A better way to install a heat pump

Short version: Install your air conditioning compressor in your unfinished basement and it will run more efficiently. Slightly longer version: Watch this video.

Lutfey Innovations is testing our theory that heat pump efficiency can be increased by installing the compressor in the unfinished basement of a house. This allows the condenser unit to draw 55 degree air all year round and will increase the life span of the unit by keeping it at a constant temperature away from the elements.



While moving the heat pump I managed to get a kink in one of the copper tubes which resulted in all of the gas escaping from the system. I’ve spent the better part of the last week taking the entire water heater apart, repairing the damage, and getting everything back to where it was before this all started.



The chart above shows the temperature of both tanks when the compressor was running for 2 hours and 20 minutes and then turned off for the rest of the test. The hot side contained 189 L of water and the cold side contained 75 L of a water/antifreeze mixture (30% water and 70% antifreeze to minimize the freezing point). So while it looks like both tanks changed the same amount, the calculations will show heat is being added to the system.

Also note that the hot side started cooling down once the compressor was shut down. If the tank was insulated perfectly this line would be horizontal. I would say that I’ve insulated the tank adequately, with definite room for improvement.

OK, now for the calculations!

Before your eyes glaze over, here are a few helpful reminders: cp is a constant in this case, so don’t worry about that too much, dt is the change in temperature, and m is the mass of what is changing temperature.

Compressor on:

Hot Tank:

E = cp dt m
E = (4.2 kJ/kgoC) (17.3 oC) (189 kg)
E   =  13745 kJ

Cold Tank:

E = cp dt m
E = (4.2 kJ/kgoC) (-22 oC) (75.7 kg)
E =  -7134  kJ

Adding these together gives 6611 kJ  (2388 kJ/hour) added to the system. This is the heat added from the basement floor warming up the cold tank.

But wait! There is more. Once I turned off the compressor the cold side kept warming up. I used the data from an hour after the compressor to find how much heat was added to the cold side.

E = cp dt m
E = (4.2 kJ/kgoC) (4 oC) (75.7 liter) (1 kg/liter)
E =  1310 kJ

So this setup can pull 2388 kJ/hour when the compressor is running and 1310 kJ/hour when the system is not running.

Water Heater Update


Here is latest version of the water heater. The entire setup fits inside a 100 gallon horse trough. The insulated tank is on the top along with the compressor. Below the Styrofoam is the non-insulated tank. The entire system can be assembled and repaired with no tools. The only exception is replacing the compressor.

The Lutfey Loop

The Lutfey Loop water heater utilizes two water tanks located in the basement of a building. The first tank is not insulated and is placed in direct thermal contact with the basement floor. The second tank is insulated like a traditional hot water heater and is placed near the non-insulated tank. A heat pump is placed between the two tanks and moves heat from the non-insulated tank to the insulated tank. The foundation of the building adds heat to the non-insulated tank passively over time.


Version 1.0: I put this together with two 50 gallon water troughs, the inside of an air conditioner, some Styrofoam, two small water pumps, and two temperature sensors.

Version 2.0: Functionally the same as the previous version, but polished up.

Version 2.0: Functionally the same as the previous version, but polished up. The cold tank is on the bottom and the insulated hot tank fits inside the 100 gallon water trough.

2.0 Fully assembled

2.0 Fully assembled

Lutfey Loop Benefits:

  • installation cost similar to electric hot water
  • 5 times more efficient than electric hot water heaters
  • 100 times less expensive to install than traditional geothermal

This configuration takes advantages of traditional geothermal heat pump configuration by extracting heat from the ground. Traditional geothermal systems have the drawback of having to bury a series of tubes underground to collect the heat. The Lutfey Loop doesn’t suffer from this drawback since the heat is drawn directly from the foundation floor.

The Lutfey Loop is also superior to traditional electric hot water heaters because it only concentrates the heat where electric systems heat the water purely from electrical resistance. This allows the Lutfey Loop to be 5 times as efficient as electric hot water heaters.


This data was collected with the setup shown above collected from the temperature sensors connected to a Raspberry Pi.

50 gallons of water was heated up 90 degrees Fahrenheit in 8 1/2 hours using 3.9 kwh of electricity. The data demonstrates that heat is added to the system even after the cold tank stops getting colder. The temperature increase in the hot tank appears to be linear. The slight drop off of the insulated tank (red) is attributed to heat loss through the Styrofoam as the temperature rises.

