[oclug]Keyboard for a Toshiba 4000CDS notebook
bstrosberg at rcpsc.edu
Fri Jan 24 12:16:50 EST 2003
Further to Rod and Charles' comments, as a former keyboard manfacturer &
designer I believe that most keyboard problems can be fixed with great
care, attention to detail and patience. Keyboards are designed for low
cost, and some of the things used are susceptible to damage. Connectors and
liquid spill residue are the most likely culprits regarding problems in the
scanning pattern - and the scan matrix layout is optimized for low cost
(i.e. minimum to no vias and optimally 1 layer) as opposed to a perfect x-y
row column array.
It does sound (without knowing more) that a single control line is
problematic in your keyboard.
Removing and reseating the connectors, and cleaning the board are the first
thing to do. After you clean the key landing pads (the interlocked gold or
carbon finger patterns), remember to clean the associated carbon puck in the
rubber keyswitch. Cleaning one side of a switch doesn't work. As Charles
indicated, "trich" is far and away the best cleaner, but almost every board
in use today is produced using water soluable solder paste, so scrubbing it
with a toothbrush and water is fine. Most initial function board-level
problems are the result of poor post-soldering cleanup of solder paste
residue. You MUST dry the board thoroughly to not leave any potential for
oxydation. No one will admit it, but most circuit board assembly houses
still have a stash of trichloroethane for cleaning really bad boards.
Checking pin soldering of surface mount packages also usually reveals things
quickly. If you've got a good ear, you can detect bad solder joints by very
lightly raking the tip of an exacto knife across a row of pins - cold solder
joints and bad connections "sound" different than proper joints ("tung"
versus "ting"). It's amazing how an exacto knife as a tuning fork catches
the same errors as a bed-of-nails JTAG test suite!
Most keyboards in laptops use multi-conductor single layer flat flex cables,
and these cables are also susceptible to microscopic dendrite growth between
conductors. You can measure the resistance between pins/conductors on these
cables, and if it isn't infinite, you've got a problem. Replacing a bad
flat flex cable usually solves otherwise unexplainable keyboard problems.
Conductivity increases over time between keyboard flex cables conductors and
is a normal, predictable degradation due to cost being more important than
quality. Industrious tinkerers can replace a flat flex cable with a ribbon
cable (costs more, requires good quality manual soldering) that will work
better than the original (if it fits in the space alloted).
Keyboard controller failures are rare, and are outnumbered by
mechanical/electrical problems 30 to 1.
The long as short of it is that like almost everything electronic (based on
personal experience gained by building hundreds of thousands of and
analyzing failures hundreds of keyboards):
- the most common failures can be traced back to "mechanical" problems
(spills, dirt, misaligned connectors);
- the next level of failures is due to voltage/grounding/static issues
(firmware corruption, etc.);
- the least common source of failures is due to catastrophic component
Laptop keyboards are the most miserable to work on, as they are fragile,
sometimes designed for only one assembly, and your first time disassembling
one is usually destructive, until you learn the tricks particular to the
design. You just can't see all the little plastic locking tabs on a plastic
injection mold parting line, until it is apart.
In summary, buying a new keyboard is the quickest way to joy. Taking yours
apart is the best way to learn why and how it failed, which usually make the
next one last forever (or a few years which is forever to computer
hardware). If you are a cheapie like me, you can probably fix the one
you've got - even to keep as a spare.
More information about the OCLUG