Native recognition of IBM scancode sets 1 and 3 on PC

User avatar
Weezer

26 Jun 2020, 00:10

I realized a few months ago that my MSI MPG X570 will interpret my IBM F 122 and my F XT without an active converter when they're plugged into the PS/2 port on the motherboard. I was under the impression that modern motherboards didn't have this capability, so this surprised me. It got me wondering how common this was. I tried to do a search for motherboard compatibility with set 1 and 3 but unsurprisingly I was unable to dig anything up. Does anyone else have a motherboard that can do this?

andrewjoy

26 Jun 2020, 16:50

Does it send the correct scancode ? And is it recognised in windows ?

Only thing i can think of is that the PS/2 is implemented via a converter and is not a true PS2 interface ( does ZEN even support native PS2 ?)

User avatar
Weezer

27 Jun 2020, 00:15

Sorry this post was just sort of a passing thought from a few months back that I'd remembered while I was on break at work yesterday. I don't think that the initial info I provided was super accurate because of that.
I did some quick tests today with my F122. I tried to get the keyboard to work on Ubuntu and in the BIOS of the motherboard and actually, neither worked. However when I hot swap the keyboard into my computer's PS/2 port while running Windows 10, Windows will beep several times in succession, like Morse code, and then I can start typing. I'm typing this post right now on my F122 without a Soarer's converter or xwhatsit. My Windows version is 1903 (pro).

While I was doing the testing, I found this interesting article:
https://www.seasip.info/VintagePC/ibm_1390876.html
The keyboard behaves the same as illustrated in this article, with the exception of the seemingly randomly assigned keys that I'll outline below.
Here is the diagram of the scancodes/outputs from the article which match the scancodes/outputs I'm getting. In the write up the author notes that set 3 is translated by the PC keyboard controller into set 2:
Image

The layout is AT as shown above. The keys that were on the AT keyboard work as you'd expect but on the top rows and the nav cluster:
F1 = F12
F2 = extra pad 5
F6 = ErEOF
F9 = Power (Computer is unaffected)
F10 = Power (computer is unaffected)
F13-24 work as they should.
Left arrow = \
Alt keys do nothing but they return both make and break codes
The rest of appear to correspond to nothing, but they generate make scancodes.

Only the modifier keys generate break codes. The rest only generate make codes.

Additionally, in the article the author notes that he was able to get the keyboard to work in Linux and DOS. He was also able to get it to work on Windows2000 by using a KVM and toggling between two computers while windows loaded. I found this really interesting because it goes against what I had known about these keyboards. The author noted that when Windows resets the keyboard as it starts it gets "flummoxed by the extra bytes which result" and is unable to interpret the keyboard. It looks like that Windows 10 supports hot swapping PS/2 and is therefore able to handle set 3.

Interesting that I wasn't able to get the keyboard to be recognized by Linux. I didn't try to hotswap like I did with Windows.

I remember windows recognizing my F XT as well but I might be misremembering this. I can't test that any longer because I no longer have a fully functional 5150 F XT. It would be interesting if someone could confirm if this works.

Also if someone could confirm my findings here that would be interesting as well. I was under the impression that these keyboards couldn't be recognized by Windows without an additional adapter, but I've just typed this whole post without one. :)

User avatar
Weezer

29 Jun 2020, 21:34

Bumping to see if anyone can confirm that their 5150 keyboard or F122 keyboard will also work on windows using a PS/2 port like mine does.

Post Reply

Return to “Keyboards”