Full Scale Test

After finishing assembly of the full scale prototype this morning I started the heat pump running so I could measure the performance. I filled both tanks with 100 gallons of water and plugged in the compressor. After letting it run all day I came up with the following data:

Time run: 9 1/2 hours

Volume of water in each tank: 378 Liters

Cold Tank: No temperature Change

Hot Tank: 4 degrees Celsius increase

Electricity Used: 3.8 Kilowatt hours

From this information the amount of energy added to the hot tank can be calculated:

E = cp dt m

E = (4.2 kJ/kgoC) (4 oC) (378 liter) (1 kg/liter)

E   =  6350 kJ => 1.76 kWh

So here is what I concluded from this test:

It appears that the cold water tank is being warmed from the basement floor. (very good result!)

The heat pump is running at only 50% efficiency (not very good :()

Since the small scale test showed a 500% efficiency with the same equipment, I assume there is a problem with the new refrigerant I put in the system after I lengthened the copper pipes. I didn’t notice any leaks in the soldered copper joints and the hot and cold pipes seemed to be in the right temperature range. I suspect that using methods based on YouTube videos might not have been the best idea.

My next step is to find an actual HVAC person to charge up the heat pump and run the test again. If all goes well this will result in a 10x efficiency improvement.




Making Progress

Strange how the temperature getting down to zero Fahrenheit motivates me to go work in the basement. So today I started to build my first full sized prototype. I hacked apart the remains of the window mounted air conditioners so that I could place the radiators inside the two 100 gallon steel containers. The tub on the left is the hot water tank and will be eventually surrounded by insulation to minimize heat loss. The tub on the right is in direct contact with the concrete which will passively warm up the water to 55 degrees. Alert viewers will notice that the copper tubing is held in place with duct tape– I assure you this is temporary until my plumber friend is available to finish the connections. Currently he is too busy repairing frozen pipes to help with my crazy water heating experiment. (see the first sentence in the paragraph)


I positioned the radiator in the hot water tank near the bottom of the tank and the cold water radiator near the top to help maximize the heat transfer.

Once I finish the tubing and insulation I will put refrigerant in the system, fill up both tanks and run the compressor. Unless something unexpected occurs, I should get the same results as in version 1.0 that I ran upstairs in the garage. After that I’m going to put together a raspberry pi to monitor the temperature of the tanks and turn the compressor on and off as needed and measure power consumption. Finally I’ll connect it to the internet so I can check the temperature of the water on my smart phone while I’m at work. No, I’m not serious about that last part. I mean really, who ever cares about her water heater until she has to take a cold shower in the morning?


Warming Up Water


My assistant stands next to the water tank. Bonus points for finding my dog in this image.

I ran a test this weekend to measure how fast the basement floor will heat cold water. I purchased 60 pounds of ice and put it in the first tank. Next I added water from the garden hose and let the mixture sit for a while. Once almost all of the ice was melted I pumped the water into the second tub and measured the water temperature in thirty minute intervals. Here are the results:

0.0   32.3
0.5   34.2
1.0   38.3
1.5   38.6
2.0   39.9
2.5   41.1
3.0   42.1
3.5   42.9
4.0   43.7
4.5   44.5
8.0   54.9

All of the measurements above are in Fahrenheit, but I converted them to Celsius for the rest of the calculations.

The tub contained 88.3 liters of water and in the first hour the temperature rose 3.33 degrees Celsius. Using the equation

E = cp dt m

E = (4.2 kJ/kgoC) ((3.50 oC) – (0.166 oC)) (88.3 liter) (1 kg/liter)

E   =  1,298 kJ => 0.36 kWh

So in the first hour the foundation of the basement added 0.36 kWh of energy to the water. Given the tank can hold up to 378 liters it has the potential to store the following amount of energy.

E = cp dt m

E = (4.2 kJ/kgoC) ((12.72 oC) – (0.166 oC)) (378 liter) (1 kg/liter)

E  = 19,924 kJ =>5.53 kWh

Estimating that the tank can heat up from 0 to 12.6 degrees Celsius in 8 hours it would have the potential to provide 500 kWh of energy per month to provide hot water.

While this test provides some very positive data I should bring up the rate of heat transfer will depend on many factors that might not be reflected in this initial test.

The surface area of the bottom of the tank: The larger the surface area the faster the heat will transfer to the water. Flooding the entire basement with water would provide the optimal heat transfer rate but could pose logistical problems for homeowners.

