From: Adam M. <ad...@me...> - 2002-06-14 01:05:07
|
I was using lirc for a while with my desktop's built-in SIR reciever. The range was crappy (1m). Then I attached a homebrew serial reciever that a friend built for me. Much better -- now the range is several meters. Unfortunately, I use a big-screen CRT projector, so what I *really* want is to be able to bounce the signal off the wall (ie face the screen, press a button, signal bounces off the wall and is received by the computer *behind* me). I'm using a radio shack 6-in-1 remote (rebranded Cinema6, I think) for both my projector and my computer, and the projector works bouncing it off the wall, so I know that the remote is strong enough. When bouncing off the wall, I still see activity in mode2, but lircd doesn't interpret it as signals. I can even train lircd (via irrecord) in bounce-it-off-the-wall mode, but when I do that, it generates a 'raw mode' config file which doesn't seem to work. If I train it by pointing the remote at the reciever, it works okay, but only when pointing it *at* the reciever (not bouncing off the wall). Anyways, I know the signal is getting to the reciever... is there any way for me to tell lircd to take the signal it gets and just guess which button's signature is closest to the signal it recieved instead of demanding a perfect match? I only programmed 6 or 8 butons, so a good guess should be good enough... Also, since the Cinema6 is pretty popular and is programmable-and-learning, can anybody else recommend a 'personality' with the remote that gets good range with lircd? I have it emulating some random Sony VCR right now. Thanks! - a -- Sick of HTML user interfaces? www.xwt.org |
From: Bill P. <goa...@ya...> - 2002-06-14 20:17:26
|
This will anger alot of people, and I AM duly impressed with Milan's IRdA driver (whoa that is cool!) but nonetheless... IRdA should never be used as a consumer IR receiver. It's like complaining to Dodge that your Viper has poor traction in snow. Yea it's a car and yea you can drive it on snow. Just not very well nor very far. :-) I'm impressed you got it to work because so many can/will not. But as you discovered it's range leaves something to be desired for a remote control receiver app. It must be said: It was never intended for this. In your case you're using it in a projection TV room, so there's probably not much sunlight here because of glare. But in a brightly lit room or for controlling your MP3 player thru the window from outside on the patio, it will never work. Don't even try, anyone reading this. (and if you are the 3rd person to tell me all the tricks and stupid stuff you did to avoid making a $5 serial receiver, don't bother. I'll say "good for you!" right here. :-) As for range with a homebrew I can easily get 6m, most units 10m... sometimes more. With a standard remote. My test zone is 6-7m... anyone that can afford a house and TV big enough to sit >6m from the screen can pay an AV consultant. IR protocols use a carrier frequency and if you match your transmitting remote to an IR protocol that's at the target bandpass of your detector module you'll get 10x the range of that IRdA port, in sunlight too. Another thing is wavelength. Most IR remote LEDs are 940nm and so are the detectors for RemCo applications. But there exist 880nm 850mn, etc. If they don't match you are severely limiting your range. An 850mn filtered photodetector will only see ~2% of the light of a 940mn LED... it might not detect with it touching the detector! Since you have a universal remote it's easy. Just set a 'device' to one of the protocols around your detector's frequency. For example, if your friend built the serial detector using the Radio Shack #276-137 40kHz module I'd suggest using a Sony IR protocol (which is also 40kHz). If you're using a 38kHz module then maybe RC-5 protocol, etc. How do you know what IR devices use what protocol, and what the specifications are for each protocol? Research. :-) And an adapter port or built-in IRdA port is, uh, less than directional. With a homebrew serial you can solder it to the end of a 50' wire and place it anywhere and much farther away and aim it... like at a wall so it can see the reflected pulses easier. Also make sure your lircd.conf has the 'frequency' entry for the remote section. (defaults to 38kHz) The dissipation and scattering of light that