"C" cell O-Sinks Images FINAL PRODUCTION UNITS

[yaesumofo]

\"C\" cell O-Sinks Images FINAL PRODUCTION UNITS

Hey guys I thought you might be interested in the fact that first "C" cell O-Sinks are being cut as I right this.

It is pretty cool. We have adjusted this design to accommodate an "e-can" should you choose to use one. It is also a little thicker giving it a little more surface area. The pedestal will be slightly taller as well. And of course the pedestal will be milled so that a Luxeon emitter is perfectly centered every time when installed properly.

If anybody is interested I will post a few pictures of the cutting and milling process's tomorrow.

I am actually having a small number made for the "OLD" MAG C aluminum.
I will also have a few anodized BLACK. As an experiment.
These will be in my hands finished Friday.
I can't tell you guys how much fun I have working with these machine shop guys that know how to think in terms of metal shapes and hardness I have learned so much about what is involved in making stuff. It is fantastic.

When people like Mr Bulk or D. Speck or Hotbaem build a light or make a part it is a hell of a lot of work. With things with small tolerances you have to be very precicse about what you are doing or the whole thing can turn into scrap metal. This sure is fun.

Yaesumofo

 

[StoneDog]

Re: \"C\" cell O-Sinks !!

Is there another thread that talks about ordering info?

Thanks!

Jon

 

[yaesumofo]

Re: \"C\" cell O-Sinks !!

Not just yet. A couple of days and it will be up.
Yaesumofo

 

[yaesumofo]

Re: \"C\" cell O-Sinks !!

Hi guys Today we did the milling of the O-Sink. I took some pictures.

The plan.

The raw material.

After the first operations.

The perfect fit.

Precision.

Before the second operation in the mill.

Durring the Milling operations. (OIL OFF For PHOTO)

After the milling operation.

I did 2 sizes.

OLD MAG ALUMINUM Only 25 of these:

New Mag Aluminum.

Final Production

The "C" Cell O-Sink is really beefy.
There is a space for an "e-CAN". I can't really see any use for this but a few people asked for it.

Anyway the final milling operations are happening now.
After milling the parts will be tumbled with some ceramic media in order to remove all sharp edges. The swirls will be gone as well. I will be picking these up tomorrow.
Questions?
Yaesumofo

 

[Rothrandir]

Re: \"C\" cell O-Sinks !!

what is that third hole looking thing in the middle?
i thought i'd seen it in other pictures but could never be sure...

also, how does tumbling affect the surface finish?
(your machine shop should be able to get a surface finish measurement with a special tool)

 

[yaesumofo]

Re: \"C\" cell O-Sinks !!

There are 3 tiny holes. I have found in my experience that these holes help the adhesive to grab the O-Sink surface. A Tiny amount of AA is required for excellent adhesion. The Three tiny holes also give a visual indicator of proper emitter alignment.
The Tumbling simply removes SHARP EDGES. The surface dulling of the swirls are a side effect. We did the same process on the "D" O-Sinks with no ill effect. Just no sharp edges.
Roth do you think we should skip this step for some reason?

Yaesumofo

 

[Rothrandir]

Re: \"C\" cell O-Sinks !!

adding the holes is a bad idea thermally speaking, as it limits surface area, which is very important in getting the heat away from the luxeon, especially at the small junction area under the slug (a very small contact area as opposed to the massive sides of the heatsinks where it contacts the mag body)
quite honestly, the area between the slug and heatsink is just plain very resistant to falling off as it is, especially when the luxeon is securly soldered in place.
baking the epoxy also adds in thermal transfer and increases the joints mechanical properties.
those three "tiny" holes are taking up a lot of precious surface area which would be better put to use in actually transferring heat.

as far as tumbling, i honestly don't know.
what you want to do is ensure that you have as smooth a finish as possibly possible, for optimal heat transfer.
have your machine shop do a surface finish reading with both tumbled and untumbled, and whatever has a better finish, that's what you'd want to use.

 

[greg_in_canada]

Re: \"C\" cell O-Sinks !!

The Intel/AMD CPU overclocker guys go to great
lengths to have two extremely flat surfaces mate together
with a minimum thickness of thermal grease/epoxy. Just
enough to fill in the valleys.

They go so far as to lap the bottom of heatsinks and
waterblocks mirror smooth. Some have lapped the top of
the CPU heatspreader.

So the concerns would be that your tumbling roughens the
surface increasing the thermal resistance between the
emitter and heatsink. Secondly, your three holes remove
low resistance aluminum and replace it with (relatively)
high thermal resistance epoxy (or grease).

What percentage of the contact area has been removed by
adding the holes?

Does anyone know the heat flux (watts per square cm)
of an overdriven 5 watt emitter compared with a fast CPU?
Just wondering if they are in the same ballpark of
difficulty.

