VOGONS


First post, by retro games 100

User metadata
Rank l33t
Rank
l33t

I got my first 386 mobo about a month ago. I'm only about 20-25 years late, but it's better late than never! To go with it, I needed some old ISA VGA cards, and so I bought a bunch of them online. Here's one of them, made by Headland. The RAM chips say -10 on them, which I think is 100ns. A bit on the slow side, perhaps? The card's in good shape, and it works. I wonder what those dip-switches do, on the right hand side next to that red "control panel" box? Maybe if I enter in a "code", I can configure the card's resources?

headland.jpg

Reply 1 of 15, by shock__

User metadata
Rank Oldbie
Rank
Oldbie

try these
http://www.stason.org/TULARC/pc/graphics-card … -VGA-1024I.html
http://www.notanon.com/retro/ibm-pc-xt-now-wi … deo/2010/06/12/

100ns latency seems realistic

Also: why not hunt down an ET4000 based card since you're going with 16bit ISA and VGA already. Those are easy to get, have quite good performance and well documented.

Reply 2 of 15, by retro games 100

User metadata
Rank l33t
Rank
l33t

Thanks a lot for the info! 😀 I do have one ET4000 card. I might get another, just for some extra fun. Also, I quite like Cirrus Logic cards. I am pleased with this headland-based V7 VGA 1024I video card, because the signal quality is good. It might be a bit slow, but for typical 386 use, it will be OK.

Reply 3 of 15, by TheLazy1

User metadata
Rank Member
Rank
Member
shock__ wrote:
try these http://www.stason.org/TULARC/pc/graphics-card … -VGA-1024I.html http://www.notanon.com/retro/ibm-pc-xt-now-wi … deo/20 […]
Show full quote

try these
http://www.stason.org/TULARC/pc/graphics-card … -VGA-1024I.html
http://www.notanon.com/retro/ibm-pc-xt-now-wi … deo/2010/06/12/

100ns latency seems realistic

Also: why not hunt down an ET4000 based card since you're going with 16bit ISA and VGA already. Those are easy to get, have quite good performance and well documented.

Take a look at the Doom benchmarks near the end of the 386-40 thread in system specs.
The ET4000 isn't much better than the ATI-28800.

Of course Doom is CPU limited but for a 386 system what is the point of tracking down a ET4000 when they are more rare and expensive?

Reply 4 of 15, by Old Thrashbarg

User metadata
Rank Oldbie
Rank
Oldbie

I've been curious about the Headland cards for awhile now... I haven't been able to find much information about their capabilities/performance. I do have a vague memory that Headland was one and the same with Video Seven and were somehow related to LSI Logic, but I don't remember all the details.

That one you have looks to be an early accelerated card, if the VRAM is any indicator. Not sure what the two chips are for, it could be that one is the VGA core and the other is the accelerator chip, though it's equally possible that the smaller one is for CGA/EGA support or something.

Reply 5 of 15, by Mau1wurf1977

User metadata
Rank l33t++
Rank
l33t++

V7 sounds like SPEA? Did they change company names or got bought out?

EDIT: Found this reference:

Video 7 Super VGA.
Bought by Headland Technologies in
Bought by SPEA Software AG in 1993.

Source:
http://pdos.csail.mit.edu/6.828/2010/readings … adoc/VIDEO7.TXT

So SPEA bought Headland. SPEA was pretty huge and had a very good reputation with cards such as SPEA mirage and many others. They also made the SPEA MEDIA FX, which used Ensoniq technology.

I believe they where more well known for their later Windows accelerator cards rather than DOS cards.

Reply 7 of 15, by DonutKing

User metadata
Rank Oldbie
Rank
Oldbie

The RAM chips say -10 on them, which I think is 100ns. A bit on the slow side, perhaps?

Sounds about right, 100ns is about 10MHz and the card is probably designed for an 8MHz ISA bus give or take

My Headland cards have a buggy BIOS that detects a monochrome monitor on faster (~486) CPUs...

You know, I have two Trident 512kb ISA cards that do the same thing. They boot ok the first time on a DX2-66 but after that they always come up monochrome. I thought they were faulty but maybe its a bug as you say. Haven't tried them on the 386 yet

If you are squeamish, don't prod the beach rubble.

Reply 8 of 15, by h-a-l-9000

User metadata
Rank DOSBox Author
Rank
DOSBox Author

Ehm the video RAM is primarily there to deliver pixels, not to fit the bus 😉

You can try this little tool to get color back on such a card. Works for me on the HT/V7 and TVGA8900 cards.

There are some though that apparently crash in their BIOS with fast/cached CPUs, especially 8-bit cards, these can't be helped.

