Does anyone know of a good on-line heatsink calculator that works with finned heatsinks and natural convection?
This heatsink is really common on Ebay, but has no spec for °C/Watt.

Number of Fins: 27
Overall Size: 150 x 69 x 36mm/5.91 x 2.72 x 1.42 inch (L x W x H)
Base Plate Thickness: 4.8mm/0.19 inch

The nearest thing to a technical spec is a helpful hint from the seller "Use with a fan, the effect is better".

 

The spacing between fins is somewhat too small for forced cooling, the back pressure will be significant. What wattage and case temperature are you aiming for?

 

It looked like an alternative to this one
https://www.fischerelektronik.de/web_fischer/en_GB/VA/SK100150ME/datasheet.xhtml?branch=heatsinks
and I wondered how it would perform.
Someone has gone to a lot of trouble to make an extrusion die for fins so fine!

 

I see very helpful the thumbrule: Natrural convectiojn in air with large between-ribs distance. The permitted overheating dT and corresponding cm2 for each thermal Watt. 10C=170 cm2; 20=65; 30=40; 40=26; 50=19.5'60=12; 70=9.8

In case when ribs are much closer but not still so dense as for forced cooling, better works formula dT=10(0.8/Vol^0.68)*sqrt(N), where N =[W] and Vol=[litres].

Less accurate but more universal is formula R(th)=400/A(cm2), where R(th)=[C/W]

 

Hi Ian.
This video has some information related to your H/S project.
Approx 16 minutes in, there is a useful H/S graph, followed by an LTS simulation.

E

 

Hi Ian.
This video has some information related to your H/S project.
Approx 16 minutes in, there is a useful H/S graph, followed by an LTS simulation.

E

Thanks. That was the opposite of what I was expecting. I thought that the natural convection would be affected less by the stagnation due to close fin spacing. I wonder if is also affected by orientation.

 

I wonder if is also affected by orientation.

Hi
For non fan applications, I have found that the Fins in a vertical orientation give the best cooling result effect.
Ideally, stood off the supporting surface using right-angled brackets, with about 1 inch above board clearance.

E

 

Hi
For non fan applications, I have found that the Fins in a vertical orientation give the best cooling result effect.
Ideally, stood off the supporting surface using right-angled brackets, with about 1 inch above board clearance.

E

Yes. I always used to do the back of amplifier cases like that.
Fins horizontal isn’t very good, but I was wondering about the efficiency with the fins vertical but the mounting plate horizontal.

Do you remember the first pin heatsink, that came out in the 80’s aimed at the amplifier market, with pins about 8mm diameter, and it looked as though some bloke had made a wooden former out of dowels then sand-cast it in his kids’ sandpit?

 

Airflow through fins is best if the base and fins are vertical - this is orientation for which JEDEC test scenarios (and official degC/W specs) are given. Base horizontal, fins vertical, will be less efficient but much depends on the heatsink design.

What are you planning to mount to heatsink and how much wattage?

 

Airflow through fins is best if the base and fins are vertical - this is orientation for which JEDEC test scenarios (and official degC/W specs) are given. Base horizontal, fins vertical, will be less efficient but much depends on the heatsink design.

What are you planning to mount to heatsink and how much wattage?

It was going vertical (fins vertical, plate vertical) but dissipation is still undecided (somewhere in the 40W to 80W region) I'm still thinking about:

• the final specification for the product
• how many MOSFETs in parallel
• whether to use a fan
• whether use a transfer bracket to a steel chassis.

The only thing I do know so far is the thermal conductivity of the steel chassis, because there are plenty of flat plate heatsink calculators around. It's 2.6K/W so I'm not going to get very far with that idea.

 

@Ian0

A quick thermal simulation (CHT) with an ambient of 25C mounted horizontally suggests something in the order of 0.4K/W based on a 100W of heat flux across the complete base surface.. For a centrally mounted 40W MOSFET in a TO-247 pack it (contact area 20 x 16mm) would be around 0.7K/W

May be worth getting one for testing?

 

@Ian0

A quick thermal simulation (CHT) with an ambient of 25C mounted horizontally suggests something in the order of 0.4K/W based on a 100W of heat flux across the complete base surface.. For a centrally mounted 40W MOSFET in a TO-247 pack it (contact area 20 x 16mm) would be around 0.7K/W

May be worth getting one for testing?

I already have some. I used one on a bass guitar combo where it would have been overspec’ed by a wide margin, so there was nothing to learn from that.
I’ll give it a whirl on the new application and add a fan if it’s too warm.