Is there any research into use of thermoelectric cooler along with part cooling fan to get quicker cooling without strong air currents that apply pressure to the still-soft material? I experimented with custom fan ducts in the past trying to get better cooling and avoid warping for printing thin layers of PLA at high speeds, but found that the concentrated stream of air blowing on the part actually deformed it before it could cool. At the time I wondered if using significantly colder air, at a much lower flow, would work better. But every time I've searched for thermoelectric (peltier) coolers with 3D printing, I've found results that are about cooling motors or the heatbreak (especially inside heated enclosures), nothing about part cooling.

If there is no research on this and I want to experiment myself at some point, are there constraints I should consider for how to mount it (in my case on an Ender 3, but also in general)? Perhaps on a separate intake duct before the cooling fan? Or between the cooling fan and hotend assembly to let the waste heat dump into the assembly that the hotend fan is already cooling?

  • 2
    $\begingroup$ The room temperature air that the fan would normally provide is already much, much colder than the molten filament. Even if you manage to cool the air by 20 C (which would be quite impressive using a TEC) that's only a ~10% increase in cooling performance. TECs are very inefficient and I think you will have a hard time cooling the air by more than a few degrees; trying to piggyback off the hotend cooling fan is foolish -- you will need a much larger heatsink and fan than that. $\endgroup$ Dec 9, 2020 at 19:41
  • 1
    $\begingroup$ You will still need a way to remove the heat from the hot side of the Peltier cooler. If not a fan then liquid/water cooled. $\endgroup$
    – Perry Webb
    Dec 9, 2020 at 20:16
  • $\begingroup$ And a water-cooler would mean a huge mass and the trouble with sealing a system right next to a 12 or 24V heater cartridge... $\endgroup$
    – Trish
    Dec 9, 2020 at 20:18
  • $\begingroup$ What if. instead of blowing air across the print, you suck the air out the top of the box with side vents allowing an even flow of air? $\endgroup$
    – Perry Webb
    Dec 11, 2020 at 17:17
  • $\begingroup$ I noticed running a box fan on low next to the printer cools way more than the little fan on the head, yet exerts less air pressure due to lower windspeed. It's also cooling all the time instead of just as it passes by. Might be all you need to push the speed up. You can also print two prints at a faster rate than one, since it's cooling for twice as long per part. $\endgroup$
    – dandavis
    Dec 18, 2020 at 22:50

3 Answers 3


On the printhead?

TECs or Peltier Elements are incredibly inefficient compared to airstream coolers. Their only benefit is perfect temperature control, from which you will have nothing because there is no firmware that cares for the temperature of cooling air or the cooling body of a Hotend. Also, a TEC creates a lot of heat on its output side - which means you heat the air just millimeters away from where you want to cool the air!

To get the heat produced by the TEC away, you either need a rather large cooling body - which is a lot of weight and space you need. As a result, you reduce the maximum print speed a lot. A water cooler isn't necessarily that much lighter, but it also gets us the trouble of having a highly conductive liquid right on the printhead.

lighter alternative: compressed air

You'd have much better efficiency by having compressed air decompress (as in: get out of a slim nozzle) slowly just a few millimeters in front of the air intake of your part cooling fan - expanding air cools down a lot, and running a compressor for a few moments takes less energy than running a Peltier element with the same temperature drop. In a pinch, a CO2 canister could provide the needed high pressure air, and a nozzle like you have it on an airbrush would work.

Move it off the printhead?

As the weight of the necessary secondary equipment is an issue, it might be better to move the Peltier element off the printhead. For example, by using a flexible hose that supplies the air to the cooling fan, and feeding that with precooled air - and now a Peltier element can shine: by having the weight be no longer a matter, we can use a rather large cooling body on the outside and cooling fins on the inside.

Sample "Intake"

  • $\begingroup$ Some further thought exchange in chat?. $\endgroup$
    – Trish
    Dec 9, 2020 at 23:09
  • 1
    $\begingroup$ Just to quantify how terrible peltier coolers are: cooling the output from a 40mm noctua fan from 25C to 0C would require 68 W of cooling. You can expect a COP of about 0.5 so you'd need 136W of electrical input power to the peltier, and need to deal with 204W of waste heat. That's three high-end CPU's running at full blast. Imagine needing to strap three CPU coolers with two 120mm fans each to your hotend to cool the peltier. Only for a ~15% increase in how much colder the cooling air is than the molten filament. $\endgroup$ Dec 10, 2020 at 8:42
  • $\begingroup$ Adiabatic cooling is worth considering. $\endgroup$
    – Perry Webb
    Dec 10, 2020 at 13:57
  • $\begingroup$ A primary consideration with adiabatic cooling is if the gas isn't dry you will get condensation or even ice crystals. An inventor tried to use adiabatic cooling for car air conditioning using the cabin air, but ice crystals made his invention useless. However, if the gas expansion is close to the heat sink so that the air never gets below the ambient temperature, then condensation isn't an issue. $\endgroup$
    – Perry Webb
    Dec 10, 2020 at 14:25
  • $\begingroup$ Whether or not there's any way to do this that makes sense, I think the estimate of benefit is erroneous. The difference in rate of cooling at the moment the material starts cooling is low but not relevant; what matters is the difference in time to drop below the glass transition temperature which is significantly higher than the ambient or cooled-air temperature. Some possibly wrong back of the envelope math suggests that time to go from 180 to 60 is reduced by 21% if your air is at 0 instead of 20, and by 23% if you want to get all the way down to 50. $\endgroup$ Dec 10, 2020 at 18:41

