9 Jun
2010
9 Jun
'10
2:30 p.m.
On 9 Jun 2010 at 20:55, Tony Duell wrote:
That is not the same thing at all. The serial
command/data transfer
to/from the system is bit-banged by the microcontroller, so the
microcontroller must be running its firmware correctly to handle that
RESET command. Yes there is such a command which initialises various
parameters of the keyboard, but it's not what I was thinking of.
You miss my point--I was pointing out that on some keyboards, it may
not be possible to determine if the KB firmware is operating
correctly until a command has been sent to reset the keyboard. A
keyboard sending an endless stream of wrong-parity NAK characters
might not seem to be characteristic of a properly functioning
keyboard.
I ran into quite a few variations when I wrote the firmware for the
PS/2/AT-to-XT translator. There was what IBM said--and then there
was what everyone else did.
--Chuck
I was thinking of the hardware reset to the microcontroller that
causes it to run the firmware in the first place. The reset pin of the
8048 is connecvted to an RC network in the keyboard, not to the
hardware reset signal in the host system (strangely, at least on XT
machines, pin 3 of the DIN connector (labelled as 'reserverd') did
carry the system reset line, but no IBM keyboard used it. I don't
think it's connected at all in AT machines.
If the microcontroller does not reset properly, then it will not be
able to accept any commands from the host.
-tony