I wonder if there is any way of hacking into the controller to find out what's really going on?
For Midea units, this may be possible, at least for some parts of the system, eg I have found the javascript behind the Midea app. But there are still a lot of black boxes involved:
What I was getting at was more trying to find out the logic behind the decisions taken by the ASHP when to start and stop and, in the case of Mitsubishi's AA, the relationship between target flow, WC settings, room temperature, ambient, etc. It must be in the FTC unit somewhere. I'm thinking of something like the equivalent of 'rooting' your phone, games console or the like and being able to see what's happening.
I wonder if there is any way of hacking into the controller to find out what's really going on?
For Midea units, this may be possible, at least for some parts of the system, eg I have found the javascript behind the Midea app. But there are still a lot of black boxes involved:
What I was getting at was more trying to find out the logic behind the decisions taken by the ASHP when to start and stop and, in the case of Mitsubishi's AA, the relationship between target flow, WC settings, room temperature, ambient, etc. It must be in the FTC unit somewhere. I'm thinking of something like the equivalent of 'rooting' your phone, games console or the like and being able to see what's happening.
If someone can extract a copy of the code in readable format, I may be able to decipher it.
I wonder if there is any way of hacking into the controller to find out what's really going on?
For Midea units, this may be possible, at least for some parts of the system, eg I have found the javascript behind the Midea app. But there are still a lot of black boxes involved:
What I was getting at was more trying to find out the logic behind the decisions taken by the ASHP when to start and stop and, in the case of Mitsubishi's AA, the relationship between target flow, WC settings, room temperature, ambient, etc. It must be in the FTC unit somewhere. I'm thinking of something like the equivalent of 'rooting' your phone, games console or the like and being able to see what's happening.
If someone can extract a copy of the code in readable format, I may be able to decipher it.
There is no firmware download that I could find, so I suspect one might have to intercept the data as a unit upgrades, which would mean running something like wireshark for an indefinite time until the next update. And then that probably only gets you the object code, not the easier to read source code.
If someone can extract a copy of the code in readable format, I may be able to decipher it.
I've got, as I say, the javascript behind the Midea app, and can make it available, but that won't have the control logic, only how the app works.
I think at least some of the logic used by the Midea units is in the technical documents, not directly and clearly stated, but it can be inferred. Taking a very simple example, the DHW heating, there is a setting which sets the temperature difference between the desired hot water temp and the actual hot water temp at which the hot water heating is activated. IF desired hot water temp - actual hot water temp > 10 degrees THEN heat up the hot water etc.
The set LWT is a straight linear relationship between the ambient and LWT between the two set points, with a flat line either side:
There is some stuff about about flow rates, but it doesn't make much (any) sense to me, because, according to the wired controller, the pump operates at around 1.42m^3/h, whatever the heating requirements are (except of course when it is off):
I'll see what else I can tease out.
Midea 14kW (for now...) ASHP heating both building and DHW
I suspect the ecodan algorithm works much as @derek-m has outlined in the past, but one obstacle to an open-source control is that it seems like the controller has access to some useful performance-improving features that are not exposed by any API (Application Programming Interface, which for the ecodan means melcloud or the cn105 connection devices) that I've seen, such as being able to run the circulation pumps while the compressor is off.
I had my first better-than-4 day for a while yesterday. I believe the key change was that I told my controls not to bother ramping up from the daytime setback as early (I currently run two setbacks: one while we're asleep and one while I'm moving around doing stuff in the daytime) in the late afternoon because I was out and it was forecast to be mild. So, it stepped up in the evening instead. The house seems to have been warm enough. At least, no complaints yet!
I might just leave the ramp later until the next cold snap, or I might add some logic to the controls to start ramping up once the outdoor temperature falls below 11 degrees after 2pm, instead of a fixed timer.
I might just leave the ramp later until the next cold snap, or I might add some logic to the controls to start ramping up once the outdoor temperature falls below 11 degrees after 2pm, instead of a fixed timer.
Well, it's an excuse to test the Blockly logic programmer in domoticz, so I did this with some slightly different trigger points:
If I need to rewrite it in Python, I can, but it'll be less pretty!
Have a look at this. The two graphs are my house after a setback, one on Mitsubishi AA, one on WC. Both have similar (but not identical) ambient temps. AA has a target room temperature and it's supposedly 'learned' the best way to achieve this. WC just gets the flow up to target as fast as possible and keeps it there.
With AA, the house is up to temperature by 9am then it's steady state (apart from HW at 3pm). With WC, it's more like 1pm (note I had HW at 9am also)). So to compare the effect of a boost, we can compare consumption 5-9am in both cases. AA was 9.1kWh, WC was 6.2kWh. So the cost of a boost to get the house up to temperature is circa 3kWh. Steady state consumption is about 1.4kW so the 5 hour (12-5am) setback saves c.7kWh.
