I'm not an expert as well, so here is a dumb question. How does a 512kbit ROM image fit in a 256kbit chip?

PD2JK wrote on 2025-06-18, 08:24:

I'm not an expert as well, so here is a dumb question. How does a 512kbit ROM image fit in a 256kbit chip?

From what i saw often the bios uses only one half of the 64KB chip memory (the majority of the dumps online are 32KB) or is written twice (one copy on the first 32KB and and another identical copy on the final 32KB) and the latter seems to be my case. But i could be terribly wrong ahah!

Runningwild82 wrote on 2025-06-18, 08:02:

One strange thing: when I read the working BIOS, the first part of the file — where you’d normally see readable strings like “copyright” and “all rights reserved” — appears corrupted, as shown in the attached picture. But the chip I’m reading from definitely works 100%, so I have no idea why this happens.

Possibly poor contact in the programmer socket.

Postman5 wrote on 2025-06-18, 09:50:

Runningwild82 wrote on 2025-06-18, 08:02:

One strange thing: when I read the working BIOS, the first part of the file — where you’d normally see readable strings like “copyright” and “all rights reserved” — appears corrupted, as shown in the attached picture. But the chip I’m reading from definitely works 100%, so I have no idea why this happens.

Possibly poor contact in the programmer socket.

Would this be possible? Wouldn't the programmer give a pin error if the contact is not stable? Also the programmer is new, and the clip seems very tight so i don't know

The read information looks distorted. The least of the reasons is a bad contact, the biggest is a programmer malfunction. You can try reading several times, shaking the chip in the socket before that, try to disable the pin detection function, perform a self-test of the programmer.

Will try, thanks!

Runningwild82 wrote on 2025-06-18, 08:02:

Hi everyone, I’m new here so I hope you’ll forgive my lack of experience! I’m trying to revive an old VLB video card, the PWA-MV […]
Show full quote

Hi everyone, I’m new here so I hope you’ll forgive my lack of experience!
I’m trying to revive an old VLB video card, the PWA-MVGA-542X (AGATech), which has a CL-GD5428 graphics chip.

The card was completely dead until I tried removing its original BIOS chip and replacing it with a BIOS from another card that uses a CL-GD5426 (an older version of the same family). With this change, the card is now detected and works fine in DOS.
However, when I start Windows 3.1 or run software that uses windows and various colors, I get graphical artifacts on the screen. I’m not sure whether this is due to the wrong BIOS version or a hardware issue with the card itself — I plan to investigate that further.

The real issue I need help with is this: when I try to copy the working BIOS (which I believe is a good starting point for testing) onto another chip, the card doesn’t recognize it at all.
I’m using a TL866II programmer with XGpro v12.9 to dump the BIOS. The chip I’m reading from is a TMS27C512 EPROM, and I’m trying to write to an AT28C256 EEPROM. I know the first is 64KB and the second is 32KB, but they should be compatible — or am I wrong?

I’ve also tried flashing some generic BIOS files for this graphics chip that I found online, but I get the same result: nothing works. Just out of curiosity, I even dumped the BIOS from a motherboard and tried replacing the BIOS chip with the AT28C256, but again, no success.

At this point, I’m wondering whether the problem lies with the chip I’m writing to (maybe it’s not compatible?), or if there’s a setting I should adjust during read/write?

One strange thing: when I read the working BIOS, the first part of the file — where you’d normally see readable strings like “copyright” and “all rights reserved” — appears corrupted, as shown in the attached picture. But the chip I’m reading from definitely works 100%, so I have no idea why this happens.

Could someone please help me figure this out?
Thanks a lot, and sorry for the long post!

Okay so, first of all here's a BIOS bump from a CL-GD5428 card that I took from one of these two cards - mine is scrambled just like yours and this ROM data definitely works:

Your read is correct *but* video card manufacturers like Trident and Cirrus Logic did odd stuff with the ROM BIOS by I think swapping around the data and/or address lines. That's compensated for in the card's PCB design so the computer itself will read an unscrambled VGA ROM BIOS:

If the ROM dump was taken from a running computer via software then it'll be the unscrambled form which is useless for replacing the physical ROM chip if it's missing or damaged.

Now the next thing is that a 28C256 isn't the same as a 27C256 and I made the same mistake before, have a read of this:
https://electronics.stackexchange.com/questio … -how-to-wire-we
I've used a 28C256 in place of a 27C256 with some minor mods: Re: What retro activity did you get up to today?

To fit my 28C256 EEPROM where the card was designed for a 27256 mask ROM, EEPROM Pin1 had to be kept out of the ROM socket, along with pin 27. Run a blue wire from the EEPROM's Pin1 goes into the socket's Pin27, then the EEPROM's Pin27 is soldered to Pin28 (VCC) for normal operation.

