Heat Exchangers and 3D Printing





Introduction: Heat Exchangers and 3D Printing

About: Engineer from the Netherlands. Projects in 3D printing, electronics, prop making, sewing and whatever interests me. (If you have a pressing question go to my site and ask it there. I am way more likely to ...

I live in a one bedroom apartment and while I love living there, there are some issues with ventilation. There are only windows on one side with small ventilation grilles above two windows. The air tends to get stale in the winter because there is not enough air flowing. Opening the windows every day would solve the problem, but that would waste a lot of heat.

There is a way around this, a 'heat exchanger' or 'heat recovery system'. Now normal people would just open a window, but I always wanted to try and make one. In this Instructable I will share a few designs I tried and share efficiencies and my conclusion for now.

All exchangers are at least in part 3D printed and there is even a completely 3D printed design.All files are available (and will be kept up to date) here:


Step 1: What Is a Heat Exchanger

Simply put a heat exchanger is a device that transfers the heat from one flow of to another (in this case air to air). They do this by making 2 flows of air with different temperatures pass one another separated by a thin wall. Heat travels from the hotter flow of air, through the wall and into the cooler flow of air. The only part that requires energy is making the air move. Everything else happens by temperature difference alone.

The heat exchanger is great for ventilation systems for homes in cooler countries because homes can then be ventilated without losing the heat. The humid, stale air of inside can be replaces with the dry, fresh air from the outside without also replacing the heat.

In an ideal world, heat exchangers can transfer up to 100% of the heat from one flow to the other, but in the real world, percentages are lower. My initial goal was to make an exchanger with 80% efficiency. Efficiency in heat exchanger is how much of the heat is transferred between the flows of air. An example:

The air outside is 0°C and the inside air is 20°C. If there is a 80% efficient heat exchanger in the ventilation system, the 0°C air from the outside will be heated to 16°C when it exits the exchanger. The 20°C air will be cooled to 4°C when it exits the heat exchanger.

As seen in the image, there are a few design options. Also there are 3 distinct flow directions.

  • Cross flow: The streams flow perpendicular to each other. Exchanging heat sideways (the 50 - 70% exchanger)
  • Parallel flow: Both flows enter from the same side and exit the same side. This way allows for at most 50% heat transfer.
  • Counter flow: The most efficient, where the flows are opposite from one another. This creates a hot side and a cool side. Depending on the design, counter flow can get close to 100% efficiency. All my tests use counter flow heat exchangers.

Step 2: Partially 3D Printed Heat Exchanger

The partially 3D printed heat exchanger has all of the complex parts printed. These are the end caps, adapters and shrouds. The main frame of the exchanger is a PVC tube. For those who want to duplicate the design, it is a 90mm PVC tube and I dare anyone to find it in an ordinary hardware store. I messed up and designed this heat exchanger around a tube that is only available on the internet.

The exchanging is done in drinking straws. Normal straws are only around 20cm long but I managed to find straws 75cm long on the internet. The straws have a wall thickness of 0.1mm and a diameter of 6.5mm. 91 were used in the current design. Effectively each straw has a surface area of around 140cm² (it is 150, but not all is used). With a total of 91 straws, this gives a total surface area of 12500cm². Over a square meter or 13.5 square feet.

To assemble the exchanger, the tube was cut to length. The straw holes in the end caps were drilled to give a loose but clamping fit. Both end caps were glued to the tube and aligned. The straws were fed through by hand, one at a time. After all straws were in place, one side was glued using super glue. Then from the other side the straws were pulled to tighten them and the other side was glued as well.

The fans are standard 60mm fans. The fans are rated at 38m²/h, but due to the resistance of the exchanger, a fraction of that is probably achieved. The fans consume 1.75W each. The fans have adapters to mount them to things like vacuum cleaner hoses. 60mm fans are not ideal. They are noisy and not that efficient. In the future I might replace them with something quieter and more efficient like 120mm case fans.

To download the files, go here:


Step 3: Completely 3D Printed Heat Exchanger

The completely 3D printed version is, as the name suggests, completely 3D printed. To make it I modified my Ultimaker with an E3D V6 with 0.25mm nozzle.

