Appliances Peltier based mini air conditioner not working

  • Thread starter jamescam
  • Start date
Trying to make mini a/c. Peltier cold site is getting colder but when I attach heat sink to cold side then cold is not transferred to heat sink.
Hi guys

I am not certified engineer but I watched many videos to make home based mini air conditioners using Peltier module. So I decided to try this as experiment.

My plan is if this small test project works well then I will proper mini a/c (for my small 10x10 feet office or at least for my office table) using 2-3 Peltier modules with radiator method instead of heatsink and fan for hot side.

What I did is, I made approximately 1 foot long and 8 inch wide and 6 inch high box with help of chip boards and extra unused stuff available at home. This box is divided into 2 parts. Both parts are separated by chip board so hot air from hot side heat sink don't flow into cold side area.

Peltier Module: TEC1-12706 (6 Amp)

Box Back side: Used computer power supply 12 volt 16 amp (16 amp in case I add one more Peltier module so same supply works fine) + TEC1-12706 Peltier Module + heat sink for hot side + 12 volt fan to exhaust heat out of box.

Box Front side: Only Cold side of Peltier module is visible in this front box + heat sink attached to extract cold from Peltier module surface. I attached one 4" x 4" 12 volt fan on front of this box to throw cold air out of box.

Its very simple I join Peltier module + hot side fan with power supply and Peltier module is working fine. Back hot side heat sink is getting hotter and fan is throwing heat out of box.

And other cold side of Peltier module is cold. I checked with digital temperature meter and temperature on Peltier module cold surface was less than 20 degree. Front box where cold side is attached is almost 8" x 6".

I put coin on Peltier module and in just just 10 minutes ice was started making on Peltier module.

But the problem is when I attach heat sink and turn on front fan then fan is not throwing cold air out of box. Although Peltier module surface is cold but heat sink attached to cold side is not getting colder.

So please help me to figure out why cold is not transferred from Peltier module surface to heat sink. I assume once heat sink is cold then fan can throw cold air out of box.

I did one more thing, I put one 3" x 3" fan on cold side heat sink that was throwing air on heat sink (getting air from front outside and throwing on heat sink) and I put digital temperature pointer in heat sink fin and temperature dropped just 1-2 degree and then stopped. So it did not work.

Another thing, since I am totally new to this so when I buy Peltier module then shop keeper gave me standard heatsink paste. (white color). Is there any kind of special heat sink paste or grease for this purpose. Is this grease or paste issue that's stopping cold from being transferred from Peltier module surface to heat sink.

Or there is any special heatsink for cold side or this purpose?

So please help me to figure it how to fix this.

If this mini project is successful I mean a small 8" x 6" front side box is cold then I planned to make proper a/c with 2-3 Peltier modules using Radiator to take off hot side heat.

Help me please.



I put coin on Peltier module and in just just 10 minutes ice was started making on Peltier module.

But the problem is when I attach heat sink and turn on front fan then fan is not throwing cold air out of box. Although Peltier module surface is cold but heat sink attached to cold side is not getting colder.

So please help me to figure out why cold is not transferred from Peltier module surface to heat sink. I assume once heat sink is cold then fan can throw cold air out of box.
Welcome to PF!

Peltiers put out very little cooling. You may want to try to calculate the maximum air delta-T; how cold the cold side air could get. Google for the spec sheet on the Peltier module, find the cooling output in watts, google for the fan airflow (if you don't already have it), and use the information to calculate the change in air temperature (I can walk you through that if you need it).

Also, could you provide a picture? The heatsink type/size matters a lot.

I've built a couple of Peltier coolers and they don't have fans on the cold side because Peltiers don't put out enough "cold" for a fan to be helpful.
Use thermal grease and also, try putting heatsink/fan on both side. That way the heat can be removed from the 'hot' side, so the 'cold' side will get colder. If no cooling on the 'hot' side then the temperature there will just get high enough to match the heat flow from the heatsink on the 'cold' side.


