ACT Apricot keyboards, or infra-red japes

[Tony Duell]

> 
> > Surely there's one encoder and IR transmitter for the
> > entire keyboard (there may be several IR LEDs, but
> > normally they're wired in series and send the same
> > information). In which case if _any_ keys work, then
> > it's very unlikely the problem is with the IR transmitter.
> 
> Not true. IR formats where the energy density per bit type
> differs can fail on specific codes due to the IR LED supply
> capacitor failing. E.g. in the Sony SIRC protocol a 1 bit is

That's why I said 'very unlikely' not 'certainly not' :-)

> transmitted as 1.2ms of carrier and .6ms gap whereas a 0 bit
> is .6ms of carrier and .6ms gap. In this case any command that
> has a sequence of two or more 1s may fail where a code with
> only single 1s way work.
> 
> Not very likely I'll admit but it can happen. The easy way to
> diagnose this is if it works with brand new cells but cells a
> couple of weeks old don't work well.

Even then it may not work, if the inductance of the supply leads is 
significant. Personally, I'd check/replace any suspect capacitors early on.

If you have a 'socpe or logic probe, see if you can find the drive signal
to the IR transmitter (often the LEDs are driven by a discrete transsitor,
in which case look at the its collector, for example). See if you get any 
signal at all from the non-working keys. If you don't then I would really 
suspect the membrane.

It's a pity you don't have a matrix layout for this keyboard. If entire 
rows or columns are missing, it could be an electronic fault -- a bad 
port pin on the microcontroller, or something.

-tony