JP1 Remotes

Okay, so the modulation number is some kind of multiplyer that effects all the duration times? And somehow, the modulation time must be different in Craig's sample, because in all the other duration times, of all the other samples have an 8, where Craigs has a 6?

So theoretically if I multiplied all Craigs 6's by his modulation number, and divided all that number by the modulation that I found in the other samples, I'd be getting 8's?

vickyg2003 wrote:So theoretically if I multiplied all Craigs 6's by his modulation number, and divided all that number by the modulation that I found in the other samples, I'd be getting 8's?

Durations aren't measured that exactly.

A difference of 2 is below the level at which the measurement is accurate. Also, the total of each pair is both more important and more accurate than the individual numbers in the pair.

Look at the most recent "digit 1" with some line breaks added for clarity:

0000 006c 0042 0009
0010 000d 0006 000d 0006 000d 0006 0016 0006 0019 0006 001b 0006 001e 0006 000a 0006 0436
0010 000d 0006 000a 0006 000d 0006 0019 0006 0019 0006 001b 0006 001e 0006 000a 0006 0436
0010 000a 0006 001b 0006 0019 0006 000a 0006 0016 0006 000a 0006 000a 0006 000d 0006 0436
0010 000a 0006 0019 0006 0019 0006 000d 0006 0016 0006 000a 0006 000a 0006 000d 0006 0436
0010 000a 0006 001b 0006 000d 0006 000a 0006 0016 0006 0016 0006 000a 0006 000d 0006 0436
0010 000a 0006 0019 0006 000d 0006 000d 0006 0016 0006 0016 0006 000a 0006 000d 0006 0436
0010 000a 0006 0010 0006 0019 0006 000a 0006 0016 0006 0016 0006 000a 0006 000d 0006 0436
0010 000a 0006 000d 0006 0019 0006 000d 0006 0016 0006 0016 0006 000a 0006 000d 0006 0436
0010 000a 0006 001b 0006 0016

We see clear sets of nine burst pairs, with the same ending value (436) for each. In the XMP we understand, that ending value should vary.
In the Nokia protocols we understand, there shouldn't be nine pairs (Nokia12 would have eight pairs, other Nokia protocols more than nine pairs).

The first pair has a larger ON value (10 vs. 6). Nokia has that feature. XMP doesn't. 10 vs. 6 is barely enough for a capture device to get right. Maybe the capture imagined that.

The middle seven pairs of each of those lines carry the real data. What are the unique burst sizes? In either Nokia or XMP the unique sizes should be 5 or 6 apart. Anything closer should be a capture error:
6-a
6-d
6-10
6-16
6-19
6-1b
6-1e

I can't make sense of that. In Nokia there would be only four different values, maybe 6-a, 6-10, 6-16 and 6-1b. But then 6-d is exactly half way between 6-a and 6-10. It could be an error for either. 6-19 is barely closer to 6-1b and 6-1e is a big error from 6-1b. It just is too evenly spread to figure out.
In XMP the total range should be higher. The largest value 6-1e is unreasonably small.

Rob once sent me a document describing some "XMP" protocols that weren't the "XMP" we knew about. But I can't find it. Maybe he'll give me another copy. But whatever it is, it wouldn't make sense for all those burst sizes to be valid 6-a, 6-d, 6-10, etc. 6-d is simply too far from 6-a to be legit as a correctable misread of 6-a in a protocol as picky as XMP or Nokia, but too close to 6-a to be legit as a different burst.

John, thank you for that comprehensive reply. I had popped in for the shorter version and was totally confused. I've often wondered how you could determine a bad learn. I'm going to have to print this out and read this very carefully to fully digest this.

Thanks

I sure hope Craig reads Binkys post and comes back with a signature and perhaps setup code.

After reading this for the 4th time, I think I see the pattern for the "misslearn". I marked them with the xs. Each line repeats with the 4th group of 4 off by -4 (except if its a 000a or 000b, in which it seems to go longer like 001e probably some kind of boolean operation that I just don't understand) and the 8th group of 4 is off by +4 in the repeat. And that holds true with ALL the learns that Craig posted.


Would that be something that the pronto would have introduced? Does that make sense with any known protocol rules?

Edit Saturday July 26 2008 8:05 am
Also, the 6th group of 4 alternates short,short,long,long,short,short or long,long,short,short,long,long for each function. Since UEI markets this protocol as XMP-4-Gamer, as opposed to XMP, XMP-2 or XMP-32, would it make sense that they'd have a parity byte every frame because there would be a lot of signals floating around in gaming? The last video game I played was PONG, so I'm not up on gaming at all. I'm just trying to get the general IR concepts.
End Edit 8:05 am

It looks like Craig got these signals from a bunch of different places, since sometimes they are capitalized and sometimes they are lowercase.

All the ones that were capitalized had those consistant 436s and most of the others vary like the XMP is supposed to. They alterntate like 019a,06bc,01c5,06bc. It looks like he may have been manipulating the code by hand and then gave us some junk.

Craig needs to post some clean learns all from one session and get the signature and setup code that Binky asked for.