics should be permitted, regardless of whether you sourced it from a manufacturer's dead FTP or that you dumped it yourself. Unless I am mistaken, people have linked to manufacturer BIOSes hundreds of times before in Marvin. So I'm not completely aware how this would be any different, other than the fact that it doesn't come from the manufacturer and people can't hold VOGONS responsible if flashing it toasts their hardware.

(However, I think that linking to firmware to obsolete arcade games for the purposes of repair and diagnostics and emulation should not be permitted. I am aware that is an inconsistent and self-contradictory policy to hold. 🤣)

What I would not want to see is "here's a post with a thousand links to every video card firmware ever made ever, 🤣!!!111eleven" or "here's a complete ROM set for MESS or PCem". But I think sparingly, in moderation, and moderated, and only for those stated purposes, there's nothing truly wrong or bad about it.

"I see a little silhouette-o of a man, Scaramouche, Scaramouche, will you
do the Fandango!" - Queen

Stiletto

Reply 16 of 31, by JayCeeBee64

User metadata
Rank Retired
Rank
Retired
CelGen wrote:

IF we ever get the hardware problems sorted I've already discovered a second issue. The drive is a Connor with a SCSI interface. It runs fine and the gasket has not failed but that I can tell so far it's been formatted and data not related to the machine has been put on it. If an unformat can be performed, we're probably going to experience some form of permanent data loss. I need to find someone else with one of these machines or ANYONE who might have the software needed to control this extremely specialty hardware. 😢

Ugh, that's bad. Really bad. A heavily customized machine like this needs very specific software to run as intended; without this, it just becomes a huge doorstop - not a desirable outcome at all 😖

I wonder if one of the previous owners (I'm assuming there was more than one) tried to use it as part of a custom arcade setup and simply failed; it would explain why the hard drive has non-related data. Unfortunately I haven't been able to find anything useful to help you out so far; an old family friend did tell me it looks like something that was probably made as part of a special order, maybe for a large corporation, government agency or even the military. If true, that would make finding the original software package even more difficult.

I'll continue to look for more info and ask a few other people I know about this unique machine; maybe I'll get lucky somehow ^^.

Ooohh, the pain......

Reply 17 of 31, by CelGen

User metadata
Rank Member
Rank
Member

Well as a mod has given an OK, here's the rar file containing the one kinda-dumped image of the video ROM, plus the two verified dumps of the SBC's AMI BIOS.

>>LINK<<

an old family friend did tell me it looks like something that was probably made as part of a special order, maybe for a large corporation, government agency or even the military.

You're not far off. It has a military asset tag attached to the back.

100_2368.jpg

I've been told by members of another forum that they have seen more units but they were all junked over a decade ago.

emot-science.gif "It's science. I ain't gotta explain sh*t" emot-girl.gif

Reply 19 of 31, by CelGen

User metadata
Rank Member
Rank
Member

That's just the laserdisc player and not the actual computer. By the looks of it someone salvaged the player and junked the rest so they could use it an LD arcade cabinet (the serial protocol is pretty standardized.
IMHO these players are terrible. Not only is the main drawer mechanism belt driven but the optical pickup is belt driven too. Really? You couldn't use a gear drive like everyone else since the beginning? I would personally recommend a commercial grade Sony/Pioneer deck if you need a serial interface. They're still pretty easy to find.

emot-science.gif "It's science. I ain't gotta explain sh*t" emot-girl.gif