I want copper copper copper! (well... maybe not)

[jakeGT]

That's not old age! That's like an abortion hahaha

 

[sinner]

Maybe a bad diode or something!! But.. The aixiz was heatsinked with a coleman max's heatsink, which is not as big but should be enough at 350mA which was set for the sole purpose of testing the maximum time to kill the diode if its left on!!

 

[Jubathoph]

I would challange the assertion that volumetric heat capacity is less important than thermal conductivity. The goal here is to keep the diode as cool as possible. While steel can hold more heat in a given volume than Cu or Al, the temperature of the diode will be hotter for a given host temperature, here's why:

Conductivity is like a thermal circuit.
Temperature= Voltage
Heat flow = Current
Conductivity = 1/ electrical resistance.

Diodes produce constant heat flow aka Q.

The heat ohms law is Q = (T2-T1)*k

Diodes produce roughly constant heat flux Q. If the conductivity is lower you need a higher delta T to get the same Q. That means that for a given host temperature your diode will be hotter in a material with lower conductivity. The only advantage to steel or iron is that your host will take longer to heat up. However your diode will be experiencing higher temperatures.

 

[tsteele93]

Even if you are being serious, what is that going to solve? You need multiple heatsink geometries, different diodes with different heat generation conditions, account for different outside conditions (ambient temperature, air movement, etc.). If you settle on a single standard, then does that prove anything as well?

Of course that will work! All RHD needs is binned diodes so he can generate similar diode wattages at similar currents.

If he can get four diodes that are close in efficiency then he can build a common host - something like an Aurora C6 and sink each one with an identically sized and shaped sink of each material.

I think the differences between some of the metals will be big enough to completely ignore any small variables like air movement. And he can control ambient with his THERMOSTAT.

All this stuff can be done mathematically. What will be more important is to determine what material and geometry is needed for specific heat generation, for expected durations, in typical or assumed ambient conditions. Maybe you really do want to simply "store" heat in the heatsink rather than facilitate its removal at the outside surface. Or maybe the opposite. Experimenting with one single condition is not helpful to anyone unless that condition is replicated.

I don't see that. 90% of the builds on this forum are very similar in style. So similar that you could easily design a setup to approximate the conditions that most of the lasers are seeing.

Basically, if you take A C6 with the typical heat sink design and try it in 4 different metals you will either find them all similar, or you might find clear winners and losers.

The only difficult thing is measuring temp at the diode, but I'm sure RHD can come up with something for that.

If it is close for each metal then we quit worrying so much about material. If one (or more) metals show a drastically lower diode temp, then you have a winner.

It will also be interesting to see how long it takes for all the different metals to achieve a steady temp at the diode, if any of them do... It is a great idea for an experiment.

I think copper will win for most scenarios. . Maybe nickel or SS will show an advantage for extremely long run times though?

 

[rhd]

Yep, it's no problem at all to do the test portion of it.

I have 4 diodes with identical efficiencies (or at the very least, within 5% of each other). I bin all my diodes, and that wouldn't be an issue. And if diode tolerance was an issue, we could use the same diode pressed into a single copper module, which could be used inside the 4 difference heatsinks (moved between them).

In fact, we could do it both ways. Try the "same diode moved" approach, as well as the "4 similar efficiency diodes" approach. The beauty of using a Saik as the test platform, is that it is super easy to actually use the exact same host, and just swap the heatsinks on/off. I'd even run the diode leads straight out the tailcap and do the driving via bench PSU.

I would test two things quantitatively:
1) The diode's optical output on an Ophir. You see symptoms of heat in the form of output power drops.
2) Directly, via a temperature probe on the diode itself. I would (for consistency) attach the probe in the same spot,.just below the barcode, on each tested diode. Again, the Saik host makes an excellent platform for this, as it's easy to string probe leads out the back.

In the end, this is actually a pretty decent setup in terms of controlling variables. If the differences are minute, then I'd say the data would be still open to question. But if I'm right, there will be major differences. I think nickel will perform MARKEDLY better than anything else. I think steel will perform a bit better than copper. Jury is out for me on where aluminum will fall in that.

