SHARP AC with CRMC-A950JBEZ remote

[menexttoday]

This is my first experience tinkering with an ESP8266 and I would like to say great job. As a beginner I would have never gotten this project finished without this code.

I would like to start by saying that the settings for the A907 remote work also for the A950. Note, I haven't tried the timers yet. The only feature missing is the COANDA which sets the vswing flap to point the air across the ceiling when cooling. If anyone has this working or any ideas I am very interested. I'll try to figure it out anyway.

Once again great job and Thank you!

[NiKiZe]

If you have anything that is unsupported you need to post the data that you get when recording that unsupported feature using IRrecvDumpV2+

[crankyoldgit]

That's pretty close, but we need to handle SwingToggle, which uses the same bits. I'll code something up that handles it better.

[crankyoldgit]

@menexttoday Actually, rather than me guess how they (swing toggle & coanda) interact. Can you please record some state[]s that show how they interact together?
e.g.
Swing all off etc. Then Coanda on, then swing toggle on (noting how coanda is affected on the device), then coanda off (noting how swing toggle is affected on the device), etc etc etc.

[crankyoldgit]

coanda OFF cool mode;

Timestamp : 003249.779
Library : v2.7.20
Protocol : SHARP_AC
Code : 0xAA5ACF10C731220A0D8006F471 (104 Bits)
Mesg Desc.: Model: 3 (A903), Power: On, Mode: 2 (Cool), Temp: 22C, Fan: 2 (Auto), Turbo: Off, Swing(V) Toggle: Off, Ion: On, Light: -, Clean: Off
uint8_t state[13] = {0xAA, 0x5A, 0xCF, 0x10, 0xC7, 0x31, 0x22, 0x0A, 0x0D, 0x80, 0x06, 0xF4, 0x71};

The value of 0x0D is unusual. I'd have expected 0x08 ... i.e. SwingNoToggle.
We currently have the swing bits as the least significant 3 bits. i.e. 0x0D & 0x07 == 0x05 or 0b101 (as you've noted)

yet when it's off in heat mode it goes to 0x08 (aka 0b000 for our purposes) which is what we would expect i.e. kSharpAcSwingNoToggle

What was the settings/config prior to you generating the Cool Coanda Off msg?

[crankyoldgit]

/// Set the (vertical) Coanda Toggle setting of the A/C.
/// @param[in] on true, the setting is on. false, the setting is off.
void IRSharpAc::setSwingCoanda(const bool on) {
_.Swing = (on ? kSharpAcSwingCoanda : kSharpAcSwingNoCoanda);
}

Per my last comment, this won't produce the same "coanda OFF heat mode" msg/state of:

uint8_t state[13] = {0xAA, 0x5A, 0xCF, 0x10, 0xC8, 0x31, 0x21, 0x0A, 0x08, 0x80, 0x00, 0xF4, 0x81};

[menexttoday]

I don't know enough to properly fix code. Last time I touched programming was COBOL and Fortran.

In regards to SWING and COANDA they control the same flap. The options are;

Clear (default start code) 0b000 (0x08)
Toggle swing: 0b111 (0x0F) (This needs to be cleared right after the command is sent to the AC otherwise it will toggle swing every time you change temperature or press another function)
Toggle Coanda: 0b110 (0x0E) (This rotates the swing flap to point up and send airflow across the ceiling when cooling. It aims the swing flap to point down when heating. There is no swing when this is active)
The NoCoanada 0b101 (0x0D) is what I see the remote sends but I would think the clear would be a better setting.

[menexttoday]

In my tests I removed the 0x0D and use the 0x08. The D sets the flap in a 45% angle instead of Up as the clear setting.

…

On Aug 30, 2021, at 6:36 PM, David Conran ***@***.***> wrote: /// Set the (vertical) Coanda Toggle setting of the A/C. /// @param[in] on true, the setting is on. false, the setting is off. void IRSharpAc::setSwingCoanda(const bool on) { _.Swing = (on ? kSharpAcSwingCoanda : kSharpAcSwingNoCoanda); } Per my last comment, this won't produce the same "coanda OFF heat mode" msg/state of: uint8_t state[13] = {0xAA, 0x5A, 0xCF, 0x10, 0xC8, 0x31, 0x21, 0x0A, 0x08, 0x80, 0x00, 0xF4, 0x81 }; — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[crankyoldgit]

Interesting. We might be able to have proper swing positions rather than "toggle" and "coanda"
Any chance you can try different values (0b000 to 0b111) to see what they do?

[menexttoday]

I’ll try it. It will take some time. I’ll get back to you.

