I had a problem with my ecodan unit freezing up having to literally put a hot air gun on Christmas Eve at 10pm 2 years ago.

Took various engineers over a month to sort out.

Low and behold after all the headaches it mounted to a dodgy temperature sensor mounted at the outside or the outside unit at the rear.

Basically it was sending the wrong signal.

good luck 

@westkent Thank you for the info, this was exactly what i was thinking about. i have a Beamex temperature calibrator and a dry well temperature generator at hand so i will check the service manual what kind of thermistors they use and then check the temperature curve and compare with the manual. Best to check them all as they seem prone to failure / drifting.

I did not have the time to check the thermistors Yesterday but i am now almost certain that's the problem. I enabled the cooling function on the heat pump and manually "defrosted" the unit by cooling the home. Worked great, had it on about 12 minutes and all the ice was gone. 

At least the 4 way valve should be ok, worst case there is low refrigerant but would not bet on that as the pump both heat and cool very good.

Also i noticed that when using weather compensation i got really bad COP, then when using room auto adapt i got much better COP which also points to the outside unit thermistors. 

This weekend i will test ALL thermistors and hopefully find a faulty one.

@thundermink can you give us an update on your tests?

@editor Sorry for the slow reply, it looks like a happy new year to me. 4 days ago it was a bit warmer so i got to start trouble shooting a bit more, i noticed that if i completely shut down the system the LEV electronic expansion valves open fully, then the unit defrosted ok for a while so and i started thinking if it is the valves that skip some steps causing the defrosting issue over time when the valves misses steps and is in the wrong position.

I removed the valve heads and noticed one was not fully in place and another one was stuck to the valve body so i sprayed some wd40 to loosen it and then cleaned the stepper motor coils, also the cables was zip tied to the copper tubing and the cable shrink seemed to have shrinked and caused very much tension on the valves wich also can have been why one of the valves was rotated out of place, 4 days now and no problem, it has not been too cold yet but it is now -4 degrees C and i have not seen any frosting what so ever on the outside unit. Also the COP is much better so it seems that this has worked. Still need to wait and see how it behave the coming days as colder weather is on its way (friday forcast -11)

Hopefully the system will work 100% for the rest of this winter 🙂 

A bit of a follow up on the progress of the Ecodan pump, while finding and correcting the issues i faced last winter we are now getting in to heating season again 😀 the repairs done this winter had a big difference at cop values going from under 2 some days to always over 2.5 and the defrost problems was minimal, but last month i thought it is time to remove the panels from the outside unit completely to get a real good look at everything. I found that Mitsubishi is not very good when it comes to temperature sensors, they where all just clipped to the pipes with copper clips, no isolation and no thermal paste at all. noticed a problem specially with Liquid sensor TH3 facing outward while it should face away from the pipe that is next to the sensor to not cause false readings. i removed all sensors and cleaned the pipes, added thermal paste and isolated the sensors using pipe foam insulation. THIS WAS A BREAKTHROUG, COP values had been 3.0-3.6 the past weeks, now after the modifications COP has been between 3.95 and 4.70!! Before the mods i had never seen cop over 3.9. Everyone owning a Ecodan unit should check the sensors and modify them and check the behaviour. atleast for the Ecodan shw112 this was a game changer. Hope this helps some of you out there.

Posted by: @thundermink

↑

A bit of a follow up on the progress of the Ecodan pump, while finding and correcting the issues i faced last winter we are now getting in to heating season again 😀 the repairs done this winter had a big difference at cop values going from under 2 some days to always over 2.5 and the defrost problems was minimal, but last month i thought it is time to remove the panels from the outside unit completely to get a real good look at everything. I found that Mitsubishi is not very good when it comes to temperature sensors, they where all just clipped to the pipes with copper clips, no isolation and no thermal paste at all. noticed a problem specially with Liquid sensor TH3 facing outward while it should face away from the pipe that is next to the sensor to not cause false readings. i removed all sensors and cleaned the pipes, added thermal paste and isolated the sensors using pipe foam insulation. THIS WAS A BREAKTHROUG, COP values had been 3.0-3.6 the past weeks, now after the modifications COP has been between 3.95 and 4.70!! Before the mods i had never seen cop over 3.9. Everyone owning a Ecodan unit should check the sensors and modify them and check the behaviour. atleast for the Ecodan shw112 this was a game changer. Hope this helps some of you out there.

@thundermink can you provide photos to illustrate your fixes please?  I wouldn't know where to start looking for temp' sensors or liquid sensor TH3 on my Ecodan. My COP is often barely over 2 and never over 3.

Pictures would be helpful if possible.

Hello, unfortunately i did not take any photos, but you should check the service manual of your pump, In the puhz-shw112yaa service manual they have photos and instructions how to remove the sensors and there you will clearly see where the sensors are placed. Here is a screenshot from the manual of TH3 liquid sensor. i removed it, put thermal paste in the sensor clip and isolated it using pipe isolation then repeated the procedure on all other sensors

I think this is an interesting observation and aligned with what a few of us have been saying for some time: the sensors are only as good as they've been installed. The best sensor in the world wrongly installed will still give you poor data. 

Interesting case study! 

Posted by: @hcas

↑

I think this is an interesting observation and aligned with what a few of us have been saying for some time: the sensors are only as good as they've been installed. The best sensor in the world wrongly installed will still give you poor data. 

 

Interesting case study! 

@hcas all well and good but the majority of us, myself included, have no idea how to even identify, locate and then fix the issue with these sensors. The issue then, were I to sign up to Havenwise, which I’m seriously considering, is that your calculations and settings to operate my system ‘efficiently’ is based on flawed information.  I haven’t a scooby where to even look for said sensors never mind decide if they’re installed correctly.

@morgan Fully agree with you. It shouldn't be the homeowner's job to do this. The installer should. When I went to site with @ashp-bobba some time ago, he mentioned he makes sure the sensors are all "installed" correctly to the best standard. He knows this based on his experience installing many Mitsubishi heat pumps before. 

When it comes to Havenwise, the accuracy of the flow or return temp measurements don't matter so much for the optimisation. Whether the flow temp is 35C or 36C doesn't matter much in practice for the operation of the heat pump.

It does matter for the reporting because small errors in measurement can have an impact on the reported COP. A 1C difference on the flow temp measurement would be 20% difference in reported COP, even though the actual COP probably was almost exactly the same.