MSZ-LN sulatushuijaus

Omat esp8266 wemos d1 mini kloonit ovat nyt mittailleet lämpötiloja yhteensä 7 vuorokautta ilman uudelleenkäynnistyksiä.

repomies sanoi:

Based on that 22:45 to 23 data I would suspect something strange is happening with the measurements.

I would suspect the shown temperature measurements to have a significant delta to what temperature the indoor unit itself perceives as the room temperature.

Perhaps the indoor unit isn't blowing air with the same speed or perhaps there is an air leak having an impact on the shown measurement.

Klikkaa laajentaaksesi..

Thank you, Interesting, however the temperature is taken from melcloud so that is the temperature the unit is seeing. I'm also logging fan speed and during that strange reading the fan speed was running constant at full speed. Well it's some kind of anomaly that's for sure.

I have been running with shorter runtimes now and more frequent defrosts and it seems like the unit is back in business again. I assume there was some severe icing it couldn't get rid of causing the power loss. It seems to have cleared that up now running shorter runs between defrosts. It's my guess.

Here are the curves for this afternoon. Red is garage temp, purple bars is power consumption, yellow is heat exhanger temperature. I don't see the same loss of heating power with defrosts running each hour. Additional power a while after defrost is reflected in garage temperature rise, (that was not the case last night)

Icing situation is when the heat exchanger temperature (yellow) falls below dewpoint temperature (blue). These periods are filled in with light blue. I'm logging dewpoint from weather.com to get an idea of the current icing conditions and when ice is being accumulated (theoretically) as that is affected by air pressure and humidity which the dewpoint data takes into consideration. So looking at the situation right now the icing periods are not that long, and I could probably relax the max runtime restrictions.

It's of course added complexity but it would be kinda cool i mitsurunner had a setting to download dewpoint data for the place where the pump is situated, and it could take this into consideration when calculating runtimes. However the current system normally works fine just as it is. I think we have had some extra humid air the last days causing extra bad icing conditions.

Mitsurunner doesn't have a lot of memory to play with at least when it runs in esp8266, so things like calculating dew point related matters would be better off done elsewhere, like perhaps Home Assistant. I suppose some sort of integral could be calculated in the microcontroller quite easily, if the current dew point could be fetched from somewhere. But there isn't much storage for history data, nor can it download big amounts of information without pushing it to the flash memory.