The Homemade Heat Pump Manifesto

[AC_Hacker]

Here's a picture of the backyard.

The loop field is complete, it's fall, the leaves are falling on the buried loop field, nature is healing her scars. It's raining every day now.

I just wanted to finish off the loop field installation thoughts and then move on to hacking the air conditioner...

Now that I have done it, I have to say that it sure was a lot of work, to put it mildly.

If you live in an are where there is a sand/clay mix and no big rocks, the mud-drilling method is a great way to go. But the mud drilling method is not suited to rocky ground.

If you have the land, and access to a backhoe or bulldozer (rent them), the trench & slinky method is a good bet, but certainly not suited to hand digging.

If you live on a small lot like me, vertical is really the only choice.

But I do realize that if you need to go vertical and have no access to heavy equipment and aren't likely to buy or build your own power auger, hand digging is possible, soil conditions permitting. A combination of a hand-turned post hole auger that allows water-pipe extentions ($35) and a beefy shop vac, such s a Rigid 5HP, or bigger ($60) would work. If you set yourself the goal of one hole per week, it can be done (one day to dig, six days to forget how much work was involved). My boreholes were 17 feet deep when I hit rocks and then hardpan. I didn't sense that 17 feet was any kind of limit to the auger or the shop vac, once I switched to 2" ABS pipe (lower friction than small pipe). The deeper the better, soil conditions permitting.

* * *

So I'd like to thank the folks who helped me in this project with their physical effort, the fabrication of parts I had no way of making, their sage advice, the loan of just the right tool at a critical time, and last but not least, the folks who helped me by simply taking an interest in such a curious endeavor:

William_Hackerson
Ann-the-Flame
Howard-the-Machinist
Distant-Relative-Ken
Bruce-the-Pirate
Rhonda-the-Grandma
Daox
Hugh Jim Bissell
M. Dumois
All the other folks who left comments on the blog
...and the all the other folks who took time out of their lives to take 9,693 peeks into this very curious project.

You have all meant a lot to me.

So, now it's time to hack that air conditioner!

Best Regards,

-AC_Hacker

[AC_Hacker]

The Air Conditioner I have chosen is a small unit, that was in good shape, actually works and was cheap. As I recall, I bought it for $25.

What I'm going to do is to remove the air-to-refrigerant heat exchangers, and in their place I will put some water-to-refrigerant heat exchangers.

This will involve:

1. cutting into the hermetically-sealed refrigeration circuit
2. capturing the refrigerant that is in the system
3. brazing the new heat exchangers into place
4. testing the system for complete hermetic seal
5. pumping out all of the air/water vapor that entered the system in the above process
6. re-charging the system with refrigerant
7. testing and adjusting as needed

There may be a few side-trips in the process, but this is the general road map I'll be following.

I have done this once before as a test, to see if I could actually work with this technology. The results were very encouraging, so much so that I'm now going to build a unit that I intend to use as a heat source for my house.

I started a thread previously that can be found here. This previous thread covered general operation of a vapor-compression refrigeration system and plenty of views of the airconditioner I'll be modifying.

There will be some tools and skills you will need for this process and I'll call them out as we go.

One of the general ideas here is that you want to have the tools and skills, and parts you require, on hand before you start. You don't want a refrigeration system to stand open, exposed to air and water vapor any longer than absolutely necessary. You will want to open the system, capture the efrigerant, remove the old parts and install the new parts in quick succession.

Another general idea is that this is not designing a system from scratch, but rather re-purposing an existing device. So there are parts in the system that you will want to handle carefully, as you will be reusing them, like the capillary tube(s).

* * *

DO NOT SILVER SOLDER REFRIGERATION EQUIPMENT, BRAZE IT.

You will be brazing the system back together once it is opened, so you should have on hand the equipment and skill for brazing before you open the system up. In some of the vapor compression hacker forums, I have heard of people using propane tourches to do refrigeration brazing. This was how I began, and it was very frustrating. I also started out using phosphor-bronze brazing rod because the price was very attractive, but still no love. Finally I got some silver-alloy brazing rod and a MAPS tourch and things began to work for me. The percentage of silver in the brazing rod really affects it's melting and flowing characteristics, and price. I have seen 5% silver rod, 15% silver rod and 40% to 55% silver rod. This stuff is really expensive the more silver, the more money. I was able to get a small amount of 56% rod and with the appropriate flux, it worked beautifully with MAPS gas.

...not pictured is the silver brazing flux, I need to get some more.

Here's a link to a brazing discussion from a HVAC blog:

Mapp gas for brazing??? - Refrigeration-Engineer.com forums

...and another:

Brazing material for refrigerant lines - Shop Floor Talk

Do some research, get to know your local refrigeration and welding supply folks, pick up some equipment and do a bit of practicing.

