peltier junction heated bed?

Anyone use a peltier junction for a heated bed? I'd think with the surface area it could be a good fit. Plus they are kinda slim.

That would be possible, though I think that a resistance heater would be a lot more efficient, but I could be wrong.

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

Do they make them that big? A quick search found some 50x50 mm Peltiers for $40, but at that price it'd cost over $600 for a full Mendel. Then you need heat sinks on the cold side, and you're still limited to a max delta T of around 60 deg C. Seems pretty expensive relative to a PCB heater.

Wade

Wade Wrote:
-------------------------------------------------------
> Do they make them that big? A quick search found
> some 50x50 mm Peltiers for $40, but at that price
> it'd cost over $600 for a full Mendel. Then you
> need heat sinks on the cold side, and you're still
> limited to a max delta T of around 60 deg C.
> Seems pretty expensive relative to a PCB heater.
>
> Wade

Price and delta t make sense but why does size come into it? Do you fully cover the bottom of a bed with resistors?

Hah, good point! You do cover the bed with a PCB heater, but not if you use discrete power resistors. I was imagining a single large Peltier element used as a build bed. You could use several smaller ones under an aluminum build plate, but you'd still need about 4 of those 50 mm Peltiers to get enough wattage, assuming you have a big enough cold side heat sink.

Still, Peltier elements are pretty expensive; compare 100 W (hot side output) Peltiers to an equivalent set of power resistors, and add in the cost of the cold side heat sinks - it's pretty hard to justify, unless you really, really need to keep electrical loads down. Peltiers are great when you need to cool something below ambient and traditional, cheaper refrigeration cycles can't be used. For simple heating loads though, they're overkill.

If the price comes down though, it'd be a great way to keep the energy costs of running a reprap down. At the moment though, I'd make a WAG that a Peltier heat bed could double the cost of a reprap.

Wade

make sure the solder contacted end of the Peltier is on the cold side. the Peltier uses a principle called the thermoelectric effect. it is also called the Peltier–Seebeck effect. i was pleasantly surprised that Peltier is quite efficient at heating.

it is describes in some detail below.

[en.wikipedia.org]


you may also want to read here with issues related to Peltier.

[www.tellurex.com]

quote from article:

For heat-only applications, do thermo-electric devices have advantages over resistive heaters?

Yes, but... Resistive devices create heat solely by virtue of the power dissipated within them. TE devices, on the other hand, not only provide this I2R heating, but also actively pump heat into the thermal load; this, potentially, makes them much more efficient than resistive heaters. Unfortunately, the need for a DC power source and the generally higher cost of TE systems compared to resistive heaters, precludes their use in most heat-only applications. Furthermore, Peltier devices have a far more limited temperature range than most resistive heaters. Generally, TE devices are only used for heating in systems that also require cooling or where efficiency is extremely important.


so i guess it is more efficient to use Peltier, than resistive heating. just make sure that the device never ever reaches re-flow temp of solder, and pay attention to monitor cold side temp and hot side temp. when things go wrong it is unpredictable what side will be hotter than hot. I would monitor both sides for temp readings.

I bought some peltier squares, 40mm, a while back from hong kong, price was reasonable, they have the advantage of heating or cooling and can emit current, anything that can do that has got to be usefull for something.

---

Random Precision

I've been using them for heated beds for almost a year and they work great on a 1/4" aluminum table, without a heatsink and without control. I posted it somewhere in the forum just haven't had the time to post some pics.

I'm using 1x 40mm x 40mm module. The size that I'm using is model 12708. Steady state it draws about 5.4A and the table temp gets to about ~90oC.

I like it because, there is no need to have any temp control circuit and NTC, just hook it up to 12V and your ready to go.

I have lots of TEC's if anyone is in need. 5 pack shipped for $50

Be careful with the Peltiers. I bought a set of 5 which were max rated to 70C. I loved how it would heat pretty quickly. All of a sudden i couldnt get it to heat. It burned out on the 3rd use. Im sure there are more capable Peltiers out there, but considering the cost difference, it may not be worth it.