 

[Wolfman29]

While I disagree with you on what the results may be, that's what experimental science is all about! Controlling variables, making hypotheses, and testing questions! I can't wait to see some results

 

[rhd]

While I disagree with you on what the results may be, that's what experimental science is all about! Controlling variables, making hypotheses, and testing questions! I can't wait to see some results

I think finding a machinist whole will make me nickel sinks is going to be the challenge. Nobody is going to want to have extra nickel rod stock left over (particularly if they think I may be wrong about my theory!). In fact, looking at pricing on the copper rod stock that would be required for a Saik-sized heatsink, I'm rethinking whether I could afford to do the test with a Saik. $150+ for the copper, probably twice that for nickel, and I'm going to have crap-loads of the stuff left over.

If I cut the heatsink diameter to 1.25" and go with a different host (I'll find something reasonable), then this may be affordable testing, depending on the price of nickel.

I may need to grab a nickel rod myself and just pass it along. If anyone here has a line of communication going with OnlineMetals, could you ask for pricing on:
Order Nickel R405 Round in Small Quantities at OnlineMetals.com

I'd compare it to copper:
Order Copper 110 Round in Small Quantities at OnlineMetals.com

 

[Wolfman29]

You think $150 for the copper? o.o Yeah, I suppose the nickel rod stock will be sort of expensive. What diameter do you need for a Saik? And why not try a standard C6? I would think that would be more representative of the more common heatsinks around here.

 

[rhd]

You think $150 for the copper? o.o Yeah, I suppose the nickel rod stock will be sort of expensive. What diameter do you need for a Saik? And why not try a standard C6? I would think that would be more representative of the more common heatsinks around here.

Because then I'd have to use a C6, and I hate the C6. This is all in the name of science, but I'd at least like to not waste the host/heatsinks and have whatever is left over be something remotely interesting to me

I think I could settle on doing this experiment with SH-032 hosts. I'd need 1" stock (or 7/8th if it existed). In each metal, I think these are the right selections:

Aluminum: Order Aluminum 6061 Round in Small Quantities at OnlineMetals.com
Steel: Order Stainless 17-4 Round in Small Quantities at OnlineMetals.com
Copper: Order Copper 110 Round in Small Quantities at OnlineMetals.com
Nickel: Order Nickel R405 Round in Small Quantities at OnlineMetals.com

My price tag minus the nickel (which isn't quotable online) is about $75. I'm ballparking another $75 for the nickel, so perhaps $150 total, before the cost of labour.

Maybe a machinist on the forum will let me "rent their time" and basically turn as much of that stuff into heatsinks within X hours as possible. At least then I'll minimize material waste, and can maybe use the extra heatsinks for something else (I guess to build more SH-032s).

 

[Wolfman29]

Haha. Alright That's not half bad, I suppose. Renting the time would sound good. I would offer my assistance, but I am sure there are better machinists here that can be more consistent with their work and be faster.

 

[cilegray]

Replying from my phone so sorry if its allready been brought up but machining heatsinks out of 4 different materials means youll end up with slightly different tolerences for your module fit into the heatsink...which will add more variables and uncertainties to the test... But I get the idea here, and its very nice of rhd to offer to do as thourough testing as possible!

 

[lasersbee]

@ RHD.... There is a Toll Free number to get prices on
items like Nickel that has an unstable price. All standard
listed parts have prices for the sizes you want to have
cut.

You can also buy smaller than 12" of stock. Look near
the bottom of the OnlineMetals pages...


A few things about the tests you want to do....

Since all hosts are made differently it is not important
to place the Test Heatsinks into hosts. As a matter
of fact your Metal Heat properties tests would be more
accurate without the hosts....

You are trying to test the heat capacity and thermal
conductivity characteristics of the different metals.
That can be done on the pure metal Heatsinks with
more precision.

The LD must be press fit into the bare metals directly..
You can not use a Copper Module if you want your
different metal heatsink tests to be valid.

You will need to test the temperature of the LD itself
and the temperarature gradiants across the Heat sink
in all directions and over time.