Science Advisor
Homework Helper
Insights Author
Gold Member
I agree with @russ_watters . I used to have a cooler which used Peltier cooling and would plug into the car cigarette lighter. It had a small fan inside, which I think was mainly to circulate the air in the cooler. It had a fan on outside to remove the heat. This was good at keeping things cold, but not great if I wanted to cool down some drinks. This particular model could reverse polarity, so I could keep hot foods warm.
Thanks everybody who replied to this thread to help me out.

@russ_watters, I think calculating this kind of stuff for a layman like me is a lots of work and beyond my expertise. I will appreciate if you can do this for me. I have uploaded few pictures as you asked.

box-min.jpg cold-side-min.jpg heatsink-paste-min.jpg heatsink-1-min.jpg heatsink-2-min.jpg

One Experiment Result:I put one steel glass filled with water directly on Peltier cold side and after approximately 20 minutes bottom (bottom side 2" high, not full glass or upper side) of steel glass was enough cold just like it was put in freezer. Also small bubbles started appearing like we see when we put cold drink in freezer. So it means in this experiment I found that cold was transferred from Peltier device to steel glass and it took almost 20 minutes to completely cool down bottom of glass. It works. Same result I want with heatsink. Steel glass was put on Peltier cold side without any heatsink, paste or grease and it was directly on Peltier surface.

So throwing cold air outside is secondary issue. First issue is why heatsink is not getting colder like steel glass experiment. I think this is the main problem, once this is solved then second step is to adjust air flow.

@Rive, I am sure "thermal grease" will play an important role in solving this issue. It looks this is improper "heatsink paste" that is stopping cold transfer from Peltier device to heatsink. In addition at Peltier hot side there is already heatsink attached with fan and fan is throwing out heat from hot side chamber. Now again I will go to market to get proper "thermal grease" instead of "heatsink paste".

Please see attached images and answer few questions.

1. Backside of box (hot side chamber) is open, does this matter? Should I also cover that back side. If I cover backside then only source left to take out heat from hot side chamber will be the fan attached with hot side heat sink. Should I keep that side open or cover?

2. Front side of box (cold side chamber) is also fully open left to attach fan (I planned big 4" x 4" fan for high air pressure but I also have 3" x 3" fan) to throw cold air outside. Should I keep this area open 100% or cover remaining area after installing 3" x 3" fan. I think here I need to redesign this chamber. There must should be some sort of air flow I mean air struck with cold heatsink and cross through heatsink fin and then front fan throw that cold air outside. I don't know how to do this!

Please see images and help me out. First step is to apply proper "thermal grease" on Peltier cold side + heatsink to make heatsink colder like my steel glass test.

I am sure its possible to make mini air conditioner with Peltier device (no matter with 1 or 2 or 3 Peltier modules) with Radiator use (instead of hot side heatsink) and put this mini air conditioner on side table to cool down temperature of 4' x 4' or 5' x 5' working area.

Thanks everybody for helping me to complete this mini air conditioner project successfully.

Last edited:
@scottdave please see this video
It works, it looks problem is using wrong thermal paste.
I think a smaller heat sink on the cold side and a much smaller fan, you could put some bottles of water in the cold side without allowing the air to be brought in or vented, the small fan would circulate the cool air around the bottles. Let it run for a while to allow everything to cool before opening any air vents, as it is I think your expecting a bit to much of the peltier.
How you insulate the hot side from the cold side especially around the peltier unit is very important, the thermal paste should be ok though don't use to much.
I've made a number of different designed coolers, with mixed results and I still believe that peltiers work by magic. :)


Science Advisor
Gold Member
2018 Award
I used to use peltiers for chilling down biological specimens under a microscope in an electrophysiology rig (since they use DC they did not produce electrical noise).

The ones we used had built-in to the heat accumulating a plate to which water tubes (coming and going) could be attached to conveniently remove the heat. Although, these were expensive when bought from a lab supplier, you might be able to fabricate one yourself and stick it on the hot side with thermal paste. I have also seen normal AC units that use running water to remove the heat.

You can also stack them to achieve even colder temperatures (as long as you don't exceed their temperature limits). I was considering making a digitally controllled a device for fish sperm freezing, a process where a well controlled temperature decrease is important.

I think you would need a massive peltier to cool a room however. That could get expensive.


