BIOS upgrade in IBM 5160 doesn't work. Help needed!

[Malc]

I had a look at some of my boards, 64-256K motherboard with original ROMS have no wiring mods, 256-640K motherboard with original ROMS have no wiring mods, A couple of 256-640K mother boards with EPROMS as in your example have the wiring mods on component side. I have used TMS27C256 EPROMS in the Boards with no wiring mods to U18/U19 with no problems.

I've noticed this wiring mod before but obviously never looked close enough, As you say it could be a problem for some Make / Models of 27256 Eproms, I'll have to have a look and see what Eproms i can find.

 

[per]

This is actually a question I've been wondering about for a long time, and that's how the XT motherboard evolved in the later years.

• First revision boards were made untill early 1984, and exactly matches the schematics in the first XT Tech.ref.
• Second revision boards had a small change to DMA timing, and some 1983 boards were patched to include the change (you see these as XT boards lots of patch wires).
• Then I have suspected that there was a third revision without A14 on pin 1 of the ROM socket. I can only assume these might have come out in late 1985 or early 1986.

Technically all XT boards can be 64-256K or 256K-640K boards, in practice it's just a jumper setting and the BIOS ROM chips that actually differ.

 

[modem7]

I intend to add something about this 'MK38000/27256' issue in the IBM 5160 to my web site. Included will be a table of 27256 make/models, and whether or not they are known to be affected by placement into a socket wired for MK38000.

I'll be sure to add a note about certain entities on eBay relabelling chips, i.e. the make/model of 27256 (or 27C256) shown may not reflect the actual make/model.

 

[mrmanse]

Wow, you guys really possess some impressive knowledge! Thank you for explaining this to me, and for the contribution to the community!

First of all, I got my new EPROMS (ST) from ebay the other day, programmed them with the Minipro programmer, and voila! It works like a charm!

On to the matter. As someone suggested, I havn't tried to read the NEC EPROM with vs without the VPP pin tied to +5V, but I believe it's now established that it's held low during read of half the ROM and that obviously doesn't work with NEC D27C256D-25's. Btw, do you also want a batch number/manufacturing date for your compatibility list, modem7? Like I said, I have 3 of them, of which none work.

The parts that DO work are:
National Semiconductor NMC27C2560-200
ST M27C256B-10F1

 

[modem7]

First of all, I got my new EPROMS (ST) from ebay the other day, programmed them with the Minipro programmer, and voila! It works like a charm!

That is the make/model of 27C256 that I use. I have lots of them.

Btw, do you also want a batch number/manufacturing date for your compatibility list, modem7?

At this time, I do not see a requirement.

I intend to add something about this 'MK38000/27256' issue in the IBM 5160 to my web site. Included will be a table of 27256 make/models, and whether or not they are known to be affected by placement into a socket wired for MK38000.

I have done this. The web page is at [here].

 

[mrmanse]

I have done this. The web page is at [here].

Nicely done!

I have now read the product ID's of my ROMS, and the programmer software says they are OK, but I can't find a reliable source to verify them with.

The manufacturing date and product ID's of my actual ROMS are as follows:

NEC D27C256D-25 (doesn't work in 5160)
8604 0x10A4
8609 0x10A4
8609 0x10A4

NMC27C2560-200 (works fine in 5160)
8742 0x8FA0

So if anyone can verify the 0x10A4 and 0X8FA0 product ID, that'd be great!

 

[modem7]

So if anyone can verify the 0x10A4 and 0X8FA0 product ID, that'd be great!

Of 10A4 and 8FA0, the first byte of each corresponds to the maker, and a cross-list of that byte to the maker is [here]. (Byte in far right column.)
So, 10 = NEC, and 8F = National

The second byte of each corresponds to the device type. I do not know of a table for those.

My EPROM programmer establishes the make/model by matching the two bytes to the ID entries in its 'devices' file, a copy of which someone has placed at [here].

 

[pcdata76]

After doing some successful updates on my 5155, i decided to upgrade its cpu with a V20 mainly to gain the possibility of running some real-mode 286 software. For the purpose, bought some v20 chips from eBay and put one of them in my 5155 but didnt post like it has a dead board. I suspected from the possibility of getting a re-badged fake chip first and checked the new chip on my turbo xt clone, it was working so that the chip was good. After some digging on the web, found a thread which explains the same symptom. On a 5160 with early revision of BIOS (1982) and with a conventional memory board AST 6pakplus, system was behaving like totally dead after the change of 8088 with V20, which is the case exactly happened on my system. Instead of removing memory board, decided to go with new BIOS since it was solving the painfully slow memory test of 5155 (especially with full-640k) which is a big plus for me in addition.

I had two Texas Instruments 27C256 EPROM. Erased and flashed with the latest 5160 BIOS on minuszerodegrees.com. Since the board of my 5155 is one of the earliest boards, it had MK38036N in U18 but hopefully TI 27C256's are worked without the need of any rewiring job. So we can add following TI 27C256 EPROM's to the working list.


These were the original ones:

 

[modem7]

I had two Texas Instruments 27C256 EPROM. Erased and flashed with the latest 5160 BIOS on minuszerodegrees.com. Since the board of my 5155 is one of the earliest boards, it had MK38036N in U18 but hopefully TI 27C256's are worked without the need of any rewiring job. So we can add following TI 27C256 EPROM's to the working list.

Thanks for the info. I have updated the [list].