…

On Aug 30, 2021, at 6:49 PM, David Conran ***@***.***> wrote: Interesting. We might be able to have proper swing positions rather than "toggle" and "coanda" Any chance you can try different values (0b000 to 0b111) to see what they do? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[crankyoldgit]

No rush. I'll work out how I'll tackle the code depending on the results. i.e. It's better if we can use positional controls if we can, even if they are undocumented ;-)

[menexttoday]

So we can in fact control the baffle. 0x08 = null (I didn’t see any change. The last setting stayed as is) 0x09 = COANDA UP 0 degrees (pushes air straight out along the ceiling in both heating and cooling) 0x0A & 0x0D = both seem to point the baffle to the last fixed toggle position after the swing toggle was set to OFF (heating and cooling) Took me a little while to figure it out since it seemed to choose different degrees at random intervals. ie: if the swing was toggled at the highest position it would return to this position. If swing was stopped at the lowest position it returns to the lowest position. It seems to remember the state of the SwingToggle OFF position. 0x0B = about 30degrees down (heating and cooling) 0x0C = about 45degrees down (heating and cooling) 0x0E = cooling is COANDA up 0 degrees (pushes air straight out along the ceiling) / heating COANDA DOWN about 75degrees down (pushes air towards the floor) 0x0F = Toggles swing ON/OFF The degrees down is a best guess

…

On Aug 30, 2021, at 6:58 PM, David Conran ***@***.***> wrote: No rush. I'll work out how I'll tackle the code depending on the results. i.e. It's better if we can use positional controls if we can, even if they are undocumented ;-) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[crankyoldgit]

Excellent news and work!

So, what exactly does the "SwingToggle"/0xF do? Does it just set the vanes swinging up & down continually, or does it do a "Change to the next position" thing? or something else?

Don't worry about the exact angle, we tend to use values/descriptions like this:

IRremoteESP8266/src/IRsend.h

Lines 69 to 80 in 999dde2

	/// Common A/C settings for Vertical Swing.
	enum class swingv_t {
	kOff = -1,
	kAuto = 0,
	kHighest = 1,
	kHigh = 2,
	kMiddle = 3,
	kLow = 4,
	kLowest = 5,
	// Add new entries before this one, and update it to point to the last entry
	kLastSwingvEnum = kLowest,
	};

0x09 = COANDA UP 0 degrees (pushes air straight out along the ceiling in both heating and cooling)

So, 0x09 is always the "Highest" regardless of heat/cold etc? Correct?

I'm going to map the values to:

Highest = 0x9
High = 0x9
Middle = 0xB
Low = 0xC
Lowest = 0xE (if Heat) or 0xC (if Cold)
Off = 0x8
Auto = 0xF (depending on the answer to the first question in this message)

Sound reasonable?

[menexttoday]

The SwingToggle"/0xF toggles between continual swing or stop swing and keeps it at current position. 0x0A and 0x0D remembers and brings it to this position. The baffle directing the air is curved. On regular operation the curve angles downward. The 0x09 rotates the baffle so the curve of the baffle angles upward. It rotates the baffle about 180 degrees so the edge that was on the outside is now on the inside. This is a new feature. It’s the first time that I see this feature. Here is a link of 4 videos and diagram explanations. The 2 videos on the top left show the operation of my units. The lower top left video is the 0x09. https://sharphvac.ca/en/avantages.html The upper left video shows the Coanda heat setting which seems to be the lowest swing setting that can only be accessed directly when heating(0x0E). The only way to access this in the cool mode is by toggling swing and stopping the swing at this angle manually. Then you can use 0x0A or 0x0D. On my unit the 0x09 always flips it to the highest position with the baffle angled upwards. Your settings sound reasonable. In my code I set up a HTML page which acts as a remote. Each time you press a button it sends across just one change. I have to set 0x08 right after I send the command 0x0F otherwise it toggles swing on/off with each ac.send that does not involve a change to the swing settings such as change temp or fan speed. It doesn’t matter with all the other swing settings. I am assuming this will only work with new AC models I have no way to test older AC models.

…

On Sep 1, 2021, at 12:11 AM, David Conran ***@***.***> wrote: Excellent news and work! So, what exactly does the "SwingToggle"/0xF do? Does it just set the vanes swinging up & down continually, or does it do a "Change to the next position" thing? or something else? Don't worry about the exact angle, we tend to use values/descriptions like this: https://github.com/crankyoldgit/IRremoteESP8266/blob/999dde2d71a35e5bbc7cd3a2931724062c3ae40a/src/IRsend.h#L69-L80 0x09 = COANDA UP 0 degrees (pushes air straight out along the ceiling in both heating and cooling) So, 0x09 is always the "Highest" regardless of heat/cold etc? Correct? I'm going to map the values to: Highest = 0x9 High = 0x9 Middle = 0xB Low = 0xC Lowest = 0xE (if Heat) or 0xC (if Cold) Off = 0x8 Auto = 0xF (depending on the answer to the first question in this message) Sound reasonable? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[crankyoldgit]

I've coded it up. Please download & test Branch https://github.com/crankyoldgit/IRremoteESP8266/tree/SharpSwingV / PR #1594 and let us know how it goes.

[menexttoday]

I want to be as clear as possible in my answer here so we all agree on the behaviour. I will first describe the OEM remote options in regards to the swing features and then how our options behave.

OEM remote has 2 buttons.
Swing - Toggles Vswing ON/OFF (Vswing never reaches the highest position that sharp calls COANDA)
Coanda - It reacts as a toggle (ON/OFF) and acts differently depending on the mode.

