2A, 1.x v drop or so across the TIP120 ( Vce(sat) is ~1.4v right?)... that's like 3W dissipated in the TIP120. You need a heatsink for that.
Or use a MOSFET, which will have a much lower voltage drop across it, and hence dissipate less heat. The TIP120 is a BJT, those are obsolete for the purpose of power switching for exactly this reason; modern MOSFETs can have an on-state resistance of milli-ohms.
No matter what it's gonna get hot if you're putting 2A through a darlington transistor - at least assuming a highish duty cycle!
Vce(sat) is like 2v (you guys made me dig up the damned datasheet) - you will drop that much voltage across the TIP120. Voltage times current equals power; 2A * 2v = 4W! 4W on a TO-220 package is gonna get hot as hell without a heatsink. This is why nobody uses BJTs to switch high current loads nowadays.
the heating element has a resisting wire with 1,4-1,8 ohms, i dont know much about ohms law and stuff but is it possible htat the heating element is drawing more then 2amps from the power supply (its not a constant current supply) ?
it cant be normal, 24w without a heatinsink is a lot but its going up to a hundred degrees within seconds
edit: i just calculated this, 1.8 Ohm at 12V equals 80W / 6.666A
When you spec a MOSFET, you don't look for the total power of the load you're switching, you look at the maximum voltage it will hold off, the current it will handle, and whether it is compatible with the logic voltage you have available to switch it (typically 5v). Then, you check the graph in the datasheet, and see what the Rds(on) would be with the gate voltage you have, apply ohm's law to see how much voltage you'd drop across it; multiply by current to get the heat dissipated, and make sure that you can deal with that (many MOSFET datasheets give a crazy high current spec - but you'd need to heatsink the bejeezus out of it if you actually want it to carry that kind of current!). Never try to use MOSFETs at the limits of their specs - that's asking for trouble.
IRF3708PBF is probably a good choice? TO220 package, works great even at 3.3v on the gate (it's about the only decent high power MOSFET that works at 3.3v gate voltage and comes in through-hole package), switches loads up to 30v.
I also just so happen to sell some sick 4-channel logic level MOSFET driver boards Though it sounds like you're only switching one channel, so that might not be a sensible option for you.
I did not know the IRF3708. It seems to be perfect in an Arduino scenario. Low gate voltage.
That MOSFET is great if you only have the choice to drive at low gate voltage. If you are using a higher gate voltage (4.3-5V) there are literally hundred of better choices.
Take the AOI510 that crossroads mentioned for example it has less than half the resistance AND its half the price ( much better choice). Same voltage limits and close to the same gate capacitance. There's no reason to pick the 3708 UNLESS you are using a gate voltage less than 4v ( its a great choice for 3.3v micros though!)
someuser:
more resistence equals less power power doesnt it?
No, you have to do the maths.
With a constant voltage supply across a resistor, more resistance = less power,
With a constant current supply across a resistor, more resistance = more power.
Regardless of all the math and details that other people have posted about, you can NEVER expect a power device like a transistor to meet its rated capacity without it being mounted on a fairly substantial heat sink, probably with a fan blowing on it....
If you have 2k in series with 12 volts the current would only be 6mA, the heat is coming from high current flowing through the 2 base-emitter junctions because of no current limiting resistor in the base circuit .
Maybe you could swap the heating element & transistor.
What is the resistance of the heating element? You should have 12V+ to heater, heater - to Collector, Emitter to GND, 2k resistor between Arduino and Base and Arduino GND to 12V -. Like others suggested, you need a heat sink for the TIP120.
i will supply a circuit soon, i just replaced the transistor with a relay and the power supply with bigger one. it works okay for now, besides the click click click of the relay.... but it will do until i resolved some programming issues, bought a mosfet and find time to map out a diagram.
westfw:
Regardless of all the math and details that other people have posted about, you can NEVER expect a power device like a transistor to meet its rated capacity without it being mounted on a fairly substantial heat sink, probably with a fan blowing on it....
Adding to that, a TO-220 can dissipate up to 2W in most cases without a heat sink (depending on the max junction temperature from the datasheet and the max temperature you want in your setup - 2W WILL already be pretty hot).
At least above that, you need a heatsink.