Heatsinking

[Benm]

Your result looks right.. with half a watt of power dissipation and a resistance to ambient of 100 K/W you would indeed get a 50 K temperature rise above ambient on the diode. A bare aixiz module would probably be more than enough to satisfy this demand.

One important issue is how hot you actually want to run that diode. I dont think 75 degrees is acceptable here. You will see the wavelength go longer (and less bright), and you should keep in mind that you are operating the diode way above the manufacturers maximum continous current. For those reasons i'd say 10 K/W, perhaps 20 K/W would be more reasonable choices for heatsinking.

 

[Morgan]

It's good to know at least I'm gettin to grips with it. The 75 deg C came from the datasheet as a maximum case temperature so it would be better to stay lower, (this would still probably be considered acceptable regards wavelength though wouldn't it? It falls within the manufacturers limits. Maxed out but they'd reduce that figure if it distorted the wavelength that much? I'll bow to superior knowledge here, that was merely my own speculation). The datasheet also gave the 205mA at max 3V as a CW running figure, (400mW pulsed). If the figures were increased to the 450mA that is considered the max for our uses on these diodes the power to dissipate goes to around 1.1W! That brings the C/W down to 45.45 C/W.

Thinking about it after I posted the last post, a tiny heatsink is in fact exactly what is used in the sleds. If I'm getting the correct feel for the size of the hardware then a 101 C/W heatsink would probably only be about that size. (I'm sure the sled comes into that particular equation too but it FEELS right.)

The next thing I'd be interested to do is actually design a heatsink to achieve a given C/W. I think this may come up in further lectures but I'm not certain as to what degree. Have to talk to the, "Materials", lecturer for help on that one I think.

Excellent, I'm really pleased with the outcome of this thread so far. :thanks:Thank you so much for your input on it... I'll be adding in due course.

M

 

[Benm]

In the sled, the diode is in a small heatsink that might be of proper size on its own, but its also connected to the whole sled taking away more heat.

As far as wavelength is concerned: This increases about 0.3 nm for every degree in temperature rise. So a diode that was 650 nm at 25C would be 665 nm at 75C. This may be within acceptable limits for the application, but would also be noticably dimmer even when the output power is exactly the same. The efficiency will also drop with temperature though, decreasing brightness even further.

Calculating heatsink thermal resistance from size and shape will probably be difficult to do accurately. Personally, i compare with heatsinks of known thermal resistance (manufacturer stated), and leave some margin for error. If you want to know extactly you can install the heatsink in the intended position, bolt on a transistor and dissipate a known amount of power. A thermometer will tell you the rest.