Re: Homebrew IR receiver: RTS at -10v

[Robert T. <mai...@fa...>]

(First off, apologies for breaking threading, but I'm new to the list 
and didn't get the original message and I can't figure out how to get a 
hold of the original message source.)

 > Monitoring the supply voltage to the circuit I see that the RTS line
 > (i.e. the +Vcc) is always at -10V... which isn't going to help!

Let me add a little "me too" to this.

I've noticed the exact same behavior with the serial ports on my 
motherboard (an Abit KR7A-133 with an onboard Winbond W83697HF super-IO 
chip driving two physical ports).

 > Adding some debug prints I can see that the command in lirc_serial.c
 > that should toggle RTS high (and hence generate +10V rather than -10V)
 > _is_ being executed.

I did this too and went one further.  I instrumented the lirc_serial.c 
driver to toggle the RTS and DTR lines on and off during init:

/* RTS off */
soutp(UART_MCR, UART_MCR_OUT2);
schedule_timeout(10 * HZ);
/* RTS on */
soutp(UART_MCR, UART_MCR_RTS|UART_MCR_OUT2);
schedule_timeout(10 * HZ);
/* DTR off */
soutp(UART_MCR, UART_MCR_OUT2);
schedule_timeout(10 * HZ);
/* DTR on */
soutp(UART_MCR, UART_MCR_OUT2|UART_MCR_DTR);
schedule_timeout(10 * HZ);

On my Abit board, the above code seems to have no effect on either 
serial port (which I've verified with both my multimeter and an inline 
serial tester), both pins/lines just sit at -11V.

On another computer I have access to (but which I can't actually use 
LIRC on), the _exact_ same code toggles the RTS and DTR lines just as I 
would expect.

Consequently, LIRC and my transceiver work just fine on this other computer.

I'm pretty certain that the serial ports on my Abit board are fully 
functional and that it is somehow possible to actually set those lines 
to a positive voltage since the stock linux 2.6.16.7 serial driver will 
load, recognize and properly present functional ports (the combination 
of a null-modem cable, a inline serial tester and some cat/echo 
redirection does show data flowing across the ports).

 > has anyone got any suggestions as to what could be the problem?

Not yet, but since the standard serial code knows the magic to make the 
port do the right things, it's gotta be buried in there somewhere.  I'm 
slowly working my way through the code, but I haven't got the full 
picture yet.  (Other hints or insight from someone with more UART 
knowledge that I would be greatly appreciated :-).

 > can't determine whether my lirc device should be sym linked to the
 > ttys0 device or not? Currently /dev/lirc is linked to /dev/lirc0 which
 > itself is set as a character device. Does this need to be linked to
 > ttys0 or not?

No, it shouldn't.  While both /dev/ttyS* and /dev/lirc* are character 
devices, they've use different major numbers (4 for "real" serial 
devices and 61 for lirc devices) so they shouldn't be symlinked together.

-- Robert