PDA

View Full Version : IBM XT Keyboard I/F Problem



Iain
February 11th, 2006, 10:27 AM
This is a diagnostic challenge directed at any IBM XT experts out there.

I started out with an IBM XT fitted with a motherboard (64/256k) that appeared to have a dead keyboard interface. The XT would boot up but would not respond to the keyboard, an original 83-key version. A second XT-compatible keyboard (one of the XT/AT switchable types) also failed to elicit any response.

Next, I obtained a second XT system unit, and this time I found that the keyboard started working correctly but after a few minutes key presses would be accompanied by beeps, then shortly afterwards the systen no longer responded to any keys.

In my search for some more XT spares I inadvertantly acquired an IBM PC. This worked perfectly; both of the keyboards I had been using appeared to work correctly with the PC.

After some deliberation I decided to send for another XT motherboard. In due course this refurbished unit (640K) arrived and I fitted it in place of the one whose keyboard interface apparently failed after a few minutes. When I powered up the XT everything started as it should, with memory checks incrementing up to 640K, followed by booting to DOS 2.10. Then, just when everything seemed to be fine, the keypresses began to be accompanied by the dreaded beeps and the system stopped responding!

Finally, getting somewhat desperate, I replaced the XT's power supply with that from the PC (also a 130W unit). Initially I thought that this had inexplicably done the trick, but sure enough, back came the beeps! The keyboard interface had failed again.

Right, after all that, I am rather puzzled as to what is causing this fault. The only assumption I can make is that all three motherboards have a similar fault, but this seems to be rather a coincidence. I am not aware of the XT having an unreliable keyboard interface - in fact the machine seems to have a particularly good reputation for reliability. Is there some XT guru out there who has experienced anything similar, and more to the point, some idea as to what the problem is likely to be?

Any assistance would be gratefully received.

Iain

Vlad
February 11th, 2006, 10:31 AM
Looking at from a electronics persons point of view, it sounds like there's a short in the keyboard thats blowing out microfuse's or just burning out the keyboard interface. Thats my guess.

-V

mbbrutman
February 11th, 2006, 10:35 AM
It's hard to imagine things that can kill a keyboard interface, but perhaps it is not the motherboards that are to blame.

I suspect that you are reusing cards each time. An IBM XT can be operated with nothing more than a video card in a slot. You can remove any hard disk controller, floppy controller, memory cards, etc. until you have a stripped down machine.

At this point the machine will only be able to boot into Cassette BASIC, which is kind of funny on an XT as there is no cassette drive interface to save or restore programs with. No matter, it's still there, and it's enough to see if the keyboard works.

If the keyboard works in this configuration, then you can start adding cards back in and see which one 'breaks' the keyboard.

And as always, look for bent or broken pins, leaking capacitors, etc.


Mike

Iain
February 11th, 2006, 12:10 PM
Thank you vlad and mbbrutman for the prompt replies to my XT quandry.

Vlad, I had wondered whether it might be the keyboard causing the problem, but the keyboards seem to work quite happily with the PC I mentioned. Unless the PC has a more robust interface than the XT I would have thought this exonerated the keyboards.

Also, as the XT works quite happily for a few minutes, this tends to indicate that the interface is not burnt out or that the microfuses have blown. (I am not familiar with the microfuses, so perhaps you could let me know where they are for me to check them).

Mike, you are correct, I have been using the same additional cards. When I have substituted the motherboards I simply refitted the selection of cards in the XT. There is one full length card for the mono monitor, one full length card for the floppy controller and one for the hard disk controller, plus a short card (with a 37-pin socket on it). I must admit I hadn't considered removing the controller cards - I'll give that a go. Have you come across extension cards giving rise to a keyboard problem like this? The problem in my particular case seems unusual in that the interface initially works correctly, it's just that it falls over after a few minutes. And doubly strange because it has happened on two successive motherboards.

Thanks again for the suggestions.

Iain

mbbrutman
February 11th, 2006, 12:43 PM
Cards on slots should be be able to interfere with the keyboard interface, as the keyboard interface has a dedicated IRQ that is not available in the card slots. But there are lots of potential things that could be going wrong, and the keyboard not working could be just a symptom of other problems. You think that the keyboard is not working, but without the ability to type how do you know that the whole machine is frozen?

My personal bet - one of the cards is bad and as it warms up it start drawing more voltage or injecting noise onto the bus.

So yes, strip it down to the bare essentials and see how it behaves.

