# how do they rate heat sinks ... what is better?

#### xtal_01

Joined May 1, 2016
169
So, I have a couple of linear regulators ... T220 case ....

Just need some small heat sinks on them.

I did a quick calculation ... if everything in my little project was on at once, you could hit 1.2W ... though this would never happen.

Anyway ... looking at heat sinks and confused by the ratings ...]

For example ...

1) heat sink rated at 56 deg C @ 2 W

2) same heat sink with just a bit higher (so I think it should dissipate more heat) 50 deg C @ 2W

So are they saying the part will keep 6 deg cooler at 2 watts? ... or am I getting less cooling at 2 watts ????

Then I see some having a rating like 28 deg C / W (natural) ... I assume this is the same at 56 deg C @ 2 watts in free air?

Don't need a lot of cooling but may as well get as much as I can.

Thanks!

#### LowQCab

Joined Nov 6, 2012
3,937
That's XXX Degrees Temperature increase above Ambient-Temperature / per Watt(s) of Heat-Input.
.
.
.

#### MrChips

Joined Oct 2, 2009
30,467
You calculate the thermal resistance required by your device.

For example, let us assume:
ambient temperature TA = 25°C
junction temperature TJ = 50°C
power dissipation PD = 2W

thermal resistance = (TJ - TA) / PD = (50 -25)/2 = 12.5 °C/W

You have to use a heat sink with a thermal resistance that is less than 12.5 °C/W.

#### xtal_01

Joined May 1, 2016
169
Thanks ... this reverse engineering is killing me ... way tooooooo many questions.

As I said in my other posts, the designer / builder retired and passed. Half the parts are obsolete.

Thanks!

#### MisterBill2

Joined Jan 23, 2018
17,790
One place I worked at took on a very qualified EE as a temporary way to work around having more work than we could handle, even with ten hour days. He designed one system and calculated the heat sink precisely, along with the rest of the system. It worked perfectly until our nasty 24 hour burn in check that every system received. The problem was just a small amount of overheating of the LM7805 voltage regulator. The heat sink calculations did not include the system enclosure. Fortunately there was room inside the enclosure, so the exact heat sink was replaced by a 2 inch by 3 1/2 inch piece of 3/16 aluminum, screwed to the system housing. Problem solved.
Of course this was not a production run, it was a one-off package for a repeat customer who expected correct solutions every time..
My point being that on many occasions using a larger heat sink than the math requires avoids problems caused by missing something.

#### MrChips

Joined Oct 2, 2009
30,467
Thanks ... this reverse engineering is killing me ... way tooooooo many questions.

As I said in my other posts, the designer / builder retired and passed. Half the parts are obsolete.

Thanks!
I do the finger test. Run the device for a while (10-60 minutes). Touch it with your finger. If you say "ouch" you need a bigger heat sink.

#### MisterBill2

Joined Jan 23, 2018
17,790
Thanks ... this reverse engineering is killing me ... way tooooooo many questions.

As I said in my other posts, the designer / builder retired and passed. Half the parts are obsolete.

Thanks!
If you can understand enough of the circuit to have an understanding of the current through whatever needs heat sinking, and see what voltage is across it, then you can know what power is converted to heat that must be carried away..
The problem becomes hard when the heatsink rating is for heat transfer in moving air and the application condition is still air.
Years ago one manufacturer posted a table of "watts of heat" that could be transferred to surrounding air at a given temperature for each of their heat sink styles. THAT was an incredibly useful table, even though it only referenced one set of heat sinks. A comparison with other brands allowed estimating the performance of other products. as well. If such a table were available today that would be the simple solution to your question.

#### LowQCab

Joined Nov 6, 2012
3,937
The cooler You keep your Semi-Conductors, the more reliable they will be.

You might be able to get away with running at ~100C for a while,
until some odd extenuating circumstances rear their ugly heads, then You have instant Blue-Smoke.
( Not to mention that, all sorts of important Specifications tend to drift with Temperature ).
.
.
.

#### MisterBill2

Joined Jan 23, 2018
17,790
Certainly posts #8 and #9 do apply. Excess heat dissipation capability is one place where more than is the minimum is the way to go.
At least within reason. And certainly the application of thermal coupling material is required.

A side note is that the recovery of heat sinks from failed electronic assemblies can be a cheap source of assorted heat sinks. A used heat sink was not subject to failure. That is to say, unlike other components they are easily re-usable and do not "burn out." Old CRT monitors and solid state/CRT televisions are fair sources of used heat sinks.

#### xtal_01

Joined May 1, 2016
169
Certainly posts #8 and #9 do apply. Excess heat dissipation capability is one place where more than is the minimum is the way to go.
At least within reason. And certainly the application of thermal coupling material is required.

A side note is that the recovery of heat sinks from failed electronic assemblies can be a cheap source of assorted heat sinks. A used heat sink was not subject to failure. That is to say, unlike other components they are easily re-usable and do not "burn out." Old CRT monitors and solid state/CRT televisions are fair sources of used heat sinks.
I did in fact add silicon compound to my list ... there was none on the original. Thanks!

#### Ya’akov

Joined Jan 27, 2019
8,973
I did in fact add silicon compound to my list ... there was none on the original. Thanks!
Be certain not to apply excess heatsink compound. Unlike applying excess heatsink, too much thermal paste is possible. The function of thermal compound is to increase the part-to-heatsink contact area by filing very small irregularities in the surfaces of each. Think tool marks and scratches on what looks like a shiny surface.

Too much compound actually acts as an insulator, thermally isolating the part and heatsink. Thermal paste cannot bridge a gap.

Put a small amount on the part side and with a flexible tool scrape it to a thin film. This will be the right amount. This is why lately I generally prefer the silicone pads that are applied and then compressed by the mounting hardware.

They aren’t subject to over-application (so long as you only use a single layer without overlap), and they have the added benefit of being much, much cleaner in application and replacement. They are also cheap, a sheet like above under $2.00USD. Thread Starter #### xtal_01 Joined May 1, 2016 169 Be certain not to apply excess heatsink compound. Unlike applying excess heatsink, too much thermal paste is possible. The function of thermal compound is to increase the part-to-heatsink contact area by filing very small irregularities in the surfaces of each. Think tool marks and scratches on what looks like a shiny surface. Too much compound actually acts as an insulator, thermally isolating the part and heatsink. Thermal paste cannot bridge a gap. Put a small amount on the part side and with a flexible tool scrape it to a thin film. This will be the right amount. This is why lately I generally prefer the silicone pads that are applied and then compressed by the mounting hardware. They aren’t subject to over-application (so long as you only use a single layer without overlap), and they have the added benefit of being much, much cleaner in application and replacement. They are also cheap, a sheet like above under$2.00USD.
WOW ... awesome to know! Didn't think too much .... Thanks!