happens on a bounce shot makes it unpredictable. Even the paint finish, color, and texture will control how far the bounce works. Detector height, aim, ambient light hitting the detector, etc. All are factors. If you're 4m away from a wall you want to bounce directly off of to go behind you another 3m, good luck dude. That (almost) never works. Just point your hand over your head backwards with that remote or use a long wire for the serial detector and run it up front near the TV unit. You can 'loosen up' lirc's detection with eps/aeps in lircd.conf but this may or may not help. Try running "xmode2 -t 1" under X and making it fullscreen. See how much noise the detector is picking up. Walk around, testing bounce shots and direct aims. At a certain point the remote's pulsetrain will get interrupted with a noise spike or not detect a weak pulse, and there's not much to do but get closer or use a stronger remote (or shield the detector better) If while holding a button press you see half of the rendered streams being hosed, that's probably your limit right there. As for replacing it in the case where an IRdA dongle was, that's a cool idea! But, you're still gonna have the problem of positioning. If you cannot point the remote at the computer and must bounce, I highly suggest a detector on a wire you can run along the carpet edge and mount 'up front' with the rest. Sorry for the book, and hope this helps. (fan of the wall-bounce and under-the-door shots too) ------------------------- --- Adam Megacz <ad...@me...> wrote: > > I was using lirc for a while with my desktop's > built-in SIR > reciever. The range was crappy (1m). Then I attached > a homebrew serial > reciever that a friend built for me. Much better -- > now the range is > several meters. > > Unfortunately, I use a big-screen CRT projector, so > what I *really* > want is to be able to bounce the signal off the wall > (ie face the > screen, press a button, signal bounces off the wall > and is received by > the computer *behind* me). I'm using a radio shack > 6-in-1 remote > (rebranded Cinema6, I think) for both my projector > and my computer, > and the projector works bouncing it off the wall, so > I know that the > remote is strong enough. > > When bouncing off the wall, I still see activity in > mode2, but lircd > doesn't interpret it as signals. I can even train > lircd (via irrecord) > in bounce-it-off-the-wall mode, but when I do that, > it generates a > 'raw mode' config file which doesn't seem to work. > If I train it by > pointing the remote at the reciever, it works okay, > but only when > pointing it *at* the reciever (not bouncing off the > wall). > > Anyways, I know the signal is getting to the > reciever... is there any > way for me to tell lircd to take the signal it gets > and just guess > which button's signature is closest to the signal it > recieved instead > of demanding a perfect match? I only programmed 6 or > 8 butons, so a > good guess should be good enough... > > Also, since the Cinema6 is pretty popular and is > programmable-and-learning, can anybody else > recommend a 'personality' > with the remote that gets good range with lircd? I > have it emulating > some random Sony VCR right now. > > Thanks! > __________________________________________________ Do You Yahoo!? LAUNCH - Your Yahoo! Music Experience http://launch.yahoo.com |
From: Adam M. <ad...@me...> - 2002-06-14 20:30:44
|
Bill Paxton <goa...@ya...> writes: > stuff you did to avoid making a $5 serial receiver, It's not the cost, it's the effort. If you don't know how to solder and don't understand electronics (like me), this becomes a really huge job. > I highly suggest a detector on a wire you can run along the carpet > edge and mount 'up front' with the rest. I've already got it working quite well with the detector sitting behind me, so I think it should still work when mounted in-case. - a -- Sick of HTML user interfaces? www.xwt.org |
From: Bill P. <goa...@ya...> - 2002-06-15 06:28:33
|
