Hello experts,

I am working on a project which observe human interaction with robots. From my part I need to keep the robot hands at a specific temperature. Therefore, I would like to use Peltier elements with TEC controller. Is there any other idea I can go with or I should go with this one, also what considerations I must take while going with this method?

 

Hello experts,

I am working on a project which observe human interaction with robots. From my part I need to keep the robot hands at a specific temperature. Therefore, I would like to use Peltier elements with TEC controller. Is there any other idea I can go with or I should go with this one, also what considerations I must take while going with this method?

Is the "specific temperature" fairly constant? I think a waterbath and a flow-through heat exchanger might be a better option. It'd be much easier to control the temperature of the water in the bath and circulation system.

But, it's possible to control temperature very precisely with a TEC if you must. I have a related completed project here you might want to look at.
https://forum.allaboutcircuits.com/threads/project-poor-mans-calorimeter.50045/

 

Welcome to AAC.

Is this a school project that will receive a grade?

 

Consider that a "Peltier" device is a heat pump and not a heat source, it may not be a good choice at all. First you need to know the desired temperature and then you need to know the ambient temperature in the area of robot application. Then you can determine if heating or cooling will be required. And probably nothong is slower than a Peltier device for either.

 

Welcome to AAC.

Is this a school project that will receive a grade?

Thank you for your response, no it`s a research project nothing to do with grades.

 

As others have mentioned, depending on the effect you intend, you are better off using a circulating liquid.

Since the idea is about human interaction, keep in mind that our hands are warmed by the blood flowing through them. Not only is a Peltier device going to have waste heat to deal with (not trivial) but it will feel nothing like a human hand, even under fleshly polymer.

 

As others have mentioned, depending on the effect you intend, you are better off using a circulating liquid.

Since the idea is about human interaction, keep in mind that our hands are warmed by the blood flowing through them. Not only is a Peltier device going to have waste heat to deal with (not trivial) but it will feel nothing like a human hand, even under fleshly polymer.

Everything just need to be installed on the Robot hand outside, what are the disadvantages if I go with Peltier element in this case?

 

Is the "specific temperature" fairly constant? I think a waterbath and a flow-through heat exchanger might be a better option. It'd be much easier to control the temperature of the water in the bath and circulation system.

But, it's possible to control temperature very precisely with a TEC if you must. I have a related completed project here you might want to look at.
https://forum.allaboutcircuits.com/threads/project-poor-mans-calorimeter.50045/

Thank you.
Not at all the temperature can be specifically constant but I would like to attain the normal temperature.

 

Everything just need to be installed on the Robot hand outside, what are the disadvantages if I go with Peltier element in this case?

A Peltier device produces a temperature differential. One side heats up and the other cools. If must be mounted in such a way that one side can dissipate heat when cooling the other side. This means they can’t be contained in the same enclosed space, or thermally coupled to the same object that is being heated or cooled.

Have you characterized the temperature change you will need to make?
How many degrees change in how much thermal mass over what time?
Will you need to both heat and cool, or only heat or cool?
Even if you planned to heat and cool, is it possible to only do one actively and the other passively?

 

A Peltier device produces a temperature differential. One side heats up and the other cools. If must be mounted in such a way that one side can dissipate heat when cooling the other side. This means they can’t be contained in the same enclosed space, or thermally coupled to the same object that is being heated or cooled.

Have you characterized the temperature change you will need to make?
How many degrees change in how much thermal mass over what time?
Will you need to both heat and cool, or only heat or cool?
Even if you planned to heat and cool, is it possible to only do one actively and the other passively?

Thank you, good questions to answer.
1. Yeah, it`s different for example in summer it can be hot as 40 while in winter it can be 5 and I the temperature I want to keep is 20 so the difference is around 20.
2. The thermal mass is like 400 g.
2. At a time I only need heat or cool

 

Thank you, good questions to answer.
1. Yeah, it`s different for example in summer it can be hot as 40 while in winter it can be 5 and I the temperature I want to keep is 20 so the difference is around 20.
2. The thermal mass is like 400 g.
2. At a time I only need heat or cool

Well, the thermal mass (it’s a loose term, I should have said heat capacity) is specified in J/K° (Joules per Kelvin). It tells you what it takes to influence a thing concerning its temperature. You need to have some idea just how much energy you are going to need to get the robot hand to the set point temperature from the ambient one.

You will also need to know how long it will be before the ambient temperature influences the manipulator temperature to the point of needing to use active temperature management again.

This is because you have to figure out if the Peltier is going to suffer self heating to the point that it can’t be used to manage the temperature any more. Think of this: the Peltier hot side is extracting heat from the cold side and everything it is thermally coupled to. If you turn off the device, the hot side will begin to heat the cold side and its thermal neighbor.

