High Level Heatsink Design

[RedCowboy]

It will vary greatly depending on ambient temp, I see a big difference in my static heat sinks between summer and winter when used outside........the best way to negate this is forced air, if you look at the projectors they use fan over fins and later models use heat piping and radiators but that's also a matter of remote locating and package but just a static heat sink would need to be quite large to dissipate 15w continuously, that's why so many things use fans......now if the max operating temp had enough differential, if the diode could run at 350F then it could get by with a smaller static sink, but as in summer time the differential is so small and you want continuous 100% duty cycle you need a fan, or compromise on your cool down........something I used to do is let one unit cool while lasing with another.

Also if you build several you can fire them all up at once
p.s. There's a beam stop attached to the door in my pic.

 

[Cyparagon]

I use these $1 12mm Aluminium rail-holders to interface with a standard 12mm diode holder and bolt them to an old CPU cooler. I've run numb44s at 4A for several hours at a time without a problem

 

[thebouljello]

I use these $1 12mm Aluminium rail-holders to interface with a standard 12mm diode holder and bolt them to an old CPU cooler. I've run numb44s at 4A for several hours at a time without a problem

That sounds like a nice setup. I assume the CPU cooler uses fans? If I could find (or design) a single heatsink to run the NUBM44 continuously, that would be ideal.

I know there are differential equations that govern heat dissipation and whether the heat emitted exceeds the heat dissipated. How to derive and solve those equations for a specific operating point are a bit out of my scope though.

 

[RedCowboy]

That sounds like a nice setup. I assume the CPU cooler uses fans? If I could find (or design) a single heatsink to run the NUBM44 continuously, that would be ideal.

I know there are differential equations that govern heat dissipation and whether the heat emitted exceeds the heat dissipated. How to derive and solve those equations for a specific operating point are a bit out of my scope though.

You can learn how to calculate what you will need or just take a stab at it and see what you get, I thought you didn't want to use a fan but if you now do then you can test a basic setup and see how well it works, the cooler I 1st recommended would likely do just fine and the linear motion mounts Cyparagon listed make good thermal contact with our favorite modules, you could add a tec pad but it's probably not needed unless you are going to be running it in a hot environment. Here's some reading if you really want to try to calculate it all out.

How Much Heat can be Extracted from a Heat Sink? | Electronics Cooling

Thermal transport from a heat sink creates a unique condition in heat transfer and fluid flow once it is placed on a PCB in an unducted fluid flow delivery system. Although a heat sink may appear structurally simple, the fluid […]

www.electronics-cooling.com

The Physics Classroom Tutorial

The Physics Classroom Tutorial presents physics concepts and principles in an easy-to-understand language. Conceptual ideas develop logically and sequentially, ultimately leading into the mathematics of the topics. Each lesson includes informative graphics, occasional animations and videos, and...

www.physicsclassroom.com

 

[Cyparagon]

Heat sinks, and thermal interfaces in general, are rated in terms of °C/W. For instance, with a 10°C/W heat sink, will rise 10C above ambient for every watt that is dissipated through it. A 5W load will be 50C rise. A 1°C/W heat sink at 5W would only be 5C rise. Fan cooling greatly reduces the temperature rise, as you'd expect, but similar results can also be obtained by just using a larger heat sink. CPU coolers regularly dissipate a 30-280W load with temperature rise of under 50C with airflow. The 18W load of a laser diode is small by comparison, and some of the larger ones will not even need a fan.

 

[RedCowboy]

On a 75F spring/summer day how hot do we want to allow our device to get, 125F ? So that's a 25C rise and without a fan I have found the heat sink saturates in time, larger = longer but in 10 minutes I have heated up a very large sink, much too large to be what our bunch typically calls hand held, granted with even a small fan it's a different story but OP was asking about continuous duty so if the constraints are a continuous 12-15w heat load and a compact portable build then I don't recommend a static sink, a small fan would be more reasonable if unwilling to compromise on duty cycle.........but again this is for a 100% duty cycle model.

I am curious, how big of a sink with how much finned surface area would it take to keep a 15w heat source below 126F on a 75F day without any forced airflow ? .......... maybe something like the sink that Drake aka Styro used in his 8 x numb05 build ? Actually for continuous duty that may well exceed 125F in 15-20 minutes, granted if the upper limit was 175F that would make it easier but the range we want to maintain is what makes continuous duty more of a challenge to keep compact without any active cooling.

If the plan is a nubm44 box build with a fan cooled finned sink then no problem, but a flashlight style with just a large sink won't do continuous even with a huge sink if no fan, not on a 75F day and the limit of 125F-continous.

 

[thebouljello]

On a 75F spring/summer day how hot do we want to allow our device to get, 125F ? So that's a 25C rise and without a fan I have found the heat sink saturates in time, larger = longer but in 10 minutes I have heated up a very large sink, much too large to be what our bunch typically calls hand held, granted with even a small fan it's a different story but OP was asking about continuous duty so if the constraints are a continuous 12-15w heat load and a compact portable build then I don't recommend a static sink, a small fan would be more reasonable if unwilling to compromise on duty cycle.........but again this is for a 100% duty cycle model.

I am curious, how big of a sink with how much finned surface area would it take to keep a 15w heat source below 126F on a 75F day without any forced airflow ? .......... maybe something like the sink that Drake aka Styro used in his 8 x numb05 build ? Actually for continuous duty that may well exceed 125F in 15-20 minutes, granted if the upper limit was 175F that would make it easier but the range we want to maintain is what makes continuous duty more of a challenge to keep compact without any active cooling.

If the plan is a nubm44 box build with a fan cooled finned sink then no problem, but a flashlight style with just a large sink won't do continuous even with a huge sink if no fan, not on a 75F day and the limit of 125F-continous.