Attachments

  • Filename
    coloron.zip
    File size
    3.78 KiB
    Downloads
    392 downloads
    File license
    Fair use/fair dealing exception

1+1=10

Reply 9 of 15, by elianda

User metadata
Rank l33t
Rank
l33t
h-a-l-9000 wrote:

My Headland cards have a buggy BIOS that detects a monochrome monitor on faster (~486) CPUs...

The reason the monochrome switching is happening is simple:

Early VGA standard defined a monitor detection with some pins on the vga connector, specifically Color / Monochrome.
Later the same pins were reused as DDC communication pins.

So if you have a DDC capable monitor and an older VGA card, the card detects 'something'. You can correct the behaviour if you cut it and configure the pin connections on the cards end to color monitor encoding.
(and leave the monitor ends DDC pins unconnected)

I was already considering, if a hack could help that changes the BIOS setting back to color if it had uncorrectly detected a monochrome monitor. But this could be different on each graphics card.

Edit: Well kind of synchroneous posting. The tool is something like I thought of.

Reply 10 of 15, by h-a-l-9000

User metadata
Rank DOSBox Author
Rank
DOSBox Author

Nope, it's not the monitor. There are two different schemes for monitor detection, the more common and more complicated one of detecting the presence of a connection on the analog lines, and the other with the ID0/1 pins.

The first one is implemented on IBM VGA and many VGAs of other manufacturers, the second originated from IBM XGA and I've found it on ET3000/4000 boards and some 8-bit VGA I don't remeber the brand right now.

You can identify old cards by looking for a LM339 chip. If one is present, the card uses the first method. Newer ones, such as S3 Trio, have it integrated. You can check these by using a monitor and cable with separate BNC connectors: unplug red and blue, switch on the computer and let it POST, then plug them back in. If its in monochrome mode now, it used the analog line detection scheme.

The malfunctioning cards mentioned here by me both use the first scheme.

Edit: early OTI / UMC apparently also use the second scheme

1+1=10

Reply 11 of 15, by retro games 100

User metadata
Rank l33t
Rank
l33t

As always, thanks a lot for the great retro info, people. I've just discovered something really interesting! Please see the photo. I plugged in that small Realtek ISA VGA card, and was disappointed to see the screen full of garbage at the BIOS POST stage. I cleaned the contacts twice, and tried it in 2 ISA slots. That didn't fix the problem.

But then I was determined to try again, and I just had a funny "plan of action". I don't know why I tried this, but I'm glad that I did, because it worked! I removed the controller card. That fixed the problem. What's going on? Some kind of resource problem? Possibly, but here's the interesting bit - I put the controller back in to the mobo, but removed the compact flash drive. I replaced it with a 2.5" HDD. The card works absolutely fine now. I removed the HDD, and put the CF back in. The problem reappears - the BIOS POST screen is full of corrupted garbage, and the system is unusable.

So, if you buy an old video card, and it shows garbage on the screen, it might be OK. Instead, it might be some resource problem. In this case, something weird is going on with the controller + CF + VGA card.

Edit: This is so weird. I've just tested half a dozen CFs, and the weird problem mentioned above only happens with one of them. My main testing CF unit! It's a 512MB San Disk. All the others I have tried don't cause this weird problem to occur. I've tried 128MB, 256MB, and 2GB CFs. They all work fine, and don't cause this issue.

PICT2060.JPG

Reply 12 of 15, by Old Thrashbarg

User metadata
Rank Oldbie
Rank
Oldbie

That's odd. How did you have the CF attached... was it on an IDE cable, or directly attached to the port on the card? Not that I necessarily think that would make a difference, but then again, I wouldn't have expected the CF card to be a problem at all.

Reply 13 of 15, by retro games 100

User metadata
Rank l33t
Rank
l33t
Old Thrashbarg wrote:

That's odd. How did you have the CF attached... was it on an IDE cable, or directly attached to the port on the card? Not that I necessarily think that would make a difference, but then again, I wouldn't have expected the CF card to be a problem at all.

Ah, but it does make a difference! 😀 I've just this minute done some more tests, this time with the CF unit attached directly to the controller card - without a cable. Now, I get no problems at all, and that's using the "suspect" 512MB CF unit. I wonder if this is some kind of timing problem, with the data going back and forth across an IDE cable? The cable's fine, because it works if I use any other CF or HDD attached to it.

Reply 14 of 15, by TimWolf

User metadata
Rank Member
Rank
Member
h-a-l-9000 wrote:

Ehm the video RAM is primarily there to deliver pixels, not to fit the bus 😉

You can try this little tool to get color back on such a card. Works for me on the HT/V7 and TVGA8900 cards.

There are some though that apparently crash in their BIOS with fast/cached CPUs, especially 8-bit cards, these can't be helped.

I know this is a retro bump but thank you for this. MY TVGA8900 was driving me nuts with a Hansol 17" monitor in B&W half the time. This is still relevant and useful all these years later!