If the hot side gets too hot, it won’t be able to provide the differential needed to cool the robot, and it will end up doing nothing or even heating it. To prevent this, the hot side needs some sort of cooling. This can be done with an appropriate heatsink, but that has to be exposed to the air so it can lose heat by radiation and convection. Since radiation is a fairly weak way to cool things, and convection in an ordinary room is limited, a fan is usually used on the heatsink.

Can you accommodate an exposed heatsink and possibly a fan on the robot where the Peltier will be mounted?

One possibility would be to remotely mount the Peltier, that is remote from the hand, not from the robot, and use a heat pipe to thermally attach it. I would recommend getting a small Peltier and doing some tests. In the absence of someone having the thermal management design skills to simulate or calculate this, empirical results will give you an idea if the Peltier stands a chance.

It might work just fine, but despite being really cool things (no pun intended) the reality of them is so often very disappointing. They don’t “just work“ as we’d like them to.

 

Ok thank you for the information.

 

Ok thank you for the information.

If you can, please let us know how you get this done, it’s very interesting.

 

If you can, please let us know how you get this done, it’s very interesting.

I will for sure.

 

The biggest problem using a TEC is that they use a LOT of power to move a small amount of heat. For instance under near-max conditions, they'll move one unit of heat for every ten consumed, eg. 10W moved from cold side to hot side requires 100W input. All the power they move and consume - 110W in our example - appears on the hot side and must be dissipated to ambient to avoid overheating the TEC. Even at a temperature that's safe for the TEC, they can only produce a limited ∆T from hot side to cold side, and so a high temp on the hot side limits how cold the cold side can be. TEC's are often used in cooling applications and this is why they almost always are attached to a big heat sink with a fan, to cool the hot side as near to ambient as practical, to allow the low side to get cool as possible.

You can look at the specs but I think you'll struggle to achieve a ∆T of 20C° versus ambient. They make multilayer TECs to achieve larger ∆Ts but this does not change the basic problem of heat dissipation.

The problem is greatly reduced in a heating application. Heating a resistive wire such as nichrome is the more typical solution for heating.

 

Everything just need to be installed on the Robot hand outside, what are the disadvantages if I go with Peltier element in this case?

Disadvantages of a peltier device: #1 They only come in a few shapes, #2 They need external heat conduction material on both sides, warm and cool, #3 For cooling they MUST have a heat sink#4 being a heat pump device, they are slow. #5 They are non-conformable.. #6 They do produce heat from the power applied, so that also must be dealt with. #7 They are rather large for the amount of heat they transfer.
There are probably additional disadvantages but these are what come to mind right away.
Embedded flexible tubing with temperature controlled fluid is a far simpler and much more capable method of temperature control.And for rapid changes, having both a hot liquid and a cold liquid source in the system, and a method to quickly adjust the ratio, is a very effective scheme, although more complex.

 

I would like to make the situation more clear.
I have a robot and we are doing experimentation on human interaction with the robot. When someone holds the hand of robot, he tells a story. First I want to measure the temperature of the person who wants to touch robot`s hand and then adjust the robot`s hand temperature similar to it.
This is the only job I`m focused on. I think the temperature difference will be around 5 degree or less. Can you advise me now what should I use to heat or cool the robot`s hand.

 

I would like to make the situation more clear.
I have a robot and we are doing experimentation on human interaction with the robot. When someone holds the hand of robot, he tells a story. First I want to measure the temperature of the person who wants to touch robot hand and then adjust the robot hand temperature similar to it.
This is the only job I am focused on. I think the temperature difference will be around 5 degree or less. Can you advise me now what should I use to heat or cool the robot`s hand.

I have a robot and we are doing experimentation on human interaction with the robot. When someone holds the hand of robot, he tells a story. First I want to measure the temperature of the person who wants to touch robot hand and then adjust the robot`s hand temperature similar to it.
This is the only job I am focused on. I think the temperature difference will be around 5 degree or less. Can you advise me now what should I use to heat or cool the robot hand.

 

OK, so the person is holding the robot's hand while talking to the robot..
I am making a guess that the robot's "hand" is not a hard metal surface item, but rather covered with some sort of material to try and approximate a human hand. with that being the case, and as the change will be a small temperature rise, the fastest scheme will be resistive heating embedded in the hand covering material. But in that case human hand holding causes an unintended temperature rise, an arrangement with fluid filled small diameter tubing embedded in the hand covering would mmore closely simulate the thermal actions of a human hand, as it could hold a temperature until a command for a warmer temperature was given , at which time warmer fluid could be circulated. OR, the circulation could be halted and some electrical heating could be initiated.

 

The robot hand is a hard metal. can you give me some more explanation or materials (site/link) on fluid filled system for temperature control?