Just for fun, here is a photo showing an Arduino 2009 in thermal infrared, showing you where the power goes. I am powering it with +10V DC, so the regulator shows a lot of heat, and a little bit from the 1N4007 diode. The AT328P cpu is just a faint smudge. There are two LEDs on, "PWR" and "L" and you can see two tiny dots for each; that is the LED and the current-limiting resistor next to it. The two 47 uF electrolytic caps at lower left aren't really hot, that is a reflection from something else in the room off the metal case top. It was running stand-alone with no USB connected, so the FTDI chip was not active. Total current draw: 25 mA.
What I find most interesting is that the power connector seems to be generating more heat than anything else. I'd think that the resistance of the connection would be low enough so that very little voltage would develop across it, and hence, almost no power.
has been running for a day with some tight loop? e.g. calculating a lot, no delay's etc.
That will make no difference to the thermal profile of this part. Calculations or delays will be much the same in terms of power consumption. This is not a PC processor.
Glad it was interesting. Would be neat to compare some other shots- but camera was onsite for one day only, and they just had a half hour free after the scheduled work so I grabbed what I could.
The camera was FLIR SC6000, which uses a mechanical Stirling cryocooler to cool the InSb (indium antimonide) sensor and comes with a support system on a luggage rack. Not quite a point-and-shoot setup. Apparently the lens alone (made of germanium) is twice the price of my car... I won't be owning one of these anytime soon.
