Laser diode cooling, would sanding help?

If I sanded the sides of my laser diode (essentially the same package as a stubby 5mm led), would it increase the surface area enough so that the cooling effects are noticeable if used in conjunction with thermal epoxy and a small active cooling system?
Datasheet: http://docs-europe.electrocomponents.com/webdocs/08d5/0900766b808d55b2.pdf

UPDATE: Ok, due to many unforeseen minor issues/problems that have arisen, I have decided to consider a flat diode package. Although the power and pulse width are somewhat smaller, It wont really hinder the end product (which will have loads of documentation on the build here).
New diode datasheet: http://catalog.osram-os.com/catalogue/catalogue.do;jsessionid=A8BC7E9CCE6FB1AA392B37F3F982A978?act=downloadFile&favOid=0200000200001347000300b6

sort by: active | newest | oldest
orksecurity7 years ago
Heat sink grease would probably help more...?
The Ideanator (author)  orksecurity7 years ago
Yes it would, but I want to add more surface area to the diode if it will help the active cooling system cool the laser faster and or more. I want to be as safe as possible since they're ~$40usd per piece
From all the pc cooling special stuff - people buff their heatsinks to a mirror shine - less inefficient grease getting in the way of good heat transfer.
The Ideanator (author)  frollard7 years ago
The sink is going to be drilled/milled where the diode will be, so it'll be pretty smooth. Do you have any suggestions as to which grease/epoxy would transfer the most heat? might the diamond based stuff work well enough ( http://www.innovationcooling.com/overview.htm )?
You aren't going to get a fraction of the dissipation through the plastic that you will get from the leads. Just look at the package !
The Ideanator (author)  steveastrouk7 years ago
What about the new possible choice? It has barely any plastic at all.
The Ideanator (author)  The Ideanator7 years ago
Besides, I was referring to the thermal grease there.
You only have to look at the spec to see the thermal resistance of the new device at 200K/W is a lot worse than the old package at 160K/W. Do what I said, forget shaving the package - it won't make any difference, and just pump the heat from leads. You won't get better than that.
Its not so much the shine as the flatness you need. You can have a surprisingly rough "polished" surface.
My best guess is that, on that scale, more surface area won't matter a heck of a lot. But that's just instinct; your milage may vary.
THE best thing is metal to metal contact. I haven't seen a plastic cased Laser diode before. What are you using ?
Found it.
The Ideanator (author)  steveastrouk7 years ago
Its a REALLY powerful laser (if attached to the right circuit)
The Ideanator (author)  The Ideanator7 years ago
I added the datasheet
Being a very very jaded user of optronic datasheets, I'll lay good odds its not as you think it is :-(
The Ideanator (author)  steveastrouk7 years ago
You are probably correct, since I know very little about them and I've never really used lasers before. I'm not really going to use it for the majority of the intended purposes it was designed for, nor is it going to be very scientific in nature. May I ask what qualms you have with their datasheets?
Any data sheet rarely covers all the bases. They tend to look at a quite select selection of devices which means that optical devices tend to be brighter or more sensitive in the datasheet than reality.
That device looks so small, I suspect cooling is dominated by the leads. Heatsink the leads with big PCB lands or something
In fact, actively chill the leads with a peltier cooler.
orksecurity7 years ago
Note that the rated duty cycle is 0.1% wjth a 100nSec pulse. (If I'm doing the match right, that means you can put out one of these extremely brief pulses every tenth of a second. Really powerful if all you need is a brief pulse for rangefinding and similar applications. Not hugely powerful if you're comparing it to lasers that can operate more continuously; over time, it's averaging 0.075W. That huge pulse does make it potentially more dangerous, of course; handle with appropriate care. Dunno whether that affects your project, but figured I should point it out in case you had missed it.
The Ideanator (author)  orksecurity7 years ago
I'm intending to run it quite a bit faster than that, but I'm not quit sure where you're going with the duty cycle bit, but further down it says: "Standard operating conditions refer to pulses of 100 ns pulse width at 1 kHz rate with 30 A operating current at TA = 25 °C." 1kHz is orders of magnitude faster then your estimate.
Granted; I misplaced a decimal point.
And those maximum ratings are for short-duration operation. In other words, this thing really does seem to be designed to put out a squirt of signal and then be allowed to rest. If that's what you want, great. If not, it may not be the component you're looking for.