Dan's Homebrew Geothermal Heat Pump
Published 11/17/08
Updated 9/01/2010
If you've read other areas of this website, you know that I don't do things the way most others do. With my knowledge of mechanical, electrical, and electronic things, if I want something I often build it myself. That way I can design it to do exactly what I want it to do, and use top quality materials so it lasts. And often, it cost less than the junk on the market today.

If you're not familiar with a heat pump, here's a few basics:

First you must realize that heat is a form of energy, and it will only "flow" from a warmer object, to a cooler object. Therefore, if it is 80 degrees inside your house, and 95 degrees outside, heat will flow naturally INTO your house. As long as it's warmer outside than inside, there is no simple way to make heat flow OUT of your house. The same holds true in reverse, if it's 30 degrees outside, and 70 degrees inside, heat will naturally flow OUT of your house. No simple way to make it flow IN.

You're all familiar with an air conditioner. An air conditioner absorbs heat from the inside air, which makes it colder. This heat is then moved outside, and it is "concentrated" to a temperature higher than the outside temperature, where it will flow to the outdoor air. This is also how your refrigerator works, heat is collected from inside the cabinet, moved out, concentrated, then released into the surrounding air outside of the cabinet. That's why there is warm air coming from the outside coils on your refrigerator. Concentrating heat requires the use of another form of energy, usually electricity, to run the machine.

A heat pump is basically an air conditioner in reverse. It can collect heat from lower temperature outside air (or water), concentrate it to a temperature higher than it is inside your house, then release it into the house air. Of course, the heat pumps are "reversible" so you can use it as an air conditioner when needed.

While it does use electricity to concentrate the heat, the operating cost can be considerably less than trucked-in heating fuels such as propane (LP) gas or fuel oil. It's important when designing a heat pump, or even an air conditioner, to minimize the amount of concentration of the heat needed, as this increases the amount of electricity needed (operating cost) to run the machine. Large coils of refrigerant tubing for both collecting and releasing the heat are important, and costly, but this one-time expense will soon be repaid with lower operating costs for years to come. As a plus, most of the electricity used to run a heat pump ends up being converted into heat itself, which gets added to the total heat output of the machine.

Heat pumps come in 2 basic styles, air-source and water-source.

An air-source heat pump collects heat from the outdoor air. In moderate climates, this can provide sufficient heat for a building. A big drawback to air-source units is that the efficiency and heating capacity drops off sharply as the outside temperature drops down below about 40 degrees. To collect heat from the air, the outside coil has to operate at a temperature considerably below the outdoor
temperature. This causes frost to form on the outside coils, reducing heat transfer. Also, the colder it is outside, the harder the unit must work to concentrate the heat up to a useable temperature. At a certain point, the amount of electricity used to run the machine will cost more than the heat produced is worth. So-called "hybrid" systems use an air-source heat pump, and the controls are programmed to abandon the heat pump when it's too cold outside, and switch over to another source of heat such as gas, oil, etc.

A water source heat pump collects heat from flowing water, instead of the outside air. If a continuous supply of water is available, much heat can be removed from it. This water could come from a well, or a lake, pond, etc. Or, the same water can be recirculated over and over again through a long length of pipe buried in the ground, called a "ground coil" or a "ground loop". Any of these methods mentioned are nowadays called "geothermal", meaning "heat from the earth". Most of these sources of water remain considerably above freezing temperatures (40-50 degrees), even on the coldest days. This eliminates the problem of the efficiency of the heat pump falling off when it's cold outside. An added bonus of a water-source heat pump is that it can be reversed to become a highly efficient air conditioner in the summer. The reason for the high efficiency is that the heat removed from the building does not have to be concentrated way up to above the outside air temperature, but instead only has to be concentrated to above the water temperature.

Now that you know the basics, lets get down to the nitty-gritty!

Seeing that I want my unit to work efficiently at low outdoor temperatures, I decided to go with the water-source system, using the buried ground loops.

I found that there is minimal accurate information available on the Web, for determining how much pipe needs to be in the ground, how deep it should be, how the buried pipes should be arranged, and so on. Sure, there's info out there, but then you go to another site, and they tell you something completely different. Good thing I like to experiment!