COOL/DRY

1. ON Flips the baffle to the highest position directing the air across the ceiling.
2. OFF Flips the baffle to the last Vsing position

HEAT

1. ON Points baffle down
2. OFF Flips the baffle to the last Vsing position

Our current available options and AC behavior.
kSharpAcSwingVIgnore - Nothing changes
kSharpAcSwingVHigh - Sets the baffle at the highest position, pushing air across the ceiling
kSharpAcSwingVOff - Same as kSharpAcSwingVLast. Sets baffle to the last kSharpAcSwingVToggle OFF position.
kSharpAcSwingVMid - Sets baffle to the mid position
kSharpAcSwingVLow - Sets baffle to the low position
kSharpAcSwingVLast - Sets baffle to the last kSharpAcSwingVToggle OFF position.
kSharpAcSwingVCoanda - In HEAT mode sets baffle to point down. In COOL mode it sets baffle to kSharpAcSwingVLow
kSharpAcSwingVToggle - Toggles Vswing ON/OFF (Works exactly like OEM Swing) Vswing never reaches the highest position(kSharpAcSwingVHigh). OFF stops Vswing and saves the position of the baffle to be accessed by kSharpAcSwingVLast.

The only comment I have is in using the name kSharpAcSwingVCoanda. Someone familiar with the remote and the Sharp official documentation may be confused and assume it should toggle between highest/lowest as the remote does. kSharpAcSwingVDown may be more appropriate. We could create kSharpAcSwingVCoanda to always issue 0b110 and keep the current functionality intact by renaming the current behaviour as kSharpAcSwingVDown. This isn't a big deal for me because I can still program the official behaviour into my remote.

My AC unit responded exactly as described. I saw no issues.

[crankyoldgit]

kSharpAcSwingVCoanda - In HEAT mode sets baffle to point down. In COOL mode it sets baffle to kSharpAcSwingVLow

Note: That's true if you use setSwingV(kSharpAcSwingVCoanda);, if you use the optional extra force argument, you can set Coanda in Cool mode. e.g. setSwingV(kSharpAcSwingVCoanda, true);, then it won't use the kSharpAcSwingVLow position, but will use kSharpAcSwingVCoanda instead. i.e. The highest position.

kSharpAcSwingVDown may be more appropriate.

There is already a kSharpAcSwingVLowest name/value, which is exactly the same value as kSharpAcSwingVCoanda.
In my mind, kSharpAcSwingVLowest is a better name, as without the force argument/parameter, it will result in the lowest position possible for any given mode.

If you think the risk of confusion is great enough, I'm happy to remove the kSharpAcSwingVCoanda name for the same value, and leave it as just kSharpAcSwingVLowest, what do you think/prefer?

[crankyoldgit]

I've updated the branch with comments/documentation that tries to explain/capture what you said.

[menexttoday]

It’s not a big issue. Coanda on the remote toggles from kSharpAcSwingVHigh > kSharpAcSwingVLow > kSharpAcSwingVHigh > kSharpAcSwingVLow. I can set this behaviour with the current setup. I think it’s best to leave it as it is. It allows the full control of Vswing.

…

On Sep 4, 2021, at 6:13 PM, David Conran ***@***.***> wrote: kSharpAcSwingVCoanda - In HEAT mode sets baffle to point down. In COOL mode it sets baffle to kSharpAcSwingVLow Note: That's true if you use setSwingV(kSharpAcSwingVCoanda);, if you use the optional extra force argument, you can set Coanda in Cool mode. e.g. setSwingV(kSharpAcSwingVCoanda, true);, then it won't use the kSharpAcSwingVLow position, but will use kSharpAcSwingVCoanda instead. i.e. The highest position. kSharpAcSwingVDown may be more appropriate. There is already a kSharpAcSwingVLowest name/value, which is exactly the same value as kSharpAcSwingVCoanda. In my mind, kSharpAcSwingVLowest is a better name, as without the force argument/parameter, it will result in the lowest position possible for any given mode. If you think the risk of confusion is great enough, I'm happy to remove the kSharpAcSwingVCoanda name for the same value, and leave it as just kSharpAcSwingVLowest, what do you think/prefer? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[menexttoday]

My remote has one more button called "multi space”. The unit acts like COANDA at kSharpAuto mode so we don't need to code it. It’s easy to set. The explanation that sharp gives is “quickly cool multiple rooms”. If anyone knows of any advantage for this setting I would appreciate the explanation. I’ll dump the remote data again in case I missed a bit that was set differently. I feel the code as is gives us control of all the documented features of the remote/heatpump. Great job. Very satisfying project.

…

On Sep 4, 2021, at 6:45 PM, David Conran ***@***.***> wrote: I've updated the branch with comments/documentation that tries to explain/capture what you said. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[crankyoldgit]

If you find it, please create an new "Issue" for it. I've merged PR #1594 in to the master branch so basically this discussion is complete.

[crankyoldgit]

FYI, the changes mentioned above have now been included in the new v2.8.0 release of the library.