This is an interesting question. I wonder how large of a heatsink will be needed to run a NUBM44 continuously at 7W? At about 4.5W? Even if we do not know for sure, I wonder what ballpark figure for the size would allow for a 100% duty cycle.

Not sure if I already mentioned it, but has anyone used graphene heatsinks? They apparently beat copper by a long shot.

 

[thebouljello]

You can learn how to calculate what you will need or just take a stab at it and see what you get, I thought you didn't want to use a fan but if you now do then you can test a basic setup and see how well it works, the cooler I 1st recommended would likely do just fine and the linear motion mounts Cyparagon listed make good thermal contact with our favorite modules, you could add a tec pad but it's probably not needed unless you are going to be running it in a hot environment. Here's some reading if you really want to try to calculate it all out.

How Much Heat can be Extracted from a Heat Sink? | Electronics Cooling

Thermal transport from a heat sink creates a unique condition in heat transfer and fluid flow once it is placed on a PCB in an unducted fluid flow delivery system. Although a heat sink may appear structurally simple, the fluid […]

www.electronics-cooling.com

The Physics Classroom Tutorial

The Physics Classroom Tutorial presents physics concepts and principles in an easy-to-understand language. Conceptual ideas develop logically and sequentially, ultimately leading into the mathematics of the topics. Each lesson includes informative graphics, occasional animations and videos, and...

www.physicsclassroom.com

Sorry for the late reply. I will definitely take a look at these.

 

[Lifetime17]

Hi,
For the NUBM44 diodes a very large heat sink is required as we all know from experience .
Rich

 

[thebouljello]

Hi,
For the NUBM44 diodes a very large heat sink is required as we all know from experience .
RichView attachment 68190

Most definitely. Can you run yours continuously in that heatsink?

 

[Lifetime17]

Hi,
Well let me be honest about duty cycles, As I said time and again we run these diodes well out of specs that they were intended for from the manufacturers point .
I have been a supporter of the 30on/30 off but sometimes you can break the rules. Treat them right and they will last with proper heat sinking aluminum, aluminum with a copper core or full copper heat sinking. and this depends on the diodes heat distribution . So there's very little math to do respectively . I hop this sheds some light on heat sinking issues.
Rich

 

[thebouljello]

Hi,
Well let me be honest about duty cycles, As I said time and again we run these diodes well out of specs that they were intended for from the manufacturers point .
I have been a supporter of the 30on/30 off but sometimes you can break the rules. Treat them right and they will last with proper heat sinking aluminum, aluminum with a copper core or full copper heat sinking. and this depends on the diodes heat distribution . So there's very little math to do respectively . I hop this sheds some light on heat sinking issues.
Rich

Thanks for the reply. It does help. I think I'll opt for a lower operating point and use a larger heatsink, which would add to the aesthetic and give the laser a good feel. You may be wondering why I would buy a NUBM44 and operate it below 7W, and it's because the highest efficiency is achieved around 3W or so.

 

[Cyparagon]

This is an interesting question. I wonder how large of a heatsink will be needed to run a NUBM44 continuously at 7W? At about 4.5W?

Have you been ignoring my posts? I've given you the things to google. You're not even trying. Here's one I found in 40 seconds on digikey:
LGA1156 cooler

I need some maths practice anyway, so here we go:

The fan-cooling dissipation isn't listed, but the natural convective specification is 0.37°C/W
A NUBM44 4.5V at 4.5A is 20watts, minus the 7W leaving as light means 13 watts of heat.
Negating the thermal transfer through however you interface the laser diode with this, that is a temperature rise of only 5C.
That's about 9F for the muricans out there. That's plenty cool enough for most people.

You can use an even smaller heat sink. In 2 minutes, I sorted the list on digikey and found something close to their minimal size
if we're allowing 25C rise as arbitrarily requested above by Mr Cowboy. This one at 76x122x33mm (~5x3x1.4 inches for the muricans)
is at 2C per watt, so we've got 26C temp rise in theory. Or 120F heat sink temp in a 75F room for instance. That's with only 10 fins.

Most CPU heat sinks will cool a laser diode fine without a fan.

 

[thebouljello]

Have you been ignoring my posts? I've given you the things to google. You're not even trying. Here's one I found in 40 seconds on digikey:
LGA1156 cooler

I need some maths practice anyway, so here we go:

The fan-cooling dissipation isn't listed, but the natural convective specification is 0.37°C/W
A NUBM44 4.5V at 4.5A is 20watts, minus the 7W leaving as light means 13 watts of heat.
Negating the thermal transfer through however you interface the laser diode with this, that is a temperature rise of only 5C.
That's about 9F for the muricans out there. That's plenty cool enough for most people.

You can use an even smaller heat sink. In 2 minutes, I sorted the list on digikey and found something close to their minimal size
if we're allowing 25C rise as arbitrarily requested above by Mr Cowboy. This one at 76x122x33mm (~5x3x1.4 inches for the muricans)
is at 2C per watt, so we've got 26C temp rise in theory. Or 120F heat sink temp in a 75F room for instance. That's with only 10 fins.

Most CPU heat sinks will cool a laser diode fine without a fan.

I'm reading your posts, but I do not want to use a fan as originally stated. It's difficult to find a heatsink with such a good heat dissipation that does not use a fan, so I'm figuring out what can be done to achieve a good heat dissipation without one.

 

[Cyparagon]

I'm reading your posts, but I do not want to use a fan as originally stated.

I just linked two examples that will do as you ask without a fan. So don't tell me you're reading my posts. You clearly have precisely no interest in "high level heatsink design"

Hopeless.

 

[RedCowboy]

Well you made me look it up and the spec sheets posted for the nubm44......not that they are even correct......show 65C as max operating temp so 149F rather than 125F