Choke and MOSFET temperature relationship in VRM system

Question

What is the temperature relationship between a choke and a MOSFET without heatsink in a typical motherboard VRM (Voltage Regulator Module)?

Context: I bought a dirt cheap CPU to extend the lifespan one of my old rigs. The motherboard is a good quality one, but it's not overclocking oriented. Since the "new" CPU is a flagship for its generation, it draws at least 2 times more power than the one I have now. I'm trying to measure what is the MOSFET temperature during the stress test to make sure the VRM can handle the newcomer. The problem I have now is there is heatsink on the top of the MOSFETs. So I'm now trying to figure out indirectly and approximately the MOSFET temperature without removing the motherboard from the case. Since I can measure the temperature of the chokes, I'm wondering if there is direct proportional relationship between the MOSFET and the choke regarding their temperature.

Edit: I'm not looking for something accurate. I just make sure the MOSFETs are in the safe zone even with the most pessimistic estimation. For example: If the temperature difference between the MOSFET and the choke in the worst case scenario is 20 °C, I know that the MOSFETs are in the safe zone if the choke temperature is below 80 °C.

Answer 1

The inductors have copper I2R losses, and core losses proportional (not linearly) to flux density and frequency.

The FETs have I2R and switching losses. But the conduction losses are distributed between the FETs depending on the duty cycle. The FETs will likely not have the same RDSon, so conduction losses may be different for the same current.

There may be some heat conduction sharing if they are located near each other, but without an electrical and thermal model of the whole system I don't think you can correlate the inductor temperature with the FET temperatures.

Also VR controllers will have built-in current limits per phase that will limit the current to a safe value for the inductors and FETs, so the symptoms of overload would not be a component burning up, it would be the processor resetting as the voltage dropped out due to current limit.