problem:
I may need a little clarification on this problem. I do not understand the term TR Power.
It seems that you bought some bad parts in that they failed immediately when installed in a plasma display you are working on.
You would like to verify component parameters by test before installing to avoid damage to the unit when the component fails.
I have heard of counterfeit parts on the market. I know of one company that is taking pictures of the die for incoming inspection by quality control.
The die size of the counterfeit parts is obviously smaller that the real parts. The top of the package is sanded off exposing the die for inspection under a microscope. Only 1 or 2 parts have to be tested in this way per lot.
This inspection process may be the most cost effective solution.
As for testing:
I think that you need to verify the power handling capacity of the part, Including thermal resistance from junction to case (verify internal bonding and structure) as well as case to heat sink ( package integrity)
Other MOSFET parameters that affect power handling capacity are Vds, Ron, gate charge, switching speed.
Testing all of these could be expensive, but as in the above inspection you may get by with testing 1 or 2 per lot.
Calculate Switching loss
Switching loss can be hard to calculate because it is a function of many variables, such as MOSFET parameters, methods of driving switching devices, snubbers,, etc.
Unless all of these factors are considered the calculation is at best a rough approximation.
For this reason I recommend the following:
Measure switching + conduction loss
The easiest way around this is to measure the temperature rise of a known good device in an actual working model.
A DC current would then be injected through the device and adjusted to give the same temperature rise on the same heat sink configuration.
The power dissipation (switching +conduction loss) can then easily be calculated by DC current * DC voltage.
You would perform the above test using a know good device in a know good system.
Once you have the DC voltage and Current and temperature rise data for a specific heat sink configuration, You can use this configuration to test new parts.
Your device tester would be a simple test circuit to supply this same DC current and voltage to the device under test mounted to the specific heat sink configuration.
This test could easily verify Vds and on resistance. Measure Gate charge with a capacitance meter and all that is left is switching speed.
Some things to be careful about:
This actual working model must contain the actual drive circuitry and heat sink as well as any other associated circuits like snubber etc.
It is very important to use the device just as it is in the system with the exact same heat sink configuration.
Even changing from thermal compound to an insulating thermal pad can alter the thermal resistance enough to invalidate your results.
Not having the correct Torque on the heat sink mounting screw can alter your results.
If you use a different heat sink or different airflow it will invalidate the results.
For example enclosed in a chassis versus external to the chassis. (invalid)
The Heat sink must have the same thermal resistance, size, mass and airflow as the actual system for this to be valid.
Tester:
The power dissipation (switching +conductin loss) is then DC current * DC voltage.
You would perform the above test using a know good device in a know good system.
Once you had the DC voltage and Current and temperature rise data for a specific heatsink configuration,
Your device tester would be a simple test circuit to supply this same DC current and voltage to the device under test mounted to the specific heat sink configuration.
Pretty simple: