lirc-list
|
From: Vince Weaver <vince@de...> - 2001-01-17 04:08:37
|
Hello is anyone using lirc_sir to transmit signals properly? I am using a laptop with SMCC-IRCC SIR. I can record signals from the remote using the irrecord program. But when I try to play them back to change stuff on the TV, it doesn't work. When looking at the signal through an IR-receiver/oscilloscope it looks like the signal being transmitted is too fast by a factor of 2. I've tried changing the driver or the frequencies in lirc.conf and it doesn't seem to slow them down. This is using 0.6.3pre3 Is this a known problem? Am I doing something incorrectly? Thanks, Vince -- ____________ \ /\ /\ / Vince Weaver http://www.deater.net/weave \/__\/__\/ vince@... Linux 2.4.0 on a K6-2+ |
|
From: <stewart@ne...> - 2001-01-17 05:10:37
|
Vince, This sounds like the problem I'm having with my laptop except that I don't have an oscilloscope for debugging. I've been told that the frequency setting has no effect (at least with lirc_sir). If you find a solution, please let me know. Stewart On Tue, 16 Jan 2001, Vince Weaver wrote: > Hello > is anyone using lirc_sir to transmit signals properly? > > I am using a laptop with SMCC-IRCC SIR. > > I can record signals from the remote using the irrecord program. > > But when I try to play them back to change stuff on the TV, it doesn't > work. > > When looking at the signal through an IR-receiver/oscilloscope it looks > like the signal being transmitted is too fast by a factor of 2. I've > tried changing the driver or the frequencies in lirc.conf and it doesn't > seem to slow them down. > > This is using 0.6.3pre3 > > Is this a known problem? Am I doing something incorrectly? > > Thanks, > > Vince > > -- > ____________ > \ /\ /\ / Vince Weaver http://www.deater.net/weave > \/__\/__\/ vince@... Linux 2.4.0 on a K6-2+ > > |
|
From: Milan WWW Pikula <www@ba...> - 2001-01-20 14:42:15
|
Hi,
I am known here for my late responses :I but I found some time to look at
the problem (without success). This is what I have been able to check.
I compared the current CVS version with my original (prehistoric) code
and (behind some cosmetic changes and one bugfix in sending routine)
found this difference:
drivers/lirc_sir/lirc_sir.c, lines 290 and 295; changed
send_space(((unsigned long *)tx_buf)[i]) to send_space(tx_buf[i])
send_pulse(((unsigned long *)tx_buf)[i]) to send_pulse(tx_buf[i])
As the size of unsigned long and int (lirc_t) are the same on Intel,
the only difference is the "unsigned" prefix, which is o.k., because
the prototypes of send_space and send_pulse are unsigned.
So I compiled the new code to find out, if the problem is not caused by
higher levels. To my surprise, it worked correctly. I was able to
send codes to my TV and it responded. I recorded the RC few times
and it worked each time for me.
IrDA hardware in SIR mode is only usable at speeds up to 115200, so
there is no "legal" way, how to set double speed. Vince, did you encounter
the double speed of carrier (almost-44kHz pseudo-wave is now almost-88kHz;),
or the double speed of symbols (e.g. the 44kHz carrier is untouched, but
shorter)?
In the first case, this might be caused by not-so-compatible SIR hardware,
which needs some strange initialization to get into the 115200 7N1 mode;
or my incorrect code for initialization of the hardware;
this might be fixed by some reset of UART or more careful initialization
of that mode.
In the second case, there is something strange about higher layers of
lirc - I looked at the "serial" driver and found the same handling
of "write", so I don't suppose this is the case.
All who have this problem:
Do the following test, please:
1. rmmod lirc_sir; insmod serial
2. run the minicom on the serial port (most likely /dev/ttyS2)
3. set the 115200, 7 data bits, 1 stop bit, no parity
4. leave the minicom using ctrl-a q or alt-q ("leave without reset" feature)
5. rmmof serial; insmod lirc_sir
6. check the transmit again
bye,
Milan Pikula
On Wed, 17 Jan 2001 stewart@... wrote:
W> Vince,
W>
W> This sounds like the problem I'm having with my laptop except that
W> I don't have an oscilloscope for debugging. I've been told that the
W> frequency setting has no effect (at least with lirc_sir). If you
W> find a solution, please let me know.
W>
W> Stewart
W>
W>On Tue, 16 Jan 2001, Vince Weaver wrote:
W>
W>> Hello
W>> is anyone using lirc_sir to transmit signals properly?
W>>
W>> I am using a laptop with SMCC-IRCC SIR.
W>>
W>> I can record signals from the remote using the irrecord program.
W>>
W>> But when I try to play them back to change stuff on the TV, it doesn't
W>> work.
W>>
W>> When looking at the signal through an IR-receiver/oscilloscope it looks
W>> like the signal being transmitted is too fast by a factor of 2. I've
W>> tried changing the driver or the frequencies in lirc.conf and it doesn't
W>> seem to slow them down.
W>>
W>> This is using 0.6.3pre3
W>>
W>> Is this a known problem? Am I doing something incorrectly?
