DIY Air-to-Water with old R407C on/off outdoor unit
 

Hi,

Decided to register because I was asked for more details about heat pump in youtube video (id _Cb635AmB1I).

Some background.

I live in northern Finland in a house that has total living area about 140m2 in two floors. The house has floor heating with water in all rooms and was originally heated by Nibe 410P exhaust air heat pump. This model was at the time quite common in Finland and Sweden.

Nibe 410P does the air exchange in the house and puts energy from the waste exhaust air to water tank with heat pump. It has 650W compressor and it uses R290 (propane) as refrigerant. Additionally there is an immersion heater to backup the compressor.

After 6 years of running the compressor bearing broke and rendered the heat pump part useless. About two years Nibe used only immersion heater to heat the water.

Family of five in the house using hot water and me looking at the electric bill I decided to reuse refrigerant loop in the water tank. Idea was to make hot water when outside temperature is above 0 degrees so I would not have to worry about defrosting.

I managed to find cheap (advertised like new!) on/off type heat pump out door unit that was capillary feed R407C refrigerant machine.

Attachment 5617

Here it is after opening the shipment. For my disappointment the unit was clearly used and refrigerants were long gone and both port valves were wide open. So there had been moisture in the system for a long time and wet POE oils in compressor. To minimize the costs I vacuumed the unit and started looking for good advice from local forums.

Continuing in next post...

In local forum I was adviced to refill the system with R290. Usually components in R407C units match quite nicely for R290. Especially when the compressor is capillary feed.

So I made a quick test setup with some indoor unit as an evaporator and it seemed to work fine. Got hot or cold depending on the 4-way valve position.

In this first setup the system was filled from 'official' R290 refrigerant bottle that had 370 grams content.

Attachment 5618

After testing I was convinced and made a stand outside of the house.

Attachment 5619

Extracted Nibe 410P refrigerant loop pipes and brazed new piping.

Attachment 5620

Attachment 5621

Attachment 5622

Simple electronics, 4-channel USB relay card. Relays control fan, 4-way and compressor. Relay card is connected to Raspberry PI.

Attachment 5623

Attachment 5624

I have second Raspberry PI that monitors various temperatures in the house using 1-wire sensor network. I added new sensors to gas and liquid lines. Raspi that controls the heat pump gets all temperature information from the second Raspi using owfs remotely.

Software in Raspi is written in Perl and it runs as a daemon. Basic operation is to start heating when tank water is at 44 degrees and stop when water is 47 degrees. Once in a week I manually switch the Nibe immersion heater on to raise the temperature to 60 degrees. This is to be sure that no legionella bacteria can breed in the tank.

In local forum it was estimated that COP of the system is between 2-3 when ambient air is around 12 degrees. Until now the ambient temperature has been from lowest 5 degrees to highest 26 degrees and there has been no problems.

Outdoor unit is rated around 800W in heating mode with R407C charge. With current R290 refrigerant charge the maximum total draw is around 850W including compressor, fan and 4-way valve energized.

I'm not sure about the final R290 charge, it may be around 600 grams. Original R407C charge has been 850 grams.

Based on the power draw the system is not under filled for sure ;)

Please feel free to comment and ask for questions!

Here you can see some graphs from the house monitoring:

Vallox 110 SE seuranta

You can see the tank temp to sail between 44 and 47. Immersion heater backup is not currently enabled.

Some legend translation:

Tekn - room with all house techinc
Meno - outgoing water to floor heating
Palu - incoming water from floor heating
Jate - waste air after passive heat exchanger
Ulko - outside (ambient) air before passive heat exchanger
Tulo - air to house after passive heat exchanger
Pois - air from the house before passive heat exchanger
Katt - tank temperature (used for air-to-water pump logic)

Passive heat exchanger is used to preserve energy from the air drawn from the house. I had to replace the house air exchanger after Nibe compressor break.

Initial testing (VILP means air-to-water):

https://www.youtube.com/watch?v=PmyOxnJt5Pg

https://www.youtube.com/watch?v=Was9ovWFPrM

Some random pics. Evacuating old system with bullet piercing valve. Original system pressure at idle was about 9 bar.

Attachment 5625

Accuracy before everything else.

Attachment 5626

Then there was a major leak in the 'new' outdoor unit :mad: Lowest pipe was crushed all the way because of icing. I had to bypass the lowest loop completely! After that I switched to 5kg BBQ bottle and person scale for charging :D

Attachment 5627

After this the unit started to really work and I got good results when looking at the rate of temp rise in tank.

Hot gas is around 60-65 degrees when measured near tank. At compressor outlet it is over 80 degrees. Pressure at unit service port gas side is about 20 bar when running in heat mode.

Here is some real running graph.

Attachment 5628

Usually the comp runtime is around 600-800 seconds. After heavy water usage runtime may be 1 to 1.5 hours. In software there is a failsafe stop at 2 hours of runtime. Never exceeded that.

The original build thread, in case you want to learn some Finnish :) There are more pictures from the build.

http://lampopumput.info/foorumi/index.php?topic=21584.0

I would really appreciate if somebody can explain theory of R290 refrigerant in R407C unit. I don't know much about enthalpy, superheat etc. This project is more or less working good by coincedence.

Originally the Nibe 410P makes over 50 degrees water from the constant exhaust air of about 22 degrees.

This system can make 47 degrees easily from 10 degrees ambient air. I wonder if I should change the software to make hotter water when ambient temp is higher? How would it affect COP? And what is theoretically possible with R290?

apophis,

Thank you for writing from Finland, a country I hope very much to visit some day.

I will let someone else comment on the theoretical aspects of R290 vs. R407C, other than to say that I know that in terms of performance, they are are very similar, and that R290 is often used as a 'drop-in' replacement for R407C.

It sounds like your unit is scavenging heat from house exhaust air and returning it to in-house use, which is a very clever design. I do not know of any designs like that in the US. It sounds like some of our guys could take this as an excellent challenge, to make their own units that operate in this manner. I have heard of industrial units that operate like this, and they are described as 'active heat exchange ventillators'.

This is a very good, to be thinking like this...

I would say that if you try to make hotter water from ambient air, that is remaining at the same temperature, your COP will decline, and the hotter you want to go, the more the COP will decline.

On the other hand, the COP, of a fixed-temperature tank will increase as ambient air temp increases.

So, when you try to do both of these things at the same time, you will have a multi-variant relationship. Which is very difficult to predict intuitively. It would be possible to predict the COP vs ambient air vs. tank temp relationship, if more data was available...

If you did actual testing, and had the appropriate data that was of the scenario wherein you had constant tank temp, and varying ambient air, it would be possible to graph a curve that would portray COP vs ambient air. From these data points you could do a regression analysis and determine the mathematical equation that would accurately predict that curve.

In a similar fashion, if you did actual testing, and had the appropriate data that was from the scenario wherein you had no limit on the tank temperature, and let the tank rise to whatever temp it might rise to after runs of the same time duration, with the external difference being only ambient temperature, you could plot this data as tank temp vs. COP. From these data points you could do a regression analysis and determine a second equation, that would accurately predict that curve.

Then you could combine these two equations, and the resulting equation would portray the multi-variant situation that you are concerned by.

This final equation could be predictively incorporated into the software of your Raspberry Pi program.

Or, you could do it by trial-and-error.

Best,

-AC