# The purpose

The main goal is creating a little hand-held device that would heat a thin film (locally) to $50-70°C$ by sliding over it. The film area is $<150cm^2$. A secondary goal would be to create a device with the same parameters that would cool the film down (locally) to $<-28°C$ to reverse the process.

# Problem domain

The problem that I need help with

1. Reasonably sized case within the range of units of centimetres.
2. $1mm \times 20mm$ heating element (1mm or less if possible to increase efficiency)
3. A circuit serving the heating element
4. Battery

I don't want to heat the material itself (or I do, but very superficially, maybe 0.001mm in depth). This device has to be powered with a low capacity battery (e.g. 1700mAh, 1.5V) for as long as possible.

I came up with a couple of ideas, but I can't figure out the specific parameters and components to use.

## 1. Termistor

As a self-regulating heating element, thermistor seems to be just perfect for this purpose. But what exact parameters should it have? Will it be effective enough? And where can I buy a thermistor with those specific parameters?

## 2. Peltier Module

Another solution that comes to my mind is using a Peltier Module. It uses a lot of power, but maybe if I make it small enough (by splitting a ready-made Peltier Module into separate rows?), maybe the input power wouldn't be so high. The questions about the thermistor apply here as well.

You could use a Polyimide etched-foil thin-film heater such as those made by Minco.

As far as refrigeration goes, if this was a serious high-volume application I'd be looking at something other than Peltier modules, probably completely out of the EE realm. Peltiers are just too inefficient and have too much thermal mass.

1) A termistor is used to sense temperature not change it. It is an input device - not an output device. I am not sure you could use it to change temperature.

2) A peltier element would work - you can find small ones say 2cm x 2cm that shouldn't consume to much power. A peltier heats/cools proportional to the current so you would need to play around with it to find a balance between power and energy efficiency.

If you only need 50 degrees C maybe take a look at buying these types of hand warmer gloves and cutting them open. Inside they have a rectangular heating element which is basically just a resistor. Overall a resistor heating element will be more efficient than a peltier for heating because the peltier spends energy to heat one side and cool the other - a resistor heating element only heats up.

http://dx.com/p/usb-powered-dual-side-warmer-yarn-gloves-black-pair-168832

To buy these devices you can get them at dx.com if you can wait or just look for something on digikey.

• As far as I know, thermistor is any resistor that is prone to increase its resistance with temperature. Therefore if you add a power source to it, it will start heating until the resistance prevents it from further heating. Correct me if I'm wrong. – Alex Nov 14 '13 at 0:02
• Since thermistors are resistive, you are correct. They are not designed to be heating elements though and are not appropriate for the task. – AngryEE Nov 14 '13 at 17:36
• @AngryEE that is not correct. PTC (colloquially 'thermistor') heaters are used very widely, for instance as so-called 'band heaters' and plate heaters in applications where temperature accuracy is not important. This is very different from NTC thermistors which exhibit much more stable temperature characteristics and are used as temperature transducers. This is an excellent example of a namespace collision between two fields of electrical engineering :) – user36129 Dec 14 '13 at 10:30

Peltier modules are inefficient, even compared to other (also inefficient) methods of cooling, but as far as I know, there is no other device to compete in small form factor. Obviously a peltier is not as small as a thin-film heater, but it has the advantage that you could use it for heating and cooling. For small heaters, where cooling is also required, barring cost (40x40mm bought in north america tends to be 20 to 40 shmeckels, and the cheapest I know of are those being sold on aliexpress for about 2.5 shmekels), a peltier is actually kind of great.

Because a purely resistive heater can be considered 100% efficient (all power used by the circuit is turned into useful output), so long as there is a free source for heat to be drawn from, a peltier can be looked at as a >100% efficient heater within the frame of the system, delivering more heat than the power the circuit is actually using. On your cooling cycle, a peltier will be inefficient, and how big of one you need will depend on the thermal load, but if all you want to do is remove the heat you had already localized there, you can probably get away with a pretty small one.

The disadvantage I see if you're trying to achieve a really tiny form factor, is that you will need a heat sink and/or fan if you intend to use a peltier for cooling, and these will likely be considerably larger than the module itself. The heat sink and/or fan will have to be enough to dissipate the thermal load plus electrically conductive losses, plus thermal conduction losses (any heat that makes its way back from the hot side to the cool side must be transferred again). IIRC because of thermal conduction losses, combined with the nature of DC current losses, peltiers tend to operate best in continuous current mode, throughout their conduction range, rather than PWM.

If you want to look into this, I'd consider using a small form factor CPU cooler. If you watch at your local computer store, you can get these on sale for far below their component value when cpu socket types become obsolete (pretty much all the time where I live). They can have features like omnidirectional heat pipes with a sintered coating or grooves inside for capillary action. The thermal performance of copper is superior to aluminum, and the thermal performance of copper heat pipes is vastly better than an equal weight of solid copper. Note that some types and sizes of heat pipes can be damaged by freezing temperatures.