dylanfm wrote on 2026-03-09, 01:13:

Is there something special about the keyboard I need? Is a common error? Any ideas about what I should do to fix this?

I got a Model 30, which should be close enough in that regard. I typically use it with a IBM Model M keyboard, which has an AT-style physical interface, but it indeed is special: AT keyboards are supposed to have three modes of operation:

1. The keyboard sends XT-compatible scan codes.
2. The keyboard sends AT scan codes that can be 1:1 mapped to XT scan codes and uses magic numbers to signify any extended keys like the dedicated cursor keys.
3. The keyboard sends native extended scan codes, in which the keys all just have a simple number and no magic is required.

The IBM AT and all AT compatibles use the keyboard in the second mode and have the keyboard controller translate the AT scan codes into XT scan codes, which are then handled by the BIOS or DOS applications that perform raw keyboard access. The IBM PS/2 Models 25 and 30 do not have a keyboard controller, and they do not implement this translation in their custom preprocessor on INT 71. Instead the BIOS sets up the keyboard to operate in the first mode instead. In case of late 90s standard OEM keyboard, it is possible that this mode is not properly implemented. On a related note: Different operating systems tried to use mode 3 because it simplifies keyboard handling, and they all backed out to use mode 2 by default (or just removed mode 3 support entirely) because of compatibility issues. So even a seldomly used mode that is actually useful is not compatible with all PS/2-style keyboards, so I wouldn't be surprised if mode 1, which is most likely exclusively used for the low-end PS/2 models is missing from a lot of "modern" PS/2 keyboards.

dylanfm wrote on 2026-03-09, 01:13:

After that, when I powered it up the floppy drive let a lot of smoke out. I read somewhere that these floppy drives are prone to needing all the caps replaced.

Yeah, those drives are infamous for dying caps. The one that emitted a burning smell on me, though, didn't have a burning cap, though, but a shorted cap likely blew an interference suppression inductor near the drive connector. So I suggest if you recap the drive, you check those inductors for continuity, too.

Thanks for the info.! It sounds like I need to find a Model M keyboard because it doesn't seem like there is a good way of knowing if they are compatible with the Model 25. Or at least I'm not finding that info with Google.

I dont think i've ever had a PS/2 keyboard that didnt work like MKARCHER described. Anecdotal evidence is anecdotal though.

Can you try the other PS/2 port? it /SHOULEDN'T/ matter, both ports should work with mouse/kb. that said, it sounds like this board has some damage that needs to be addressed before going forward. Any chance we can get a picture?

BTW the model 30 is the same motherboard as the model 25. 😀

luckybob wrote on 2026-03-09, 18:41:

I dont think i've ever had a PS/2 keyboard that didnt work like MKARCHER described. Anecdotal evidence is anecdotal though.

Can you try the other PS/2 port? it /SHOULEDN'T/ matter, both ports should work with mouse/kb. that said, it sounds like this board has some damage that needs to be addressed before going forward. Any chance we can get a picture?

BTW the model 30 is the same motherboard as the model 25. 😀

I just tried a keyboard in the other PS/2 port and it works! I still got the floppy disk error and now a 1701 error which I'm guessing is a HD error, but I was able to press F1 to continue. It booted to basic and the keyboard seems to work. So it looks like I have a problem with one of my PS/2 ports. Just to be clear, based on the model 25 technical reference manual I found, I was calling the keyboard connector the one closer to the outside edge of the board and the earphone connector. That is right, right? I will try to T/S that connector. I'm pretty sure it is getting power because the LEDs on the keyboard flash when it is powered up. But when I plugged the keyboard into the other port, after the LEDs flashed, the num lock light stayed on.

Attached are a couple of bad pictures.

luckybob wrote on 2026-03-09, 18:41:

BTW the model 30 is the same motherboard as the model 25. 😀

Are you sure the model 30 and 25 motherboards are the same? Attached in the connector page from Model 25 technical reference showing the keyboard connector as J2 and a snippet of the Model 30 schematic showing the keyboard connector as J1. I will try to check the components associated with the PS/2 connectors, maybe one of the inductors, pull up resistors, or diodes has failed.

yes. you can just swap the boards between the systems. The only change of note is the ISA riser card. The model 25 has just 2 ISA slots for the model 30's three. You can also just swap in the 286 version too.

dylanfm wrote on 2026-03-10, 04:32:

Are you sure the model 30 and 25 motherboards are the same? Attached in the connector page from Model 25 technical reference showing the keyboard connector as J2 and a snippet of the Model 30 schematic showing the keyboard connector as J1. I will try to check the components associated with the PS/2 connectors, maybe one of the inductors, pull up resistors, or diodes has failed.

I guess the J1/J2 mixup is just an inaccuracy of the schematics. It doesn't matter that much, anyway, as the ports are technically equivalent and the BIOS automatically handles routing the data appropriately. In the BIOS, the ports are scanned in order, and the first port on which a keyboard-like response is received will become the keyboard port, the other one will become the mouse port.

When you check the components, also try to check whether the connector itself is good. You should see different impedances betweend data (or clock) and ground or Vcc with and without the keyboard plugged in.

So far as the J1/J2 mixup, I think the schematics are right and it is the technical manual that is wrong. Or maybe it is a difference between the layouts of different versions of the board. I think I have Planar 1. But when I trace the lines through, it looks like J1 is going to the keyboard pins on the 72x803 and J2 is going to the mouse pins on the 72x803.

It seems like the components associated with J2 in that schematic snippet I posted are fine and the signals are connected to the 72x802. So maybe the 72x803 chip is just bad? Maybe it is something less obvious I am missing?

Is it possible that the ability to auto detect which connector the keyboard is plugged into is dependent on which version of BIOS you have? It looks like I have ver 1.02A.

dylanfm wrote on 2026-03-10, 20:58:

Is it possible that the ability to auto detect which connector the keyboard is plugged into is dependent on which version of BIOS you have? It looks like I have ver 1.02A.

The BIOS I reverse engineered, which definitely does the autodetection stuff, has the IBM chips labeled 61X8939 and 61X8940, dated 02/05/87. I can find the string 61X8940 at the beginning of the F segment, but I did not see any version number like "1.02A".

My ROM chips are labeled 00F2093 and 00F2093. My chipset chips are 72x8202, 72x8203, 72x8205, and 72x8300. I dumped the ROMs which are attached.

dylanfm wrote on 2026-03-11, 02:21:

My ROM chips are labeled 00F2093 and 00F2093. My chipset chips are 72x8202, 72x8203, 72x8205, and 72x8300. I dumped the ROMs which are attached.

The BIOS calls itself 00F2092, dated 06/26/87, which is later than my BIOS. It has the core autodetection function at F000:5DDC, so it looks like a hardware issue on your system, unless that BIOS is buggy.

Wow! Thanks for looking into it. I will do some more checking to make sure that I'm not missing anything else, but if it is that chipset chip that is bad I guess I will just have to live with it.