AFAIK this is not a rebranded unit but Matsushita's own. Either way, the Panasonic PNA46xxM series of detectors work great. There are two types, bare or shield/holder. ex. a PNA4602M is a 38kHz w/o shield, and PNA4612M is 38kHz with shield. It seems most of the older bare models are discontinued, though. The difference in range or noise between shielded and bare isn't that easy to determine... this is one quiet receiver! I had to try to interefere and couple noise to them intentionally. That's a good sign. If you're used to the RS unit this thing is astronomically clean. Even overhead task halogens shining right on it will not cause mode2 detected noise. (whereas that RS unit goes nuts and makes stable detection very limited.) Shielded one is rated at 16m, bare at 10m. I don't have the space! I tested both at 7m line-of-sight and 3/3m reflected, all units worked fine on both tests. The datasheets are available at: (sorry about multiline url) http://www.semicon.panasonic.co.jp/cgi-bin/ds/search.cgi?search_word=PNA46&lang=1 Digikey stocks both models but only the 38kHz one in bare. (PNA4602M-ND, $1.53/ea) The shielded ones have 5 options for shield height and mounting orientation, please grab the datasheet or the Digikey catalog page pdf to see. ($1.72/ea) ---------- __________________________________________________ Do You Yahoo!? LAUNCH - Your Yahoo! Music Experience http://launch.yahoo.com |
From: <col...@hi...> - 2002-06-16 07:47:20
|
Hi! Bill Paxton "goa...@ya..." wrote: [...] > Shielded one is rated at 16m, bare at 10m. I don't > have the space! I tested both at 7m line-of-sight and > 3/3m reflected, all units worked fine on both tests. The pulse width is given with 200-600 usecs. What protocols have you tried this receiver with? Christoph |
From: Bill P. <goa...@ya...> - 2002-06-16 15:41:22
|
Aw crap. Disregard the suggestion. I was using irrecord to generate a file from a universal remote with all the stuff I wanted to test, and then playing it back. In that respect, they work fine. Mitsu, NEC, RC5, Sony... But! It detects RC5 as space encoded, and while it will reliably decode all keypresses thinking that, it will *NOT* decode RC5 from any pre-made config. Sony 12-bit works well with pre-made configs. This (obviously) requires more testing. Although, since this would be a kludge and the configs would be unique and non-transportable... we don't need another format. Live and learn. :-) Good catch Christoph! --- Christoph Bartelmus <col...@hi...> wrote: > Hi! > > The pulse width is given with 200-600 usecs. What > protocols have you tried > this receiver with? > > Christoph > __________________________________________________ Do You Yahoo!? LAUNCH - Your Yahoo! Music Experience http://launch.yahoo.com |
From: <col...@hi...> - 2002-06-22 13:32:29
|
Hi! karl bongers "ka...@tu..." wrote: [...] > Hmmm, makes me think about a correction... > How about: > pulse_stretched=100 > Then just subtract this from every pulse and add it to every > space. > > I think lirc could use a place to put global config options. > The only place used now is the command line option, but perhaps > we can put global(non-remote) settings in lircd.conf? I think this should be a kernel module option. This is not really something lircd should have to handle itself. Christoph |
From: karl b. <ka...@tu...> - 2002-06-17 04:57:11
|
No, not aw crap! I've been using this module and it works good on what I've used it on. It is a good idea to question this 200min/600max us spec. I'll look into this, I have a bunch of the 38khz modules. <..2 hours later..> I just did some tests where I derive 38khz from some code out to the DTR with an IR LED pointed at a Panasonic module. Looked at the signals with a scope. It picks up long and short pulses/spaces fine, but it adds about 100us to any pulse signal. So this brings us back to errors in the reception due to "asymmetric latencies". So maybe this explains the RC5 errors Bill noticed. This would make sense, as RC5("biphase") appears to use the same length pulses and spaces. Hmmm, makes me think about a correction... How about: pulse_stretched=3D100 Then just subtract this from every pulse and add it to every space. I think lirc could use a place to put global config options. The only place used now is the command line option, but perhaps we can put global(non-remote) settings in lircd.conf? We shouldn't have to recompile to switch on software modulation and so on. I would think it would be best to migrate away from pre-compile configure options. On Sun, Jun 16, 2002 at 08:41:16AM -0700, Bill Paxton wrote: > Aw crap. Disregard the suggestion. .... > > The pulse width is given with 200-600 usecs. What > > protocols have you tried=20=20 > > this receiver with? > >=20 |