Building a Peltier cooler-based cooling box - Second iteration [Part 3/6]
Вставка
- Опубліковано 25 січ 2025
- In this video I continue my experiment with a cooling box, or refrigerator. This cooling box uses two TEC12710 Peltier coolers as well as 2 Arctic Cooling Freezer 7 Pro CPU coolers for cooling the hot side. I had to apply some custom solutions for CPU the cooler because the size of the hot plate is only 3x3 cm while the size of the Peltier is 4x4 cm.
I tried several ways in this video. I changed the fans on the cold side heat sink and I also clad the inside of the box with some aluminium foil. The results are promising, I almost reached 4°C.
Join my channel membership: / @curiousscientist
Previous videos:
Part 1: • Building a Peltier coo...
Part 2: • Building a Peltier coo...
4 channel thermometer video: • 4 channel temperature ...
You can find the chart on my website: curiousscienti...
If you want to support my work, please consider buying the parts using my affiliate links:
curiousscienti...
![Assembling the dual TEC12710 cooler [Part 4/6]](http://i.ytimg.com/vi/rfbnSMU9vjM/mqdefault.jpg)
![Assembling the dual TEC12710 cooler [Part 4/6]](/img/tr.png)
![Peltier based cooling box - Refrigeration test [Part 6/6]](http://i.ytimg.com/vi/0lzlDSDp_sU/mqdefault.jpg)
Thank you for posting so much detail about these Peltier devices. It's been a great resource for helping me figure out how to build a micro-fridge to chill a single can of beer using the 36V battery on my solar powered ebike (my channel has videos if you're interested).
The goal is to keep the device as small and lightweight as possible for touring. Battery voltage is 33-42V and can deliver up to 30A continuously so I'm thinking I'll wire 3-6 Peltiers in series to avoid the weight of a DC-DC converter. I'm machining custom aluminum heatsinks to tightly couple the cold side directly to the can. I've ordered some small 40x40mm heatsinks with matching 40mm fans for the hot side. They weigh only 44g each. But now I'm starting to think that these will be the limiting factor in my design as they won't be able to move enough heat. I'm just guessing based on your results and the size of my heatsinks.
The "roughly 2kg" of heatsinks you're using here are definitely too heavy for my application. Any thoughts on how to get the most cooling from the lowest possible weight?
Ideally, I'd like to turn a warm beer (25-35°C) into a cold one (4°C) in less than one hour. As a stretch goal, it would be amazing if I could freeze a water bottle so I could keep food that doesn't fit in the "fridge" cold.
I realize Peltier devices are inherently inefficient so that will be a consideration as well in deciding if such a device is worth bringing on a long distance trip. Every watt-hour consumed is less distance I can travel in a day.
I've got parts on order and will start experimenting in the next few days. Cheers! -Mark
Hi! I am glad that you found my videos useful. Unfortunately, you haven't mentioned which Peltier you are going to use, so it will be extremely difficult to say anything. But in general, you will need to dissipate around 100 W per device (a typical Peltier works in the range of 8-15 V (optimal conditions) and the current depends on the device (anything from 3 to 15 A)). So that 40 x 40 mm heat sink will be most probably useless. Look up these "COB LED heatsinks" on Google, they are much more suitable for cooling the hot side of these thermoelectric devices and you can mount multiple devices on them. Regarding the required power, you can calculate it easily. You just defined the required time (1 h) and temperature difference (~30°C), and you want to cool water or beer, so you know the specific heat too. It is easy to calculate the required power to achieve this cooling. Freezing the water will be extremely energy demanding due to the released latent heat that you also need to remove during the freezing process. I discuss and demonstrate it in my freezing experiments.
> Unfortunately, you haven't mentioned which Peltier you are going to use
Argh! I'm so annoyed with myself now. I answer a lot of questions in comments and emails about solar ebike conversions and the "not enough info" issue frustrates me as well.
I get some TEC1-12710 units, naively thinking that since power supply should not be an issue, I would have better chance of freezing water by using Peltiers that can handle more current. As I do more research, I'm starting to better understand the importance of the heatsink and why small heatsinks won't work. In one of your videos you increase current through the Peltier to try to increase cooling and get the opposite effect. Very instructive.
Thanks for the tip on the COB LED heatsinks. I'll look into it. To meet my poorly-defined "low weight" goal, I may need to rethink this a bit and maybe use only one Peltier with a properly-sized heatsink and size it to run on the 12V 5A DC buck converter I already have on the bike for accessories and lights. It probably won't freeze a water bottle in a reasonable amount of time but it might be enough to chill a single can. Even if it takes 2-3 hours, that would still be worthwhile.
I saw your energy calculations for cooling water and the additional energy required for the phase change and those seemed reasonable to me in terms of energy consumption compared with how much energy it takes to boil water using an immersion heater.
