Stacking 5 Peltier Modules, and Reached -90℃

Hello, today I would like to explore the limits of cooling using Peltier devices. In the previous video about Peltier devices, I generated power using them. This was a very interesting experiment, but Peltier devices are not originally used for generating electricity. The power generation efficiency was only around 0.6%. This time, I will use Peltier devices for cooling, which is the main purpose of Peltier devices. I'm a beginner of Peltier devices, but let's see how much temperature I can decrease. The goal is minus 60 degrees Celsius. Let's start the experiment. A Peltier device has an interesting characteristic. When an electric current is passed through a Peltier device, one side cools down and the other side generates heat. Now, I will pass current through the Peltier device. This side is getting cold, and the other side is getting hot. It actually feels quite cool to the touch. It's around 16 degrees Celsius. This characteristic is due to the Peltier effect. A Peltier device contains P-type and N-type semiconductors. The semiconductors are connected with metal to compose a Peltier device. The metal on this side is deprived of heat, which is called endothermic effect, and the heat will escape to the metal on the other side. So this side gets cold, and the other side gets hot. A Peltier device creates temperature gap. This characteristic of Peltier devices is used for portable refrigerators and semiconductor cooling. There's one thing you need to be careful about when using Peltier devices. It's that Peltier devices generate heat themselves. To operate a Peltier device, a current must be passed through it. That power is calculated by P equals to Vi. The Peltier device generates heat equivalent to the power. So using Peltier devices for cooling requires cooling of the Peltier devices themselves. I started passing current through the Peltier device. The temperature is getting lower. Let's see what happens to the temperature over time. Gradually, the minimum temperature is rising. It's 16. Now it's 17 degrees. Let's look at the other side. Here, too, the temperature is rapidly rising. It's now higher than 65. A Peltier device creates a temperature difference, but because it generates heat itself, even the side that is supposed to be cold gets hotter. The temperature of the cold side has become higher than the room temperature. For today's experiment, the Peltier devices are stacked to make three stages. The one placed at the bottom has the largest surface area, then this one is for the middle stage, and then this one is put on the top. The top side of the top one is the cold side. I'll explain why the Peltier devices are stacked. The simplest way is to use a single stage, but there's a reason for using multi-stage Peltier devices. It's because it cools well that way. This is the bottom Peltier device. Let's say its hot side is the room temperature. When you pass a current, the cold side cools down. Then place the middle Peltier device on top. This allows you to apply a temperature difference to the middle Peltier device using the temperature of the cold side of the bottom Peltier device. This means the cold side of the middle device is cooled even further compared to when using only one Peltier device. If you put another Peltier device on top of these, the cold side of the top device will get even colder. This is the advantage of configuring multiple stages. Today I'll use three stages. The important point in using Peltier devices is how much you can cool the hot side. Let's use this diagram to think about it. Power P1 is applied to the top Peltier device. It absorbs heat Q1, then heat Q2 is released from the hot side. Q2 is the sum of the absorbed heat Q1 and the input power to the Peltier device P1. The heat generated by the top Peltier is absorbed by the middle Peltier. In order to transfer this heat from the middle Peltier, you need to input power P2. Then the heat Q3 is transferred to the bottom Peltier. In the same way, you need to release the heat Q4 in the bottom Peltier by applying power. The heat that needs to be released from the bottom Peltier is the sum of the heat that was initially absorbed at the top and the power used for all the Peltier devices. Top, middle and bottom. All of them. It is necessary to dissipate the enormous amount of self-generated heat of the three Peltier devices. Therefore, today I prepared this giant heat sink. I'm going to use it to cool Peltier devices. This is used for cooling industrial power semiconductors. The datasheet shows how much heat is transferred when how much power is input to the Peltier device. It is possible to calculate how much the temperature can be reduced theoretically using the data on the datasheet. However, in reality, there are unexpected flows of heat. The actual amount of temperature reduction is determined by the balance of heat and power when these three Peltier devices are stacked. Now I'm going to place the Peltier device on the heat sink. But in this state, the heat won't be transferred well because the adhesion is poor. So I'll use thermal grease. I chose these two thermal greases. One is a cheap thermal grease used for cooling of power semiconductors, YG6260. The thermal conductivity is 0.84 Watt per meter Kelvin. The other is a fairly high performance thermal grease, TC31G. The thermal conductivity is 8.5 Watt per meter Kelvin, which is 