@russ_watters, I think calculating this kind of stuff for a layman like me is a lots of work and beyond my expertise. I will appreciate if you can do this for me. I have uploaded few pictures as you asked.
Naw, if you're smart enough and confident enough to try to DIY an air conditioner with a Peltier, you certainly can understand the thermodynamics of it. It's not hard and like I said, I'll walk you through it.

First, google for the data sheet for your Peltier device and for your fan(s). To get our feet wet, we'll start with a simplified, best-case calculation. Find for me the Qmax (maximum heat transfer) of the Peltier and airflow of the "cold" side fan off their data-sheets.

[FYI, I've already run a quick calc of my own using a guess for the fan airflow. The results are...underwhelming.]
First issue is why heatsink is not getting colder like steel glass experiment. I think this is the main problem, once this is solved then second step is to adjust air flow.
Usually the issue is a mismatch between the heatsinks. The hot side gets hotter and the cold side gets colder, but if you don't remove the heat from the hot side fast enough it leaks back onto the cold side and the cold side warms up. The hot side will always be hotter, but if you can only get, say , a 50F delta-T and the hot side is 150F, the cold side will be 100F. Cooling the hot side better and reducing that to 80F yields 30F on the cold side.

The opposite "problem" can happen on the cold side: if you remove (er; add) the heat too fast, the temperature difference can't be maintained and the air won't be very cold.
@russ_watters + @BillTre + @Laroxe Thanks for contribution to help me out. Thanks everybody, each comments give me new idea and help me in this project.

@Laroxe when you replied to this thread exactly at same time I was working on this project and I have great success (definitely this is not great for you expert people, but for me it was great).

My current problem was why cold was not being transferred to cold side heatsink and I solved this problem.

My Findings.
My designed box is totally failure, that kind of box can work for small water cooler i.e. put one steel pot (insulated) on Pelier cold side with flat heatsink. But to get cold air flow this design will not work, there is no airflow space at cold side box. So I will redesign this box according to new findings and material size.

Cause of Cold Transfer Problem
1. Hot side fan was approximately 1-2 inches away from heat sink however it should be tied with heat sink. So in fact heat was not thrown out of box/heat sink.
2. Peltier was not well tied with cold side and hot side heat sinks. Heat sinks were just put firmly at bod sides with Peltier module but not tied very well. So in fact heat at both sides were not transferred from Peltier module. That's the reason cold side of Peltier was cold but cold was not transferred to heat sink.

What I did?
I took out all things from failed design box and attached in open area. During last 2-3 days after posting here I searched a lot and found some ideas how this will work. Chinese products gave me idea, these are ready to use but expensive and will take lot of time to arrive however I can make same product at almost half price. Another great benefit of making product myself is, I will gain experience and can modify it as I needed and make it again and again for others too.

As new experiment I tied both heat sinks tightly with metal wire and also tied one fan with hot side heatsink. Although fan was big but I covered extra part with hard paper so air don't flow towards cold side heatsink.

I used same old white heatsink paste so thermal grease or heatsink paste was not an issue.

Here are heatsink temperature figures
Before Start - Hot side = 30.40 degree
Before Start - Cold Side = 30.50 degree (it should same might be, I just wait a little bit to match both)
After 12 Minutes - Hot side = 37.50 degree (almost 7 degree up)
After 12 Minutes - Cold Side = 25.70 degree (almost 5 degree down)
So it makes 12 degree heat difference at few inches gap.

Hot side temperature before start = 30.40 degree

Cold side temperature before start = 30.50 degree

Hot side after 12 minutes = 37.50 degree

Cold side after 12 minutes = 25.70

Water drops on cold side pipe after 12 minutes

This was just open area experiment and results can be improved by:
1. Proper separation of hot and cold sides heatsinks (space)
2. Using powerful (multiple) fan for hot side heatsink
3. Much better result if use water cooling radiator to get rid off heat
4. Proper Insulation of cold side heatsink area so cold remains in that area (as mentioned in above comments)
5. Adding fan for cold side heatsink to throw out cold from heatsink
6. Using Peltier Module foam insulators
7. And finally adding multiple Peltier modules, I will use 4 modules like in above mentioned product.