In the Fall of '08, I got the project under way. I decided I was going to bury 3600 feet of one-inch polyethylene pipe. This is the black stuff that comes in rolls, often used for lawn-sprinkler systems, and distributing water to buildings on farms. I had thought about using 1-1/4" pipe, but that size has a heavier wall than the 1", which would reduce the heat-transfer between the ground and the water somewhat. I also decided to make the underground "loops" into 3 seperate "circuits", each one 1200 feet long. This results in 6 pipes (both ends of each of the 3 circuits) coming into the basement, with no underground tees. My main reason for doing it this way was to allow for more experimenting, comparing the differences in performance with more or less of the loop in operation. Another reason is that when you "tee" several loops together underground, and bring in only 2 pipes to the building, it is possible for one or more of the circuits to be "air locked" or damaged and not functioning, but not easily detected. In my setup, I can visually confirm that water is circulating in all of the circuits, without any high-priced "special equipment" or "trained technicians".

Another big secret of the trade is splicing the pipes together underground. Although farmers have been accomplishing this task with a 29 cent coupling and two hose clamps for the last century, for reasons unknown to me they nowadays insist upon "thermal fusion" or some other high-tech sounding name, that amounts to melting the ends of the pipe and then pushing the two ends together. When it cools, you have a splice! Wow. I found that obtaining the pricey "special equipment" to do this was not an easy task (only available to "trained technicians"). Of course I got out my ma's old electric iron, played around with scraps of the pipe for an hour to determine the perfect temperature for "kitchen-table-thermal-fusion". I made some neat-looking splices that proved to be "stronger than the pipe itself" as they claim. But I decided to use the farmer's method anyway. I was able to get the pipe in 300 foot rolls, so I only had to put 3 splices in each 1200 foot circuit. Just to be on the safe side, I used 4 all-stainless steel clamps on each of the 29 cent plastic couplings. Then I put a 9 inch long piece of 2 inch heat-shrink tubing over the whole works. It was a bit of a challenge to shrink the heavy-wall tubing without melting the poly pipe, but I got 'r done. The heat-shrink tubing gives extra assurance that some strange substance in the clay won't devour the hose clamps, prevents water from getting into the joint from the outside which could possibly damage the joint if it would freeze, and lastly the shrink tubing alone would probably be sufficient to take over the joint's job even if the clamps or coupling failed.

Then there's the excavation. I knew that I was going to be confronted with pretty much solid clay for the big dig. That's a good thing for a heatpump, clay can transfer heat better than sand. But it's not such a good thing when it comes to actually creating the hole in the ground. The empty field in the front yard is about 100 feet wide by 300 feet long, and I decided upon laying each pipe seperately in it's own trench, roughly six feet deep. Each 1200 foot pipe was planned to leave the basement, go up, down, then up, down again the 300 foot length of the field, then return to the basement. The 12 trenches would be 18 inches wide, and about 7 feet apart. To avoid totally destroying the yard near the house, only 2 trenches would actually start at the house, to a distance about 50 feet away. There it would fan out to all the seperate trenches. Each of the 2 trenches from the house would have 3 pipes in them, one trench  for the "out" and the other for the "in" pipes. I'm aware that on many installations, they dig one big hole, and lay several rows of overlapping coils of pipe, similar to "slinky" being slid sideways off of it's stack. I'd guess that the object here is to get as much pipe in contact with the ground as possible, in a smaller area. My approach is to get a similar amount of pipe in contact with the ground, but spread out over a large area, thinking that this won't cause the ground to cool off as much as the heat is being removed from it. The warmer the ground, the warmer the water will be, and each degree warmer the water is, the more efficient the heat pump will be.  But I can't say that I know that my approach will work any better!