The heatsinks should also be mounted to a non-heat
conducting mount in a stable temperature environment
with no air current if you want the results to be valid.


Jerry

 

[tsteele93]

Perhaps, but if you want to go 60mph, it's easier and probably cheaper to buy an appropriate vehicle than to buy and modify an electric scooter. Either way, the copper "go faster stripes" won't help much. If you absolutely need a high duty cycle, use large fins or forced air. And here's a goddamn bat-shit-crazy idea... build a lab unit. :shhh:

I think you are confused! LPF is not LAB UNIT FORUMS.

The Lab units are so hard to carry in your pocket.

 

[tsteele93]

@ RHD.... There is a Toll Free number to get prices on items like Nickel that has an unstable price. All standard listed parts have prices for the sizes you want to have cut.

You can also buy smaller than 12" of stock. Look near the bottom of the OnlineMetals pages...

Heck, you might even check eBay.

Since all hosts are made differently it is not important to place the Test heat sinks into hosts. As a matter of fact your Metal Heat properties tests would be more accurate without the hosts....

I disagree. As long as you use the same host for each one. Especially if you use a common host that has seen a lot of use.

You are trying to test the heat capacity and thermal conductivity characteristics of the different metals. That can be done on the pure metal Heatsinks with more precision.

That's just it, we already know all these things. What we want to know is which of those characteristics makes a better sink for our application. It is no problem to find out the thermal conductivity and heat capacity of the metals. We're just trying to see which combination of specific heat and thermal conductivity works better to keep the running temperature of a high power laser diode lowest.

I bold that because Cyparagon's sig is a good point, "A problem well stated is a problem half solved."

The LD must be press fit into the bare metals directly.. You can not use a Copper Module if you want your different metal heatsink tests to be valid.

Again, this is not true for what we are trying to do. We aren't testing the metals, we are testing the different metal's ability to heat sink our application.

You will need to test the temperature of the LD itself and the temperarature gradiants across the Heat sink in all directions and over time.

That might be ideal, but I think the differences are either going to be big enough that it is clear that one or two metals are doing such a better job that you won't need that level of detail, or you will have them so close that it will be clear that it just doesn't matter which metal you are using.

It is like your laser power meters. Ideally you should have different calibrations for different wavelengths, but we all accept the industry standard +/- 5% and ignore that if we are testing red or green or violet, that we aren't in the ideal zone of ~445. (I think that is right.)

We just accept that it is close enough for our application.

The heatsinks should also be mounted to a non-heat conducting mount in a stable temperature environment with no air current if you want the results to be valid.

As long as he doesn't have a ceiling fan cranked up in one test or an ambient of 60F one day and 80F another, I don't think air currents are going to be abig factor.

With all due respect (and I'm not being sarcastic - I mean this honestly) I think you are applying some of the things that are important in designing and using your power meters to a test here that is much less delicate.

I think temp gradients are going to be significant enough that those elements will be negligible unless there are big differences in ambient or air flow.

:beer:

 

[tsteele93]

I think finding a machinist whole will make me nickel sinks is going to be the challenge. Nobody is going to want to have extra nickel rod stock left over (particularly if they think I may be wrong about my theory!). In fact, looking at pricing on the copper rod stock that would be required for a Saik-sized heatsink, I'm rethinking whether I could afford to do the test with a Saik. $150+ for the copper, probably twice that for nickel, and I'm going to have crap-loads of the stuff left over.

I will take this to PM, but if you can get me a cad file or a very good drawing, I think that I may be able to get this done.

You have a PM coming.

 

[tsteele93]

Replying from my phone so sorry if its allready been brought up but machining heatsinks out of 4 different materials means youll end up with slightly different tolerences for your module fit into the heatsink...which will add more variables and uncertainties to the test... But I get the idea here, and its very nice of rhd to offer to do as thourough testing as possible!

Doesn't matter. That is the beauty of this test. We want to know in the real world which works better. In fact, this is what we want! If that makes a difference then we want to know that.

We want to know - in the real world, in our application - which metal performs better.