Hi! There is no way you can keep the TEC12710 at proper temperatures at large currents with the 40 mm x 40 mm heatsinks. I struggled to keep them cold with a large CPU cooler which is rated for >150 W TDP CPUs. Also, the energy consumption will be high. You can run them at ~7 A with sufficient cooling. At that point, they will produce 80-90 W waste heat plus you can _ideally_ move 30-50 W heat from the cold side to the hot side. So you will have to dissipate 110-140 W heat from the hot side. If you run them at 5 A, you should expect around 100 W heat on the hot side (~60 W Joule heat + 40 W heat from the cold side). (This is a just quick, back of the envelope calculation)
But hey, please check this video instead: ua-cam.com/video/trUrsUQDfOU/v-deo.html
Ok. I've rewatched a few of your Peltier videos and each time I understand a little better. I'll return the 40mm heatsinks and I've ordered a couple of larger ones to test out. One is a "100W" COB LED heatsink which I suspect will weight over 300g and the other is a 180g "slim" CPU cooler which claims "up to 75W TDP." I'll try them both at several lower current levels around the suggested 5A level to see if I can find the sweet spot of power consumption, cooling time and total device weight. Thank you for the great advice.
The TEC1-12710 units just arrived. I tested them out with 3 in series, applying only 3V with no heatsink just to verify that they weren't dead on arrival. So cool to instantly feel one side getting warm and the other side getting cold. I imagine the thrill of discovery Jean Charles Peltier must have felt in his lab in 1834 when he first observed the phenomenon that bears his name.
5:05 The copper plate is a good idea, as peltiers don't conduct heat well sideways through the peltier element. Every part of the peltier surface should be in contact to use the full capacity.
Thank you! Yes, this was my concern that the heat transfer sideways is poor. If the cooler was let's say 38 mm wide in both directions, I would not have cared about that missing 2 mm. But 5 mm is too much. It survived more than 4 hours of testing and the temperatures were very stable, so I am happy about this solution. I will publish a video in about 2-3 hours where I show how I assembled this cooling system. You will see how I sandwiched the things together.
Good results
Please use fan in apposite direction maybe it will get better results
Just rotate fan into apposite direction
Hi! Thank you!
I can try, but based on my previous experience, the push or pull configuration does not really change anything. I can test it, but I do not expect too much difference.
What do you think the actual hot side peltier temperature is? Do you think slower fan speed would help or just allow ice to build up? Do you think the aluminum touching the sensors showed colder temperature because of it's thermal conductivity?
The real hot side temperature can always be determined from the current and voltage which is running through the Peltier. This is one of the reason why I always note them down, so I can check the differences. If you check the assembly video of this particular dual TEC12710 cooler, I actually answered the question. So, I don't think, but I know that the hot side is around 60°C.
Slower fan speed would allow ice build up because the heat sink can go below 0°C. On the other hand it would also mean lower Qc because there is less heat exchange with the environment due to the reduced airflow. Since the heatsink goes below 0°C, DT increases which results in decreased Qc.
If you mean the sensor placed on the heatsink, then no. The system had 1,5 hours to pick up the temperature change. I think that there is still some heat leakage.
I will retest this box with a variable power supply, probably during the weekend. I think that I did a little overshooting by applying 12.17 V, so there was a tiny bit of unnecessary and avoidable Joule heating involved. I will go down to around 10.5-11 V and there should be some improvement.
@@CuriousScientist thank you. Is there no fan on the hot side heatsink?
Each CPU cooler has a 8 cm fan installed on it.
Very interesting. You could try to put One artic cooler on the cold side in order to try to increase the thermal Exchange. I also think that the speed of the fan on the cold side should not be too fast
Thanks! I am extremely sorry, but there's no way I disassemble this cooler again, haha. On the other hand, I will play with the fans, because I agree that they might be a bit too aggressive and that results in a worse heat exchange. I will see what I can do.
@@CuriousScientist A wonderful 3d printed duct to canalize air through the cold side heat sink giving the time to the air to refrigerate
Now we are one step closer! I can print a duct easily, I have a printer.
@@CuriousScientist Another idea Is to activate the fan on the cold side only when its temperature Is in a choosen range below 0 in order to create a greater difference between the heat sink temp and intake air flow
This is also a great idea, thanks! Actually, it is not a big deal to write a small Arduino code to turn the fans on and off with a relay, based on a temperature value. I will see if I can do it soon. I have a lot of things piling up.
I cannot see where hot heat sink (CPU Cooler) installed, i think install behind the cooling box close to the wall which will not good ventillation enough. Hot heat sink should be verticle installed because of heat pipe fuction. I hope your next experiment can down to 4°C.
It is on the back of the cooling box. Also, the heat-pipe does not care about the vertical/horizontal installations. CPU coolers are doing just fine installed sideways in your PC, for example.