# Do I need a heatsink? How big?

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
(*steve*) submitted a new resource:

Do I need a heatsink? How big? - A simple inroduction to heatsinking components

0. Introduction
• Did you burn your finger on a transistor?
• Are you wondering why your voltage regulator drops to a low voltage after a while?
• Did the smoke come out of a 100W transistor when it was only dissipating 30W?
• Did someone say "You need a heatsink?
And...
• You don't know why.
• What is heat and how does it relate to temperature?
• What a...

#### Arouse1973

Dec 18, 2013
5,178
What example can you show where heat is dominant over temperature for components? I guess you mentioned this for a reason.

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
The example will be when you add a heatsink. Not only does it increase surface area for dissipation of heat, but the entire system now has more thermal mass meaning that it will take more heat to reach a given temperature even if we disregard the dissipation of heat into the environment.

Heat is never really dominant over temperature (did I say that in the resource?) but the larger the thermal mass, the slower the system will react to a given heat input. This is good for things like intermittent operation. Whereas thermal resistance can be seen as resistance, thermal mass looks like capacitance. So with these resistors and capacitors we end up with a low pass filter. If the resistances are low enough and the capacitances high enough, they will load down the input. Not good for an amplifier, but great for a junction with a peaky power dissipation.

Will I cover this? I'm not sure. Maybe it will get edited out if I don't.

Care to comment on phase change being listed as a 4th method under convection, radiation, and conduction? It's a bit of a stretch (and arguable whether "phase change" is the best term).

#### Arouse1973

Dec 18, 2013
5,178
The example will be when you add a heatsink. Not only does it increase surface area for dissipation of heat, but the entire system now has more thermal mass meaning that it will take more heat to reach a given temperature even if we disregard the dissipation of heat into the environment.

Heat is never really dominant over temperature (did I say that in the resource?) but the larger the thermal mass, the slower the system will react to a given heat input. This is good for things like intermittent operation. Whereas thermal resistance can be seen as resistance, thermal mass looks like capacitance. So with these resistors and capacitors we end up with a low pass filter. If the resistances are low enough and the capacitances high enough, they will load down the input. Not good for an amplifier, but great for a junction with a peaky power dissipation.

Will I cover this? I'm not sure. Maybe it will get edited out if I don't.

Care to comment on phase change being listed as a 4th method under convection, radiation, and conduction? It's a bit of a stretch (and arguable whether "phase change" is the best term).

No you didn't say heat was dominant, it's just you said is heat our enemy? Which I took as it could be at some point. Regarding phase change, well I suppose you could say it does release heat. After all steam when it condenses releases heat, the same heat that caused it to change state in the first place. But I am no metal expert but I assume it's game over for the component once it has changed state (Phase change) because this would be at a very high temperature. One thing I am not sure on is what happens to a metals thermal inertia when it's changes phase, I don't mean changed to liquid but at certain temperatures the crystalline structure can change several times before it reaches melting point.

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Each of these phase changes will have a latent heat associated with them. The heat is dissipates by breaking bonds between molecules.

In a very loose way, things like causing ions to diffuse through the semiconductor could be seen as a type of phase change (perhaps)

Still wondering if there's a better way to put it.

Anyway, I've added a little more, but there's plenty to go.

#### Arouse1973

Dec 18, 2013
5,178
Looking good Steve.

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Almost half way there...

#### KrisBlueNZ

##### Sadly passed away in 2015
Nov 28, 2011
8,393
This is great stuff Steve! I'll post some minor comments when I get to the end.

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Thanks Kris. I have some more stuff lined up to put in whenever I get time to scratch

Jun 21, 2012
4,886

Moderator
Jan 21, 2010
25,510
ok...

Jun 21, 2012
4,886

May 20, 2017
1,217

Replies
22
Views
254
Replies
6
Views
172
Replies
5
Views
245
Replies
10
Views
273
Replies
4
Views
194