In my case, even with a boost, I'm saving c.4kWh. I think AA uses quite a bit more energy overall but that's another story.
-- Attachment is not available -- -- Attachment is not available --
Real life data on a 16c setback from 9pm to 7am whereupon it was adjusted to 17.5c
Here are 2 MELcloud pictures of our 8.5 Ecodan performance over a 24 hour period. We are 180 sq Metre 4 bed house. -part 18th century, part new extension. Internally insulated and renovated. But with 1 run of bifold doors, one uninsulated front door. 14 radiators and a 250L cylinder with Solar thermal coil a tank low level.
system running on Auto Adapt with 55c maximum temperature restriction. DHW is set to 48c with 11c drop to reheat.
1. The interior room temperature picture shows the setback period start at 9pm where the HP switches itself off then returns to on at 16c at about 6.30 am. Room stat is then manually increased to 18c at 7am and 19.5c at 9am.
2. The energy delivery picture shows the AA flow temperatures, the DHW reheat and the off period.
We think we are seeing a very slow cool down overnight and a relatively effective re heat in the following morning. The 24 hour period was with a cool 4c overnight period and a mild 8c daytime temperature. Total consumption for the 24 hour period is 14kwh, delivering 52kwh providing a COP of 3.72. -which also included 2 tank reheats...
So for those who are used to reading these energy graphs could it be that the amount of insulation work better with lower overnight set back temperatures to make it a viable and efficient way to operate?
Real life data on a 16c setback from 9pm to 7am whereupon it was adjusted to 17.5c
Here are 2 MELcloud pictures of our 8.5 Ecodan performance over a 24 hour period. We are 180 sq Metre 4 bed house. -part 18th century, part new extension. Internally insulated and renovated. But with 1 run of bifold doors, one uninsulated front door. 14 radiators and a 250L cylinder with Solar thermal coil a tank low level.
system running on Auto Adapt with 55c maximum temperature restriction. DHW is set to 48c with 11c drop to reheat.
1. The interior room temperature picture shows the setback period start at 9pm where the HP switches itself off then returns to on at 16c at about 6.30 am. Room stat is then manually increased to 18c at 7am and 19.5c at 9am.
2. The energy delivery picture shows the AA flow temperatures, the DHW reheat and the off period.
We think we are seeing a very slow cool down overnight and a relatively effective re heat in the following morning. The 24 hour period was with a cool 4c overnight period and a mild 8c daytime temperature. Total consumption for the 24 hour period is 14kwh, delivering 52kwh providing a COP of 3.72. -which also included 2 tank reheats...
So for those who are used to reading these energy graphs could it be that the amount of insulation work better with lower overnight set back temperatures to make it a viable and efficient way to operate?
With the inevitable caveat - there is no panacea!
I'm sure your setback is saving money. You don't lose too much heat overnight and it recovers quite quickly. Why do you increase the temp in two phases? Is it to stop AA bumping up the flow too high?
Have a look at this. The two graphs are my house after a setback, one on Mitsubishi AA, one on WC. Both have similar (but not identical) ambient temps. AA has a target room temperature and it's supposedly 'learned' the best way to achieve this. WC just gets the flow up to target as fast as possible and keeps it there.
With AA, the house is up to temperature by 9am then it's steady state (apart from HW at 3pm). With WC, it's more like 1pm (note I had HW at 9am also)). So to compare the effect of a boost, we can compare consumption 5-9am in both cases. AA was 9.1kWh, WC was 6.2kWh. So the cost of a boost to get the house up to temperature is circa 3kWh. Steady state consumption is about 1.4kW so the 5 hour (12-5am) setback saves c.7kWh.
In my case, even with a boost, I'm saving c.4kWh. I think AA uses quite a bit more energy overall but that's another story.
-- Attachment is not available -- -- Attachment is not available --
i like your reporting presentation. What software are you using? How was it to set up?
It uses the csv file downloaded from my MMSP monitoring. It's done in python/github and runs on my laptop. It logs into Melcloud, calls the csv file routine extracts the data. I can select a date range and it produces the graph and calculations.
@kev-m thanks for your obvservatiins kev. Regarding the 2 step temp change…yes that was my thinking - to keep the target temperature east to reach on the assumption that AA will attempt to modulate its output…. Its just a guess that it might slow things down….
@kev-m thanks for your obvservatiins kev. Regarding the 2 step temp change…yes that was my thinking - to keep the target temperature east to reach on the assumption that AA will attempt to modulate its output…. Its just a guess that it might slow things down….
I've limited the max temp to 48; I might drop that to 43, which is the highest setting when I'm running WC. I'm not that bothered about the house heating up quickly, especially as ours is set to 21. AA can be a bit too aggressive for me.
I might try the two step approach like you some time.
Thinking about installing a heat pump but unsure where to start? Already have one but it’s not performing as expected? Or are you locked in a frustrating dispute with an installer or manufacturer? We’re here to help.