The walls of the exchanger are 0.3mm thick. The outside dimensions of the exchanger are 15x8x7cm but it has an internal surface area of around 1000cm² (1/10th of a square meter or about a square foot). It is printed in PLA and takes around 10 hours to print at 0.16mm layer thickness. With special adapters it can fit 60mm fans and all the other adapters I have designed.

special adapters were printed to connect the 60mm fans to the 3D printed exchanger. The accessories used on the partially 3D printed heat exchanger also fit on the completely 3D printed version.

To download the files, go here:


Step 4: Measuring Exchanger Efficiency 1

Making a heat exchanger is meaningless without determining how efficient it is. The test is fairly straight forward and gives a decent ball park efficiency figure. In the real world efficiency is also dependent on flow, but for this test the heat exchanger will be ran at one speed only.

The basics of the testing equipment is an Arduino Uno with 4 10k thermistors (NTC) and an SD card for logging. Thermistors are not known for their accuracy, but with some calibration I got them to operate within +/- 0.5°C of each other. The logger takes a sample of each airflow every few seconds and stores it on an SD card. These values can be imported into excel to make a functional graph.

The flow will not be equal on both sides. I currently have no way to measure actual flow. To still get a good efficiency reading, both in and output flows were measured. By calculating both efficiencies, I can average them out and get a decent ball park on the efficiency.

Efficiency is measured as full potential energy vs. energy actually being used. In the case of the heat exchanger it is total temperature difference vs. transferred heat. This is assuming equal flow and equal mediums. The 4 temperatures (unit does not matter):

  • Hot in (the warmer air that enters the hot side of the exchanger)
  • Hot out (the warmer air that exits the cool side of the exchanger)
  • Cool in (the cooler air that enters the cool side of the exchanger)
  • Cool out (the cooler air that exits the hot side of the exchanger)

Hot flow efficiency: (Hot in - Hot out) / (Hot in - Cool in) x 100%

Cool flow efficiency: (Cool out - Cool in) / (Hot in - Cool in) x 100%

Step 5: Measuring Exchanger Efficiency 2

To measure the actual efficiency a one of the inputs was suspended over a heat source (ie. a heater) and the other input was in a cooler place. Both setups have 4 fans, 2 for each flow. The heat source is approximately 35°C while the cooler side is around 19°C.

Partially 3D printed exchanger

The first graph is a short test of the partially printed exchanger inside of my house. Every bottom number is 1 second, the total test was around 18 minutes. From top to bottom, the lines are: Hot in, Cool out, Hot out, Cool in. The hot air was removed too soon, the line was not yet stable. The stables situation was: HI: 34°C, HO: 23°C, CI: 18°C, CO: 32°C. Putting these values through the efficiency formulas gives 68.75% for the hot flow and 87.5% for the cool flow. This averages to 78.125% efficiency. The difference in efficiencies is due to the different flows and measuring errors. An interesting this to see is the thermal 'lag' of the exchanger. It takes over 10 minutes for the exchanger to reach a stable temperature.

The second graph is a longer test in my local (cooler) hacker space with an electric heater for a heat source. Every bottom number is 6 seconds, the total test lasted about 2,5 hours. Again from top to bottom, the lines are: Hot in, Cool out, Hot out, Cool in. Interesting to see here is that the temperature of the room slowly rose over time, due to all the electric heaters used. The heat was periodically removed to see the effect. 2 stable lines can be found, one at 450 and one at 1000.

450 -> HI: 35°C, HO: 20°C, CI: 14.5°C, CO: 30.5°C, giving 73.2% for the hot flow and 78.0% for the cool flow

1000 -> HI: 31°C, HO: 19.5°C, CI: 15.5°C, CO: 28°C, giving 74.2% for the hot flow and 80.6% for the cool flow

The average efficiency of the partially 3D printed heat exchanger is 76.5%.

Fully 3D printed exchanger

The third and final graph is the fully 3D printed heat exchanger. The fans were run at 10V to reduce the flow. Every bottom number is 6 seconds, the total test lasted just over an hour. Again from top to bottom, the lines are: Hot in, Cool out, Hot out, Cool in. The thermal 'lag' of this exchanger is a lot lower. The 2 times used will be the peak at 120 and the valley at 210.