10 times more efficient than the other one. However, the minimum operating temperature is minus 40 degrees for both thermal greases. So I can't say for sure if they can maintain their thermal conductivity when the temperature is below minus 40. I did a quick preliminary experiment and found that TC31G decreases more temperature, so I'll use this one. By the way, liquid metal grease has a high thermal conductivity, but I'm not going to use it today because it freezes at low temperatures. Let's start the experiment. No current is flowing now, so it's not cold yet. First I'll try with about 20% of the rated current. I'll start passing the current. It's getting colder. It's cold. It's getting pretty cold. Let's measure the temperature. Current is being passed through the Peltier devices. It became cold as you can see. The temperature of the top surface is about minus 5 degrees Celsius. And it's getting even colder. I'll leave it for a while. It's reached minus 30 degrees. It was quite fast to reach minus 30 degrees. There are temperature differences between the stacked Peltier devices. The temperature of the aluminum heat sink at the bottom is room temperature. The temperature gets lower in the order from the bottom Peltier device to the top. There are temperature differences between each Peltier device. The temperature of the cold side of the bottom device is the highest. The temperature of the middle device's cold side is lower than that. The temperature of the cold side of the top device is the lowest. I just found out that this thermal camera cannot measure temperature below minus 30 degrees Celsius So I'll use a thermocouple instead. I decided to fix the thermocouple with water droplets It freezes automatically when the Peltier gets cold. It's perfectly fixed The thermocouple says the temperature is minus 55 degrees now This is a temperature that gets your finger frostbitten if you touch the Peltier. It's extremely cold The natural air cooling of the heatsink is far from enough So I attached a powerful fan here and we'll start forced air cooling. The fan started running. It's so noisy It cools the heatsink very effectively I'm going to change the current to apply to the Peltier devices to find the best operating point for maximum cooling Before we start there is one more thing to do. In this state the cold side of the top device cannot be cooled at best efficiency It's because air convection is occurring around the Peltier devices. To get the best results You need to prevent the air convection around the Peltier devices and insulate them to stop the heat transfer The most ideal environment is to create a vacuum only around the Peltier devices This is technically possible, but it's extremely difficult when you actually try to carry it out. So I'll let go of the idea Instead I'm going to use this insulation material. This is styrofoam It's typically used for home insulation. Its thermal conductivity is 0.036 watt per meter Kelvin or less It's quite low. At first I considered using these polystyrene foam and urethane foam as well But considering availability, workability and stability, I concluded that styrofoam is the best. Now, I'm making an insulated case out of styrofoam I need to leave some space between the styrofoam and the Peltier devices so they don't come into contact It's because I think it's better to put some air insulation between them. The thermal conductivity of air is 0.024 watt per meter Kelvin and that of styrofoam is 0.036 watt per meter Kelvin. Air transfer is about 30% less heat than styrofoam However, even in an insulated space, air convection and thermal gradients may occur So I'll place another styrofoam to prevent air convection. I made it so it minimizes the contact with the Peltier devices I haven't checked whether it's better to leave some space or cover completely Now current is applied to the insulated Peltier. The temperature has dropped quite a lot. It's minus 52 degrees Celsius Currently the top Peltier device is not running. Now, I'll pass current to the top device The temperature is dropping rapidly. It's dropping so quick It became minus 60 now. As you can see, insulation is very effective Now we can start searching for the optimal operating point of the Peltier devices that makes them the coldest. Today I'm using three Peltier devices. So there are three adjustable currents I'm going to change the currents and measure the temperature each time so I can find the best Combination of currents that obtains the lowest temperature. As I said earlier, you can calculate the temperature theoretically But in reality, there are many paths through which heat escapes I performed theoretical calculation today, but it was only useful as a reference value Heat experiments are quite time-consuming Once you change the current settings, it takes about two to three minutes for the temperature to equilibrate There are about 150 current combinations that are expected to be effective and I tested them all. It was a very time-consuming experiment These are the current combinations I tried. After that I tried finer adjustments around combinations that got good results to achieve the lowest temperature Now, I'm going to start experimenting with the current combination that achieved the lowest temperature I just started passing current through the top Peltier device. The starting temperature is about minus 54 degrees It's now minus 60. The power