0
\$\begingroup\$

I am trying to understand different manufacturers datasheets PD calculations for mosfets in a similar class.

The first is ON Semiconductor NTD5865NL.

enter image description here

The second is Diodes Inc DMNH6021SK3Q

enter image description here

What does the 71W value really represent, and how would one go about comparing apples to apples here?

\$\endgroup\$
3
  • 2
    \$\begingroup\$ In one case you are magically holding the case at 25 C. In this case, there is 2.1 C/W rise from case to junction, and 71 W implies an added 149 C to the 25 C ambient, if I'm reading it right. So they are saying they can handle a junction of 175 C, or so. In the other case, it's mounted on FR-4 and exposed to ambient air. It also gives a 175 C maximum, but they are accounting for a different circumstance. Don't you see this? \$\endgroup\$
    – jonk
    Commented Aug 19, 2021 at 19:52
  • \$\begingroup\$ You can compute that big number from OnSemi:$$\frac{175\:^\circ\text{C}-25\:^\circ\text{C}}{2.1\:\frac{^\circ\text{C}}{\text{W}}}\approx 71.43\:\text{W}$$And use a similar formula for the Diodes' part if you want. It's not very useful because you aren't likely to clamp a bar of aluminum around it with circulating liquid nitrogen to cool it. \$\endgroup\$
    – jonk
    Commented Aug 19, 2021 at 20:34
  • 1
    \$\begingroup\$ "What does the 71W value really represent?" : Virtually nothing. \$\endgroup\$
    – dandavis
    Commented Aug 19, 2021 at 21:07

3 Answers 3

4
\$\begingroup\$

The footnotes and thermal resistance are key to comparing performance.

The ONsemi has a lower RdsOn, while the Diodes In has a lower Rjc which when mounted on 2oz copper gives it also a better Rja 40’C/W vs 49 (Onsemi).

The 71W infinite heatsink test of 25’C defines the max power before gradient thermal stress may cause damage. That might be useful for relating transient SOA spikes of power but never for steady-state use.

The reliability is affected by the maximum junction temp rise you choose, usually below 100’C. Then you design the heatsink according to the enclosure and board copper area and thickness.

To compare apples multiply the RdsOn by the Rja for the layout you choose.

For example 2oz copper FR4 per description details:

          Ron (10V).   Rja (2oz).     Product.  
 ONsemi.    16m       49’C/W.(650mm2)   784
 DiodesInc. 23m.      40’C/W.(1sqin.)   920      

This makes the ONsemi a slightly cooler junction for same current squared.

The units of the product are mOhm ‘C/mohm-A^2 or \$degC/A^2\$ for the junction rise above ambient.

\$\endgroup\$
3
\$\begingroup\$

71W represents the power dissipation of the MOSFET when it reaches the maximum junction temperature at a fixed case temperature of 25°C. The calculation is quite simple: $$P_D=(T_{max}-T_{case})/R_{th,jc}$$

For a comparison of thermal performance of different components make sure that the conditions for the power dissipation are the same or compare them based on thermal resistance \$R_{th,jc}\$.

\$\endgroup\$
3
\$\begingroup\$
  • 1st case is the maximum allowed power dissipation if the case temperature is 25C. Now how do you cool the case to be steady 25C is one another question, probably by using a large copper plate with cooling pipes and a chiller.

  • 2nd case is a transistor mounted on a PCB with minimal footprint. It's more realistic than 1st one.

The meaningful data is the Rth junction-to-case, then it's up to you to make enough large copper footprint or to mount the heatsink - Rth case-to-air

  • 1st RJC 2.1 °C/W
  • 2nd RJC 1.8 °C/W

You can see that the second one has a better thermal conductivity, even if at first look (as you found) it seems worse.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.