120 -> HI: 38°C, HO: 27°C, CI: 22.5°C, CO: 33°C, giving 70.9% for the hot flow and 67.7% for the cool flow

210 -> HI: 33°C, HO: 25.5°C, CI: 21°C, CO: 30°C, giving 62.5% for the hot flow and 75% for the cool flow

The average efficiency of the Fully 3D printed heat exchanger is 69%. An important note to make here is that in a smaller (unlogged) test the 3D printed heat exchanger ran at 12V, the efficiency was closer to 50 - 60%.

Step 6: Making It Useful

The initial goal was to make this exchanger part of an active ventilation for my house. The goal is to properly ventilate my house without pumping all the heat from my home. Because I lack a centralized ventilation system and this is a rented apartment, I decided the ventilation above my window was the best place for the ventilation. One side of the ventilation grille is used to draw new air in, the other side is used to vent old air out. The rest of the grille is taped off.

This setup has a few drawbacks. First, the fans are quite noisy. This is not a setup I can run while I am at home. The second is that the hoses leak heat. They are not insulated and will preheat and precool the air. The test was done with outside air at around 0°C, yet the incoming air was already 6°C after only 50cm of hose. The longer hose has a difference upwards of 10°C. The Third thing is that I cannot really close my curtains with the tubes in the way. I can remove the hoses to close my curtains while I am home. It will not run when I am home anyway.

With those drawbacks out of the way,

Does it work?

The answer, YES. After running for over 8 hours while I was at work, the air was a lot fresher. Usually when I come home there is a certain staleness to the air, but now I came home to nothing. Just nice air.

I had the logger running for the entire time. The test started around 8 o'clock, every number on the X is 6 seconds. There are 3 zones of interest.

0-3000: Here the air outside is slowly heating up. Temperatures around this point are: HI: 17°C, HO: 10°C, CI: 6°C, CO: 14.5°C, giving 63.6% for the hot flow and 77.3% for the cool flow, averaging 70.5%.

3000-4000: Here the sun hits the window and there is a spike in temperature. No useful data can be gathered from this time.

4000-6000: The air outside is slowly cooling. Temperatures around this point are: HI: 17°C, HO: 12°C, CI: 8°C, CO:
15°C, giving 55.6% for the hot flow and 77.8% for the cool flow, averaging 66.7%.

Step 7: Conclusion

There are 3 conclusions I can draw from this project so far

The project as a whole was a success. I got fresh air into my home and the heat exchanger exchanged a large percentage of the heat. The execution of the ventilation could have been quieter and more compact, but as a whole, it works better than expected.

The big partially 3D printed version works remarkably well. Running at 70 - 80% efficiency, it is good for one of my first attempts. It does have some issues with unequal resistance and pressure drop. The side that goes through the straws has less resistance than the side around the straws and it takes quite some pressure to get real flow out of it. The 60mm fans can get air to move, but they are struggling hard. Whether the efficiency drops significantly if the flow increases much is not yet known.

The 3D printed heat exchanger works surprisingly well. I did not expect it to have the efficiency 69% it had. That said, from other smaller test it does seem that if the flow goes up, the efficiency goes down quickly (for the fully 3D printed version at least). The big logged test I did was at a lower flow. When I later preformed the test with a higher flow, I got 50 - 60%. Not bad, but less than initial testing.

Step 8: Futere Tests / Improvements

What my tests so far have shown me is that I know too little. I need better measurements, more different measurements and a more controlled test. What I want to do is design a better test setup.

I want to manually heat the air using electrically heated wire. This way I cannot only control the exact temperature, I can also determine the flow by comparing the energy I added to the air to the temperature rise of the air. I want to use better temperature sensors that are both calibrated and more accurate. This way I can get more precise results.

For the test itself I do not only want to determine efficiency at one temperature and flow. I also want to see what the influence is on the efficiency if the temperature difference (ΔT) becomes bigger and smaller, and how the efficiency changes when the flow is changed.

In this setup I want to test small test exchangers, so I can see the effect of wall thickness, flow paths, surface area and more. If I learned one thing from the tests I have done so far, it is that there is a lot more to know here than I already do.

While I cannot promise when I will do these tests, I really want to do them.