For cooling the printed object:

  1. Combined with air compresser

What seems the most practical is to use an air compressor with a tank large enough to ensure that the air in the tank has time to cool off. This gives you the option of adding an air dryer if needed. You could cool the compressed air just before blowing on the print though a Peltier cooler and get additional cooling as the air releases toward the print.

Information link:



  1. Printer Ceiling

You could put a Peltier cooler on the ceiling with a heatsink/cold-sink covering most of the inside ceiling and fans only on the external (outside the ceiling wall) heated part of the Peltier. Convection would move the air. Without and enclosure, positioning the Peltier to use convection would be good because the fans, especially those removing the heat would not be fighting convection.

  1. Printer Bed

You could use a Peltier cooler to both heat and cool the print bed. All you need for stitching is to change the polarity of the voltage on the Peltier. A Peltier could handle those temperatures well. While Peltier heaters/coolers aren't energy efficient, because they can do both, they have good and quick temperature control.

If you wanted to get fancy, you could have multiple Peltier devices across the bed, so that you could keep heat under the print and have all the other devices cool. You would need to know the fastest rate you could change temperature without cracking the glass on a glass bed.

Cooling the heat sink on the extruder. (answering a different question)

The most practical way to use a Peltier cooler is to take advantage of its temperature differential over a short distance. One could put it between the heater block and the heat sink, requiring a hole in the cooler for the filament to feed through. The Peltier effect has a limit of a 70°C maximum differential between the hot and cold side. Another limit is manufactures list a maximum temperature of 200°C on the hot side. This is usually because of solder joints.

  1. The Peltier cooler would need to be customer made to fit between the heater block and heat sink a) hole in the middle, b) designed to withstand heater block temperatures >200°C.

  2. The heat sink probably still needs a fan due to the maximum temperature differential of 70°C. A 70°C maximum temperature differential makes it difficult to be the only source of cooling. Rarely are Peltier coolers the only source of cooling. Only when they are in an insulated barrier such as in the wall of a ice chest. Other wise the heat from the hot side mixes with the cold side. Peltier coolers move heat from one surface of the material a short distance to the other side of the material. If you don't cool the hot side of the material, when using it in an open area such as an extruder, the heat will circle back around to the cold side.

Its main advantage would be a fast drop in temperature between the heater block and heat sink. However, it is an expensive project, and one needs to evaluate of this could be better achieved by other methods.

Links using Peltier cooling with 3D printers:





Visual example:

Cold plate enter image description here

  • $\begingroup$ I don't see how this answer pertains to part cooling. $\endgroup$ Dec 10, 2020 at 14:57
  • $\begingroup$ I don't see how your comment pertains either. $\endgroup$ Dec 10, 2020 at 18:24
  • $\begingroup$ Since you seem to have misunderstood, this question is about cooling the extruded material after it's deposited as part of the print, not cooling the part of the toolhead above the melt zone. $\endgroup$ Dec 10, 2020 at 18:32
  • $\begingroup$ This hopefully fixes my answer. $\endgroup$
    – Perry Webb
    Dec 11, 2020 at 0:00
  • $\begingroup$ I still don't understand how you intend for it to be relevant. You do not cool the bed to cool the print. Bed must remain at the same (usually heated) temperature you started at to prevent loss of adhesion, and cooling it wouldn't help cool anything beyond the first few mm of a tall print anyway. $\endgroup$ Dec 11, 2020 at 19:01

It is true that if you try to do bridges with a very hot filament the cooling air will deform or push away the hot filament if it's set at high speed, or it won't cool it enough if you run the fan slower.

I experienced it with PETG at 245 °C while performing a parametric optimisation as described in:

Still, the TEC are inefficient and they require a bulky heatsink to cool the hot side. Not only that, you also need a heatsink with thin fins to cool the air, which cannot be too small too because the air is flowing relatively fast.

Overall, it's clearly not practicable so I doubt you will find studies to confirm what is obvious.

  • $\begingroup$ I find it happens not just in bridges but in moderately steep overhangs. $\endgroup$ Dec 9, 2020 at 22:35

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .