[c-nsp] Splicing a roll-over cable

[Tomas Daniska]

> On Sun, Jul 19, 2009 at 05:08:34PM -0400, Steve Bertrand wrote:
> > Essentially, I'd like keystrokes to be seen on one monitor that is
> > connected to the console that is typed on another device connected
to
> > the same console port.
> 
> This direction should work (having two "receivers" on one "sending"
> line), if the cable is not too long.

Take care to avoid overloading the tx circuitry. Two receivers over a
short distance should be ok. When slightly overloaded, the port may fail
after a longer time (e.g., using a long cable run without line
conditioning circuitry).
 
> The other way ("typing on both machines will end up on the router") is
> not going to work due to the signalling used on RS232 - there would be
> two transmitters fighting each other.

I see no reason why splicing with diodes wouldn't work here, provided
that you avoid sending from both terminals simultaneously. RS232 has
pretty vague voltage levels, so the voltage drop should not be an issue.
The HW flowcontrol signals can spliced this way as well, IMO. Or you can
use an electrical switch to choose which of the two terminals has the
active keyboard.

Even without the diode protection the TX circuits should not fry each
other if connected directly and could work. You try at your own risk ;-)
Cisco console ports used to be tough to burn in older HW (I wouldn't say
this about other Cisco async ports).

> As a corrollary, you can't just "splice all 8 wires", but you'd have
to
> extract RXD (as seen from the host) and GND.

--

deejay
 

__________ Informacia od ESET NOD32 Antivirus, verzia databazy 4259
(20090719) __________

Tuto spravu preveril ESET NOD32 Antivirus.

http://www.eset.sk