# Designing a peltier cooling system

#### vrasp

Joined Feb 1, 2020
5
Hi all,

First of all, I m not sure where this thread should go. My apologies if this is the wrong section.

I'm currently studying solvent crystallization in materials and would like to create a discrete cooling system that I can place under a microscope to observe crystallization. To be more specific, I need to create a temperature gradient with freezing temperatures (ideally down to -30°C) on one side and room temperature on the other. After some research in the literature, I found exactly what I need:

https://hal.archives-ouvertes.fr/hal-01902742/document

Now, what they built is rather complex. Fortunately, I don't need the whole thing. The very first step and most important one is to build a system with two peltiers that are temperature controlled, connected with a temperature sensor to place on the material to be cooled.

I know the very basics of electronics and enjoy tinkering with stuff. My hopes was that you guys could help me answer some questions and guide me through the process.

I m gonna get the same peltier modules they got. The peltier modules are 3.9 A and 15.7 Vdc. The temperature controller they use is $1000. It's a TEC-1122 Dual Thermo Electric Cooling Temperature Controller. I m not sure why they got such an expensive device. My guess is that this is for lab use and it's super accurate which usually equals super expensive. This is not a requirement in my case. Can I use a more reasonably priced temperature controller that has an output around 15Vdc ? I was looking at the inkbird ITC-106RL which is around$25. I would need two of them to control the two peltier modules.

I have more questions but I think this is a good start.

Thank you!

Vrasp

#### MrChips

Joined Oct 2, 2009
27,695
The efficiency of a Peltier TEC is a function of the temperature difference between the cold side and the hot side of each TEC.
Cooling down to -30°C from room temperature will be quite a challenge with limited financial resources.

In order to achieve this goal you need to define the thermal mass of your object. Then you need to do it in stages and with the appropriate amount of TEC power plus a method of removing the thermal energy.

Finally, you need a temperature controller in order to regulate the target temperature. This is accomplished with a PID temperature controller.

To summarize, what you need are the following:

1) Knowledge of thermal mass of object
2) Multiple stages, two or three stages of TEC.
3) High power TEC
4) Water cooling
5) PID temperature controller

If not done properly you run the risk of burning out the TEC and not being able to reach your target temperature.

#### vrasp

Joined Feb 1, 2020
5
Thanks the the detailed answer.

My samples would be mainly made up of 95%-99% water. The mass would be no more than a few grams. I would say 5 g max.

Water: 4.184 J/g C°. Therefore if I want to cool 5 g of water from 20°C to -30°C I would need to transfer 4.184*5*50 = 1046 joules from that material.

Our lab has plenty of water-baths. I can hook one up to a heatsink.

To be honest, making a system with a temperature gradient from -10°C to 20°C across the Peltiers will be a good start. That can be my proof of concept and I could probably get some funding for making something more powerful.

Instead of a temperature controller, can I make this easier by getting an adjustable power supply and finding the amount of current needed to get to the desired temperature? It's ok if there are temperature fluctuations at first as long as once I get my sample to the desired temp it stays constant and small fluctuations of a few 0.1°C are acceptable.

Last edited:

#### MrChips

Joined Oct 2, 2009
27,695
If you are willing to take the time to experiment, you can do this on a tight budget.

Start off with a relatively low cost TEC1-12706 12V 6A TEC that is readily available.

Attach a water cooled heat sink used for cooling the CPU in PCs.

Use water flowing straight from the tap. You don't need high water flow or pressure.

Shop around for a kit of components.

An adjustable bench power supply is not enough. You will need to regulate the target temperature. You can use a simple ON/OFF temperature controller. A simple PI control is easy to implement on an Arduino or similar MCU platform.

As an alternative, you can re-purpose a TEC cooling system from a picnic cooler. It wouldn't get you to -10°C.

#### vrasp

Joined Feb 1, 2020
5
Thank you for your help.

I ordered a cheap temperature controller, a heat sink, Peltier module, a 12v 6A power supply, and a thermal pad to put between the Peltier and the heat sink. Once I get it, I will put it all together with one of our waterbaths in the lab that goes to -15°C and see what is the lowest temp I can get to.

#### MrChips

Joined Oct 2, 2009
27,695
Thank you for your help.

I ordered a cheap temperature controller, a heat sink, Peltier module, a 12v 6A power supply, and a thermal pad to put between the Peltier and the heat sink. Once I get it, I will put it all together with one of our waterbaths in the lab that goes to -15°C and see what is the lowest temp I can get to.
Good. I would love to hear your results.

I forgot to add that you should use silicone thermal conducting paste or heat sink thermal pad.
https://shop.sayal.com/products/qpn-1011

#### jrb_sland

Joined Dec 24, 2021
10
Good. I would love to hear your results.

