Sign in with
Sign up | Sign in
Your question

Peltier for Heating, yes, HEATING!

Last response: in Overclocking
Share
December 14, 2006 12:59:28 AM

Well guys, I have a question and figured someone on here would know enough to explain a few things to me.

My boss is working on building a machine to do medical research that requires a heat source. He cannot use pumps, so peltier is where its at.

He needs to heat about 10 ml of water to 60 - 65 *C.

Here's the question. There are peltiers rated at 320 watts (68*C); If one side of the peltier plate ( the hot side) faces the fluid, and the cold side faces the open air (about 24*C), can the plate continually heat the fluid by drawing heat from air?

(I know this isn't exactly computer related, so don't shoot me :lol:  )
December 14, 2006 1:14:19 AM

Dood... thats really silly actually.

I assume if u have a big enuf hsf on the cool side, yes, u 'CAN' do that.

However for your purpose, why the hell not use a simple heating element and thermostat!!!???

Peltiers are heat pumps, they punp heat away from one side.
What this means? To cool one side, u gotta keep the hot side cool. Or to heat something, u gotta keep the cool side colder.

Totally stupid 4 your purpose IMHO.
December 14, 2006 1:23:51 AM

Although I disagree with the method the previous poster put it to you, I do agree with the conclusion. It'd be much more efficiently and controllable to use a heating coil or element to produce heat.
Now if you want to do it the hard way (which is actually kinda fun), it is possible, in special situations, to pump more watts of heat than the watts of electrical power input. Coefficient of power will depend on the application, heat pumped, and temperature differential required. Typically, the coefficient of performance, heat pumped then divided by input power, is between 0.4 and 0.7 for single stage applications. However, higher or in your case lower, COPs can be achieved with optimized custom TECs.
Just giving you an idea of what you'll need to do.
Related resources
Can't find your answer ? Ask !
December 14, 2006 2:10:47 AM

Quote:
Although I disagree with the method the previous poster put it to you, I do agree with the conclusion. It'd be much more efficiently and controllable to use a heating coil or element to produce heat.
Now if you want to do it the hard way (which is actually kinda fun), it is possible, in special situations, to pump more watts of heat than the watts of electrical power input. Coefficient of power will depend on the application, heat pumped, and temperature differential required. Typically, the coefficient of performance, heat pumped then divided by input power, is between 0.4 and 0.7 for single stage applications. However, higher or in your case lower, COPs can be achieved with optimized custom TECs.
Just giving you an idea of what you'll need to do.



We need something that has a small footprint and actually generates a lot of heat and can be shut off and on at the touch of a switch .It takes time for a heating element to heat up and is not practical for our needs.


I was trying to explain to him that you need a high heat source on the other side of the peltier (the cold side) so that the peltier will functoin as a heat pump (its purpose) and move the heat from one side to another.

He was attempting to compare a peltier unit to a vapor phase compressor, although I tried to explain that it can't pull more heat out of the enviornment if the place where the heat is being pumped to is hotter than ambient temp.

The space we are trying to heat has about a surface area of 20 square inches with about 10-15 ml of fluid to warm. B/c of the space requirements, there's nothing really as small. The entire machine has to stay small. Its a problem really.

Basically, you guys told me what I already knew. I guess I'll be looking up literature to figure out how to explain to him what needs to be done.

I told him he needs a heat source ( something maybe 100*C) to pull heat to the other side.

Eh, this will be interesting....
December 14, 2006 5:33:06 AM

Hi Synthetic, or Killer if you prefer...

If you are using the Peltier to pump heat from the cold side to the hot side, and you want the hot side to be about 60 degrees, you do not have to have the "cold" side hotter than 60 degrees. The pump will work when the cold side is cooler. However, the warmer you can get the cold side the better. I would draw some kind of analogy to the way an air conditioner works, but air conditioners are fluid devices. I believe the Peltier effect is actually less efficient than the fluid heat pump, because you can get very good COP's from air conditioners (greater than 1, which means you get more heat moving than work input).

However, for your case, I'd say give the Peltier a try. You should be able to build one to your liking yourself... just look up which two materials are good for the job... and have a good variable DC power source available with surge protection! If you find you're not getting enough heat through, then you might consider heating the cold side... nice thing is that the cold side can be fairly far away from the hot side... you just need to pick a good length of wire. Of course, you'd need to do more homework before building.
December 14, 2006 4:13:41 PM

Quote:
Hi Synthetic, or Killer if you prefer...

If you are using the Peltier to pump heat from the cold side to the hot side, and you want the hot side to be about 60 degrees, you do not have to have the "cold" side hotter than 60 degrees. The pump will work when the cold side is cooler. However, the warmer you can get the cold side the better. I would draw some kind of analogy to the way an air conditioner works, but air conditioners are fluid devices. I believe the Peltier effect is actually less efficient than the fluid heat pump, because you can get very good COP's from air conditioners (greater than 1, which means you get more heat moving than work input).

However, for your case, I'd say give the Peltier a try. You should be able to build one to your liking yourself... just look up which two materials are good for the job... and have a good variable DC power source available with surge protection! If you find you're not getting enough heat through, then you might consider heating the cold side... nice thing is that the cold side can be fairly far away from the hot side... you just need to pick a good length of wire. Of course, you'd need to do more homework before building.


