same heatsink, different datasheets

[jazz]

Hi,

I'm a bit puzzled by different heat resistance ratings for virtually identical heatsinks from Seifert and Fischer.

The sk-130 from Fischer and the kl-240 from Seifert differ only in the thickness of the baseplate as far as I can tell from the datasheets. Yet Fischer gives a heat resistance rating which is about double that of Seifert (~,5 K/W versus ~,25 K/W for 100mm heigth).

Which am I to trust?

regards,
Joris

Link:
http://www.fischerelektronik.de/index.php?id=114
http://www.digtion-medien.de/seifert...punkte/pdfs/46[0].pdf

Edit: can't seem to get the links working directly, sorry

[AndrewT]

you must look out the information sheets from both manufacturers where they define the test conditions used to rate their heatsinks.

The biggest difference is made by choosing a different DeltaT s-a.
But this, with other conditions, could account for this apparent 2:1 discrepancy.

BTW,
I could not link to the digtion site.

[Mr Evil]

HTML entities let you link to files with stupid characters in their names.

Anyway, I thought it was the other way around and manufacturers chose a particular power to apply to the heatsink and then measured delta-T, rather than choosing delta-T.

Another factor that can make a huge difference is the area they apply the heat to. If they simulate a tiny TO-220 package, the thermal resistance will be higher than if they use a larger area like a TO-264.

[EUVL]

You have luck, because this is exactly the heatsink profile I used extensively for my Class A power amps, such as here :

http://www.diyaudio.com/forums/showt...t=#post1744005

I used Seifert, but it is exactly the same as Fischer. The Seifert thermal data are attached.

http://www.seifert-electronic.de/en/index.php

(you need to do some surfing to get to the right page for KL240)

[EUVL]

The fischer :

http://www.fischerelektronik.de/index.php?id=114

(just type sk130 and search)

[EUVL]

The difference can easily be explained by different assumptions such as dT (heatsink temperature above atmospheric).
Neither figure is close to reality if you are trying to operate at say 55°C.

A third supplier who produced the same profile was kind enough to do a full blown simulation for my application (length 550mm, free convection, black anodised, vertical), and you can see how things depend.


Patrick

[AndrewT]

Originally posted by Mr Evil
Another factor that can make a huge difference is the area they apply the heat to. If they simulate a tiny TO-220 package, the thermal resistance will be higher than if they use a larger area like a TO-264.

indeed.
Most manufacturers will try to show their product with the best figures. They achieve this by taking the whole backplate surface up to the test temperature (isothermal) and then finding what power the sink can dissipate.
It would be very unusual for big heatsink to use published data based on a single device that left the remaining surface exposed to ambient.

Simulators can do this modeling for you if you want to see how badly the sink can dissipate when it has a few hotspots where the device/s are located.

[EUVL]

All manufacturers assume that the flat side has uniform heat flow. This is of course not reality. It is up to the user to provide further heat spreading devices in case such a large area heatsink is used with few heat generating spots. But this is easily done.


Patrick

[jazz]

Hi,

Thanks for the answers and the working link.

I searched a bit more for additional data on measurements and ratings but I could only find a statement on the K/W rating for Seifert.

They state that the rating is for black anodised, free standing with the fins vertical in static air conditions. The ratings given should be good for practical situations they say. This is the same condition I intend to have and I suppose it means I should be able to use them.

I intend to dissipate 100-130 W per heatsink and allow for a 30deg.C temperature rise. I will also use a copper heat spreader of 20/8/1cm (w/h/d) located at one end of the heatsink. Each Fet will dissipate about 20W. The backplate of the heatsink will not be in free air but the case will probably make up for that.

I'll just have to hope it works out ok. The data from Fischer have me doubting quite a bit though.

regards,
Joris

[EUVL]

In my experience, the values provided by Fischer is on the optimistic side if you want to work with a 30°C temperature difference. And you would almost certainly NOT be able to reach the values stated by Seifert.

Even with a height of 550mm instead of 100mm, the heatsink I used can only dissipate 220W with a temperature rise of 40°C, or 0.17K/W (proven by measurements). I know that height above 150mm does not bring a lot, but to hope to get 0.23K/W with a height of 100mm is IMHO a bit unrealistic.

But prove me wrong, by all means.


Patrick


PS A fan running at below nominal voltage will work wonders and can hardly be heard even at close distance. Try one of these :

http://www.noctua.at/main.php?show=p...s_id=26&lng=en

Even blowing on the flat surface (where the transistors are mounted) will help quite a bit.