W>>
W>> Thanks,
W>>
W>> Vince
W>>
W>> --
W>> ____________
W>> \ /\ /\ / Vince Weaver http://www.deater.net/weave
W>> \/__\/__\/ vince@... Linux 2.4.0 on a K6-2+
W>>
W>>
W>
W>
--
Milan Pikula, http://www. Finger me for Geek Code.
http://fornax.elf.stuba.sk/~www, www@...
.. dajte mi pevnu linku a pohnem zemegulou ..
|
|
From: <stewart@ne...> - 2001-01-20 15:42:58
|
Milan,
No luck on my end, but thanks for looking into it. If you can think
of anything else, I'm game to try.
Stewart
On Sat, 20 Jan 2001, Milan WWW Pikula wrote:
> Hi,
>
> I am known here for my late responses :I but I found some time to look at
> the problem (without success). This is what I have been able to check.
>
> I compared the current CVS version with my original (prehistoric) code
> and (behind some cosmetic changes and one bugfix in sending routine)
> found this difference:
>
> drivers/lirc_sir/lirc_sir.c, lines 290 and 295; changed
>
> send_space(((unsigned long *)tx_buf)[i]) to send_space(tx_buf[i])
> send_pulse(((unsigned long *)tx_buf)[i]) to send_pulse(tx_buf[i])
>
> As the size of unsigned long and int (lirc_t) are the same on Intel,
> the only difference is the "unsigned" prefix, which is o.k., because
> the prototypes of send_space and send_pulse are unsigned.
>
> So I compiled the new code to find out, if the problem is not caused by
> higher levels. To my surprise, it worked correctly. I was able to
> send codes to my TV and it responded. I recorded the RC few times
> and it worked each time for me.
>
> IrDA hardware in SIR mode is only usable at speeds up to 115200, so
> there is no "legal" way, how to set double speed. Vince, did you encounter
> the double speed of carrier (almost-44kHz pseudo-wave is now almost-88kHz;),
> or the double speed of symbols (e.g. the 44kHz carrier is untouched, but
> shorter)?
>
> In the first case, this might be caused by not-so-compatible SIR hardware,
> which needs some strange initialization to get into the 115200 7N1 mode;
> or my incorrect code for initialization of the hardware;
> this might be fixed by some reset of UART or more careful initialization
> of that mode.
>
> In the second case, there is something strange about higher layers of
> lirc - I looked at the "serial" driver and found the same handling
> of "write", so I don't suppose this is the case.
>
> All who have this problem:
> Do the following test, please:
>
> 1. rmmod lirc_sir; insmod serial
> 2. run the minicom on the serial port (most likely /dev/ttyS2)
> 3. set the 115200, 7 data bits, 1 stop bit, no parity
> 4. leave the minicom using ctrl-a q or alt-q ("leave without reset" feature)
> 5. rmmof serial; insmod lirc_sir
> 6. check the transmit again
>
> bye,
> Milan Pikula
>
> On Wed, 17 Jan 2001 stewart@... wrote:
>
> W> Vince,
> W>
> W> This sounds like the problem I'm having with my laptop except that
> W> I don't have an oscilloscope for debugging. I've been told that the
> W> frequency setting has no effect (at least with lirc_sir). If you
> W> find a solution, please let me know.
> W>
> W> Stewart
> W>
> W>On Tue, 16 Jan 2001, Vince Weaver wrote:
> W>
> W>> Hello
> W>> is anyone using lirc_sir to transmit signals properly?
> W>>
> W>> I am using a laptop with SMCC-IRCC SIR.
> W>>
> W>> I can record signals from the remote using the irrecord program.
> W>>
> W>> But when I try to play them back to change stuff on the TV, it doesn't
> W>> work.
> W>>
> W>> When looking at the signal through an IR-receiver/oscilloscope it looks
> W>> like the signal being transmitted is too fast by a factor of 2. I've
> W>> tried changing the driver or the frequencies in lirc.conf and it doesn't
> W>> seem to slow them down.
> W>>
> W>> This is using 0.6.3pre3
> W>>
> W>> Is this a known problem? Am I doing something incorrectly?
> W>>
> W>> Thanks,
> W>>
> W>> Vince
> W>>
> W>> --
> W>> ____________
> W>> \ /\ /\ / Vince Weaver http://www.deater.net/weave
> W>> \/__\/__\/ vince@... Linux 2.4.0 on a K6-2+
> W>>
> W>>
> W>
> W>
>
>
> --
> Milan Pikula, http://www. Finger me for Geek Code.
> http://fornax.elf.stuba.sk/~www, www@...
> .. dajte mi pevnu linku a pohnem zemegulou ..
>
>
>
|
|
From: <columbus@hi...> - 2001-01-21 09:55:30
|
Hi! Milan WWW Pikula "www@..." wrote: [...] > In the first case, this might be caused by not-so-compatible SIR > hardware, which needs some strange initialization to get into the > 115200 7N1 mode; or my incorrect code for initialization of the > hardware; I think the problem is that the duty cycle becomes much too low at 115200. The frequency that you currently get is 115200/3=38400 which is ok, but the duty cycle is 3/16 (standard IrDA pulse) / 3 = 6.25%. Most receivers need 25% - 33 % duty cycle. In my opinion using a baud rate of 38400 might help. Ok, you won't get the original signal due to the stop bit but nethertheless it might be worth a try. And disabling all interrupts during transmission is a good idea too. Christoph |