Best Regards,

-AC_HAcker

%%%%%%%%%%%%%%

[filipetri]

hi,
congrats for your project.
although I bought the heat pump off-the-shelve I´ve a setting similar to yours, however my heatpump is a bit too unintelligent. I was wondering what did you use as its "brain", what controller did you use for it?
thanks,
fsr

[AC_Hacker]

Originally Posted by filipetri

hi,
congrats for your project.
although I bought the heat pump off-the-shelve I´ve a setting similar to yours, however my heatpump is a bit too unintelligent. I was wondering what did you use as its "brain", what controller did you use for it?
thanks,
fsr

Very astute question...

I actually need help on this part. If any of the readers are knowledgeable about refrigeration/AC control, I could use some help.

[* EDIT (11-11-11): Since this post, an inexpensive controller with a web interface has been found. Xringer has posted a thread where he gives a pretty good introduction to the controller including sources of interface devices that make it very useful. *]

I'm not using a variable frequency compressor, so things will be simpler.

I'm also building a heat-only heat pump, so that also makes things simpler.

filipetri, I hope that you understand that you are reading about a project that is in development. Although I have done loads of research and also have done some experiments to verify that what I'm attempting is in fact possible, I have not yet completed a fully reliable functioning unit that is now heating my home. By the time Christmas time comes around, I hope to have it all together.

But I have examined the controllers from a de-humidifier and an air conditioner and I have found that they each look for some control temperature and when it is reached, they cut power to the compressor.

In the case of the de-humidifier, there was a circuit board and one thermistor that had leads that went to the circuit board. The thermistor was fastened firmly in place to the copper tube that went to the evaporator. When the evaporator got to a cold enough temperature for frost to form over most of the evporator's fins, the power to the compressor was cut. This allowed the hot refrigerant that was in the system to migrate to the evaporator and quickly melt the frost. When the temp got high enough, the power to the compressor would resume. This cycle continued until the container that captured the melting frost from the evaporator, got full, and power to the unit was cut, until someone came and emptied the container.

The control for the air conditioner was similar, but it was mechanical instead of electrical. Again, the sensor, which was some kind of metal 'bulb' was fastened to the fins of the center of the evaporator and when the temp got sufficiently cold, the power to the compressor was cut. When the temp got high enough compressor power would resume. On the air conditioner, the water from the melting frost was allowed to drivel out the back of the unit, and the unit would continue to run until the power was manually turned off. It was interesting to me that the temperature of the air was not measured, but rather the temp of the evaporator.

In my experiments, I stripped out the controller altogether. I learned very early that the heat pump has the ability to remove heat (BTUs or watts) at a certain rate from a heat source.

If the rate of heat transfer from the heat source is lower than the rate of heat extraction of the heat pump, a freezing water condition in the heat exchanger will occur. This would surely damage the evaporator heat exchanger.

I did reach this freezing water condition, but I was sitting right there as the water spectacularly and almost instantly turned to ice. I was able to yank the cord before damage could occur.

If the rate of heat transfer from the heat source is higher than the rate of heat extraction of the heat pump, the freezing water condition will never occur.

As things now stand, I calculate that the heat pump I am now building will extract heat at about the rate of 1600 watts (5459.2 BTU/hr), the rate of heat transfer from my loop field should be about 3940 watts (13444 BTU/hr), so I might not need the freezing water protection circuitry. On the other hand, if for some reason, the water leaked out of a loop, or if a circulation pump failed to operate, I'd have a serious problem on my hands. So I believe it would be a good idea to monitor the refrigerant line going into the evaporator heat exchanger, similar to what I found in the de-humidifier. And if it got too cold, it would cut power to the compressor.

Also, in the room that I will heat, I will want to monitor the temperature, and when it gets to the level of comfort I wish, it would cut power.

There are other conditions I may want to monitor, but those are the ones I can think of right now.

Ward-the-Wiki-guy has volunteered to help me make performance information available in real time on the Internet. Here is a photo of the electronics I have purchased to make this possible:

I got both of these boards from Spark Fun. The on on the left is the Ethernet Shield and the one on the right is the Arduino board.

The Arduino board is a microcontroller board, it will monitor several 1-Wire temperature sensors. The other board will interface the Arduino to the Internet. Ward has a similar setup monitoring certain aspects of his home, so what I'll be doing here is very similar. I'm pretty sure that the Arduino will be able to monitor and respond (cut power) if certain 'out-of-bounds' conditions occur.

At this point, I don't know much about the Arduino, or power control, but there is a community of talented tech-types here in Portland that meet regularly to drink beer and swap stories and info. I've been going to the meetings and sharing what I'm doing with the Homemade Heat Pump Project, to great interest.

So that's the controller story at this point. I'm open to suggestions, and I will certainly be posting the details of the controller as it develops.

Hope this has answered your question.