I forgot to add that you should use silicone thermal conducting paste or heat sink thermal pad.
https://shop.sayal.com/products/qpn-1011
Thanks the the detailed answer.

My samples would be mainly made up of 95%-99% water. The mass would be no more than a few grams. I would say 5 g max.

Water: 4.184 J/g C°. Therefore if I want to cool 5 g of water from 20°C to -30°C I would need to transfer 4.184*5*50 = 1046 joules from that material.

Our lab has plenty of water-baths. I can hook one up to a heatsink.

To be honest, making a system with a temperature gradient from -10°C to 20°C across the Peltiers will be a good start. That can be my proof of concept and I could probably get some funding for making something more powerful.

Instead of a temperature controller, can I make this easier by getting an adjustable power supply and finding the amount of current needed to get to the desired temperature? It's ok if there are temperature fluctuations at first as long as once I get my sample to the desired temp it stays constant and small fluctuations of a few 0.1°C are acceptable.
As others have said, you'll need both water cooling & a controller. I built some precision coolers for creating "ocean-bottom" 2 degC temperatures by superinsulating {with 2" polystyrene slabs} an inexpensive little beer cooler { 8 liter nexXtech 6118241, no longer available?}, adding a 1/4" copper tubing loop drilled through the cooler's air-cooled heatsink {discard the cooling fan}, hooking up a low-flow water feed from a siphon tube installed in the tank of the toilet in the bathroom above my basement workshop supplying fairly constant +12 C water, and adding a Peltier-specific controller made by Oven Industries Inc. https://www.ovenind.com/product/5r7-350/ Lots of other details, but you'll benefit most by figuring it out yourself. Beware Oven's non-standard glgass-bead 15k thermistors - I had two fail after only a few months in 2007/8, and I substituted two more easily available 30k thermistors wired in parallel. To get higher resolution for my fixed target temperature, I changed some of the fixed resistors on the circuit board to make the control range much smaller. Oven Industries' documentation is excellent for setting up the control PI. I used a thermally conductive RTV to improve the thermal connection of the copper tube to the fins, and to glue the heatsink of the controller atop the finned heatsink of the chiller unit.
BTW, maximum thermal efficiency of Peltier devices occurs when delta-T is minimized, and when the raw DC supply voltage to the controller unit is only slightly above that needed to maintain your temperature. I used a Texas Instruments PTN78060WAH to make about 10.8 volts from my main power supply's 13.8 volts {which keeps my lead-acid backup batteries fully charged - I was calibrating/tweaking the rate of real-time clocks by monitoring them for at least a month, & uninterrupted thermal stability was important}.
My results - I used about 11 watts total to keep the beer cooler's interior at 2 C +/- 0.1 C while the DUT circuit boards inside the box were dissipating ~ 250 mW total, with a modest {< 1 gallon/minute} flow of 12C water. Your application with stacked Peltiers will need more power, of course, and likely more water flow, although if your "water" bath can get down to -15 C to start with, that's better than I can do. I've attached some photos of my chillers during construction in early 2007. Enjoy !

#### Attachments

• 747.6 KB Views: 31

#### Bordodynov

Joined May 20, 2015
3,025
This information may help you (if you have LTspice). You have to use two Peltier plates connected in series for heat dissipation and in parallel for power.

#### Attachments

• 182.9 KB Views: 22
Last edited:

#### conciseusa

Joined Dec 13, 2019
2
Not sure if this of use, but I am in the middle of getting a proof of concept design working for temperature control with an Arduino Uno. Just as big of a concern for me was monitoring the temperature over time, and the duty cycle of the elememts so I have an idea of where I am at in terms of how hard the system is working to hold the set point. I used an Arduino, a shield I designed that adds a real time clock, signal processing and glue logic move data around and hook up a LCD display. The Uno pushes out JSON packages with the data, and a Python program processes them and can push them to the cloud for remote monitoring, or put in a CSV file for viewing and charting in a spreadsheet progrom. If this is of use, let me know and I will get what I have posted.

#### Attachments

• 1.3 MB Views: 19

#### vrasp

Joined Feb 1, 2020
5
Hey guys,

Thanks for your help!

I put everything together and checked how low I could get. With the waterbath set at 2°C I was able to get to -25°C (first picture). I tried to go even lower by setting the waterbath to -5°C. It eventually got to -29.9°C (second picture)! The waterbath temp could not go lower than -4°C, probably due to the amount of heat generated by the Peltier.

I m quite happy with this. Aside from the waterbath, this whole setup cost less than \$100. I wonder how low I could go with a better Peltier.

#### MrChips

Joined Oct 2, 2009
27,695
Again, it is not about a better TEC. It is more about the temperature differential and how effectively you can remove the heat from the hot side. That is why a 2-stage setup is better than one. Sure, a higher power TEC is warranted but the previous statement still applies.

Congratulations on your successful attempt the first time!