The guy I'm working for was considering putting the entire unit in an incubator (37*C) while it runs. This would keep the are around peltier plate warm on the cold side. If that works, the only problem will be not burning out the incubator.

We'll most likely end up playing around with it before actually finalizing the product.

Medical research can be a real b*@!% sometimes.
December 15, 2006 3:04:15 PM

I am thinking, and I can't imagine what it would be for?
December 15, 2006 3:44:51 PM

Yes, what you propose would work. I have a portable cooler that uses a peltier for cooling and when you reverse the connection it becomes a warmer. I won't go into what would be more efficient because I have no knowledge in that area. Suffice it to say that what you propose would work.

hball



Quote:
Well guys, I have a question and figured someone on here would know enough to explain a few things to me.

My boss is working on building a machine to do medical research that requires a heat source. He cannot use pumps, so peltier is where its at.

He needs to heat about 10 ml of water to 60 - 65 *C.

Here's the question. There are peltiers rated at 320 watts (68*C); If one side of the peltier plate ( the hot side) faces the fluid, and the cold side faces the open air (about 24*C), can the plate continually heat the fluid by drawing heat from air?

(I know this isn't exactly computer related, so don't shoot me :lol:  )
December 15, 2006 3:56:30 PM

A peltier keeps a theroetically fixed temperature differential between its hot and cold sides.

For the sake of argument, lets say this factor is 30°C, and the ambient air temp is 15°C.

If this theoretical peltier is turned on in the air, it would take the hot side to ~35°C and the cold side to ~5°C - it "pushes" the temperatures apart 30°C, and its own power dissipation adds to it a little. This is why it is not 30°C and 0°C.

Now, if I use a heatsink or other cooling system to cool either side by 10°C, then the peltier will (assuming it is a high enough rated one that it can cope) cool the other side by 10°C. Heatsinks and radiators cannot cool below the ambient temp, and this is the purpose of peltiers in computer cooling. (as a heatsink could cool the hot side to ~20°C, taking the other side to -10°C. If the peltier were reversed, the heatsink could still only cool the cold side to ~20°C with a 15°C ambient temp, leaving the hot side at 50°C, this is why reversing them is bad in computer cooling.)

Of course, the opposite is also true, and the peltier will do its best to maintain this 30°C delta between the sides if I heat either side (up to its rated maximum of course).

A peltier on its own though is no better than a heating coil in terms of how quickly it can heat, as it still takes time to create its thermal delta. The only purpose of this would be to use a peltier WITH a heating coil to lift the temperature above the maximum capability of the heating coil, but as heating coils can happily hit hundreds of degrees, the peltier would catch fire before this.
December 17, 2006 5:38:22 AM

Quote:
...

We need something that has a small footprint and actually generates a lot of heat and can be shut off and on at the touch of a switch .It takes time for a heating element to heat up and is not practical for our needs....

The universe of heating elements is extremely diverse. If you need something thin, you can get "tape" form factors. You can get elements with high power densities.
Plenty of PCR machines use heating elements, so I would think you could find something to fit your need.
December 20, 2006 1:15:27 AM

Quote:
...

We need something that has a small footprint and actually generates a lot of heat and can be shut off and on at the touch of a switch .It takes time for a heating element to heat up and is not practical for our needs....

The universe of heating elements is extremely diverse. If you need something thin, you can get "tape" form factors. You can get elements with high power densities.
Plenty of PCR machines use heating elements, so I would think you could find something to fit your need.

I've been busy. Haven't seen the forum.


You're right about the PCR machine, or as others call them, thermal cyclers.

He's basically dead set on peltiers. We found a 320 watt peltier for 20 bucks, but now I can't even find the site that offered it. :( 

I think we could play around with it and see if we can get temperatures within tolerance ( 60-65 C).

Thanks.


Btw, if anyone has any good links to peltier products ( around 320 watts or so) and power supplies that are on the small side, I'd appreciate it.
December 3, 2009 12:54:15 PM


Btw, if anyone has any good links to peltier products ( around 320 watts or so) and power supplies that are on the small side, I'd appreciate it.[/quotemsg]

http://www.birkmfg.com/stockheaters.html

sorry -- this is not a peltier product, but I think it'll work for your medical application. It's a kapton heating element (etched foil encased in a flexible plastic). If you sandwich it between two thin pieces of sheet metal, it heats up pretty quick (the speed at which it heats up depends on how much power you dump into it and how you heat sink it: to heat up faster, apply more power and/or use a heat sink with a higher thermal resistance, i.e. thinner, less surface area). I was using these for an incubator, and was very satisfied with them, but ended up switching to good old nichrome wire because it's so much cheaper.

To reach your steady-state temperature quickly, I recommend using a microcontroller to output a PWM signal which drives a FET to control the power to the heating element. You can mount a temperature sensor to your sheet metal (e.g. LM35) for feedback to the microcontroller. Using a "take-back-half" algorithm to control the duty cycle, you can reach steady-state with almost no overshoot.

a c 86 K Overclocking
December 3, 2009 2:26:58 PM

3 years late, look at the post date.
!