Has anyone ever measure the emitter temp rise relative
to the heatsink? (Maybe by measuring the temp of one
of the leads.)

Greg (the question man) /ubbthreads/images/graemlins/smile.gif

 

[yaesumofo]

Re: \"C\" cell O-Sinks !!

[ QUOTE ]
Rothrandir said:
adding the holes is a bad idea thermally speaking, as it limits surface area, which is very important in getting the heat away from the luxeon, especially at the small junction area under the slug (a very small contact area as opposed to the massive sides of the heatsinks where it contacts the mag body)
quite honestly, the area between the slug and heatsink is just plain very resistant to falling off as it is, especially when the luxeon is securly soldered in place.
baking the epoxy also adds in thermal transfer and increases the joints mechanical properties.
those three "tiny" holes are taking up a lot of precious surface area which would be better put to use in actually transferring heat.

as far as tumbling, i honestly don't know.
what you want to do is ensure that you have as smooth a finish as possibly possible, for optimal heat transfer.
have your machine shop do a surface finish reading with both tumbled and untumbled, and whatever has a better finish, that's what you'd want to use.

[/ QUOTE ]

I agree that surface area is everything when moving heat from the slug to the sink. These shallow and tiny holes do not impede the transfer of heat at all. If anything the increased surface area between the adhesive and the aluminum sink can potentially improve heat transfer. I have found that when I use AA between the emitter and these heat sinks, that heat transfer excellent. I have also found that adhesion is excellent as well. Needless to say alignment is perfect.

Having modded quite a few lights, so far I have never had a Mag Mod type light fail due to thermal issues. So far.

I have had adhesion issues. As well as alignment issues. While it may not be a perfect design (show me one) the current design of the O-Sink reflects careful consideration of Thermal, Adhesion, and Alignment issues that I have encountered in the past and does a fine job of balancing these and other technical constraints.

The tumbling process smoothes the surface of the O-Sink. The surface making contact between the sink and the tube is very smooth making for very good heat transfer indeed.


Roth, if you would care to demonstrate the thermal effects of the 3 holes I would be happy to send you an O-sink or two for experimentation purposes with the agreement that you publish the results here.
Just let me know.
Yaesumofo

 

[Justintoxicated]

Re: \"C\" cell O-Sinks !!

technicaly the holes would Impede on the transfer of heat. The whole reason for thermal compounds in the first place is to fill in the microscopic pores between the slug of the LED and the HS, Paste does not actualy transfer heat as well as one would think...In fact it's not very good at all. metal to metal with a very thin layer (somethign like AS5 ) to fill in the microscopic pores would give the best results. Of course with epoxy, you have to use enough to actualy hold the LED on anyways... but forcing it to transfer heat through extra expoxy to fill in the holes IS a bad idea if it can be avoided. At least thermaly speaking.

On the other hand...I'm not sure it would really matter, but it could help the emitter to stick better (more area for the epoxy) but I have not had a problem with this before, actualy it's usualy the opposite problem.

 

[moraino]

Re: \"C\" cell O-Sinks !!

With the tall LIP it would be easy to revove the sink. Does that also make the length of the light longer than necessary at the focused point?

I would like to screw the head all the way down possible.

Henry

 

[yaesumofo]

Re: \"C\" cell O-Sinks !!

Well Greg if I were trying to move the massive amount of heat that overclockers are trying to deal with (50-60+ watts of heat in an area 4 or 5 times the size of a luxeon emitter slug) I would use every available technique at my disposal to do it as well.
But here we are making or modding flashlights. We are at the moment talking about somewhere between 1 and 6 watts of total energy not all heat energy. When it comes to the design of the emitter to heat sink connection designed into the O-Sink I can say that it is a much better than that same connection on a luxeon star (especially considering the material they use to conduct heat from the emitter slug and the aluminum star).

The Tumbling creates the same surface that you have seen on the "D" cell O-Sinks that you have (or have coming).
It smooths the surface of the O-Sink. It is a good thing.

Yaesumofo

[ QUOTE ]
greg_in_canada said:
The Intel/AMD CPU overclocker guys go to great
lengths to have two extremely flat surfaces mate together
with a minimum thickness of thermal grease/epoxy. Just
enough to fill in the valleys.

They go so far as to lap the bottom of heatsinks and
waterblocks mirror smooth. Some have lapped the top of
the CPU heatspreader.

So the concerns would be that your tumbling roughens the
surface increasing the thermal resistance between the
emitter and heatsink. Secondly, your three holes remove
low resistance aluminum and replace it with (relatively)
high thermal resistance epoxy (or grease).

What percentage of the contact area has been removed by
adding the holes?

