Write parameter via MQTT unreliable about p1p2mqtt HOT 9 CLOSED

Indeed I meant that each publishing on P1P2/W, even if a duplicate of an earlier value, leads to a "parameter write" on the P1/P2 interface for the main controller; but as this is not a real change (unless there was a manul change in between) this does not lead to a real parameter change in the indoor unit. I think this can be filtered in the P1P2MQTT interface but I am not sure it will work in all circumstances. For now I would recommend to filter it at the source: in HA.

from p1p2mqtt.

@Arnold-n I just sent you a mqtt log. ;)

from p1p2mqtt.

This is intentional - P1P2Monitor does not allow more than 1 parameter write per hour to avoid too much flash wear: see README and this discussion.

You should see an error message on P1P2/S/# that you are out of write budget.

In a next version I will make the limit flexible but allowing more writes is "at your own risk".

from p1p2mqtt.

@Arnold-n hm ok, understood. but in my case Ill do writes only 4 to 8 times a day, and there is a 3 hour pause to the next write. so it should work an do not harm any budget.

is there a counter or sensor regarding the write budget? in my case for example the A8_Writes_Refused sensor is always 0.

second question: when is the p1p2 module writing to the daikin? only when the write parameter in the mqtt broker changes or every time when a value is publishd (even if its the same value)?

from p1p2mqtt.

@nicx Can you check the P1P2/W/# topic for write instructions? In your log of less than 1 hour there are really 7 parameter writes (and perhaps more failed attempts): 2 to switch heatpump on (35:2F -> 01/01) and 5 writes to change the DHW setpoint (36:03 -> 40.0/35.0/30.0/43.0/30).

You can make the write-budget visible by adding 0x20000 to the J-mask, and restarting mqtt-discovery for HA with D3.

Writes_Refused only counts writes which are refused due to command overruns (which will be solved by buffering); perhaps another counter is useful. There is a counter "Counter_Parameter_Writes" but by default it is not output unless you add the 0x20000 mask option.

from p1p2mqtt.

@Arnold-n you are right, these multiple writes where my manual tests... but that is not the normal case ;) i didnt now the budget-feature before.

could you give an answer to my second question? because maybe that is my problem: i am setting the write parameter in mqtt regularly, even if ther is no change. for example the on/off setting you can see in the log, the daikin is on, i am setting the write parameter to turn on. the parameter changes max twice a day. but it is set every hour.

from p1p2mqtt.

The module always performs a parameter write action when asked to with the 'E' command. This could be filtered but so far I did not see the need for it.

from p1p2mqtt.

@Arnold-n Unfortunately, I don't quite understand: what do you mean by "asked for it"?

example: i publish the value "E 35 2F 1" in the topic P1P2/W at 10:00 a.m. the same happens at 10:30 a.m., i.e. the identical value is published again. between 10 and 10:30 there is no publishing of any value.

do 2 write processes then take place? is the time of the publishing the write trigger and not, for example, a value change?

from p1p2mqtt.

I changed the automation in HA for now and will monitor the behaviour. Thanks so far @Arnold-n for clarifying.

from p1p2mqtt.