Rate of heat transfer: This value will depend on many factors such as the materials used in the tank, the amount of fluid movement in the tank, and the rate of heat transfer in the foundation.

Volume of water in the tank: I didn’t completely fill the tank for this test, so the rate of heat transfer will be different for a larger volume of water. Adding water in the tank would allow the compressor to run less frequently, but also adds complexity to the system in terms of structure and cost.

Version 2.0


After showing several people version 1.0 and processing their feedback I came to the following conclusions:

1. If you bend the copper pipes too far they rupture and all the refrigerant leaks out. This is bad. (To be honest here I figured this out on my own.)

2. Don’t position a compressor sideways. Apparently there is some sort of lubricant that depends on the force of gravity to keep the compressor functioning properly.

3. Ripping apart all the wires with the intention to reassemble them later surpasses my skills as an electrician.

So I drove back to the Habitat store and purchased another used window mounted air conditioner. Fortunately there was a different person at the register so I didn’t have to explain why I keep buying air conditioners in the middle of October. Next I drove to a farm supply store and bought two appropriately sized containers that would be both waterproof and yield to the will of my hand saw.

I was much more careful in ripping apart the unit this time around. Other than cutting the wires leading to the fan I left all the other electronic components alone. This AC unit was larger than version 1.0, but the copper tubes were thicker which I believe is a good thing. I used a heat gun on the copper wires before I started to move the radiators with the hopes that I wouldn’t break anything. Finally I made sure to keep the compressor upright the whole time.

Next I cut some holes in the tubs to allow the radiators to sit inside with the goal of minimizing the movement of the remains of the AC unit. Finally I used duct tape and silicone adhesive to repair the sides of the tubs. (or, if you are a fanatic of “The Office,” I filled the holes with my caulk– THAT’S WHAT SHE SAID!)

That is as far as I was able to get this week due to previous commitments with my employer and family. Once the silicone dries I’ll fill both tubs with water and run the compressor. This will let me measure the rate of heat transfer between the two tanks and the amount of electricity required from the compressor. Armed with this data I can make plans for version 3.0– the full scale basement test.

Ground Source Heat Pump Water Tank Version 1.0

I spent a good portion of the weekend building a prototype water heater. My main goal here is to build a functional unit so I can measure stuff like how fast heat is transferred from the foundation to the insulated tank and how much electricity the compressor consumes.

Here is a rundown of the materials used for this project:

$40 used window mounted air conditioner: I purchased a unit at the local Habitat For Humanity Rebuild store. Shopping for this item in mid October is highly advisable. They will not, however, give you a partial refund for all the pieces of the unit that you will end up throwing away.

$6 lumber: I consumed about 16 linear feet of two by fours: These are used to build the stand to mount the compressor and heat exchangers.

$10 uninsulated water tank: This can be any water container that has maximum contact with the floor and minimal insulating properties. I’m keeping my eyes open for a small kiddie pool with a flat bottom.

$20 insulated water tank: Lots of choices for this piece. For testing purposes I’m not so concerned about the insulation, but I would like to have a container that is large enough to simulate the volume of a typical residential hot water tank.

$4 nuts, bolts, and miscellaneous parts: This is all stuff that I already have in my basement, so I’m not going to give a ballpark estimate here.

After crunching all the numbers here I spent $80 on all the materials.

So here is how I got it all together:

Prepare the air conditioner: Remove the outer case and fans from the unit. Most of this can be done fairly quickly with a power drill and an adjustable wrench. Remove the wiring for the fan and any controls that are not needed. This image shows functionally how the system works with the uninsulated tank on the left and the hot water tank on the right.


Build a stand: This allows the cold side heat exchanger to be a few inches off the ground and the hot side exchanger to be on the bottom of the insulated tank. The compressor is placed in between the two tanks.

Mount the compressor and exchangers: The best advice is to plan ahead to minimize how much these components are moved around. The tubing will bend a fair amount, but if you pull too hard the refrigerant will start leaking out. Not a good situation.


Place the stand in the uninsulated tank: The lower exchanger should be completely submerged in water.

Place the hot water tank on the top of the stand: The upper exchanger should be placed near the bottom of the tank. This allows the heated water to passively rise up and the cooler water to be heated.

Once I get this version up and running I’ll be able to measure how well this setup performs. Stay tuned.