Mini Air Conditioner Target
My target is not to cool down temperature of whole room, no, I know this can't be done with Peltier module a/c. My primary target is to make one mini a/c that I can put on my office table or side table to get cool air (30 degree is enough for 2-4 square feet area). And in same way at night if I put this mini a/c at side table then I can get cool air flow at my bed. So this is not very high or unachievable target.

4 Peltier modules device can (as in above mentioned product) collect enough cold air in one insulated cold air chamber and if one medium size fan is attached at front of that cold air chamber then that fan can throw cold air at 2-3 fee distance. I am sure this can be done. :)

Now I am working on this, will buy proper heatsinks, fan and more Peltier modules to make all this and will update. Please see pictures of my recent experiment.

Please post if you have more idea on this.

The ones we used had built-in to the heat accumulating a plate to which water tubes (coming and going) could be attached to conveniently remove the heat. Although, these were expensive when bought from a lab supplier, you might be able to fabricate one yourself and stick it on the hot side with thermal paste. I have also seen normal AC units that use running water to remove the heat.
You are talking about this product and I am going to make this exactly at home.

Click to see 4 Peltier Modules Water Cooling Kit - Pet Area Cooling - Mini Air Conditioner.


I will reply to other friends later, I need to leave now. :smile:

4 Peltier modules device...
I can understand why do you fancy Peltiers, but for the log I would like to mention that ~ for the same price and consumption of those four Peltier modules (and the PSU) you can get four 40cm standing fan, and it is really hard to compare the efficiency...

Ps.: Peltiers would be more efficient if used as a dehumidifier.
If I'm reading the data sheet for your Peltier modules correctly, they are capable of transporting something like 50 Watts (best case). The heat produced by your body will require (approx) 2 modules, just to maintain equilibrium. all of the power supplied to your cold side fan(s) is (ultimately) heat added to the room. If you have a computer and a light in the room, your other 2 modules are consumed. If the power supply for the peltier modules is in there with you... Enjoy the project, but don't expect to outperform an open window for cooling your room.
@Rive Peltier module device will not only supply air but that will be cold/chilled air. This is the question for how long distance this cold air can be thrown by this device. So there is difference between standard fan and Peltier based device.

@Dullard When I started this project then I don't knew that there are already ready to used devices (same kind) available in Chinese market like at above mentioned link. After starting this project and posting here then I found that these kind of devices are available and now I am confident that I can make the same device using multiple Peltier modules (definitely one module will not work here). And second you misunderstand, I am not trying to cool down whole room like split or window a/c does. I am trying to throw cold air at 3-4 feet distance using this device and I am sure this device can do this.

These devices already available in market (see at above link and you will find more and more powerful devices using 8 Peltier modules) and work fine. These devices are used to cool down small Pet areas, electronic cabinets and many other applications.

As long as heat is concerned now I am using Water Radiator method to get rid off heat. By this method all heat will store in water. You can google and see this method or above linked product also use same method.

So cooling down temperature of whole room or getting cold air just for 3-4 feet distance requires different resources. I am sure, this can be done.

@russ_watters Thanks for your help in understanding thermodynamics stuff. You are right, if heat is not removed from hot side perfectly then cold side will not give required results and today I practically did this and you can see my previous post with pictures and results. I will try to get data as you asked me.

I have some idea/plan in my mind and to implement this idea today I bought another Peltier TEC1-12706 module, one water cool radiator (80mm for 2 Peltier modules) and one superb heatsink for cold side which will extract/store cold from cold side. This heatsink for cold side is excellent having tunnel for airflow. I will put fans at back and front sides and I am sure I will get cold air from front side. Hopefully it will take 5-10 minutes to chill this heatsink when both Peltier modules are working.

Right now I will use 2 Peltier 12706 modules + Water Cool Radiator for better results to transfer heat. And this method will not throw heat in air (I mean at same place) instead Radiator will store heat in water.

