From: Andreas Grosche <ten@ru...> - 2009-04-08 21:07:19
# Christoph <lirc@...> wrote 2009-04-08 17:02:
> Do you have anything connected to DCD on that serial port?
No, as I wrote the effects occur even if nothing is connected at all
(which is why I haven't just ordered an
http://iguanaworks.net/product2.psp), so it can't be an issue of any
circuit physically overloading RTS.
Usually, one of the standard (unamplified)
http://lirc.org/receivers.html is connected.
> Looks like the input is floating.
> I suggest that you connect it to a defined signal level, e.g. DTR.
For the above reasons, I'd suspect process timing issues rather than that.
My current layout is this incarnation
http://lirc.org/transmitters.html on a tiny 8x8 strip grid board,
requiring just 2 cut traces, and also suitable for directly soldering
Not sure what you are proposing about DTR: just to use that instead of
RTS (why should it make a difference?), or is there any other
Moreover, I do have +5V from a USB anyway for the transmitter, so if
anyone could point me to the suggested wiring of the TSOP 1738 in this
setting (not shown on http://lirc.org/receivers.html), I would
certainly like to try if powering it on these improves reliability.
BTW I would very much like to find a way to share the port between
LIRC and lcdproc (but couldn't find documentation on how to do that),
as the latter only requires TxD for most cashier displays (VFD/LCD)
that are a great match for "VDR" Linux media servers/DVRs and MP3
jukebox applications, and often have cases that could well accomodate
a LIRC transceiver.
> Probably that won't solve your sync problems,
> because lirc_serial disables interrupts during transmit.
Does it have to, at today's CPU speeds? Shouldn't there rather be a
fast timer clocking LIRC's transmissions?
Or could you propose anything with a hardware oscillator that standard
LIRC might use to send without running into this kind of issues (I
recall even venerable Z80s could drive circuits like these, at some
loss of flexibility, but then again the TSOP 1738 probably won't
properly decode anything far off 38 kHz anyway)?
PowerTOP currently also shows LIRC grabbing 66% of interrupts during
"irsend SEND_START Device Command" transmissions, where there would
otherwise be some 70% of activity required for SAA7146 (DVB) and EHCI