Aluminium heatsink question

[catalin gramada]

Hello
I don't have any appreciation criteria. May I dissipate 45W continuous on this aluminium piece of heatsink if I put a fan on the top please ? What rough temperature to expect please ? Thank you.

[Tom Phillips]

Heat sink design with known heat sink parts such as the photo you posted is done with equations that calculate the temperature rise after you input such parameters as the energy produced by the device, the thermal resistance of the specific heat sink and the ambient environment (Temperature and airflow) and more..

Sometimes it is easier to go with a gut feeling or experiment for simple purposes. If you want to see some design examples I'd suggest looking at some heat sink catalogs which would probably include an application notes section.

Here is a link to google search results that should have what you need.
https://www.google.com/search?as_q=h...filetype=&tbs=

[Helmholtz]

Or mount a power resistor to the heat sink (using thermal grease), feed it 45W (via a lab supply or a variac) and measure temperature.

[catalin gramada]

Is on the way. I ordered already. It have 600 square cm. and want to mount on it two 50w zeners at 180v for no more than total 45w dissipation both. I used 300 square cm of copper fins 0.8mm thick on 12pc 5w / 30v zeners and get 30* C for the same 45w dissipation under fan convection. This is the only term of comparison I have... (meant 3.2w dissipation of a single diode mounted on 25 square cm copper fins)
Aluminium is worse in term of thermal conduction and dissipation compared with copper but a double surface area should compensate... will see...
I ordered also 1mm thick electrotehnic copper sheet to build a proper sink just in case the aluminium profile will not work. A copper sheet letter size is just enough for 40w continuous dissipation with moderate fan convection to keep it at 30C.

[Helmholtz]

For temperature calculation look for the thermal resistance Rth of the heat sink (should be specified by the supplier or datasheet).
Multiplying the Rth (given as K/W or °C/W) by your power gives you the temperature increase above ambient in °C.

[catalin gramada]

Datasheet shows Rth 1.5-0.55K/W for this model depends of the length. For 100mm the graph shows 0.75 value. That meant 30C over the ambient for 40W. Say 60C around with natural convection , vertical disposed.

[Helmholtz]

Seems to make sense for heatsink temperature. Depending on thermal contact the zener temperature will be somewhat higher.

[trobbins]

Heatsink thermal resistance would depend on heatsink orientation, as well as fan forced air-speed. Is that information not known, or can't you identify a similar heatsink with specs that include that information ? Also not identified is the thermal resistance between device and heatsink, or the thermal resistance of a device.

[tedmich]

There have been some good advances in thermal transfer compounds; dont believe ALL the " ceramic nano tubes/ dark matter" hype from the OC computer chip guys but go for the best thermal conductivity (W/m-K) at a reasonable price. Here's a recent review https://www.tomshardware.com/best-pi...-thermal-paste

Most are oils mixed with metal oxides, like beryllium, which conduct heat well but notably NOT electricity
Metals like Ag and Cu do conduct great (thermally and electrically)
Diamond conducts heat best (2000 W/mK! or 5x Cu) but is a bit ...abrasive

[catalin gramada]

Didn't found any infos regard junction to case thermal resistance for my 50w nte diodes. But using a standard mica washer insulator with Rth 0.4K/W seems will raise the case temperature at least 10C for 20W dissipation around

[catalin gramada]

Quick question please : May I use just a 0.1 mm spacer between diode case and heatsink and fill the gap with thermal paste ? I just thinking to get rid of the washer insulator and use just thermal paste as an interface between heatsink and case for better heat transfer but keeping the device electrical isolated, if it have any sense...

[trobbins]

Does the 50W nte diode have a part number or a case style? We haven't yet been informed about the device that is going to use the heatsink, so imho all the queries are rather moot.

[catalin gramada]

Does the 50W nte diode have a part number or a case style? We haven't yet been informed about the device that is going to use the heatsink, so imho all the queries are rather moot.

Yes, thanks. Diodes are NTE5294A. DO-5 case type stud mount. Maybe is better to cut the sink in two pieces and mount direct the diode on it....
The sink is a Fisher sk57 100sa
Will be no more than 22w continuous dissipation on each diode.

[Tom Phillips]

Quick question please : May I use just a 0.1 mm spacer between diode case and heatsink and fill the gap with thermal paste ? I just thinking to get rid of the washer insulator and use just thermal paste as an interface between heatsink and case for better heat transfer but keeping the device electrical isolated, if it have any sense...

You should not try to use thermal paste as an insulation barrier. The mounting stud of the NTE5294A Zener diode is connected to the anode. Therefore, unless you can let the whole heat sink be at the same electrical potential as the anode, you need to use all the parts of a proper mounting kit along with the thermal paste.

[trobbins]

DO-5 zener package has Tj-c of about 125C/50W=2.5C/W, based on ideal heatsink with 0 C/W thermal resistance, and 1N3325 datasheet from American Micro.

DO-5 insulator thermal resistance is circa 0.6 C/W from AN1040.

I'd suggest Tj is kept below 125C, and your ambient may be as much as 50C, indicating a temp rise of 125-50=75C, with a thermal resistance of 2.5+0.6+0.75= 3.85 C/W, so circa 20W.

Imho your main uncertainty is Tj-c, as that can vary with anode or cathode to stud, and it is not specified by the manufacturer (a big red flag).

This scenario is also for one zener per 100mm heatsink, mounted in the 'middle'.

You may want to use your fan airflow margin, however you need to have a good handle on your ambient, given that you may have multiple zeners heating up the local air.

20W with a 150V zener is 130mA, which is a lot more than the reference current (circa 85mA), so you may be noticeably over 150V.