Does anyone know the heat flux (watts per square cm)
of an overdriven 5 watt emitter compared with a fast CPU?
Just wondering if they are in the same ballpark of
difficulty.

Has anyone ever measure the emitter temp rise relative
to the heatsink? (Maybe by measuring the temp of one
of the leads.)

Greg (the question man) /ubbthreads/images/graemlins/smile.gif

[/ QUOTE ]

 

[yaesumofo]

Re: \"C\" cell O-Sinks !!

Henry, on my test light the optimal focus point is around 3 turns of the head adding less than 0.250" to the overall length of the light.
Yaesumofo

[ QUOTE ]
moraino said:
With the tall LIP it would be easy to revove the sink. Does that also make the length of the light longer than necessary at the focused point?

I would like to screw the head all the way down possible.

Henry

[/ QUOTE ]

 

[StoneDog]

Re: \"C\" cell O-Sinks !!

Three holes or not, I want a couple.

Jon

 

[greg_in_canada]

Re: \"C\" cell O-Sinks !!

[ QUOTE ]
yaesumofo said:
Well Greg if I were trying to move the massive amount of heat that overclockers are trying to deal with (50-60+ watts of heat in an area 4 or 5 times the size of a luxeon emitter slug) I would use every available technique at my disposal to do it as well.
But here we are making or modding flashlights. We are at the moment talking about somewhere between 1 and 6 watts of total energy not all heat energy. When it comes to the design of the emitter to heat sink connection designed into the O-Sink I can say that it is a much better than that same connection on a luxeon star (especially considering the material they use to conduct heat from the emitter slug and the aluminum star).

The Tumbling creates the same surface that you have seen on the "D" cell O-Sinks that you have (or have coming).
It smooths the surface of the O-Sink. It is a good thing.

Yaesumofo

[/ QUOTE ]

I'm not trying to be a trouble maker, but this is an
interesting subject to me.

A CPU can run for years with a junction temperature of
75 or more degrees celcius. A luxeon will degrade
quickly if operated above 25 degress C. So maybe they
are similarly difficult. Don't know, that's why I asked.

The LED is only about 20% efficient, I'm guessing, so most of
the power goes into heating it up.

I haven't seen a Luxeon star, so I'll take your word
that the O-sink is better.

On the tumbling: smooth to the touch isn't necessarily
smooth enough for optimum heat transfer. Though I don't
think lapped heatsinks will be necessary until they make
a 20 or 50 watt luxeon (though I'm sure they would
change to a larger slug if they did).

What percentage of the contact area has been removed
by the three holes?

Thanks - Greg

 

[Justintoxicated]

Re: \"C\" cell O-Sinks !!

Yes I hear what you are saying about them not being CPU's with small surface cores and lots of power. But what is their purpose? If none, it would be better to leave them out IMO.

 

[yaesumofo]

Re: \"C\" cell O-Sinks !!

It is a design choice. Their primary porpoise is to give the adhesive something to hang on to.
It is done. They are there.
They will not have a negative effect.

I will happily send 2 "C" cell O-Sinks for experimentation to anyone who can show demonstrate any effect that the the holes may have.
Yaesumofo

[ QUOTE ]
Justintoxicated said:
Yes I hear what you are saying about them not being CPU's with small surface cores and lots of power. But what is their purpose? If none, it would be better to leave them out IMO.

[/ QUOTE ]

 

[pbarrette]

Re: \"C\" cell O-Sinks !!

Hi all,

This thermal discussion is interesting to me from a purely scientific viewpoint.

On the one hand, Arctic Silver epoxy has a thermal conductance of ~7.5W/mK while aluminum ranks in at ~215W/mK and copper at ~395W/mK. This means that the epoxy used is extremely less efficient at moving heat from the Lux to the heatsink than, say, a straight Al to Al contact.

On the other hand, that pure metal to metal contact can be near impossible to achieve which means you would likely have air (0.026W/mK) trapped between the two, degrading performance more than your thermal epoxy.

Now to me, the question about the holes under the lux heatsink lug is this: "Will the small holes allow me to use much less epoxy between the lux-lug and the heatsink to keep the same, or better adhesion?"

The answer to this question seems likely to be "Yes."

If this is true, that means there will be less epoxy between the actual heat transfer surfaces than a setup without the holes while maintaining roughly the same adhesion level. At these lower epoxy thicknesses at the transfer surfaces (roughly .001 in.) the heat transfer losses of the holes may well be made up for by the better surface to surface contact.

According to Arctic Silver Corp. the thermal conductivity of AS5 thermal grease is roughly 350,000W/mK (1W/mK = 1W/mC) at that thickness. This massive difference in conductivity is due to the extremely low volume and surface area since thermal conductivity is measured as the amount of thermal power that is transferred over a given area.

