From: Karl B. <ka...@tu...> - 2003-06-13 05:19:53
|
That doesn't look good at all. The 39's are not a good thing, typical pulses are around 800 to 2000 in size. 39 is way to short, and indicates the lirc_sir driver gets a single interrupt at the start of a pulse, but then gets no more until a new pulse begins(and maybe some spuradic ones inbetween which is what makes it unusable.). What is supposed to happen is that a bunch of back to back interrupts are generated throughout the pulse period, the lirc_sir driver measures that time. We know that IRDA doesn't work for all hardware. Another strike against IRDA I guess. I'm certain that you would not see this problem with the home-brew serial circuit. There is other hardware you can buy as well with varying degrees of tradeoff on features. On Thu, Jun 12, 2003 at 07:07:35AM +0200, Michael Reinelt wrote: > >>I think there's something strange with the serial port on the EPIA,=20 > >>maybe there's some kind of "buffer" which adds some delays, which in=20 > >>turn struggles lirc. Is this possible? How could I debug this? > >=20 > > If you get a chance, send us some mode2 output of the bad readings so > > we can learn more about it. You should be looking at mode2 output, > > it has to be consistent. Try the -m option to make it fit on a screen. >=20 > Here you are: >=20 > first, this is my (working) lircd.conf from the ASUS board: >=20 >=20 > begin remote >=20 > name CV_250-2_P188 > bits 15 > flags SPACE_ENC > eps 30 > aeps 100 >=20 > one 39 2044 > zero 39 1016 > ptrail 39 > gap 46026 > min_repeat 1 > toggle_bit 0 >=20 >=20 > begin codes > POWER 0x00000000000041A2 > .... > end codes >=20 > end remote >=20 >=20 > the mode2 output from the working ASUS board, where I pressed the=20 > "Power" Button twice: >=20 > 16777176 >=20 > 39 2062 39 1048 39 971 > 39 999 39 1040 39 992 > 39 2029 39 2073 39 1040 > 39 2026 39 1062 39 970 > 39 997 39 2035 39 1040 > 39 46018 39 2129 39 946 > 39 1067 39 972 39 968 > 39 2069 39 1041 39 969 > 39 2058 39 1042 39 2024 > 39 2052 39 2063 39 1039 > 39 2047 39 1704313 >=20 > 39 2061 39 1044 39 982 > 39 999 39 1056 39 971 > 39 2053 39 2055 39 1039 > 39 2040 39 1015 39 1040 > 39 999 39 2018 39 1039 > 39 46006 39 2051 39 1039 > 39 1002 39 1040 39 991 > 39 2023 39 1017 39 1038 > 39 2007 39 1040 39 2025 > 39 2129 39 2010 39 1040 > 39 2014 39 >=20 >=20 > this looks good to me, as far as I understand this (and your information= =20 > on http://www.turbobit.com/lirc2.html was very helpful!) >=20 > and here comes a mode2 dump from the EPIA board (I pressed the "Power"=20 > button three times) >=20 > 16777176 >=20 > 39 2063 39 1022 39 176 > 39 779 39 1003 39 1003 > 39 176 39 807 39 2065 > 39 2056 151 885 39 2078 > 39 990 39 176 39 788 > 150 911 39 2064 39 989 > 39 176 39 45858 39 176 > 39 1836 39 1014 39 176 > 39 782 39 1015 39 1006 > 150 1953 150 897 39 990 > 39 2063 39 1004 150 1953 > 39 180 39 1848 39 174 > 39 1863 39 989 151 1952 > 151 1753984 >=20 > 39 175 39 1849 39 176 > 39 807 39 989 39 1015 > 150 899 39 990 39 176 > 39 1864 39 2064 39 990 > 39 176 39 1863 39 176 > 39 774 150 911 39 990 > 150 1966 39 176 39 775 > 150 45959 151 1952 154 896 > 39 993 39 174 39 802 > 150 896 39 2065 39 989 > 39 176 39 800 150 1944 > 151 910 39 2051 39 2064 > 39 2057 150 910 39 2055 > 39 1213609 >=20 > 39 2044 150 911 39 989 > 39 176 39 807 39 1007 > 39 990 39 175 39 1850 > 150 1964 39 176 39 775 > 150 1966 151 885 39 1018 > 39 176 39 774 150 1953 > 150 897 39 46073 39 2064 > 39 990 39 1015 150 899 > 39 1001 39 176 39 1849 > 39 993 150 911 39 2064 > 39 1003 39 2064 39 2051 > 39 2067 39 1015 39 176 > 39 1849 39 176 39 >=20 >=20 > I can see two things: > - the value "39" is somethimes increased to 150 > - the number of bits differ >=20 >=20 > I could think of several reasons: >=20 > - the serial port acts "strange" > - the internal timing (rdtsc?) isn't exact enough > - there are some other interrupts which disturb LIRC >=20 > some other maybe helpful hints: >=20 > - the machine is diskless, it boot sand runs with nfsroot > - at the time of testing, there were no DVB drivers loaded and no VDR=20 > running > - the machine was idle > - I'm using the first serial port, without the serial driver loaded (so= =20 > no "setserial uart noe" is necessary) > - the serial port is at 0x3f8/irq4, the IRQ is not shared (at least=20 > /proc/interrupts says ist only used by the lirc module) >=20 > I hope we can find the reason for this strange behaviour. I want to=20 > finish by VDR PC :-) >=20 > Maybe I could try to assemble a simple serial receiver, but I'm afraid I= =20 > will run into the same timing issues there.... >=20 >=20 > Many thanks in advance! >=20 > bye, Michael >=20 > --=20 > netWorks Vox: +43 316 698260 > Michael Reinelt Fax: +43 316 692343 > Geisslergasse 4 GSM: +43 676 3079941 > A-8045 Graz, Austria e-mail: re...@eu... >=20 >=20 >=20 > ------------------------------------------------------- > This SF.NET email is sponsored by: eBay > Great deals on office technology -- on eBay now! Click here: > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5 >=20 |