There are some other ROMs you can try here, from what I recall they've been read from real ROMs so they are scrambled and can be written to a ROM & work: http://chukaev.ru54.com/video_en.htm

If you're seeing artifacts when the accelerated driver is running but it's fine in basic VGA, I recall reading a post from mkarcher a while back that pointed out that the memory timings / memory speed are more relevant when doing windows acceleration compared to VGA. IT's possible a more standard CL-GD5428 BIOS will have better timings or memory speed configured for your card and allow windows acceleration to work properly.
If the artifacts persist, remove the DIP memory if it's fitted, maybe it's not fast enough for the card.

Thermalwrong wrote on 2025-06-18, 22:26:

Okay so, first of all here's a BIOS bump from a CL-GD5428 card that I took from one of these two cards - mine is scrambled just […]
Show full quote

Runningwild82 wrote on 2025-06-18, 08:02:

Hi everyone, I’m new here so I hope you’ll forgive my lack of experience! I’m trying to revive an old VLB video card, the PWA-MV […]
Show full quote

Hi everyone, I’m new here so I hope you’ll forgive my lack of experience!
I’m trying to revive an old VLB video card, the PWA-MVGA-542X (AGATech), which has a CL-GD5428 graphics chip.

The card was completely dead until I tried removing its original BIOS chip and replacing it with a BIOS from another card that uses a CL-GD5426 (an older version of the same family). With this change, the card is now detected and works fine in DOS.
However, when I start Windows 3.1 or run software that uses windows and various colors, I get graphical artifacts on the screen. I’m not sure whether this is due to the wrong BIOS version or a hardware issue with the card itself — I plan to investigate that further.

The real issue I need help with is this: when I try to copy the working BIOS (which I believe is a good starting point for testing) onto another chip, the card doesn’t recognize it at all.
I’m using a TL866II programmer with XGpro v12.9 to dump the BIOS. The chip I’m reading from is a TMS27C512 EPROM, and I’m trying to write to an AT28C256 EEPROM. I know the first is 64KB and the second is 32KB, but they should be compatible — or am I wrong?

I’ve also tried flashing some generic BIOS files for this graphics chip that I found online, but I get the same result: nothing works. Just out of curiosity, I even dumped the BIOS from a motherboard and tried replacing the BIOS chip with the AT28C256, but again, no success.

At this point, I’m wondering whether the problem lies with the chip I’m writing to (maybe it’s not compatible?), or if there’s a setting I should adjust during read/write?

One strange thing: when I read the working BIOS, the first part of the file — where you’d normally see readable strings like “copyright” and “all rights reserved” — appears corrupted, as shown in the attached picture. But the chip I’m reading from definitely works 100%, so I have no idea why this happens.

Could someone please help me figure this out?
Thanks a lot, and sorry for the long post!

Okay so, first of all here's a BIOS bump from a CL-GD5428 card that I took from one of these two cards - mine is scrambled just like yours and this ROM data definitely works:

The attachment romdump-27c512or256.zip is no longer available

The attachment cl-gd5428-repaired (Custom) (1).jpg is no longer available

Your read is correct *but* video card manufacturers like Trident and Cirrus Logic did odd stuff with the ROM BIOS by I think swapping around the data and/or address lines. That's compensated for in the card's PCB design so the computer itself will read an unscrambled VGA ROM BIOS:

The attachment romdump.png is no longer available

If the ROM dump was taken from a running computer via software then it'll be the unscrambled form which is useless for replacing the physical ROM chip if it's missing or damaged.

Now the next thing is that a 28C256 isn't the same as a 27C256 and I made the same mistake before, have a read of this:
https://electronics.stackexchange.com/questio … -how-to-wire-we
I've used a 28C256 in place of a 27C256 with some minor mods: Re: What retro activity did you get up to today?

To fit my 28C256 EEPROM where the card was designed for a 27256 mask ROM, EEPROM Pin1 had to be kept out of the ROM socket, along with pin 27. Run a blue wire from the EEPROM's Pin1 goes into the socket's Pin27, then the EEPROM's Pin27 is soldered to Pin28 (VCC) for normal operation.

There are some other ROMs you can try here, from what I recall they've been read from real ROMs so they are scrambled and can be written to a ROM & work: http://chukaev.ru54.com/video_en.htm

If you're seeing artifacts when the accelerated driver is running but it's fine in basic VGA, I recall reading a post from mkarcher a while back that pointed out that the memory timings / memory speed are more relevant when doing windows acceleration compared to VGA. IT's possible a more standard CL-GD5428 BIOS will have better timings or memory speed configured for your card and allow windows acceleration to work properly.
If the artifacts persist, remove the DIP memory if it's fitted, maybe it's not fast enough for the card.

Hey mate... i don't know how to thank you, your post is pure gold to me, really.... tried to copy the scrambled bios to the 28C256, made the pin mods and... it works! Thank you, really! Now i'll be able to make some tests with timings and memories to get rid of those artifacts! Talking again about the bios chips, maybe do you know which eeprom chip is directly compatible with the 27C256? Thank you again my friend, you saved my day ahah!