Bug in IR7 b3

[PhileFriendly]

I'm not sure where I'm supposed to post this, so apologies if this is the incorrect forum.

Environment:
URC 8820
JP 1.2 Serial cable
Windows Vista RTM
IR7 beta 3

Background:
Yesterday I got my remote talking to the Vista box with no problems back and forth, so things basically seemed to be working.

Steps:
1. Use URC 8820 to learn about 25 commands from 2 different devices (Optoma HD-70 projector, Panasonic receiver)
2. Boot IR7
3. Choose "Download from remote"

result:
expect: download succeeds
actual: alert, "The learned command references an internal burst table, but the index is out of bounds. This could be caused by an incorrect LearnedAddr in the RDF."

this is followed by a ReadAV at 00490658

[PhileFriendly]

As a followup, I re-learned a subset of the commands, and things are working fine now... could it be that I had just tried to learn too many at once?

However, I do see that there is inconsistency in the protocols that the learned commands are saved with. For the Optoma projector with about 10 signals learned, I see protocols NEC1, NEC2, SONY9, and an internal decoder with the following info:
Sent Once:
+2684 -8538 +516 -96506 +8902 -2320 +516 -96506 +8902 -2320 +516 -96506 +8902 -2320 +516 -96506 +8902 -2320 +516 -96506 +8902 -2320 +516 -96506 +8902 -2320 +516 -96506 +8902 -2320 +516 -96506 +8902 -2320 +516 -96506 +8902 -2320 +542 -130526

Shouldn't all of the learned signals be from a single protocol?

Thanks!

[The Robman]

Sounds like you have a new protocol there, which tends to confuse the decoder and cause bad decodes. You should post your IR file in the Diagnosis Area and then start a new thread about this in the Protocol Decodes forum.

[PhileFriendly]

OK, I went back over and re-learned everything, and I think that maybe there is just some flaky/garbled data every once in awhile. I'm now able to get everything to be filed as either NEC1 or NEC2... is that consistent enough?

[The Robman]

Yes it is. Actually, i should have looked at the timing data that you posted more closely as it's obviously just the tail end of an NEC1 signal, which tends to imply that you are pressing the teaching remote button too soon in the learning process.

You should...

1) Put the JP1 remote into learning mode.
2) Press the JP1 remote button that is to be programmed.
3) Wait a second.
4) Press and hold the teaching remote button until the JP1 remote's LED flashes twice.

[PhileFriendly]

Thanks for all the advice. Thus far I've been unable to repro the crash I was seeing previously; perhaps it was some garbled learned signal or something. At any rate, I went through and copied all of the buttons into RM for the Optoma HD70 and uploaded it... it seems to work fine for me (I left the protocol as NEC2).

Thanks!

[The Robman]

I'm surprised that using NEC2 isn't causing you problems. That snippet of code that you posted earlier tends to imply that the original signal is NEC1. If you use NEC2 on a device that expects NEC1, it often has the effect of pressing a button in very rapid succession, which can sometimes lock up the machine.

I'd need to see the learns myself to be able to offer a more informed opinion as to whether you need NEC2, NEC1 or one of the many other variations.

[jetskier]

I found that I get that error on my 10820 when I try to learn to an upgraded remote. If you 981 reset it then relearn, you'll get better results to find the device.