Someone had mentioned to me that a Peltier may be an excellent heater/cooler if applied to glass or maybe aluminum. By heating it you minimize warping and help adhesion to the surface. Then by cooling it rapidly (Reversing polarity) you can quickly pull heat off the surface and plastic parts should in theory pop right off the surface due to differences in the rate in which the plastic and surface shrink. Could be an interesting experiment.

[en.wikipedia.org]

Edited 2 time(s). Last edit at 03/09/2011 03:59PM by CidVilas.

I forgot to mention I'm making an extruder plate using a TEC to cool the PTFE.

@CdnReprap: You could create a thermal loop. Attach a TEC onto copper pieces(for maximum thermal transfer) have one side attached to the heater barrel and the other to the insulator. So it will heat the nozzle and pull heat transferring into the insulator back into the nozzle. This is taking into account that TECs are really "heat pumps". Just an idea.

See attached picture, and i apologize for the quality of the image. It was drawn in Paint using a touch pad. Bleh..

Update: After thinking about it some more, this may actually not be possible.

Edited 1 time(s). Last edit at 03/09/2011 05:01PM by CidVilas.

@CidVilas: If we used the TEC in that manner we would be making electricity with it!

Interesting to say the least.

I think I'm gonna go the resistor route for now and experiment with TECs later on.

Whats the typical resistance and wattage and # for for a typical mendel? Running them isn't a prob, have more then enough stuff here to run it solid.

I'm not sure what this type of approach is trying to address. The efficiency of converting electricity to heat using resistance is already 100%. And the circuit is very straightforward.

The efficiency of converting electrical energy to heat with a Peltier junction exceeds 100%. But as Wade pointed out, it's really not worth the cost to use them this way in most applications. Maybe if you were trying to run a battery-powered reprap it would be worthwhile.

100% Efficiency? Not for a Peltier... Electrical efficiency doesn't really tell you the whole story. Heat pumps have their own terminology - CP (sometimes COP) - coefficient of performance. Basically, how much heat (energy) does it take to move one unit of heat. If we ignore the environment, then it looks like efficiency is greater than unity - but without an environment, we don't have a heat pump.

Peltier junctions have a COP_cooling around .5 (cheap ones being lower, expensive ones up to .7). When used as heaters, they can be greater than 1.... You'll need to insulate the heater from the cold side, however.

That said, if you exceed the deltaT (temp difference between hot side and cold side) of a peltier, you basically have an expensive resistive heater and your COP falls back to 1 (it will heat up slightly faster, however). I suspect that is how CdnReprap achieved ~90C. At 5.4A/12V... P=~65W. That's more than enough resistive power to hit 90C. At least it passes the thermodynamic gut check as Makerbot did it with less than 30W, Up! with less than 60W and Mk1 PCB Heater with way less than the 90W peak. Dare I say, if steady state for the Peltier was drawing 65W, that sounds pretty high.

------
At the end of the day.... Resistors are the way to go unless you need to cool the bed (inducing rapid/nonuniform shrink to make it pop off the bed doesn't sound like a feature I'd want). Resistors have a COP of 1 in every condition I can think of. They're cheap, fairly reliable and scale cheaply. The only thing more reliable than a wire wound resistor is a heat cartridge - that comes at a cost.




jbayless Wrote:
-------------------------------------------------------
> The efficiency of converting electrical energy to
> heat with a Peltier junction exceeds 100%. But as
> Wade pointed out, it's really not worth the cost
> to use them this way in most applications. Maybe
> if you were trying to run a battery-powered reprap
> it would be worthwhile.

What is your supplier? I may be able to put those to use, but I am in the market for raw TE material that I can build my own experimental units out of. I have seen iron/constantin wire that comes as two-wire pairs. However I can't source that online. Any leads? Feel free to shoot me an email. I'd like to get some semi-conducting p-/n-types if possible that have a higher Z than iron constantin.

BMK

I make the TEC's myself, I use bismuth telluride as the semiconductor material.

If you don't have edm and cnc wire cutting your going to have a tuff time making modules with a tight spec. You can however modify/disassemble cheep TEC's into whatever you need.

If you ask nicely you might be able to buy some of the processed bi te material ready to cut from Marlow if they are still making modules in the USA.