consumption of each Peltier device is 100 watts for the bottom 24 watts for the middle and 5 watts for the top. The temperature is decreasing rapidly. It's reached minus 70 degrees Now the temperature is decreasing less rapidly Gradually, the temperature is equilibrating. The temperature is minus 74 degrees. This is the lowest temperature it can get So I was able to achieve a temperature lower than minus 70 degrees Celsius I'm a beginner when it comes to heat, but I tried my best and got a certain level of result By the way the current room temperature is 26 degrees. If it's lower to about 20 degrees, the temperature of the Peltier device drops by 2 to 3 degrees. If the temperature of the air used for cooling is low, the temperature of the Peltier devices will also be lower accordingly I guess the temperature would reach around minus 80 degrees if you perform this experiment in winter Now that I can perform low temperature experiments, I'll try using it right away First, I'll cool a 1 kilo ohm resistor. Resistors have a temperature coefficient which means the resistance rises and falls following the rises and falls of the temperature. The resistance at room temperature is 991 ohms. I'll use four terminal method. The thermocouple is attached in the middle. Now, let's start The temperature is rapidly decreasing. The resistance is increasing. The temperature has just dropped below zero degrees Celsius The resistance value was initially 991 ohms, but has increased to 995 996 ohms. Therefore we can tell that this resistor has a negative temperature coefficient Let's have a look at the graphs. The top one shows the temperature and the bottom one shows the resistance. As the temperature decreases, the resistance increases. It's great that I can easily perform low temperature experiments now The temperature has almost reached the minimum. The resistance has become larger by 17 ohms than in room temperature. Quite a big increase Next, let's cool down an electrolytic capacitor. When the electrolyte inside is frozen, the capacitance decreases I want to try lowering the temperature enough to freeze the electrolyte The capacitance at room temperature is 53 microfarads. Let's start. The temperature has dropped below zero What about the capacitance? It seems decreasing little by little. Because the capacitor is big, it may take a while to freeze the electrolyte I'll fast-forward the video The temperature fell below minus 55. Minus 55 degrees is the lowest temperature where this electrolytic capacitor works It's reached minus 60. The capacitance is rapidly decreasing. Let's leave it for a while It's minus 70. If the temperature doesn't decrease further, the capacitance won't decrease. It's hard to decrease anymore The temperature of dry ice is minus 79. At that temperature, the capacitance decreases a lot But minus 70 seems not enough. The capacitance has reached 48 microfarads The capacitance won't decrease anymore if the temperature doesn't decrease anymore However, the operating waveform looks totally different from when it's at room temperature This is the waveform of the electrolytic capacitor at room temperature And this is the waveform when it's at minus 70 degrees. The angles at the rising and falling parts are almost perpendicular So it's obvious some kinds of changes are occurring inside. Today, I performed experiments to see the minimum temperature I can achieve using Peltier devices. As a result, by stacking the Peltier devices in three stages I was able to get it down to minus 74 degrees Celsius I'll use this module when I perform low temperature experiments in the future However, from what I have seen in my experiments, it is not possible to cool large components It can be used to cool components the size of a fingertip By the way, I prepared a reference model for this experiment. A Peltier device that is already stacked in three stages This module costs about $200. The three-stage module I made today cost about $85 Ready-made modules that are already stacked are more than twice as expensive However, this ready-made module was easier to use and it cools efficiently The temperature drops to about minus 70 degrees The main difference between ready-made and handmade modules is the structure and power consumption My handmade module consists of, from the top, ceramic, semiconductor, ceramic, thermal grease, ceramic, Semiconductor, and so on. On the other hand, the ready-made module consists of ceramic, semiconductor, ceramic, Semiconductor, and so on, which uses less thermal grease and ceramic. This makes the ready-made modules thermal resistance much smaller The design is also optimized so it reaches low temperatures efficiently with low power consumption. That's the difference I tried combining the ready-made device and the bottom and middle stages of my handmade device To my surprise, the temperature went down to minus 92 degrees Celsius. That was unexpected. This is pretty amazing There are probably no videos on YouTube that brings the temperatures down to minus 92 or 93 degrees Celsius using Peltier devices Maybe this is a new record on YouTube However, although it does achieve very low temperatures I don't think it's very practical in actual experiments because when you try to cool large components I don't think it reaches minus 90 degrees, around minus 80 at best. If this video was helpful Please comment, like, and subscribe to the channel. Thank you for watching