Sharp air conditioner

[Mark Pierson]

I just got a couple of Sharp A/C units that came with remote controls. Of course, before I even installed them, I dug out the remotes and attempted to learn the signals!

As luck would have it, IR w/DecodeIR.dll v2.14 doesn't decode them at all. No information for Protocol, Device, etc. Worse yet, all the signals appear to be identical based on the Sent Once and Sent Repeatedly data., and they do not work (the A/C doesn't respond).

I found the Sharp Air Conditioner.txt file that includes a custom protocol (based on Kaseikyo) with 5 of the 6 functions my remotes have. This device/protocol upgrade DOES operate the units, so I'm heading in the right direction.

So now I'd like to find the proper command for the SLEEP function I'm missing. I've uploaded Sharp_AF-06ERL_Air_Conditioner_1994.txt to the Diagnosis area. The signal I need is learned to the AUX1:1 button. Since the learned signals all appear to be the same, I can't find any correlation between them and the data in the upgrade file, maybe one of the learned signal decoding experts can help me out?

[jon_armstrong]

Mark,

I suspect the learning is bad. You might try learning with a different remote. Try just tapping the key on the OEM remote rather than holding it down. But I can't see why a 1994 would have trouble learning Kaseikyo that is the same envelope as Panasonic (just different calculation of the check byte). I do seem to recall that some of the air conditioners use two way IR communication and that may be causing the problem or confusing the 1994.

But like we always say, make sure to align yourself with magnetic North and don't forget the aluminum foil over your head.

There is a sleep command in the device upgrade and I found a cml file with a Sharp A/C with the same command (OBC=25). So you might try that again.

[Mark Pierson]

I suspect the learning is bad. You might try learning with a different remote.

I thought that as well, so I learned to an 8810w and a 2104, all with the exact same results. I just tried again with another 2104 and brand new batteries.

Try just tapping the key on the OEM remote rather than holding it down.

Tried that too. Press and hold, quick press, alternating fingers... made no difference.

I do seem to recall that some of the air conditioners use two way IR communication and that may be causing the problem or confusing the 1994.

But the upgrade file works... so I don't think that theory applies here.

But like we always say, make sure to align yourself with magnetic North and don't forget the aluminum foil over your head.

Did all that,and even stood on my head, too!

There is a sleep command in the device upgrade and I found a cml file with a Sharp A/C with the same command (OBC=25).

On my units that's actually called the OFF TIMER. This A/C also has a SLEEP function (it doesn't turn the unit off after a period of time; it simply raises the preset temp by 2' after 1 hour).

[jon_armstrong]

It seems like that if the OEM remote works, and the upgrade works, then somehow the learning process is being corrupted. It is unlikely that the Air conditioners respond to two different IR protocols.

Maybe try the opposite. Depress the button on the OEM remote and keep it depressed, then start the learning process on the OFA remote.

If that still doesn't work, you could send one to me and I can use Tommy Tyler's IR analyzer or Rob can use his Pronto and see what is going on.

[Mark Pierson]

somehow the learning process is being corrupted

I've tried several different variations on several different remotes and always end up with the same results.

It is unlikely that the Air conditioners respond to two different IR protocols.

I would tend to agree with that assumption as well. I thought there might be a combination of IR and RF, but can't find an FCC ID# anywhere, so I doubt that.

You mentioned 2-way IR in your previous reply, so I took my camcorder and viewed the IR window on the A/C itself to see if there's any activity, and there isn't.

Maybe try the opposite. Depress the button on the OEM remote and keep it depressed, then start the learning process on the OFA remote.

Tried that and failed. The JP1 remote doesn't learn anything that way (the rapid LED flashing times out followed by a single blink).

If that still doesn't work, you could send one to me and I can use Tommy Tyler's IR analyzer...

I'll keep that offer in mind. Maybe someone else will jump in here with additional theories in the meantime.

[Mark Pierson]

I searched the old Yahoo messages and found this post where Dan Arnold mentions being helped by Jon and John to create the upgrade and protocol:

I have a Sharp air conditioner at home that John Fine and Jon Armstrong helped me add to my URC-7800.

Subsequently, I tried searching for messages pertaining to that help, but can't seem to find them (no surprise given Yahoo's wonderful forum seach engine).

Does any of this ring any bells, Jon?

Here's another interesting detail: I took the 1994 with the working upgrade, and learned the signals into another remote. Guess what? They're identical to the signals learned from the OEM (i.e. no decode info in IR), and the learned signal doesn't work the A/C! Very strange...