I will assemble all these and post results here. Right now both Peltier modules will be on top of heatsink i.e. at same place. I just put Radiator and modules on heatsink and took picture. Don't assume I will fix these in same way. :)


My next plan is to use 3 Peltier modules on this heatsink for cold side. I mean 1 at left side, 1 at right side and 1 at top side so this heatsink get cold from all 3 sides and when air flows from this chilled heatsink tunnel that will be cold enough. Proper insulation of this cold heatsink is necessary so cold is not wasted in air and remains in heatsink. I am thinking to use same foam that used for split a/c pipe cooling that's attached with outdoor unit. That foam type stuff keeps cooling pipe cool.

Please guide me what kind of material to use to insulate this heatsink to avoid cold leakage?

In coming months my office room temperature will be 42-43 degree (inside room, outside it should be 50) and lot of humidity and I can't sit. So if I get 30 degree cold air from this device at my office table that's enough. So please don't assume I am trying to make mini air conditioner to cool down whole room like split or window a/c, I know that's not possible with Peltier modules.

Let's see what result I get with my latest parts.

Got it. You can certainly achieve the temps that you cite (at the heatsink). The rub is how much air you can actually cool. My rough estimate (for ideal performance with 4 modules) is < 20 LPM of air flow.
Peltier module device will not only supply air but that will be cold/chilled air.
Air movement and (with regards of the supposedly minuscule amount of cold air) exposed skin surface has a definitive effect on the subjective thermal comfort. You will just feel far better with a decent fan for the same price/power (in usual circumstances).

Again, this is just for the log. Your project is not an ideal solution for fighting high temperature, but it is still valid.
You've made some good progress, but I'm in agreement with others that this is unlikely to provide much localized cooling at all. And that will be offset by the extra heat being generated in the room, which will raise the incoming air by some amount. Not just the heat the Peltier has moved, but the heat from efficiency of Peltier, and heat from the power supply. I've experimented with Peltiers, and they are fascinating, and tempting, but overall pretty weak. They are excellent for some of the small scale uses that were mentioned in this thread, where a compressor is just not practical. But cooling the air around a person is, I think, asking too much of these.

May I suggest an alternate approach? Freeze water in some smallish (1/2L?) metal containers (do an image search on "beer in aluminum screw cans" for examples, and leave some room for expansion). Place them in a plastic bucket, and set up a fan to blow air through that, and onto you. Have a piece of plastic with spacers set a few inches above the bottom of the plastic bucket to form a shelf to set the frozen cans on. It should fit fairly tight, but allow the condensation to drain down. This will help provide some de-humidification, as the air won't flow over the collected water. Mulitple smallish cans will provide more surface area for better heat transfer, and more care to collect condensation.

Do some math on the amount of energy a few Liters of ice going through a phase change can absorb, versus what those Peltiers can move. I expect it will be a large difference. Plus, the cold cans will cool and de-humidify the room, instead of heating it up (I assume your freezer is not in the same room? And your freezer will be more efficient than any Peltier anyhow, so less heat gain regardless). It is cheap and easy. Have enough cans so you can swap them out, freezing one set while the other is in use.

I actually did some calculations on this a few years back, and it seemed like it would work well as a personal cooler. You can find examples on youtube, mostly using a cooler with ice, but I like the idea of cans (re-freezable), and a 'water trap' to reduce humidity.

Good luck!
You can also consider getting some 'ice gel packs' and make a cooling vest of some kind.
These packs are often used to keep sensitive medical materials (like vaccines and so) cold during delivery, and they are often discarded. Maybe worth asking about it in the local pharmacy?

Actually for the price of a Peltier system you can get a ready made cooling vest too...
@NTL2009 & Rive, you both gave me new idea i.e. using cold drink cans and ice gel packs, I will work on these but these ideas requires adding cold material again and again. I will give this try to see what happens. Freezer is not in my room and if you are concern about device power supply heat then I can keep power supply 5 feet away from device. I mean radiator water bucket (which will store heat from Peltier device) and power supply can be at same place i.e. 5 feet away.

I already have some different idea to use cold drink can in my existing Peltier device idea. I will also give that idea a try.

