Questions regarding John Fine's DecodeIr.html

[Alphix]

I have two questions regarding the protocol descriptions at:
http://www.hifi-remote.com/johnsfine/DecodeIr.html

First, NECx1 has the following IRP notation on the above mentioned page:

Code: Select all

{38.4k,564}<1,-1|1,-3>(8,-8,D:8,D:8,F:8,~F8,1,^108m,(8,-8,D:1,1,^108m)*)

However, testing with makehex seems to generate IR codes that look like:

Code: Select all

{38.4k,564}<1,-1|1,-3>(8,-8,D:8,D:8,F:8,~F8,1,^108m,(8,-8,1,^108m)*)

(i.e. no "D:1" in the repeat message)

Which one is correct? MakeHex or the NECx1 description?

Second, RC5x has the following IRP notation on the above page:

Code: Select all

RC5x: {36k,msb,889}<1,-1|-1,1>(1:1,~S:1:6,T:1,D:6,-4,S:6,F:6,^114m)+

But testing the RC5x generated by my Pronto seems to suggest that it should be:

Code: Select all

RC5x: {36k,msb,889}<1,-1|-1,1>(1:1,~S:1:6,T:1,D:5,-4,S:6,F:6,^114m)+

(i.e. D:5, not D:6 in RC5x before the mid-message gap, thus giving RC5 and RC5x the same format before the gap).

Which one is correct?

Regards,
David

[johnsfine]

Which one is correct? MakeHex or the NECx1 description?

I'm pretty sure that NECx1 description is correct (there is an extra bit in the repeat part). I'll try to remember to review that MakeHex .irp file. (Use of MakeHex output would not break badly enough for people to report if that .irp file had that detail wrong).

(i.e. D:5, not D:6 in RC5x before the mid-message gap, thus giving RC5 and RC5x the same format before the gap).

I don't recall. I have decoded a lot of rc5x with decodeIr.dll and that difference would break all those decodes if DecodeIr source code didn't match the true signals. Also MakeHex has been used a lot by other people for RC5x and that would break if MakeHex were wrong. But DecodeIr.html isn't checked by anyone, so it could easily be wrong. I'll try to remember to check the three against each other when I have time.

[The Robman]

First, NECx1 has the following IRP notation on the above mentioned page:

Code: Select all

{38.4k,564}<1,-1|1,-3>(8,-8,D:8,D:8,F:8,~F8,1,^108m,(8,-8,D:1,1,^108m)*)

However, testing with makehex seems to generate IR codes that look like:

Code: Select all

{38.4k,564}<1,-1|1,-3>(8,-8,D:8,D:8,F:8,~F8,1,^108m,(8,-8,1,^108m)*)

(i.e. no "D:1" in the repeat message)

Which one is correct? MakeHex or the NECx1 description?

The second description is the right one. The repeat trail of NEC1 look like this "+9002 -2266 +550 -96716" so you need the "1" to get the +550 part but you don't need anything from the device code.

[Alphix]

I just checked makehex again for the RC5x definition...it says

Code: Select all

Form=;1,~S:1:6,T:1,D:5,-4,S:6,F:6

So my own decoding and makehex agrees while DecodeIr.html disagrees.

As for NECx1, makehex's NECx1.IRP says:

Code: Select all

R-Suffix=~D:1,1,-170

But thats not what's generated when I use the result with my pronto (the correct 1,-170 sequence seems to be generated). Which would explain why noone has had any problems with the makehex output?

In summary it seems as though DecodeIr.html that is wrong on RC5x and makehex is wrong on NECx1?

[johnsfine]

I'm back from vacation but too swamped to look seriously at any IR questions.

For a superficial answer: The repeat part is often ignored and/or unnecessary, so if some tool has the repeat part of NECx1 wrong, not many things would care.

But I think Rob answered the wrong question. He gave the correct info for the repeat part of NEC1, but we are talking about NECx1, which has a strange repeat part.

[The Robman]

But I think Rob answered the wrong question. He gave the correct info for the repeat part of NEC1, but we are talking about NECx1, which has a strange repeat part.

You're right, I was talking about NEC1, I missed that NECx1 was being discussed.