2 People Made This Project!


  • Oil Contest

    Oil Contest
  • Water Contest

    Water Contest
  • Creative Misuse Contest

    Creative Misuse Contest

27 Discussions


16 days ago

I have built a heat exchanger, for a friends terrarium.
I took an used heat exchanger from a dead tumble dryer, so I had only to design the enlosure and fa connections. I did not measure the efficiency.


OMG Thank you thank you thank you !!

i live in a tiny box room and its FREEZING COLD this time of year - well the only reason i hate UK is because we got 9 MOnths OF WINTER AND ONLY 3 months OF SUMMER... SO....

This will help greatly when i do my soldering marathons and still keep the heat in my room while bringing in fresher air, well, i hope i get it right when my 3D Printer arrives and i get used to printing with it lol

Gathering up them straws now !

Congratulations on your great work!

A couple of years ago, I myself designed an air to air heat exchanger that was supposed to be 3D printed and be the core of an HRV system for small spaces. My company, together with a fellow company were thinking to set up a crowdfunding campaign, in order to develop a customisable commercial solution. We were focusing on using standard size (10 of 15 cm), circular diameter pipes with insulation material encasing the heat exchanger and computer fans for moving the air (like you did). We were thinking we could achieve air flow values of 30 m3/hour. Unfortunately, I was then told that my heat exchanger design would cost nearly 500 euros to print, even in polymer. That caused us to cancel the initiative.

Since our company has been working on engineering projects to improve the indoor environmet, I would be fascinating to be able to provide a customizable solution with a cost of, let's say, 100 euros. I hope that we will get the opportunity to set up this project again in the future. Any ideas?


1 year ago

I'm going to try to build one of these using my library's 3d printer! Newb questions. I will have to do the partially 3d-printed version. So, question about sourcing the other items. I think I've found straws that will work. Is 90mm the inside or outside diameter of the pipe? I'm in the U.S. and wondering if I can fit 3in pipe, which has an outside diameter of 3.5in/89mm.

2 replies

iisan7, did you have any luck building it with 3 inch PVC pipe? Planning on making one too.

90mm is the outside diameter. 89mm should fit.

Hey, very interesting project, but I have one remark. You're not using your ventilation power to calculate the efficiency;)

an other remark is about your ventilation. You're using computer ventilators, they're made to give speed to your air, not pressure(that's what you need). You could achieve this maybe, I'm not sure, by putting two ventilators in serie.

You could also use the building to help you. If you live in a windy area it can suck the air out by itself, you just need to make some changes ;)

1 reply

I am currently only focusing on thermal efficiency. I would love to include the fans into the equation, but for that I also need the flow rate, and that I lack. My future setup will include an airflow meter.

PC fans were what I had at hand. I knew from the beginning they would be sub optimal.


2 years ago

Interesting! I've had a vague plan to build a flat panel cross-flow heat exchanger with laser-cut acrylic spacers and aluminium foil for a while, but yours looks way nicer and probably works better too. It's interesting to hear the separator doesn't need to be thermally conductive, that opens up a range of new possibilities.

As someone living in a cold damp country I'm most interested in HX for ventilating bathrooms and kitchens to deal with humidity inside. Have you had any issues with condensation in yours? I'd imagine the outgoing air side could have some problems with condensation forming as the air cools down.

1 reply

From what I have learned so far, the hardest thing is getting the heat to move from the air to the wall and vice versa. If the wall is not too thick, it should not be the bottleneck. More thermally conductive materials do create less resistance through the walls, but I have yet to find out if the added transfer outweighs the extra cost of using metal.

Thank you very much for this instructable.
Have you thought of making a cross-flow heat exchanger using stacked coroplast?
The black PP version has the best lambda (heat-transfer) value..

3 replies

I might try one for testing when I have my test setup, but for right now I am not that interested. It is a very common design that a lot of people have already made. It is good enough for the materials available, but I think I can do better with 3D printing.

Sure, you guys want to play with your 3D printers ;)
Keep in mind that in moist applications (bathroom, kitchen) the coroplast stack can be placed in such a manner that the moist-and-cooling airflow is pointed downwards, therefore gravity will help to move condensate dropplets downwards :P, where a mass-triggered bleed-drip can remove the water without (air) pressure loss.
Obviously you can tackle the same problem with your 3D print design in a similar manner.
In professional HRV designs sometimes porous material is used, so the condensate can migrate 'back' to the other air stream, again without pressure loss.
Please keep us posted, it's an intriguing project.

