Max Thermal Mass Mag Heatsink Design and SST-90 Build

[gt40]

I have been working on a new design for an all copper heatsink for mags that would provide maximum cooling and thermal mass and using a modded mag switch. It could be adapted to a c size light and probably provide enough mass to cool for high heat emitters like the sst-90 but I am building the first one for a mag d running 3 or 4 CTA NIMH 5500 mah batteries.

The design incorporates an adjustable 5/8" insert that has a set screw to lock it in place. The heat sink is larger then anything else I have seen on here and I am hopeful I can run an sst-90 at at least 10 amps without heat issues even with extended run times. I have also found some silver flat bar that is approximately 1/8" by 1/16th to bend into upgraded terminals to replace the stock mag switchs ones.

Basically they are u shaped and by making them much thicker and out of silver, I should get much lower resistance and headroom.

Here is a diagram of the design:

Here are some pics:

Weight in ounces:

Weight in grams:

I am going to try to wire it up today and will post more pics of the build after I get it together.

Materials:

approximately 3" x 1 3/8" round bar scrap from local ims
2" x 5/8" round bar scrap for insert

Silver flat ribbon approximately 1/8" x 1/16" thick for modding mag terminals

Tools:

Drill
hacksaw
file

 

[Tekno_Cowboy]

I've been working on something surprisingly similar. It stops at the original switch (tower removed), and is made of aluminum. I threaded the insert on mine, but I like your idea of using a set screw instead. With a tight slip-fit and some thermal paste, I'll bet it works better than my threaded version.

I think that your heatsink may be just a hair bit of overkill. Needless to say, I love it :devil:

 

[The Dane]

I've been working on something surprisingly similar. It stops at the original switch (tower removed), and is made of aluminum. I threaded the insert on mine, but I like your idea of using a set screw instead. With a tight slip-fit and some thermal paste, I'll bet it works better than my threaded version.

I think that your heatsink may be just a hair bit of overkill. Needless to say, I love it :devil:

Not so hasty.

A tight mecanical contact is always better than a contact through another medium. A rod inside a cylinder only has mecanical contact at one point opposide the set screw. The rest is thermal grease transfer.

BTW where do you get high bin SST-90's?

 

[fppf]

Yeah, for a sustained thermal load its not about the mass of a heat sink. Mass only effects how long it will take to saturate a heat sink, which is good for low duty cycle type apps. Like if you only use your light for 1 minute every 10.

You don't really need mass, you need to transfer and shed the heat. Skip worrying about how much mass is there and press fit the copper into the tube. That will give much less thermal resistance and turn the whole tube into the heat sink.

You then need to calculate the resistance to heat of each joint, starting at the LED die. The mfgr will spec how many watts the LED will consume, and how many degrees the temperature will rise for each watt. There is a lot to thermal modeling.

 

[HarryN]

Nice work and ideas.

What might be more helpful is to re-name the part a "heat spreader" instead of "heat sink". Thinking of it this way is a bit more realistic of what is really happening.

Everything helps, but in the end, the heat has to go somewhere, which is either into your hand, or into the air. Your body can handle about 10 - 15 watts before it overloads on a normal day. In the winter, or while drinking an ice cold drink - a bit more.

Any heat above this either has to go into the air somehow, or the temperature will rise to a point of discomfort. This is why you often see people attempt to fin their lights. Personally, I think it is worth considering a heavier duty host, like a streamlight or barbolight.

IIRC, an SST-90 is something like 30+ watts, so just plan for it.

 

[gt40]

Not so hasty.

A tight mecanical contact is always better than a contact through another medium. A rod inside a cylinder only has mecanical contact at one point opposide the set screw. The rest is thermal grease transfer.

BTW where do you get high bin SST-90's?

I was planning to make the copper is a really snug fit in the tube so it would seem to be transfering along its total length. I agree about mechanical contacts though.

FYI I got two GN100 stars and two GN200 bare emitters from Avnet:

http://avnetexpress.avnet.com/store/em/EMController?langId=-1&storeId=500201&catalogId=500201&term=SSR%252D90%252DW65S%252DR11%252DGN100&N=0&action=products&x=24&y=6


http://avnetexpress.avnet.com/store...2D90%2DW57S%2DF11%2DGN200&N=0&action=products

Its weird: if Search by sst-90 or ssr-90 only they didn't come up but the actual part code it comes up:

2SSR-90-W65S-R11-GN100 Luminus Devices
PHLATLIGHT WHITE SST-90 LED ON STAR BOARD, CCT 6500K, CRI 70

22SST-90-W57S-F11-GN200 Luminus Devices
PHLATLIGHT WHITE LED, 1000 LUMENS, CCT 5700K, CRI 70

 

[lucca brassi]

Optimal heatsink- cooller should have :

MINIMAL WEIGHT AND MAXIMAL SURFACE & proper orientated cooling by (convection , conduction )

because heat should not store itself in the weight of material , you know equation from school:

Q=m*cp(Cu)*(T2-T1) ; Q=A*t

cp=specific heat

so if you want follow your design , Cu cylinder must be hollow

 

[HarryN]

Optimal heatsink- cooller should have :

MINIMAL WEIGHT AND MAXIMAL SURFACE & proper orientated cooling by (convection , conduction )

because heat should not store itself in the weight of material , you know equation from school:

Q=m*cp(Cu)*(T2-T1) ; Q=A*t

cp=specific heat

so if you want follow your design , Cu cylinder must be hollow

I am not sure that I agree with you on this. If he wants to move heat down the spreader, then it needs maximum cross section. Reducing thermal mass only changes the time constants of transient heating effects, not the thermal conductivity, which he wants to maximize.

BTW - nice machine work on that copper - it is such a PITA to work with for me.

 

[Aircraft800]

There is a great thread here from long ago that showed Thermal Images of different heatsinks. I wish I could find it, it would help you with your design. Anyone remember where it is? It was from back in 2005 or so.

There is another thread I bookmarked about different materials I thought you'd find interesting:
https://www.candlepowerforums.com/threads/63990&highlight=thermal

 

[lucca brassi]

http://www.integratedsoft.com/products/celsius/default.aspx

some of free demo software - celsius , there is many on the net

 

[ktronik]

There is a great thread here from long ago that showed Thermal Images of different heatsinks. I wish I could find it, it would help you with your design. Anyone remember where it is? It was from back in 2005 or so.

There is another thread I bookmarked about different materials I thought you'd find interesting:
https://www.candlepowerforums.com/threads/63990&highlight=thermal

Do you mean Modamag's thermal profile??

https://www.candlepowerforums.com/threads/115232

K

 

[Aircraft800]

Do you mean Modamag's thermal profile??

https://www.candlepowerforums.com/threads/115232

K

Thanks ktronik!! That's the one I was looking for. It shows some interesting info about finning and size of heatsink.

 

[gt40]

Do you mean Modamag's thermal profile??

https://www.candlepowerforums.com/threads/115232

K

Great find...

Update:

I got the light together running 4 c CTA batteries in a mag 3d with the voltage down a tad after letting them sit. Hit 10 amps on the tail and it was only warm to the touch after 10 minutes so I am really happy with the heatsink.

Going to try the heatsink with (2) imr 26650 batteries now in a 2d body wired in parallel. I tested it with the heatsink out of the host with the batteries soldered parallel and hit 11 amps!

It isn't hard to make a battery pack with these IMR's and I think I can get a more compact light albeit with less run time with the imr batteries. 2 of them fit fine in the 2 d host. I will post some pics and results later.