the most ocmmon one is ahort somwehre in the keyboard memberan matrix, or equvalently a key stuck down. I did investigate this some years ago. I started fof by taking the keyboard apart : Unplug the keyboard cable at the keyboard end, remvoe the black push-in 'feet' (if present) and the 2 screws udner them. Then take off the top part of the caseing. Free the locking tak and slide the guts to get it out of the botto mcase. OK, the /guts' conissts of the keybaord membrane assmebly linked to a little encoder board. The ones I saw had an 8051 microcotnroller on them. I am told oterhs were used too. The encoder is fized to the keybaord by a metal clip on the tap of the votlage regualtor and t least one screw. With those out, you can move the encoder away. It's connected to 3 'tails fro, the memberan =assmebly, which are held in place by metal clips. Bend up the tabs on these (on the print seid of the PCB), remvoe the clips and separate the encoder and embrane assemnly. I foudn that if you conencted the bare encoder board to a VT220 (or whatever( useing the standard cable, it would power up with no errors. And after doing a lot of circuit-tracing had the conenctions to the memorbrane assmbly worked out. Shortign the approrpriate contacts o nthe PCB (where the tails connected) simulated a keypress, and I found that worked too. The encode board was not the problem. The keycaps come off quite easily (you need to twist them to free the lockign barbs, this is obvious if you are doing it. But that didn't help with the fault. The membrane assembly is heat-staked togetheer. I tried cutting off th=e stakes. It'll all come apart easilty, but I never foudn a way to get it to stay back together agina for any length of time. -tony