I do like playing with 3D printers yes. One problem is that coroplast is a bit harder to find here in the Netherlands. Another is that is already widely explored by a lot of people. The only reason I would do it is for reference. I want to see if there are some 3D printer specific designs that might offer specific benefits. Maybe they don't exist, but I want to have tried at the very least.

A thing I like about the design of coroplast HRV's is how cheap they are. Coroplast is inexpensive and the design gives enormous surface for little money. It does come with the disadvantage that the heat needs to travel from the air, to the coroplast, through a layer of glue, into another layer of coroplast and then back to the air. There are a lot of transitions.

The moisture thing is something I will keep in mind.

This is very interesting. Most of the homemade HRVs (heat recovery ventilation) exchangers I have seen were made from clorroplast. The idea of 3D printing the heat exchanger means that you can make the exhausted air stay in the exchanger for a longer duration. By making the exchanger with turns and kickbacks inside (for the exhausted air) it should heat the unit up faster. I would think it would be similar to the effect of electricity traveling through a conductor that is too small. The conductor gets hotter, faster. It may require a larger ducted fan to push the air through on the exhaust side. The down side is that condensation could build up inside the exchanger and not be able drain out very well. You could always use the arduino to time out and backflush the exchanger once a day to clean it out.

Cool project! Do you think you achieved high enough air speeds to get turbulent flow? If not you might think about adding baffles to your setup to combat the stagnant film that forms at the surface of the straws due to the laminar flow.

Very interesting! I never thought to print a heat exchanger before. I am surprised you had reached this high efficiency nevertheless the heat conducting material is plastic. We have to wait for metal-like material to print with our printers. Maybe it will be a good idea to add a filter to prevent the unit from clogging by dust and decrease it's efficiency. Anyway a good job and a way to made air heat exchange possible for less money.

3 replies

The interesting thing is that metal may even work detrimental. I spoke with a manufacturer of ventilation heat exchangers when I was at school. All commercial units are made with thin sheets of plastic. He said that metal did not really make it more efficient. What I gathered from it is this.

The biggest challenge is getting the air to move to another material. At the wall thicknesses used it doesn't take any time at all for the heat through the material and it needs to be transferred to the air again. The main challenge is getting a lot of surface. Metal might even conduct so good that it no longer transmits to air directly, but travels through the material itself. The heat then travels to an area cooler and messes up the transfer.

I will try metals when I get to doing the big test, I still want to see for myself.

And filters are probably a good idea if I am going to run this for longer periods of time. I will add them when I replace the fans for something more quiet.

Nevertheless i work in customer service for a big HVAC company, i never knew that the trick doesn't depend on metal. At least the cross-flow HX is the most used in our ventilating and dehumidifying systems, also it doesn't need another supporting energy like pumps or motors.

Your fully printed HX is very interesting but i know that printing with a 0,25 nozzle could lead into a long challenge.

As member TheBlackBird has noted it would be a good idea to use turbine-like radial fans like in vacuum cleaners or the real ventilating systems. Maybe a nice project to develop for printing. I think a pc-fan quickly gets into an air short- circuit if it has to build pressure.

It was only one guy. I am not expert myself on the subject. When I get to making a good test bench I will give metals a good try as well to see if they are worth the extra cost or not.

The fully printed HX was the reason why I designed the partially printed HX, it took 10 stressful hours for only 1000cm2. Once I had that I started looking for better ways.

There are special high pressure PC fans for stuff like radiators but those are only 3mmH2O, or 0.0003 Bar (that is high pressure!?). Radial fans do have the pressure, but they are noisy and power hungry. The reason I didn't use them on my tests is because I only have 1 and I needed at least 2 for it to work.

I have been thinking of developing a mixed flow fan for 3D printing for various of my projects. I have designed and printed RC propellers and turbines before, it is really fun to do. For domestic use the noise level needs to be extremely low and all motors I can use have drawbacks. For my better test setup I think of using RC BLDC because I precisely need to control the power.