Passionate Eng said:
because as temperature increases the resistance decreases, so reducing the power absorbed.
is that true?
In general no. Let's assume you have a resistor with a negative temperature coefficient so that your premise is true. Let's also make it a 10 Ohm resistor being fed by a 10V source. The current will be 1 Amp so it will be using 10 Watts of power. If the resistance fell by half because the resistor became hot then the current would be 2 Amps and the power would become: (2Amps)^2*5Ohm = 20 Watts.
Negative temperature coefficients can cause runaway thermal currents. As some devices get warmer their electrical resistance gets lower and it allows more current to flow which creates more heat etc. It's an engineering problem.
Manufacturers will sometimes give you a
thermal resistance in a spec sheet for a given part. You can plug that and the ambient temperature into a formula to estimate how hot the device will become for a given load. It gets complicated though. The device could have multiple paths to dissipate heat. There could be a thermal resistance for the device package to ambient air, thermal resistance to the printed circuit board, and thermal resistance from the circuit board to ambient air. All of those figures depend on how the designer implements the device.
Anyway, the formula is straightforward.
$$ Temp = P_{dissipated}θ_{ThermalResistance} + T_{ambient}$$
Remember that this is just an estimate. It's used to ensure that you are not overheating components in a design. It's not intended for temperature control.
When you do want some kind of temperature control it's more common to use feedback. A temperature sensor could send a signal to increase power to a heating element when the temperature is lower than you want.