I rented the "mini excavator" that the rental company recommended for the job, for a week. Soon after beginning the excavation, I realized that the clay was much harder than expected, and this wasn't really big enough of a machine. But I decided to do what I could with it anyway. I made the two 18 inch wide trenches from the house to the edge of the field, plus one run up and down the length of the field, which used up the alloted 40 hours on the machine. While attempting to find a larger machine for rent, I came across a contractor willing to come and do the excavating for me. Even his much larger machine had a rough time with the clay, so we changed the plan to dig only 4 more trenches, 4 feet wide, and put two pipes in each trench, one on each side, which took about 2 days. So I ended up with a shorter, 700 foot loop in the 18 inch trench, and two 1200 foot loops in the 4 foot wide trenches. The rental machine cost $800, the contractor cost $2200, including backfilling and leveling off. The pipe and all the fittings only cost about $500. So I have about $3500 invested in the ground loops. See pictures below.
This is a finished splice.
Here's the hired, bigger backhoe digging a 4' wide trench. One pipe being is being laid in each corner of the trench.
That's me with the small backhoe digging the first 18" wide trench.
One more point on burying the pipe. Many years ago, I buried a 400 foot or so water line to supply water from our well to a neighboring property. That job involved a hired backhoe also, and the experienced operator warned me that steps have to be taken to avoid the newly-installed pipe from being damaged from rocks and such, as the trench is being filled back in. His recommendation was to lay the pipe against one side of the bottom of the trench, and with a shovel, cover the pipe with several inches of ground. We did that again on this project. Additionally, I connected a vacuum cleaner to one end of each of the pipes in the basement, blowing air through them as the trenches were being filled in. I frequently checked to make sure that the air was returning through the opposite ends of the pipes. Although this was no guarantee that a pipe didn't get damaged, at least I knew that a big rock didn't completely flatten it out!

In the Fall of 2008, I threw together a "temporary" homebrew heat pump unit, as follows:

An open-top 15 gallon blue plastic drum sits on 2 cement blocks on the basement floor. From the bottom of the drum, a pipe goes to the "inlet" side of a Grundfos 1/25 hp. cast-iron circulator pump, scavenged from a hot-water heat system. The "outlet" side of the pump connects to a manifold made from 1" threaded plastic "tee" pipe fittings, to create the 3 feeds for the three underground loops. Each of the three loop feeds has a 1" ball valve to regulate the flow. The 3 "returns" from the loops go to a similar manifold, minus the ball valves, to make a single pipe carrying the water coming from the loops. This single pipe is a flexible 1" I.D. hose, which enters the plastic drum over the top, positioned such that the water coming out of it creates a "swirling" motion of the water in the drum. The drum is filled with water, to within about 8 inches of the top. The pump moves the water at about 6 gallons per minute (total) through the loops. This completes the "water" portion of the system.

The refrigeration system which extracts the heat from the water, then concentrates it up to a useable temperature, consists of:
A Thermostatic Expansion Valve meters the liquid  refrigerant (Freon-22) into top of the "inner" evaporator coil, which consists of about 25 feet of 1/2" O.D. copper tubing wound in a spiral cylinder (about 8" inside diameter), then it flows upwards through the "outer" evaporator coil, consisting of about 40 feet of 5/8" O.D. copper tube, also wound in a spiral cylinder (about 12" inside diameter). This entire coil assembly is submersed in the water in the blue plastic barrel. As the liquid refrigerant makes it's way through the evaporator coils, it "boils" into a vapor, absorbing heat from coil, which in turn absorbs heat from the water in the barrel. The refrigerant vapor from the the evaporator then flows into the compressor, which is a 1 hp. unit I removed from a walk-in cooler at a bowling alley. After being compressed, the now HOT refrigerant gas is piped half-way across the basement in an insulated copper tube, to an "A-coil" (also called a "condenser" in this application, and intentionally way over-sized) located on top of my oil-fired furnace. In the A-coil, the hot refrigerant gas gives up it's heat into the house air, which is being blown over the fins on the A-coil by the furnace blower. As the heat is removed from the refrigerant vapor, it "condenses" back into a liquid. This liquid refrigerant (which is now at room temperature) is piped back over to the evaporator coil, where it enters the expansion valve, and the process repeats continuously. That completes the basic heat pump.


When I first fired it up in the fall, the water returning from the underground loops was about 58 degrees F (a little higher than expected). As the winter went on, this temperature slowly fell, finally reaching about 38 degrees near the end of the season (a little lower than expected). To my surprise, this had very little relationship to whether I was drawing heat from the ground or not. I didn't start using the 3rd loop, which is located well away from the others, until late in the season, and the temperature was almost the same as the loops I used all season. The heat transfer from the ground to the water in the pipes is much better than expected. It also appears that the 700 foot long loop performs as well as the 1200 foot ones. (Much easier to pump the water through too!)

About half way through the season, ice started to form on the evaporator coils. I had to add 10 gallons of RV anti-freeze to the water (system holds about 150 gallons of water). That solved the problem.

