This video card has 8 RAM chips : Mosel Vitelic V53C464AP80L

Each of these is 64K by 4 bits wide or 32 Kilobytes in capacity.

8 times 32 Kilobytes = 256 Kilobytes

So there is apparently only 256 Kilobytes of RAM on this card (unless there is more on the other side).

This is insufficient for pretty much any SVGA modes except 800x600 in 16 colors (which is not an official VGA mode is debatably not really SVGA either).

EDIT : https://www.dosdays.co.uk/topics/Manufacturer … oa.php#SVGA6200 mentions a max of 512 Kilobytes of memory being possible on that card, but the card in the photo there has the same memory configuration as yours, hence 256 Kilobytes, which is the same a standard VGA .

Lack of memory is a great point.

I am kind of hoping to see how to activate the SVGA modes that VGADOC does list, like 640x400 256c and 800x600 16c, and then expanded text modes, like 132x(25-60), or some of the quirkier modes, like 912x480 16c, that the VGADOC article hints that might exist.

darry wrote on 2024-02-10, 23:24:

EDIT : https://www.dosdays.co.uk/topics/Manufacturer … oa.php#SVGA6200 mentions a max of 512 Kilobytes of memory being possible on that card, but the card in the photo there has the same memory configuration as yours, hence 256 Kilobytes, which is the same a standard VGA .

Yeah.. I think I've actually acquired that very card that was pictured by DOS Days, there is a vertical black scratch (for example) on the right side of the white sticker going through the dot on the word Inc. that matches my card.

A single frame of 640x400 256c would fit in 256 Kilobytes. No idea if the onboard VBE actually allows that with 256KB RAM .

If UNIVBE/UNIVESA supports this chip, it might be worth a try to see if it exposes 640x400 256c and/or 800x600 16c as VESA modes .

clb wrote on 2024-02-10, 23:35:

darry wrote on 2024-02-10, 23:24:

EDIT : https://www.dosdays.co.uk/topics/Manufacturer … oa.php#SVGA6200 mentions a max of 512 Kilobytes of memory being possible on that card, but the card in the photo there has the same memory configuration as yours, hence 256 Kilobytes, which is the same a standard VGA .

Yeah.. I think I've actually acquired that very card that was pictured by DOS Days, there is a vertical black scratch (for example) on the right side of the white sticker going through the dot on the word Inc. that matches my card.

Wow, it's a small world sometimes.

UniVBE at least didn't support this, but gave just

Good idea to try out UniVESA, maybe that might do something different.

QPV, when configured for Genoa 6200 with 256 KB, provides support for 800 x 600 x 16 as mode $6a or $79 - run SETUP.EXE, select the appropriate adapter, and (optionally) edit QPV.CFG

Grzyb wrote on 2024-02-10, 23:52:

QPV, when configured for Genoa 6200 with 256 KB, provides support for 800 x 600 x 16 as mode $6a or $79 - run SETUP.EXE, select the appropriate adapter, and (optionally) edit QPV.CFG

I presume you mean Quick Picture Viewer ?

https://www.sac.sk/download/graph/qpv17e.zip

Exactly!

@clb

https://dosdriver.de/graph.php might have some interesting stuff

EDIT: specifically,

Thanks, tested out both QPV and VESADRV.. but oddly, nothing seems to work with this adapter: https://oummg.com/dump/GenoaSVGA6200.mp4

QPV seems to mis-think that the video mode would have successfully changed, and it doesn't seem to verify that it actually did - so it writes pixels in the new assumed video mode without realizing that the old 640x480 video mode is still in effect.

SNOOP scans all the INT10h video modes 00h-7Fh and finds nothing above mode 13h. The Genoa VESADRV does add some VESA modes (including a 640x480 256c which the adapter would not have enough memory for), but fails to set any of them.

There are three jumpers on the board (no DIP switches at the back), marked IRQ2, 16bit, and "AUTOH/C". For good measure, I flipped the state of all of these three in the hope that one of them might do a "SVGA unlock" of sorts (I recall Headland HT-208 had such a functionality), but that didn't do anything either.

Genoa is a sad panda.

Generally speaking, 640x400 @256c and 800x600 @16c should work with 256KB of video RAM.

If not, the video RAM can perhaps be increased by using piggy-back method.

Both are valid VBE modes, also.
640x400 256c is VBE mode 100h. Not all VBE BIOSes support it, though.
800x600 16c is VBE mode 6Ah and 102h. It's also Super VGA (Paradise video mode wasn't too uncommon).

I'd recommend to try picture viewers CompuShow 2000 (2SHOW) and PV. ^^
Re: Software for my 286/8 1MB ET4000 SBpro

PS: The Windows 3.1 800x600 SVGA driver has code for both VBE as well as as contemporary SVGA chips.

If that isn't working, maybe a missing jumper setting is the culprit.
Could be that the monitor type jumper block was being removed for cost saving measures.

PS: Oh, and please don't overrate UniVBE so much. ;)
It's, um, not too great with ISA VGAs. Better try the original VBE 1.x TSRs of the chip maker, too.
UNIVBE 4.0 and old hardware
Re: OAK OTI-037c - 800x600 mode ?

