# Calculating Temperature Rise inside an enclosure

Discussion in 'General Electronics Chat' started by MSD, Jun 9, 2009.

1. ### MSD Thread Starter New Member

Jun 9, 2009
2
0
I'm kind of stuck on this problem. Perhaps someone here can help. I'm working on some calculations (thermal, derating, junction temperatures, heatsinking, etc) and some of these depend on the temperature rise, inside an enclosure. That's where I'm stumped... for now. Never was very good with thermal calculations. Assume uniform distribution of power across the circuit board which is also the same diameter as the enclosure.

I'm trying to research this problem but I'm having a bit of trouble finding what I'm looking for.

Problem:

I have to determine the temperature rise, inside a metal enclosure, with no ventilation. This enclosure is exposed to ambient air of 70 deg. C.

The metal enclosure contains a small circuit board with some electronics. The electronics produces about 5W of power, inside the enclosure. The enclosure is anodized aluminum and the size is cylindrical (2" dia, 1" tall). And the electronics are surrounded by a potted compound where I don't know the thermal resistance.

I don't need a specific numerical answer but what I'm trying to arrive at is an equation that can calculate the temperature rise for this situation.

Some questions:

1) Is it possible to even calculate the temperature rise, inside the enclosure, without knowing the thermal resistance of the compound?

2) I assume it's possible, if one makes an assumption about the thermal resistance of the compound. But if that can't be determined, what is an appropriate assumption to make.

3) Most importantly: If perhaps I can find the thermal resistance of the potting compound, can someone provide the equation for calculating the temperature rise, using the thermal resistance of the compound and knowing that it's surrounded with an anodized aluminum enclosure defined above?

If it's not asking too much please post the equation, definitions of the terms, assumptions made and such to avoid me having to ask more questions.

Thanks,

Nov 20, 2008
639
110
3. ### MSD Thread Starter New Member

Jun 9, 2009
2
0
So... let me see if I have this right. I will make some comments and someone can tell me if I have the correct understanding.

Refer to Figure 3 only, in the link in the previous post, as an example of a generic thermal system.

Statement 1 (determining Tj):
Tj = Ta + θca*Pd + θn*Pd + ... + θ2*Pd + θ1*Pd (is that correct, for the example shown?)

Statement 2:
I realize that that model is for a power module. But if one wanted to apply that same model to a single device, could θ1 + θ2 + ... + θn be summed up to represent the thermal resistance that might be what a component manufacturer calls out as θjc (junction to case thermal resistance) and θca might be the case to ambient thermal resistance, including a heatsink? (would that be a correct model).

Statement 3:

In statement 2, I defined that θca "included a heatsink" but if it didn't, would θca be a higher value and adding the heatsink would be a means to reduce θca? (is that correct?).

Statement 4:

If a manufacturer's datasheet has Rθja maximum of 110, then as long as I can keep the total of θ1+θ2+...+θn + θca (or rather θjc + θca) less than 110, then I should be ok.

Statement 5:

If the previous statement is correct and I find that the total is greater than 110, then I would have a problem (correct?). If I understand this model, then anything I add (that would be a series element) would add thermal resistance and wouldn't help reduce Tj. So, if that is correct then I would then have to reduce one of the existing thermal resistance parameters (i.e. θca). I might do that by adding some sort of heatsinking device or cooling. (Is that correct?)

I guess what I'm asking with this statement is, is it correct that anything added adds series thermal resistance and the only thing one can do is to lower one of the areas where the thermal resistance is too high. And one way of lowering the thermal resistance is by adding a heatsink somewhere? (is that a correct interpretation).

Statement 6:

In a theoretically perfect world, if all of the θ numbers were 0 (zero), does that mean that Tj would just be the ambient temperature. I know this is never reality. Just trying to understand everything that has an impact on Tj.

Feel free to split hairs on what I have said. If I've misunderstood even the slightest part, it affects my ability to understand the topic.

Thanks,

4. ### ifixit Distinguished Member

Nov 20, 2008
639
110
Hi,

Re Statment 2:
θca is case to ambient. No heatsink.

Re Statment 3:
Ambient usually refers to ambient air temperature. Adding a heatsink improves the effiency of the transfer of heat from the case into the ambient air around the heatsink by increasing the contact area with the air.

Re Statment 4:
Yes. However, when you operate components at the limit then the life (MTBF) will be shortened.

Re Statment 5:
Correct. To reduce Tj you could; use heatsink, use a better potting compound, reduce power dissipated in j, reduce Ta, etc...

Re Statment 6:
Correct, but Ta must be be kept at or below its maximum. You could use fans to bring in cooler air. The complete thermal solution must consider how the heat in the ambient air will be removed from the building (if this is relevent).

Re Question 1) in your first post...
FR-4 PCBs (Woven glass and epoxy) should not be operated at more than 105C for reliability reasons. If Ta is 70C then the θ of the potting compound will have to be less than 6C/W if 5 Watts is dissipated on the PCB evenly. 6C x 5W = 30C + 70C = 100C.

5. ### sanjayakp New Member

Apr 15, 2010
1
0
I am totally in agreement with ifixit.

You would add thermal resistance of heat sink or the potting material, when present to the total, however the θha(heat sink to air) or the θpa(potting to air) would be small enough as compared to the θca, due to the increased area, that it will overcompensate for the θch(case to heatsink) or θcp(case to potting material).

Please let me know if I am correct.

Last edited: Apr 15, 2010
6. ### retched AAC Fanatic!

Dec 5, 2009
5,201
313
Sorry sanjayakp,

That thread is 10 months old. And the Original Poster hasn't been back since June 10th 2009.