I have repaired several more of these keyboards in recent months and have noticed two primary types of failures among these keyboards. The first common problem of stuck LEDs is in fact caused by the dried capacitor(s) mentioned in this thread (in my experience, a dried 2.2 uF capacitor causes the LEDs to stick). The second problem deals with stuck keys.
For those who have the M2 keyboard and are experiencing the 'stuck key' problem:
This problem is ==>NOT<== caused by the rubber mat, or the 'membrane,' as it is sometimes called, and if the mat has any remote effect, it is minor. I would highly recommend not wasting time trying to find a replacement rubber mat or 'membrane.' You will only get more frustrated. The rubber mat has two jobs:
1. Transmit the force of the spring/pivot assembly (the key force) to the two plastic circuit sheets in order to adjoin them (thereby closing the circuit and generating a keystroke).
2. Provide a friction surface for the spring pivots (the tiny plastic piece that the spring is attached to) so that they don't slide around.
On this keyboard there are two thin plastic printed circuit sheets that contain the key switches: an upper sheet for the upper halves of the key switches and a lower sheet for (obviously) the lower halves of the key switches. Between these switch sheets lies a critical sheet that maintains a very small gap between the upper and lower key switch sheets. This "gap sheet" just has a bunch of holes punched through it, and the holes align with each of the switches.
to illustrate (horribly) with ASCII characters: Here is a schematic of a cross-section view of the rubber mat and circuit sheets:
-----------------------
rubber mat
-----------------------
++++++++++++++
upper switch sheet
++++++++++++++
~~~~~~~~~~~~
GAP SHEET
~~~~~~~~~~~~
++++++++++++++
lower switch sheet
++++++++++++++
The gap/spacing between the circuit sheets is critical. Normally, the gap sheet does its job and provides a large enough switch gap that no keystroke is generated when the keyboard is not being used. Over time, and with frequent assemblies and disassemblies, the gaps between the switches can shrink enough to cause one or more key switches to remain closed, always registering a keystroke. It is more difficult to diagnose when the , or other non-lettered keys are stuck, and it takes considerable testing to figure out which specific keystroke circuits are affected. Generally, if one key is affected, it is entirely possible that those near it are also affected.
Common symptoms of stuck keys are:
1. indefinitely repeating characters
2. more than one character generated in a keystroke
3. keystrokes generating the wrong characters
4. keystrokes registering before the click of the buckling spring
5. completely non-functional keyboard (you see the LEDs properly flash once plugged into a computer, but keystrokes do nothing)
To test this out, disassemble the keyboard according to the instructions in this thread (remove the screws, pop the keys off, push the snaps aside, pry it in half), and remove the rubber mat. Plug the PS/2 connector into a PS/2-to-USB active signal converter and plug it into a modern PC. Use some sort of text editor and press the individual dots on the key switch circuit sheets. Each key should register properly. The problem occurs after you re-assemble everything. You plug it in, watch the LEDs do their usual flash, and then you experience one of the symptoms above. This is where the gap problem arises.
Try to determine what area of the keyboard is affected. This may take a few assemblies and disassemblies, but if you are able to locate the affected area, here's what I did:
You solve this problem by increasing the gap between the upper and lower switch circuit sheets. The upper and lower circuit sheets have the printed dots and lines on them (this is really just one single sheet that is folded in half). When you determine the affected area, place thin pieces of scotch tape on the center or bottom sheet (the sheet with holes in it, or the sheet with the bottom switch halves) around each affected switch, being cautious to apply tape **around** but not **directly between** the switches (if you tape directly between the switches, you completely prevent the switch from closing--never do this). If you want to avoid this problem, apply tape only to the gap sheet and DO NOT cover any of the holes. The tape effectively increases the gap, which allows the key switch to remain open when not in use.
For example, on one M2 keyboard I repaired, I noticed that the a, s, d, z, x, and c keys were acting strange. I'd plug the keyboard in and see one or more of these keys repeated in my text editor window, or I'd not get a keystroke at all, or I'd get two or three characters rather than the one I asked for. At times, I would barely press a key and still get a keystroke. I disassembled the keyboard and placed strips of scotch tape around each of the switches in that particular area (the entire lower left area of the keyboard). I put everything back together and there were no more stuck keys.
This has worked every time without failure.