The original BIOS image shows that the bytes are only EVEN. Maybe the first half of the BIOS image is composed of just EVEN bytes and the other half from ODD bytes?

Thanks so much for your reply!

Ah!! So perhaps this board was originally designed to have two EPROMS (odd and even) and later it was squished down to one?

How would I go about formatting my BIN to comply with that standard?

It looks like the EPROM socket address pins are not wired correctly. Possibly A0 and A14 are swapped. Try putting all the even bytes first, then all the odd bytes.

It is not uncommon for the pin layout to differ from expected on EPROM sockets. This is one reason why software dumps can be problematic, as they'll always see things as if the layout is correct.

In which case for a purely hardware solution I could correct the bad address lines on the card, then allowing me to run the standard ROM?

Some scalpel action and a couple of bodge wires... 🤔

Probably be easier to tweak the binary. There is a split utility included with the Turbo XT BIOS. You can do something like:

1win32/split bios.bin even.bin odd.bin
2copy /b even.bin+odd.bin newbios.bin

Ah ha! Awesome, thanks so much for that!

I'll give it a whizz tomorrow 👍

Plasma wrote on 2022-01-11, 22:13:

It looks like the EPROM socket address pins are not wired correctly. Possibly A0 and A14 are swapped. Try putting all the even bytes first, then all the odd bytes.

This has to be it, every odd byte would be read starting from 4000H address in the ROM, which makes sense.

So was this by design or do you think the manufacturer made an arse of the PCB that they 'fixed' with this bodge? Maybe it was to combine from an older design with two ROMS?

The "homemade" EPROM makes me think it was probably a mistake. Maybe an initial run of boards with the error that they salvaged to sell. Usually when it's done intentionally, the data lines are scrambled to better obscure the contents.

It's not easy to swap A0 with A14 around just by a mistake either, perhaps they simply didn't want 3rd party to tamper with the BIOS contents and today's post just proves that such tactic was somewhat successful, no? 😁

Pretty weak for security but IDK. Maybe a cat walked on the keyboard 🤔

Well I'll let you all know exactly what j discover tomorrow when I give the new BIOS a whizz!!

Yay great success!

It was exactly as you all theorised, ODD and EVEN lay on top of each other. I used the split command and then combined them back, wrote it to an EPROM and it worked first time. I've attached the new (1.30) BIOS that I made in case anyone stumbles across this thread in the future.

Next thing I'll look to do is pull the 1.41 BIOS out of one of the .EXE bios updates and see if I can burn that in. I'm probably being naive, but I'm hoping it's a cut and paste of the newer code into the BIN file.

Thanks again everyone!!

Hmmm, I've had no luck getting the 1.41 BIOS out of an EXE and into the ROM for my card today. Many highly musical 'NO VGA' post beeps was the result, 🤣.

Still, got 1.30 working yay!!

It might need the checksum calculated. Do you have a link?

Yeah I wondered about that, but a bit outside my experience of hex editing!! I attach the executable that has the 1.41 BIOS in it.

This runs great from DOS and all, but of course I figure better loaded from ROM, and if it's in the ROM it should work for any hard coded BIOS routine calls.

Oh I thought it was an actual BIOS update for an EEPROM. This is like a TSR. It can be made to work but will require more modification.

Yeah.... Ah well, not sure it would make a huge difference anyways!

Thanks a lot for taking a look. 😀