Tony Duell wrote:
The
reason is that they don't want the expense of adding a diode in the
matrix along with each switch, to prevent the phantom keys.
I suspect it's
also the physcial problem of adding all the diodes. It's
OK if the kkeyboard is separate switches soldered to a PCB, but most are
membrand devices, and it's non-trivial to add idodes to a membrane.
As I said, cost. It's not that they don't do it because it is
difficult; they don't do it because more difficulty means higher cost.
OK, I think we agree. The point I was making is that the cost is not just
the cost of 100 diodes and soldering them in.
It is difficult (not impossible, but difficult) to add components to a
membrane. And a lot of keyboards, including the Type M and some horrible
plastic modern thing I have here depend on the membranes being flat with
no bumps (e.g. diodes). Yes, it's possible to get roudn both of those
problems, but it adds to the complexity and cost.
It is clearly possible to make a keyobard with full n key rollover But it
would probalby be done most easilty (read cheaply) by not modifying the
design of a normaal chap PC keybaord but by going to soemting like
seaprate swithces on a PCB, or contacts that short traces on a PCB (like
some old Commodore keyboards).
One keyboard that certainyl could do full rollover (I ahve no diea if the
interface microcontrolelr handled it properly) was the one for the
HP9816. This is the only keyboard I have seen where hte switches are
indicidually wired to IC inputs (in this case inputs on 4051
multiplexers) rather than being wired in a matrix.
If customers weren't willing to buy it without
n-key rollover, they'd
manage to include the diodes, by surface-mounting them on the back side
of the PCB. However, the vast majority of customers are willing to buy
the cheapest crap they can get.
Unfortuantely true... Something has to be wrong when I can buy a
complete PC keyboard (101 keys + encoder electronics) for less than the
cost of half a dozen (reasoanble quality) keyswitches.
-tony