Undoubtedly, a straight metal to metal contact would be a better conductor, but an extremely thin layer of thermal epoxy would perform much better than: A) Using nothing and leaving the air gaps in, or B) Using a thicker layer of epoxy to improve adhesion.

So, my guesstimation on the holes is that they will allow a much thinner layer of epoxy to be applied at the thermal transfer points, thus offsetting the thermal insulation created by having the holes to begin with.

Since the holes are already there, and yaesumofo feels that they aid in adhesion and alignment, then I, personally, wouldn't worry too much about it.

Just remember that when mounting a lux in this, the idea is to fill the holes with your thermal epoxy and spread as thin a layer as possible on the remainder while maintaining complete coverage.

Just my $0.02 (US$, pre-tax),
pb

 

[Rothrandir]

Re: \"C\" cell O-Sinks !!

i don't have either the means or desire to test the thermal differences between the holes and no holes, but can tell you that adding them greatly reduces the already too small surface area in the slug/heatsink joint.

will adding the three holes improve adhesion?
quite possibly, but in my opinion, the benifits of this are not only so minimal as to be irrelevant, but better results could be achieved by other means, such as using thicker portions of adhesive around the black area (not the slug), and baking the epoxy.

having the holes will not result in a thinner "overall" joint either, all it will do is greatly increase the joint size between the bottom of the holes and the slug itself.

i don't know what i'm doing wrong, but i've never had an adhesion problem with any mod i've done, in fact, once when i was trying to remove an emitter from a heatsink, i ripped out everything but the slug /ubbthreads/images/graemlins/icon15.gif

just out of curiosity, does anyone know the surface area of the bottom of the emitter slug, and does anyone know how much surface area is being reduced by adding the 3 holes?

 

[yaesumofo]

Re: \"C\" cell O-Sinks !!

In terms of operating temperatures all of the flux specs that LUMILEDS give are based on a nominal junction temperature of 25C.
On the technical data sheets from LUMILEDS is a graph that shows degraded performance above 25C but it certainly is not a deal killer. +- 30% @ 120c. I do however see a absolute maximum rating of 135c.
Unfortunately we all will see a degradation in emitter performance no matter what we do. What if it is a hot night in the desert where the ambient temperature is 35c?
I don't think the emitter will ever operate at or below ambient 25c in a flashlight. (Actually I can imagine some ambient weather conditions where this might be possible. At that point I am sure that the main issue would rapidly would shift from Heat sinks to batteries)


http://www.lumileds.com/docSearch.cfm
Will take you to the Luxeon thermal design guide.
I have read this document.
When it all comes together ANY FLASHLIGHT design will be thermally compromised.
Here we are sticking heat source that produces a little light into an enclosed area...and hoping it won't get too hot. Forget it. It is going to get hot.
The O-Sink does a great job of getting heat away from the emitter. It is not perfect. Show me one that that will fit in a MAG that is. The O-Sink works. That is the point.
The percentage of removed material under the slug? 10% -+.
These holes are more like divots. Just deep enough for the adhesive to have something to grab. They also provide a greater surface area to the adhesive for heat transfer.
Yaesumofo







[ QUOTE ]
greg_in_canada said:
[ QUOTE ]
yaesumofo said:
Well Greg if I were trying to move the massive amount of heat that overclockers are trying to deal with (50-60+ watts of heat in an area 4 or 5 times the size of a luxeon emitter slug) I would use every available technique at my disposal to do it as well.
But here we are making or modding flashlights. We are at the moment talking about somewhere between 1 and 6 watts of total energy not all heat energy. When it comes to the design of the emitter to heat sink connection designed into the O-Sink I can say that it is a much better than that same connection on a luxeon star (especially considering the material they use to conduct heat from the emitter slug and the aluminum star).

The Tumbling creates the same surface that you have seen on the "D" cell O-Sinks that you have (or have coming).
It smooths the surface of the O-Sink. It is a good thing.

Yaesumofo

[/ QUOTE ]

I'm not trying to be a trouble maker, but this is an
interesting subject to me.

A CPU can run for years with a junction temperature of
75 or more degrees celcius. A luxeon will degrade
quickly if operated above 25 degress C. So maybe they
are similarly difficult. Don't know, that's why I asked.

The LED is only about 20% efficient, I'm guessing, so most of
the power goes into heating it up.

I haven't seen a Luxeon star, so I'll take your word
that the O-sink is better.

On the tumbling: smooth to the touch isn't necessarily
smooth enough for optimum heat transfer. Though I don't
think lapped heatsinks will be necessary until they make
a 20 or 50 watt luxeon (though I'm sure they would
change to a larger slug if they did).

What percentage of the contact area has been removed
by the three holes?

Thanks - Greg

[/ QUOTE ]