Replace a 27c1001 eprom with a 29c020 eeprom possible?

[poot36]

I have an old 486 motherboard (MSI MS-4144) that I want to try some different bios files on and don't want to mess with UV erasing a 27c1001 chip.  I have doubled up the bios image that I have dumped from the original bios chip to fit on the 29c020 chip using Notepad++ and have also pulled address line A17 low and the system will not boot with the 29c020 chip.  Got any ideas what I am doing wrong here?  Is the board smart enough to know that this chip is not the original one and therefore refuses to boot?

[CJay]

Try this:

Read the 27C1001 into the buffer with your programmer, without changing anything plug the 29C020 chip into your programmer, run a verify.

Betcha it fails.

[fzabkar]

I don't know what you're doing wrong, if anything, but here is a free hex editor:

https://mh-nexus.de/en/hxd/

[Kjelt]

What are the required access times.
Just a quick check for two random 27c1001 from ST and an 29C020 from Atmel, the eprom has 35ns accesstime and the eeprom 90ns.
So investigate what is required and if the new chip meets that requirement.

[poot36]

The 27C1001 chip has a 150ns speed rating and the 29C020 has a 120 ns speed rating.  The 27C1001 is from ST and the full part number is m27C1001-15F1 and the 29C020 is from Winbond but is marked on the chip as a ASD brand AE29F2008-12 but passes the electronic ID check in the programmer as a Winbond W29C020C-12.  I am wondering if the 27C1001 has the pins A16 and G bar swapped as the datasheet I found for it mentions a date of 1998 and the chip has a date code of 95 on it.  How would I be able to test that?

@fzabkar I have that program on my system as well and I may have to give it a try but when I was making test EEPROMs for a SNES I used Notepad++ to double up (or some times more) a smaller file and I had no issues with that.

[fzabkar]

You can double up a file in a DOS box as follows:

copy /b half.bin + half.bin double.bin

Can you upload the EPROM dump? These chips are checksummed, so you can verify the checksum of the binary image, assuming we can work out the algorithm.

I also have some tools which can extract various modules from certain BIOS types.


Edit:

I found an AMI and an Award BIOS update here:

https://web.archive.org/web/19970707123816/http://www.msi.com.tw/bios/486/4144/4144.htm

https://web.archive.org/web/19970707123816/http://www.msi.com.tw/bios/486/4144/af54s.exe (AMI)
https://web.archive.org/web/19970707123816/http://www.msi.com.tw/bios/486/4144/wf53s.exe (Award)

If you unpack the EXEs with 7Zip, you can extract the 128KiB ROM images (AF54S.ROM and WF53S.BIN).

[poot36]

Here is the dumped file from the original bios labeled ms-4144-1.4 (it is from the rev 1.4 board).  I have also attached the doubled up version labeled ms-4144-1.4-2.  I had to change the file extension from bin to hex if that makes any difference to upload it here.  I got the other bios files from this websight at he bottom of the page: https://soggi.org/motherboards/msi.htm.  I first tired the bios update files and having no luck with them decided to try flashing the dumped file to the 29c020 chip as non doubled up and then doubled up and it did not work.  I do know that the bios dump is good because a few years ago I did program it onto a M27C1001-10F1 ST chip from (if the date code is to be correct 2013 (this may be a fake chip)) for a rev 1.3 board of the same model and it worked.

[fzabkar]

Your -2 image is not correct.

Code: [Select]

Offset(h) 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F

00005220  15 F7 FB 2A F5 B4 35 FE 15 0D 7C DD 0D 7E 8E 84

Code: [Select]

Offset(h) 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F

00005220  15 F7 FB 2A F5 B4 35 FE 15 0D 0A 7E 8E 84 22 09

Notice that 0D (carriage return <CR>) becomes 0D 0A (<CR> <LF>). That's an artifact of a text editor.

[poot36]

Thanks for figuring that out,  I will use the HxD hex editor to double it up and try again.  I have attached the doubled up file using HxD.  Now just to wait 5+ min for the programmer to boot up and I will be able to give it a test.

[poot36]

Well it is working now with the original ROM.  Now to try the updated ones.  Thanks for the help.

[asis]

Hi,
You connect the address A17 at 29C020 to GND when you put it on the PCB.
Your volume is different, because A17 hangs in the air (logical 1).

[poot36]

Have found out that A17 is being held high (5V) by the motherboard so that works as well (selects the upper bank).  I have so far had luck getting both the AMI bios's working with my ATI Radeon 7500 PCI video card (thread about me trying to get it working here: https://www.eevblog.com/forum/vintage-computing/486-motherboard-compatability-with-newer-pci-video-cards/).  I also tried the WF54S Award bios and it did boot with a older video card but it did not display with the Radeon card (bios settings perhaps sounded like it was booting though?).  Will perhaps try the WF53S bios later.

[fzabkar]

Those BIOS images would have an area that is reserved for PnP stuff:

https://en.wikipedia.org/wiki/Extended_System_Configuration_Data

I expect that the ESCD would be recreated by the BIOS, but you may need to go into BIOS setup to do so. It may be that the ESCD created by AMIBIOS confuses an Award BIOS, resulting in a black screen???

Edit:

The "extended" area in CMOS RAM has chipset related configuration data, at least in some BIOS-es. You may need to clear CMOS RAM in such cases.

[asis]

You, yourself, pin-30 (29C020) bend so that it does not fit into the DIP32 socket and connect with a jumper to Vss pin-16.

[fzabkar]

You, yourself, pin-30 (29C020) bend so that it does not fit into the DIP32 socket and connect with a jumper to Vss pin-16.

There is no need to do this. The same data have been copied to the upper half of the IC, so A17 can remain tied to Vcc.

In fact, if the OP has two preferred BIOS versions, they could be stored in separate halves of the ROM and then selected via an SPDT switch.

[asis]

I agree.I did so.

[poot36]

I have tried the WF53S bios and it also works.  At some point I will try the WF54S bios again after clearing the cmos settings.  My end goal with this is to have the fastest video card possible in a 486 if possible.