Latest experiment with 2 Peltiers on Top of heatsink
Actually I already did experiment with 2 Peltiers on top of heatsink with Water Radiator cooling method and in just 15 minutes heatsink temperature dropped from 29 to 19 degree and water temperature was up from 31 to 41. Then I put two fans at back and front of heatsink and I started getting cold air. Heatsink was cold down from all 3 sides i.e. top, left and right. 10 degree drop in 15 minutes with 2 Peltiers is good progress.

Unfortunately water pump stopped working after 15 minutes (this was very old windscreen pump of my car. I think this kind of pump is not made to keep running for long time i.e. continuously 15 minutes) and I have to shutdown device otherwise I keep device running for at least 30 minutes to see what happens. I also insulated heatsink with simple cloth to stop cold leakage. This was an open environment experiment and definitely when I have good progress then I will design proper chamber for this purpose and properly insulate heatsink to stop cold leakage.

I visited freezer and air conditioner repairing shop to see split air conditioner cold pipe (attached to outdoor unit) insulation material and I found that it was same material/foam that's used to lay down under carpet. Same kind of foam. So I will get that sheet to properly insulate heatsink.

I bought Water Cooling Radiator for 2 Peltiers (80mm), that was mistake however I should buy 2 radiators of 40mm each so I can attach both Peltiers on 2 different sides of heatsink for better results so cold cold spread in heatsink from 2 sides.

Now I have planned to arrange 3 Peltiers + 3 Radiator + new water pump + insulation sheet for final experiment. If this works (which I believe will work) then I will start work on proper body for this mini air conditioner.

I already have rough idea to use metal body like APC 1000VA UPS. That kind of body is good for this device with 4" x 4" fan at front to throw cold air at 2-4 feet distance. (I already have this fan and this fan throws air very at very good speed). Three chambers in that body. Back chamber is for power supply and any other electronic kit, etc. Middle chamber is for cold generation unit (heatsink, etc.) which will move cold air into front chamber. And front chamber with insulated steel box will receive cold air from cold generation unit and throw outside. I will use insulated steel box in front chamber so when cold air will pass on from this place then that steel box will absorb/store cold (I mention at start that I have an idea to use cold drink can in different way) before cold air is thrown out of box. So as long as device is running 20 degree cold air is flowing from front cold box that box will keep cold and help to throw enough cold air at 2-4 feet distance. Its just my idea, don't know it works or no.

So this kind of idea but all depends on my final experiment with 3 Peltiers on all 3 sides of heatsink with powerful fans at front and back. If 2 Peltiers on top of heatsink can drop heatsink temperature from 29 to 19 in just 15 minutes then 3 Peltiers on all 3 sides will give much better result.

Let's assume if heatsink temperature is dropped to 10-15 degree in just 15-20 minutes and when air will flow through this heatsink, will that not give enough cold air? No matter if cold air flowing out from front side is 25 degree cold (temperature drop during flow due to any reason), its sufficient.

Heatsink is very good and I will looking for this kind of heatsink that absorbs cold from all side with cross plates so air struck with these plates to get cold. Heatsink outside wall width is 1cm and inside plates width is 1/2cm. Excellent piece to absorb cold even device is shutdown for few minutes.

19 degree temperature of heatsink in 15 minutes with 2 Peltiers I achieved yesterday.


I still believe this will work.

Brave work.
You may find you must spiral-wrap that flexible tubing to prevent even slight kinkings, lest they throttle the inherently limited performance.

Some of the HPLC equipment I used had a Peltier column oven module. IIRC, each drew almost as much power as the rest of its stack combined, blew a LOT of waste heat out the back to keep but a foot of slim steel pipe cosy...

When I peeked under lid, the circuitry resembled a heavy-duty H-bridge motor controller, albeit with a non-inductive load...

I've one of those in-car heater/coolers, too, with the house-power option for pre-cooling. When our kitchen's big fridge died, the Peltier was fetched from box-room. Running full-out, it kept kitchen's milk, spread and other to-hand consumables just cool enough for a few days' shelf-life. When became apparent that the service contract repair on fridge was going to take rather longer than hoped, I rushed out and got a 'worktop' larder fridge...

Want to reply to this thread?

"Peltier based mini air conditioner not working" You must log in or register to reply here.

Physics Forums Values

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving