Create a Custom Protocol for Ducasa AC

[ZoiloMora]

Hi! I've been using the software for a few days and it seems amazing!

I have already been able to reverse engineer several devices with recognized protocols.

Now I am trying to read the codes of some remotes for air conditioners of the Ducasa brand but they appear to me as RAW.

I have read in the manual that it is possible to create custom protocols in the software .ini file. I have tried with values that I have seen in the signal but it does not recognize it.

I do not know what I'm doing wrong. Can anybody help me?

[AnalysIR]

Hi

yes we usually add in additional protocols for users by updating the ini file. (within reason)

So just follow the instructions here: https://wiki.analysir.com/index.php?tit ... r_AC_units

We will issue a new ini file once we get them, but it may take a few days.

[ZoiloMora]

Thank you so much for the quick anwser!

So just follow the instructions here: https://wiki.analysir.com/index.php?tit ... r_AC_units

I will read all the information in the link.

Yes we usually add in additional protocols for users by updating the ini file. (within reason)

We will issue a new ini file once we get them, but it may take a few days.

I would like to learn how to create new protocols myself instead of asking for help.

For now it's time to learn.

[AnalysIR]

I would like to learn how to create new protocols myself instead of asking for help.

Sadly we don't publish the syntax as it would generate too much followup & gets complex very quickly.

That is why we offer to add support in for users, with the bonus that everyone gets access to all the new protocols.

PS: Some users have managed to figure out the simple part of the syntax which should be obvious.

[ZoiloMora]

I have already sent the session with the IR codes!

Sadly we don't publish the syntax as it would generate too much followup & gets complex very quickly.

What a pity I think it would save them work. Many more protocols could be generated and loaded by the community itself in a common repository for all users of the program.

PS: Some users have managed to figure out the simple part of the syntax which should be obvious.

Looking at the TOSHIBA_80_96_144_160AC protocol in the AnalysIR.ini file and comparing it with the signals, I think I know what many of the parameters mean.

Code: Select all

[CUSTOM14]
altName=TOSHIBA_80_96_144_160AC
Generic=True
Syntax=HB096THB096T,HB080THB080T,HB072THB072T,HB072THB062T,HB056THB056T
Marks=543,543,4400,543,0,0,0,0
Spaces=1623,543,4400,0,0,0,0,0
Header=4400
HeaderSpace=4400
Mark0=543
Mark1=543
Space0=543
Space1=1623
Delta=200
Bits=0
Carrier=38000 

I created a new custom protocol with the fields Marks, Spaces, Header, HeaderSpace, Mark0, Mark1, Space0, Space1 but for now I don't know what the other fields mean.

Code: Select all

[CUSTOM69]
altName=DUCASA
Generic=True
Syntax=HB096THB096T,HB080THB080T,HB072THB072T,HB072THB062T,HB056THB056T
Marks=470,470,8400,470,0,0,0,0
Spaces=1630,580,4200,0,0,0,0,0
Header=8400
HeaderSpace=4200
Mark0=470
Mark1=470
Space0=580
Space1=1630
Delta=200
Bits=0
Carrier=38000 

I also don't know how the software works since those µs marks are almost never exact. There are always variations up or down...

Sorry if i'm meddling in your work

[ZoiloMora]

I just saw in the user manual a very interesting section on the description of the fields of the configuration file.

I have copied them in a gist to be able to make a quick translation to any language in case someone is interested:

Appendix B – Configuration File, ANALYSIR.ini

[AnalysIR]

OK I got your file and sent you a new ini file with instructions. It does indeed look like a new protocol to add to the list.

The timings of IR signals always vary due to the distortion created by IR receivers. can be +/- 200uS.

However, Our LearnIR module usually captures to within a few uSecs.

re GIST: Please add the version# of the manual & a link back to our website contact page or this forum topic to avoid any possible confusion in the years to come.

In the session history you sent, there we several bad signals. I have added a note to the wiki to keep the remote about 1->2m away from the receiver and point directly at it, when recording signals.

Your A.IR Shield Nano was shipped today & when you get it please submit a new set of Signals to check/validate the timing accuracy etc.

[ZoiloMora]

OK I got your file and sent you a new ini file with instructions.

I have not received any email.

It does indeed look like a new protocol to add to the list.

That makes me very happy!

re GIST: Please add the version# of the manual & a link back to our website contact page or this forum topic to avoid any possible confusion in the years to come.

Added!

In the session history you sent, there we several bad signals. I have added a note to the wiki to keep the remote about 1->2m away from the receiver and point directly at it, when recording signals.

Do you need me to recapture the signals in Arduino with that distance?

Your A.IR Shield Nano was shipped today & when you get it please submit a new set of Signals to check/validate the timing accuracy etc.

Of course, as soon as it gets to me I will! Thank you!

[ZoiloMora]

OK I got your file and sent you a new ini file with instructions. It does indeed look like a new protocol to add to the list.

Having to add the new AnalysIR.ini file and see the AnalysIR updating config file link I have felt very stupid to see that the tests I was doing before opening this post were being done in another directory...

When testing the new protocol, I suspect that it is inverting the bits, seeing many 0xFF

Code: Select all

cold-26c
C7 FB D5 FF FF FF FB 7F 27

I have tried MSB, LSB, and LSB8 to get the checksum algorithm with no success.

I have tweaked the protocol settings a bit to invert the bits.

Code: Select all

[CUSTOM83]
altName=DUCASA72AC
Generic=True
Syntax=HB072T
Marks=470,470,8400,470,0,0,0,0
Spaces=1630,580,4200,0,0,0,0,0
Header=8400
HeaderSpace=4200
Mark=470
Space0=580
Space1=1630
Delta=160
Bits=72
Carrier=36000 # TBC

This is the result:

Code: Select all

cold-26c
38 04 2A 00 00 00 04 80 D8

Find the checksum algorithm by setting the signal format to LSB8.

checksum.jpg

Code: Select all

cold-26c
1C 20 54 00 00 00 20 01 1B

SUM (nibble): 1B

I would like to hear your opinion about it in case I have screwed up something.

[AnalysIR]

Looks good...

I also came to the same conclusion on Checksum.

However I made changes to the ini file as follows:

Code: Select all

[CUSTOM83]
altName=DUCASA72AC
Generic=True
Syntax=HB072T
Marks=530,530,8480,530,0,0,0,0
Spaces=1590,530,4240,0,0,0,0,0
Header=9000
HeaderSpace=4500
Mark=530
Space0=1590
Space1=530
Delta=160
Bits=72 
Carrier=38000 # TBC

If temperature is in Celsius it looks like the LSB bit of nibble5 and the top 3 bits of nibble6 with the result added to 15 to get the actual Celsius value.

It doesn't all match, but that could be due to operator error when capturing the signals.