The compressor draws almost exactly 1000 watts of power.  I've calculated that I'm getting about 14,000 BTU's of heat out of it. That equals a Coefficient Of Performance (COP) of about 4.1, which means that compared to using regular electric heat, I'm getting more than 4 times the heat for the money! (Better than expected.) And that's late in the season, when the ground has cooled a lot. The efficiency will be even better in the fall.

Here's the operating cost comparison to using my fuel-oil furnace:

With my 80% efficient oil furnace, one gallon of oil will provide approx. 112,000 BTU's of useable heat. The December 2009 cost for a gallon of oil was $2.38, so it could be said that 112,000 BTU's costs $2.38.
With my heat pump producing 14,000 BTU's per hour, each 8 hours it also produces 112,000 BTU's of useable heat. Using 1000 watts of electricity, in 8 hours it will use 8 kilowatt-hours, at a price of 12 cents each, costing only 96 cents!
So even though it has run non-stop for the last month, adding about $86.00 to the electric bill, it has saved about $213 in oil costs. That means a net savings of $127.00 for the month. Not bad!

I was also a bit surprised how much you can actually heat a house with only 14,000 BTU's per hour, considering that the oil furnace puts out 70,000 BTU's. The heat pump alone can keep the house (not well insulated, old windows) at 72 degrees all the way down to about 35 degrees outside, at which time the oil furnace has to start helping out. Last season (2008-2009) I used my little "test" heat pump all season, with the intention of building a bigger one, which never happened. This season, I relocated the A-coil on the furnace from the "outlet" (hot-air) side of the furnace, to the "inlet" (cool, return-air) side of the furnace. When the A-coil was on the hot side of the furnace, I had to arrange for the heat pump to stop while the furnace ran, and until it cooled down (not good to run hot air into an operating heat pump coil). Now with the A-coil on the cold side, the heat pump can continue to run while the furnace is helping out. The heat pump A-coil increases the air temperature by 15 degrees. So the 68 degree return-air gets preheated to about 83 degrees going into the furnace.

Next year, I plan to add a second stage to my unit, so it'll handle most all of my heat needs. I also plan to build my own geothermal heat pump water heater, to replace the electric unit I have now. It should cost about 1/3 as much to run! This will be a seperate machine, but utilizing the same ground loops. I anticipate being able to easily draw at least 36,000 BTU's per hour (3 "tons") from my existing ground loops.

One last thing. Last summer I had a brainstorm to test out air-conditioning the house by using ONLY the ground-loop water circulating through a coil on the furnace, with NO compressors or Freon. I reworked an old 5 ton freon A-coil to be a water coil instead. This coil has 8 circuits of 1/2" tubing, which allows the water to flow quite freely through it.  I didn't get around to hooking it up until July, when the ground had already become quite warm. It actually cooled the house fairly well, but wasn't quite cool enough to remove the humidty from the air. But I decided that for the pennies on the dollar that it cost to run, it was good enough! I still have it in place, so I can test it in the Spring of 2010, to see how it works with nice cold water.

ADDED 8/19/10: I used my "free" air conditioning this summer. As expected, it worked very well at the beginning of the season, the 38 degree water did a good job at both cooling and dehumidifying the house. As July came around, the water temperature reached about 58 degrees, which reduced the cooling, and especially the dehumidification capabilites. Although at this time it is still doing 80% of the cooling, my window A/C finishes the job on the real hot days.


I would bury the pipes deeper, perhaps 8 to 10 feet, to reduce seasonal fluctuations in temperature. I would use shorter loops, 600 to 750 feet each. And I'd rent a really big backhoe!

A little more experimenting needs to be done with the design of the evaporator coil. At first I tried using 50 feet of 1/2" O.D. copper tube, but it had more pressure drop than desired. Then I tried using 40 feet of 5/8" tube, but the freon seemed to be able to shoot right through down the middle of the tube, with poor heat transfer. So I added 1/2 of my first coil, to the second coil, and have a fair compromise in heat transfer vs. pressure drop. Yeah, I know, one could buy a pre-made heat exchanger, but what fun is that? And if one manages to freeze it up (easy to do), it would instantly be junk! I froze my home-brew coil several times with no damage.

I'll add more detail to this as I learn it. Later, ---Dan.

PS: Be sure to check out my
HOME page for more interesting articles!