Best Regards,

-AC_Hacker

[filipetri]

I´m amazed how close we are to each other.... I´ve been through all that: almost instant freezing inside the heat exchanger, studying ACs and de-humidifiers... and finding arduino (and all the other duinos in the family). for a while I thought I could do it with arduino too, but after some playing around over a year ago. but it lacks the internal clock, the watch dog hardware the control functions (not enough). I´m looking at these ones now: AMD® LX800 PC/104-Plus CPU Module - PCM-3353 - Advantech But still havent got the time to go deeper into studying the thermodynamics of the process, in order to optimize it. I found out however that I can keep my water just above freezing temp by keeping a certain level of water flow, while not letting the water at the water tank fall bellow 25Cº. Somehow the highest the condensation temp at the "hot" side the highest the evaporation temp at the "cold" one. Keeping them stable on both sides around those temps seams to maximize my COP. Which makes a lot of sense given that the cooler the water loop goes into the soil the higher deltaT, the higher the energy transfer into the system. Now, given that I live in the Algarve (sunniest and probably with the most stable themperature in Europe) my heating and cooling needs are much different from yours (I´m guessing!). I also need to cool the house in the peak of summer > 37Cº everyday and quite sunny. Actually my heating and cooling needs are almost the same over the year. I also have around 20 M2 of solar panels to heat the house´s water, which cover almost all our needs in the winter and is well beyond what we need in summer. I want to coordinate all my energy generation (the heat-pump and solar panels) and its uses (house heating and cooling, swimming pool heating, 1000lts thermal accumulation tank, 300 lts hot water tank). BTW, here we at night electricity costs 1/2 of its day price. I´m considering in running LinuxMCE to manage my house, BTW. Getting back to "our" controller: In order to maximize the energy captured from the soil (BTW, what is its winter temperature?) the water should leave your heatpump as cold as possible (~0Cº). right? For that its important to keep the water pressure as low as possible at that particular point. Instead of messing with the compressor I think I will play with the water flow instead (knowing its not easy to accelerate the water column that extends over 1KM of berried pipe, though). The accumulation water tank is very important to maximize the COP. Professional heat pumps have a water pressure sensor inside the heat exchanger + a termocouple + a device to check if there is water flow, in order to prevent freezing (and evaporation on the hot side).
A friend said he is going to fetch me the drawing of an actual pro HP.I´ll be more than happy to share it here when that happens.

I´ve splited the room themperature control from the heat pump control (at least for now). So, I am running low cost thermostats on each room controlling the electric valve that opens the relevant underfloor circuit + circulation pump. This circuit pumps the hot water from the accumulation tank. Only when the temp inside the top part of the tank go bellow the set temp (25Cº) the heat pump will start and then stops when it reaches 30Cº (normally it takes almost an hour, if the underfloor circuit is pumping too). So when in reaches 30Cº in would stay OFF for around 2 to 3 hours (<25Cº) cuz both the tank and all the concrete on the floor are warmer than 25Cº. This setting has prooven quite efficient for the heating, I should say. TO improve that I wish to manage the withdraw some extra energy from the swimming pool too.


You can easily post readings from arduino to the internet:
Web-Connected Light Sensors With Pachube & Arduino
Web-Connected Light Sensors With Pachube & Arduino

But in my opinion ARduino as it is wont do the trick... low memory and lacks the internal clock and watch dog for instance. I know you can buy one and add to it but it will take most of the I/O ports.
At leats these duinos are a bit bigger:
Sanguino: Arduino's Big Brother on Vimeo
Arduino - ArduinoBoardMega

But they still lack, I think I´ll go for a pc-104.
What do you reckon?

[Xringer]

If I were doing this kind of project, I would add a basic thermostat input.
That will allow you to micro-manage the exact room temperature without the need of resorting to a program/keyboard.
A good programmable thermostat isn't real expensive and some are very versatile.

Another input that might be useful is a motion sensor. Or human activity detector.
With this type of detector in the control system, you could kick the daytime
heating/power usage down a notch, if no one has been walking around the house for the last 3 hours.
And, when you come back from bowling, the system senses you and returns back to the normal heating program.
Sorta like when you put your PC in powersaver mode..

Add a blinking status LED that can be installed in the heated room.
Using solid ON or a blink rate, it can inform you of basic system status at a glance.

Add a beeper output, to alert you of any problem that might occur.

I'll bet you are planing on monitoring the temperatures of the
compressor and any critical motors in the system..


One other thing, which I know almost nothing about is the 5 minute delay.
Almost all AC systems have them for some reason. Some kind of system recovery is going on with the refrigerant?
I don't know, but I would look into adding a don't re-start failsafe routine, if it's needed on your system.

Hope something here is helpful..

