AC IR code decrypting

[siklosi]

Maybe this will work

[dpressle]

@Nca78 Thanks for the answer.
The brand of the AC is a local brand called "Tadiran" (not sure if its world distributed).
I am capturing the IR code with following arduino code, please note you need this code because AC have very long IR code since every button push the whole state is sent i.e. fan speed temp mode etc.
Now i know that the code i am capturing is good because i can successfully transmit it back using any IR library.

I agree on the uneven bytes and could not figure this out as well, i also tried the XOR checksum and couldn't get it to match.
The output i am getting from the code is of time delays which i later convert to binary i use the follow mapping to convert to bin:
HDR_MARK 9000
HDR_SPACE 4000
BIT_MARK 620
ONE_SPACE 1600
ZERO_SPACE 540
MSG_SPACE 19000

Please note that there is a message space of around 19 ms that separates the message in to 2 parts, the second part does not change beside the last 4 digits which i think is the checksum. the long pause is right after the 3 digits, so for an example look at the code i got below for a certain operation, they all got the same checksum:
cold, 26 C
Fan:
Auto
100100000101000000000110000 0101001000000000000001000000000000001111
Fan1
100110000101000000000110000 0101001000000000000001000000000000001111
Fan2
100101000101000000000110000 0101001000000000000001000000000000001111
Fan3
100111000101000000000110000 0101001000000000000001000000000000001111

I appreciate the help and hope we can resolve this as i am about to give up...

#define IRpin_PIN PIND
#define IRpin 2


// for MEGA use these!
//#define IRpin_PIN PINE
//#define IRpin 4


// the maximum pulse we'll listen for - 65 milliseconds is a long time
#define MAXPULSE 65000


// what our timing resolution should be, larger is better
// as its more 'precise' - but too large and you wont get
// accurate timing
#define RESOLUTION 20


// we will store up to 300 pulse pairs (this is -a lot-)
uint16_t pulses[300][2]; // pair is high and low pulse
uint8_t currentpulse = 0; // index for pulses we're storing


void setup(void) {
  Serial.begin(9600);
  Serial.println("Ready to decode IR!");
}


void loop(void) {
  uint16_t highpulse, lowpulse; // temporary storage timing
  highpulse = lowpulse = 0; // start out with no pulse length
  
  
// while (digitalRead(IRpin)) { // this is too slow!
    while (IRpin_PIN & (1 << IRpin)) {
     // pin is still HIGH


     // count off another few microseconds
     highpulse++;
     delayMicroseconds(RESOLUTION);


     // If the pulse is too long, we 'timed out' - either nothing
     // was received or the code is finished, so print what
     // we've grabbed so far, and then reset
     if ((highpulse >= MAXPULSE) && (currentpulse != 0)) {
       printpulses();
       currentpulse=0;
       return;
     }
  }
  // we didn't time out so lets stash the reading
  pulses[currentpulse][0] = highpulse;
  
  // same as above
  while (! (IRpin_PIN & _BV(IRpin))) {
     // pin is still LOW
     lowpulse++;
     delayMicroseconds(RESOLUTION);
     if ((lowpulse >= MAXPULSE) && (currentpulse != 0)) {
       printpulses();
       currentpulse=0;
       return;
     }
  }
  pulses[currentpulse][1] = lowpulse;


  // we read one high-low pulse successfully, continue!
  currentpulse++;
}


void printpulses(void) { 
  // IR Remote Results 
  Serial.println("For IR Remote Arduino sketch: ");
  Serial.print("unsigned int raw[");
  Serial.print(currentpulse*2, DEC);
  Serial.print("] = {");
  for (uint8_t i = 0; i < currentpulse; i++) {
    if(i!=0){
    Serial.print(pulses[i][0] * RESOLUTION, DEC);
    Serial.print(", ");
    }
    Serial.print(pulses[i][1] * RESOLUTION, DEC);
    Serial.print(", ");
  }
  Serial.print("};");
  Serial.println("");
  Serial.print("irsend.sendRaw(raw,");
  Serial.print(currentpulse*2, DEC);
  Serial.print(",38);");
  Serial.println("");
  Serial.println("");
}```

[dpressle]

@Yveaux Yes, but i want to be able to control each of the parameters separately and not have to record many operations

[pjr]

Most likely its manufactured by GREE just like most of local branded stuff ;)

[dpressle]

@siklosi Thanks but this is for a different AC with different protocol.
The biggest problem with these AC companies is that each one creates its own protocol and makes our lives hard...

[dpressle]

@pjr Dont know this one, not sure your local is my local :wink:

[Nca78]

Can you please indicate in one of the messages where the 19ms "space" exactly is ?
Maybe the 3 extra bits are related to this space and they should be discarded before/after the space ?

Never heard about this brand, where do you live ?

[dpressle]

@Nca78 The space is where the space is in the binary i wrote (its actually hard to see):
100100000101000000000110000 19 ms space 0101001000000000000001000000000000001111

I am from Israel.

Thanks.