[johnsfine]

I took the 1994 with the working upgrade, and learned the signals into another remote. Guess what? They're identical to the signals learned from the OEM (i.e. no decode info in IR), and the learned signal doesn't work the A/C! Very strange...

That's the sort of thing I like to see in a bizarre symptom: A process any of the experts can try for themselves that should duplicate the symptom without needing the orriginal equipment.

I'm of course too busy to try it, but I assume some other expert will.

If I understand you correctly, you have an upgrade for a protocol structurally identical to Panasonic; The upgrade generates correct signals, but UEI remotes can't learn those signals correctly.

As Jon already mentioned, there is nothing about the Panasonic structure which should make learning in any way difficult.

[The Robman]

Hey Mark, try this upgrade in your 15-1994. I have programmed some of the nearby OBCs to numbers 1 thru 7.

Upgrade Code 0 = 0F CF (Cable/1999)
24 00 D0 00 05 55 5A F3 08 00 00 28 89 A8 81 68
8D E8 85 B8 80 78 8C F8 84 00 00 00 00 D8 86 18
8A 98 82 58 8E 38 88
End

[jon_armstrong]

I don't have any better ideas, but I did update the upgrade to use the "official" Kasiekyo protocol that is one byte rather than the old one that I kludged together with PB that had two variable bytes. There was also a minor mistake in the old one because the lead-out used then wrong on-time.

So maybe we just let the learning issue rest and you try all 256 possible commands using the new upgrade.

[jon_armstrong]

I took the 1994 with the working upgrade, and learned the signals into another remote. Guess what? They're identical to the signals learned from the OEM (i.e. no decode info in IR), and the learned signal doesn't work the A/C! Very strange...

That's the sort of thing I like to see in a bizarre symptom: A process any of the experts can try for themselves that should duplicate the symptom without needing the orriginal equipment.

I actually did that using my 15-1994 and the original upgrade won't decode in IR, but that is because the lead out On time must have been set for the option of the lead in On time rather than the One On time in Protocol builder. All 48 data bits are there however. I changed it to the Kasiekyo protocol and retested and all is well.

[Mark Pierson]

That's the sort of thing I like to see in a bizarre symptom: A process any of the experts can try for themselves that should duplicate the symptom without needing the orriginal equipment.

That's what keeps you guys honest!

If I understand you correctly, you have an upgrade for a protocol structurally identical to Panasonic; The upgrade generates correct signals, but UEI remotes can't learn those signals correctly.

That appears to be the case.

I changed it to the Kasiekyo protocol and retested and all is well.

In the lab, perhaps. Here in the real world, it doesn't work!

I have programmed some of the nearby OBCs to numbers 1 thru 7.

None of them do anything, Rob.

[jon_armstrong]

I changed it to the Kasiekyo protocol and retested and all is well.

In the lab, perhaps. Here in the real world, it doesn't work!

OK, it's because I'm dyslexic and transposed the OEM code to 190 70 instead of 170 90. That is now fixed. Try again -- same link.

[Mark Pierson]

I'm dyslexic

Te moo!

Try again

That's more better... it works!

I also found the missing SLEEP command via an EFC search: EFC=087, OBC 042. The complete list of functions is:

Code: Select all

Function  OBC EFC
power     026 247
temp up   027 251
temp down 024 245
selector  028 244
off timer 025 249
sleep     042 087

The only problem now is all the functions repeat 2 to 4 times with what I consider a "normal" button press. I just need to peck the buttons to get a single command sent. Since we've come this far, can this be easily tweaked?

[Mark Pierson]

FYI, using Jon's latest upgrade with the regular Kaseikyo protocol (and the proper OEM codes ), my 8810w succesfully learns and decodes the signals from the 1994. Perhaps that will give John a better understanding (if he's even interested?) why the original upgrade with the Panasonic-like signals didn't?

[jon_armstrong]

This is the difference between the first upgrade and the Kaseikyo protocol. The last burst pairs:

Old upgrade
+3824,-127246

Kaseikyo
+436,-70156

So, it looks like there was a much bigger gap. The +3824 is wrong (and why I think decodeIR.dll didn't decode it), but if Rob or John could tell you what byte in the protocol to tweak to increase the gap to the max (FF, I believe), then I think that will solve your repeat problem.

I just went back and found Dan Arnold's original learned commands and they do repeat but the last burst pair is:

+422 -130648

IR does decode those as Kaisekyo.