Heat sink thermal resistance when there is no force air cooling

[akh_power]

Hi,
I want to select a heat sink for a DPAK MOSFET. The MOSFET is in a PCB which is inside an enclosure without any forced air cooling. So, I think the air flow rate is 0. From the datasheet of one of the heatsinks, I see the following thermal resistance chart.

The top horizontal side of the box is for air flow scale. The right side vertical side for the thermal resistance. From this graph it can be seen that the thermal resistance data is not available for air flow rate less than 200 LFM for the dotted thermal resistance line. Similarly, for the thick thermal resistance line the data is not available for the air flow rate approximately below 70 LFM. What does it mean?

 

[FvM]

Read the diagram so that thick bottom line is describing temperature rise versus power dissipation without forced cooling.

 

[cupoftea]

Yes, the little wee orthogonal arrows point to the scales you should be looking at for each line.

 

[dick_freebird]

There could be (or not) some convective ("chimney") cooling
in a non-forced case but it's liable to be "not much", though
non-zero.

There will be an internal heat capacity in the sink body, for
transient thermals. This chart looks like steady-state with
no pulse-width / duty-cycle family.

 

[FvM]

Said chimney effect can be seen in the non-linear performance curve of post #1 diagram. I presume it's measured with infinite space above the heatsink. In a small enclosure, convection can't build with same intensity up as it does in open space. Thus thermal resistance will be larger in small enclosure. Increase of internal enclosure temperature is another effect.

 

[D.A.(Tony)Stewart]

Good for < 0.5W with no heatsink and not enclosed and 2.5W with high velocity air flow. using 80 rise max.

 

[KlausST]

Hi,

even if no "forced" air flow, there usually is "gravitational" air flow. (Except in non gravitational environment).
And there is conductive thermal transport via PCB and copper and so on.

You say enclosure, this may be a low thermal conductive plastics one, or a high thermal conductive metal one.
The enclosure could be big with much space for air flow, or it could be small with reduced air flow inside.
It could be vented or not. We don´t know.


Klaus

 

[akh_power]

Read the diagram so that thick bottom line is describing temperature rise versus power dissipation without forced cooling.

My mind was stuck on the thermal resistance. Now looking again at the diagram, I understand that there is either the temperature rise information or the thermal resistance information for all the air flow rate and power. And either one of them is enough to select the heat sink. Thanks for the help.

 

[D.A.(Tony)Stewart]

Hi,

even if no "forced" air flow, there usually is "gravitational" air flow. (Except in non gravitational environment).
And there is conductive thermal transport via PCB and copper and so on.

Klaus

I never used to think about the forces of convection heat being anti-gravity, which it is, as the less dense gas rises as long as there is a free air supply beneath for the chimney effect to be effective. Using this effect will enhance air speed greatly with vents for in and out in a vertical orientation. A spiral intake enhances this even more.

Sealing it tight is like making an oven.

A lighter might produce a ball of flame in a space capsule and sputter out with a lack of airflow.

--- Updated Aug 25, 2024 ---

Using 3.5m/s or 700 LFM air speed over the fins does not need high volume flow just speed, but will reduce the temperature rise as much as reducing >=95'C/W to 12'C/W for a D-PAK SMD. I suspect even better with a TO-220.

Consider 3.5m/s without heatsink Rja = 31 'C/W and without forced air is 95'C/1W
But with heatsink 3.5 m/s is Rja= 12'C/W and without forced air is 40'C /1W

I discovered this benefit during a design of a 1U high Rack with a 180 W PSU using an eddy current intake to lower the gap of air rushing too high over the hot spots to effectively cool them, e.g. SMPS power transformer was reduced by 50'C with dual 1 7/8" fans with high-velocity airflow at the hot surface.