Cheers,
Rich

[filipetri]

Thanks.
I think concrete floor has lot of thermal mass which gives it a very high thermal mass. it may take up to a day or two from when warm water starts to circulate until the room reaches the intended temperature. so normally its better to keep it always warm (smart energy stratagies for this kind of heating are a bit more tricky and take a bit more of calculation).

[AC_Hacker]

filipetri,

Do you have any photographs of your heat pump system?

Did I read correctly that you have over 1 kilo-meter of pipe buried in the ground? In what pattern have you placed the pipe, and at what depth? What kind of pipe and what diameter? Did you do it yourself?

What kind of pumps are you using? How many watts? Are you trying to pump fast enough to cause turbulent flow? Are you using antifreeze?

How many watts does your vapor-compression unit consume?

What kind of COP are you able to reliably get?

What kind of heat exchangers? How big are your heat exchangers?

How many meters square is your house? Is it well insulated?

Can you describe your hydronic floor?
* How thick?
* How many meters?
* How close did you space the pipe?
* What diameter pipe?

Nice to know someone else who is working with heat pumps & hydronic floors.

Best Regards,

-AC_Hacker

%%%%%%%%%%%

[filipetri]

Hi AC_Hacker

Let me try to answer all your questions:

The system I describe is at my summer house, which used to be my full time house until recently and for a couple of years. Let me emphasise that I do not complain neither about the house thermal confort nor about the energy efficiency to warm it up (23Cº stable all winter long). My worries are more about cooling it and make it smarter.

This is the house:
Offre locative n° 275593
Total living area: ~ 300 M2 (no idea whatsoever of how much that will be in imperial units)

Insulation is very good (has EU Class-A certific -http://en.wikipedia.org/wiki/Energy_Performance_Certificate) , this is the system I´ve used:
+ externally insulated walls with 50mm poliurethane boards all around the house including the roof similar to this: http://www.knaufinsulation.pl/files/...20_lowres_.gif
+ 100 mm rockwool between the plaster boards and the external wall on the inside
+ 100mm rockwool over the plasterboards on the ceiling
+ double glazing (14mm air between glasses) with thermal blocking

This is where it is located:
lagos, algarve, portugal - Google Maps
Definitely not a harsh place.
http://re.jrc.ec.europa.eu/pvgis/cou...blications.png

My heat pump is the SRS-100P5A/T (a packaged model) from Shanghai Qisai Electric Industrial Co., Ltd.
Its a very simple unit, comprising a scroll compressor working at 230VAC 50Hz and consuming 2650W maximum ever measured. Now, I´ve realised that the COP is maximized (ie, current consumption lowers) when the hot water leaves the system at around 25Cº and the cold water leaves at ~0Cº. for some reason the compressor works with less effort (which makes a lot of sense BTW), in this case it can seat work at around 1950W while still generating > 10KW of heat.

Because here the temperature almost never goes < 0Cº (even < 7Cº is rare) and we have a big state we could buried the outside pipe at a quite random depth (sometimes is less than 20cm, other places is > 1mt). We burried the pipe while we were building the house foundations, the utility trench and the garden. We burried 4 circuits with 250mts each of 25mm low cost high density polimer pipe.

Inside the house we burried 22mm PEX-b pipe with 10 cm space (cuz we also use it for cooling) for heating I read somewhere that 16mm at 20cm is enough.
We have underfloor heating for the whole house (one circuit per room, 3 circuits for the living room + kitchen).

I´ve programmed the whole system before I built the house. So I could buy from China through a local import agent all the components that were cheaper to buy from there than to buy locally (heat pump, solar panels, heat exchanger for the pool, pex-b pipe, underfloor heating manifold controllers etc, willos replica or circulation pumps, etc).
In the US I think you can find chinese cheap equipment at places like this:
http://chinabestproducts.com/

NOW I NEED to make it more intelligent. to manage all aspects seamingly (specially for cooling)

[Maxis]

Hi AC_Hacker

I was searching internet for long time to find someone who is making DIY Heat Pump!

Now finally I found this page(forum) and I see that you are quite far with your project! Congratulations!

I also had idea to make it by myself, but than because of good deal at ebay I decided to buy it. I still made the instalations partly by myself - I installed ground loop and heat pump in basement + did underfloor heating system instalation.

I have 13 KW (about 4 Tons ) HP and 450m (1500 feet) of horizontal ground loop (40mm in diameter).

It seems that it's not enough for my HP and I will have to expand it. It seems as good idea to make drilling instead of digging more trenches.

I have some questions for which I didn't found the answer in your previous posts:

* what is the diameter of your drilling pipes - as I understood you should fit two 3/4 pipes in drilling hole?

* how the swivel is build from inside - how water gets into the pipe if it goes through swivel?

* what did you use to fill drilling holes when pipes were installed? As I know the bentonite should be used.

Thanks in advance!

Best regards
Maxis