[NTLK] MP2x00 Internal Interconnect Breakout

[Jake Bordens]

>SerPortSelect is an output. I assume that this pin goes low (high?) for a 
>few cycles to test if a modem card exists. During those cycles, the modem 
>card must pull some other line low (or high?) that was previously high 
>impedance (or high, or low?). Question iss, which line(s), and what 
>logic. I took a good look at the SER-001, but I could not figure it out. 
>And again - it did work once for me, but I could not repeat it! Dang it!

I'm not sure about pulling something high or low to inform the newton that 
there is a modem. I always thought it was the opposite, these select pins 
allowed the newton to say "I'm addressing the internal modem now" or "I'm 
addressing the external port".  If there happens to be something there, 
then good... Otherwise it would be sending serial into the void.

My design used a bus buffer with tristate enable (something like an 
73LCX126).  The enable pin was attached to the SerPortSelect.  When 
disabled, this would isolate the modem module from the main board.  For 
boards like the SER-001, the LTC1323 has its own enable lines, and I 
believe that the SerPortSelect line is connected to these enable lines on 
the SER-001 but I haven't looked.

Unrelated note on inputs:  The external port has a serialGPI input.  I was 
surprised to see that this is actually mentioned in the Newton Programmers 
Reference.  RegPowerHandler can tell the handler that the newton was woken 
up by a signal on GPIO (pin 7) of the connector.  Unfortunately, there 
doesn't seem to be an equivalent for the modem port.  I wish I knew more 
about the GPIO pin (26) that can be used as either an input or output.  In 
software, I don't want to probe randomly at all the GPIOs just to find 
what GPIO# is is.