ya neither the current rule or yours will work for me as even with a 10 degree difference in temp vs set temp the ac does not stay on constantly and cycles so that the time will never hit the 30 minutes or whatever i set.
your are also talking about 2 minutes. i was talking about 30 minutes and at least two subsequent refreshes of the therm state which is what is causing the issues.
i think i just need to ignore the time on for the therm and just go with temp setting below the outside temp .. door open too long and ac on cooling..
next attempt since i cannot rely on determining how long the ac is running since it cycles on and off even with a wide temp differential.. so now just determine if it is running or will run at all with the door open for a certain amount of time.
Can you elaborate? Are you saying that the thermostat's state in Hubitat appears to cycle or the thermostat itself is cycling on/off as it hits temp/loses the target?
If the approach we've tried thus far doesn't pan out, it'd be trivial to track the total runtime of the AC across an hour--I already have a more complicated version that tracks my ACs month over month for total cooling/heating time each. That would be easily adaptable.
It is cycling itself and since it is in the cloud and only polled every 10 to 15 minutes or so i cannot get accurate runtime either
Ah, I see, I use Ecobees--they are cloud-backed but there's eventing happening, too, so my runtime readings are fairly accurate (a few mins drift on occasion).
Can you infer the compressor/fan are running based on a power meter (if you have one)?
i dont have power meters here as no plug to install near or convenient to breaker box.
But this is a honeywell t6 cloud connected therm and i do have and have modified and shared the code for this.
so i am implementing a running statistics in the code for the last hour based on the refresh time. so for instance every 15 minutes will be the last 4 polls , every 10 minutes last 6 etc.
i will keep track of the percentage of time the system was running ie (cooling heating) for that number of polls.
i can use this instead of time in the rule.
obviously this is not 100% fool proof as it could poll everytime between cycles and be inaccurate but i am hoping the likelihood of that is slim.. so say i kick the rule in if i get a 50% on time.
Ack--agreed; if you've got the driver to hand, amp it up to so what you need.
I don't know when they'd start throttling any increase in polling rate but I'm sure you'll account for that assuming the return code is meaningful.
so far new rule in conjunction with therm stats seems to be working (in my 2 condos) in the negative case where runtime seems to be between .33 and .5
now opening door to try and trigger it
]Download the Hubitat app