[ToniA]

Would these help:

https://github.com/ToniA/arduino-heatpumpir
https://github.com/ToniA/Raw-IR-decoder-for-Arduino

[dpressle]

I can send the timings i get from the code before i convert to bin, since i am converting it with notepad++ i might be wrong.

[dpressle]

@ToniA I love your library but non of your pre-defined AC's worked for mine.
I will defiantly use it once i crack the code of this AC and will pull it to your repository.

Thanks.

[ToniA]

How about writing a new decode module into Raw-IR-decoder-for-Arduino? I find it very useful in trying to figure out how a certain A/C protocol works.

[pjr]

Gree is the most largest manufacturer of AC devices and Tadiran outdoor unit looks quite a lot like Gree.. at least I'd try that first.

e: Oh. Looks like Carrier also has some relation with Tadiran.

[dpressle]

@ToniA can you elaborate?

[Nca78]

@dpressle said:

@ToniA can you elaborate?

From wikipedia :
"Tadiran Appliances (also known as Tadiran Air Conditioners), was acquired by the Carrier Corporation, the world’s largest manufacturer and distributor of heating, ventilating and air conditioning (HVAC) systems, itself a subsidiary of United Technologies Corporation (UTC) of the USA and a constitute of Dow Jones Industrial Average. Carrier initially acquired 26%of the company in 1997 and took full ownership of it in 2004.[2]"

So looking for Carrier AC infrared should probably help.
Is the reference of your remote control (on the sticker at the back) YB1FA ? As it seems to be shared by gree/carrier/tadiran/ferroli/condor

[Nca78]

From the temperature codes, the checksum is obviously a sum, you have to reverse the bits before you sum, the MSB is on the right in both data bits and checksum.

16° 0000 checksum 0101
17° 0001 checksum 0110 = 0101 + 0001
18° 0010 checksum 0111 = 0101 + 0010
19° 0011 checksum 1000 = 0101 + 0011
20° 0100 checksum 1001 = 0101 + 0100
...
27° 1101 checksum 0000 because 0101 + 1101 = 10000 and you keep only the last 4 bits

Now we just need to understand what bits are included or not in the sum, it seems the fan speed is not, but we need more messages in different modes and with different settings to be able to know which bits are used or not in the calculation.

I take your example code :
100100000101000000000110000 19 ms space 0101001000000000000001000000000000001111
Split in 4 bits words :
1001
0000
0101
0000
0000
0110
000 19 ms space => I ignore this part as it's not matching the 4 bits words, and only 0s anyway
0101
0010
0000
0000
0000
0100
0000
0000
0000
1111

Invert bit order and try to guess what words are included or not.
If you make the sum of included words with binary calculator, it matches the checksum but we must validate with more parameters changed, try with horizontal/vertical swing if you have these on the remote control, then any other specific function.

1001 => on + mode, included as it's the only one with LSB = 1 and we have LSB = 1 in checksum
0000 => fan speed, not included as shown in your examples on fan speed
1010 => temp 16°C + value, included as shown in your examples of temperature
0000 => ??? cannot know because 0
0000 => ??? cannot know because 0
0110 => I think included else I don't have the right checksum
000 19 ms space => I ignore this part as it's not matching the 4 bits words, and only 0s anyway
1010 => I think not included else I don't have the right checksum
0100 => I think included else I don't have the right checksum
0000 => ??? cannot know because 0
0000 => ??? cannot know because 0
0000 => ??? cannot know because 0
0100 => I think included else I don't have the right checksum
0000 => ??? cannot know because 0
0000 => ??? cannot know because 0
0000 => ??? cannot know because 0
1111 => binary sum of all included words, only the 4 LSB bits are kept

[dpressle]

@Nca78 I think you are on to something, thanks a lot. I will investigate it further, but in the mean time here is some more data that i took:
mode fan temp
cold auto 26
1001 0000 0101 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 1111
cold auto 25
1001 0000 1001 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 0111
cold auto 24
1001 0000 0001 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 1011
cold fan1 24
1001 1000 0001 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 1011
cold fan2 24
1001 0100 0001 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 1011
cold fan3 24
1001 1100 0001 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 1011
dry fan1 24
0101 1000 0001 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 0111
air fan1 24
1101 1000 0001 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 1111
heat fan1 24
0011 1000 0001 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 0000
OFF cold auto 26
1000 0000 0101 0000 0000 0110 0000 1010 010 - 0000 0000 0000 0100 0000 0000 0000 1110

[dpressle]

@Nca78 can you please explain why do you add 0101 to the checksum in the temperature example?

[Nca78]

@dpressle
I add it because it is the initial value of the checksum when temperature is 16°C = 0000
I just wanted to show than an increase of 1 in temperature was producing an increase of 1 in checksum.

I think your minus sign (= 19ms pause ?) is in wrong position as it is supposed to be after the 6 (and not 8 ) first words ?

Please take some values by changing the horizontal and vertical swing so we can try to change the second part of the message. After that try some other functions if you have (anything like "eco"/"quiet"/"turbo" options for example) and also if you have it the timer.

[dpressle]

@Nca78 OK, i will generate more commands and see if it makes more sense, Thanks for your help.