Quick update. I was wrong, there might be jumpers.

https://wiki.preterhuman.net/Genoa_Super_VGA_6200

Jumper J3 might be worth a try ("Monitor is VGA or XVGA" and "Monitor is CGA").

I know of XGA, but I'm not sure what exactly is meant by XVGA. Probably eXtended VGA?

Edit: The way I read it, jumper set=CGA monitor, jumper removed=VGA/XVGA monitor

Edit: I see that you already tried these jumpers.. Hm.

clb wrote on 2024-02-11, 12:41:

SNOOP scans all the INT10h video modes 00h-7Fh and finds nothing above mode 13h.

Congratulations, you've got a pretty unique thing: a vanilla VGA card!
Yes, such cards are pretty rare, even very early VGA clones usually include the 800 x 600 x 16 mode, which essentially makes them SVGA.

Though that Genoa may support one feature that the original IBM VGA lacks: register-level CGA and Hercules modes.
So it would be nice to find the appropriate software.

Would a vanilla VGA card have been marketed as a SuperVGA card?

VGADOC documents that the card should support different SVGA modes, and that QPV viewer also. And there is that VESA driver for the card as well, which attempts to set SVGA modes but fails. So there are at least three independent sources of information that say this card should have SVGA modes.

I am leaning towards investigating the idea that someone might have flashed some other VBIOS onto this card that should not belong together with this card, and that might be causing the card to misbehave with setting SVGA modes.

The main chip is GN007001-B, right?

That chip supports up to 512 KB, so it's totally SVGA.
But this particular card doesn't support more than 256 KB, therefore SVGA modes got disabled in the BIOS.

I suspect it can support 800 x 600 x 16, but it would require some software to directly program the registers.
If not UniVBE, then perhaps some old version of XFree86 ?

You can also try the BIOS from there - http://vgamuseum.info/index.php/companies/ite … gn007001-b-gvga

Some SVGAs do memory interleaving and other weird stuff.

They require memory expansion to properly work in higher resolution/colour depths.

I've seen this with Windows 3.1 graphics drivers:
Some resolutions/colors depths are available twice.
One is higher performance, but needs a higher memory expansion.

Personally, I really think that piggybacking is worth a try.
I often thought about doing that myself with an old Paradise card..

But be careful, there are two meanings of the piggyback method.
a) Stacking RAMs atop, but with the address lines (RAS, CAS etc) re-routed to be in series.
That way, the RAM logically appears same as if being installed properly in an empty expansion socket.
b) Just stacking RAMs; it's being used for troubleshooting faulty RAM. Both RAMs are working in parallel, no increase in memory.

Edit: If the VGA BIOS is the culprit, it's also possible to use a RAM BIOS version of it:
It's a TSR version of a BIOS that can be loaded in DOS. It installs like a mouse driver or like the VESA VBE TSR and can be found on the VGA card's utility/driver diskette.

The neat thing is that this VGA BIOS *can* be different to what's burned into the. ROM chip on your VGA card.

So ideally, for upgrading your VGA card firmware you don't need to resort to use an EPROM programmer and a blank EPROM.
For testing, that utility might be fine enough. It wastes some memory, of course. 33KB or so.

But that depends on the way the TSR works.
If it's big (24 to 32KB or 48KB) it will contain the whole VGA BIOS.

Otherwise, it's just an utility that copies the ROM version to RAM, for increased performance.

That's what the RAMBIOS was being made for originally, after all.
To not have to call the VGA ROM code over 8-Bit or 16-Bit bus each time.

Edit: But personally, I'd just make sure the VGA card isn't the culprit.

- Are other Genoa users have same problem?

- Are 512KB really required for 800x600 16c on this Genoa model?
Most (S)VGA cards support that mode, even with 256KB. It's the most popular SVGA mode. It was the very reason for VESA VBE to exist.

Thanks, very good points there.

On the video link I posted above, I was using the card's own VESA driver (to my understanding).

And not sure what to make of the fact that Genoa's own VESA driver makes the card advertise that 800x600-16c mode, but then actually fails to initialize it.

I'll have to look around. If I try to flash a new BIOS, definitely won't overwrite the original one, but will get a fresh BIOS chip to try it out on.

Hi, I (we) wish you good look! 🙂👍
We'll be are around here, please keep us uptodate.

clb wrote on 2024-02-12, 16:15:

I'll have to look around. If I try to flash a new BIOS, definitely won't overwrite the original one, but will get a fresh BIOS chip to try it out on.

That's a good idea. You can use OTP EPROMs, if needed.
These are one-time-programmable ROMs.

Electronically, they're working by same principle as vintage EPROMs but don't have the quarz window for UV erasure.

So they can't be cleared and re-used. Except by using x-rays or something. ;)
The window was being removed to make production as cheap as possible.

They're like those disposable cameras of the 90s, so to say.

Anyway, their electric specs aren't bad. Their access time can be very good, in fact.
So they might be less prone to fail in faster systems than period-correct chips.

Edit: It might be possible to use modern EEPROM or Flash replacements, too, but I've little to no experience with them.
Some replacements may need adapters.