Also, find a copy of the diagnostics diskette. It should be with the Guide To Operations. And if that doesn't flesh any problems out, find the Advanced Diagnostics Disk. As they say over here across the pond, 'Rinse, Lather, Repeat'. But do this with as few things plugged in as possible until you start narrowing down the failing component. I doubt that your motherboards are bad.

One other thing - a 5160 motherboard is basically the same thing as a 5150 motherboard, with the exception of the number of card slots and the lack of a cassette interface. Neither will be more robust than the other ..

Terry Yager
February 11th, 2006, 04:31 PM
Sounds like virus shenanigans to me. Try booting from a known clean disk, and look up 'Keypress Virus' at your favorite virus information website.

--T

Vlad
February 11th, 2006, 04:33 PM
A virus on an XT machine??

I doubt it.

-V

Terry Yager
February 11th, 2006, 04:41 PM
A virus on an XT machine??

I doubt it.

-V

Oh yeah, Keypress has been around for a long time:

http://vil.nai.com/vil/content/v_672.htm

--T

Terry Yager
February 11th, 2006, 05:02 PM
Oh yeah, when you go to swapping daughterboards around, keep in mind the dreaded 'slot-8 bug'. Certain kinds of boards can cause strange behavior when installed in slot #8. Try to avoid using it if possible.

--T

itch
February 12th, 2006, 07:49 AM
Does anyone know where to get a real IBM AT keyboard?

Terry Yager
February 12th, 2006, 09:01 AM
Does anyone know where to get a real IBM AT keyboard?

Which one? Model Ms are pretty easy to find, but the AT-style cable is a little more scarce. If you want one of the older 84-key style, you'll have to look long and hard, and probably pay dearly when you do find one.

--T

itch
February 12th, 2006, 09:20 AM
Yes, I am looking for the 84 key original AT, not the M series.

Terry Yager
February 12th, 2006, 09:43 AM
Found another page about another strain of Keypress, which describes pretty much the same behavior noted above:

http://www.f-secure.com/v-descs/keypress.shtml

Keypress kicks-in on a timer, so your computer will boot up and act right for a time before it starts to interfere with key input.

--T

Iain
February 13th, 2006, 07:05 PM
Thanks for all your additional comments.

I've now had an opportunity to investigate the problem a little further.

Firstly I took note of Terry's information about possible 'keypress viruses'. (I hadn't associated XT's with viruses). I booted the machine off a good DOS 3.2 floppy disk I had available. Unfortunately, after about three or four minutes the keyboard stopped responding again.

Next, I considered Terry and Mike's advice about the additional cards in the system, and the warnings about slot 8. As I had originally considered the problem to be associated with the keyboard interface I had ignored the other cards in the system. This notion was reinforced by the fact that I had recently acquired the XT and the previous owner was confident it had all worked correctly. I assumed it had picked up a keyboard glitch of some sort.

So I removed the hard disk interface, disconnected the power from the hard disk, and removed the single short board. I had described the short board incorrectly in my opening description - it actually has a 25-pin socket and a 15-pin socket on the edge connector (keyboard and joystick?). Anyway, once I had removed these boards I powered up the system with the DOS 3.2 boot disk in the floppy drive. Everything started up as usual, but lo and behold, no beeps and no locking up!

I refitted the hard disk controller and hard disk connections and repeated the exercise. Again, the system ran for half an hour without a hitch.

Finally I refitted the small board to the slot nearest the power supply from where I had removed it. Some five minutes after switching on again the keyboard failed to respond. I relocated the card to another vacant slot, but with the same results. When I left the card out of the system and powered it up the XT performed perfectly!

So, this short card appears to have been the cause of my troubles. The main components on the card are three 40-pin chips: Winbond W86C450, W86C451 and W86C456A devices. The first is mounted in a socket, and a closer inspection of the pins indicates some suspicious post-production soldering of its pins, but no obvious shorts. It also has two blocks of ten pins each, marked ASYN1 and ASYN2.

I may try this card in the PC to see if it has the same effect, but for the moment I am more than happy to have a fully working XT.

Thanks again to all who offered advice on this problem. I was really impressed with this forum and how quickly you responded to the problem with my XT.

Iain

mbbrutman
February 13th, 2006, 07:52 PM
I'm glad the offending card has been identified.

It's an old trick .. if the problem follows you around, you are probably carrying it with you. Stripping the machine to the bare essentials will help isolate the problem.

Enjoy the XT ..

Terry Yager
February 13th, 2006, 08:02 PM
Yeah, sounds like an I/O card to me...and the 'slot 8' thinggy (slot #8 is the one closest to the PS).

--T