Thermal Properties of Titanium - by Prof. McGizmo

[Kiessling]

In the end it looks like before reaching steady state it is impotant to get the initial heat away from the LED quickly and transfering it to the whole body as quick as possible to avoid a heat-buildup right at the slug ... for example with a generous al heatsink with good and big contact area to the ti body.
Then we (or ... YOU :nana: ) would have to mke sure that the heat will be able to spread throughout the body quicker than the heat-buildup at the LED to avoid the magic smoke.
Only after that will the steady state be important and as a consequence, the heat removal from the light itself via air, hand, or whatever means.

So ... from my limited technical understanding it owuld seem that steady state thermodynamics are most certainly the most important factors in those lights, but we*'d have to make sure to be able to reach steady state at first.

bernie

 

[AilSnail]

Kiessling, maybe you have a bad electrical contact between the tube and the head. that could make heat. Thought you said the heat was located around there.
Maybe you just think the ti light is hotter because the head is. What about the tail of the Ti light. is that hotter than the alu tail?

don, your highway analogy is rong. this is because, in a congestion, or a traffic light or single lane road or whatever, the cars does not move any faster whether there are 10 cars behind them or a 100. Contrarily, If there are 100 drugged phonies pushing, temp would rise and the transmission of energy to the air would increase! (which would decrease the amount of congested phonys backwards in the queue.) or what?

at the end of turbo's experiment, there was a 35 degree difference between the head and the tail. ambient-head difference was 85 deg, and ambient-tail difference was 50 deg. that means the head dissipates 1,7 times more heat per square inch than the tail. What if the conductance of the shell was 12 times worse? You could probably in good concience disregard the whole tail from the heat emission equation - only the head would dissipate any significant amount of heat. so if the head is 3sqin, and the tail is 6sqin, you're requiring the head to dissipate more than twice as much as before! Which again means, the temp gradient ambient-head will be doubled... and the surface temp of the head is now 250 deg.

I feel that the threads usually are a more significant hindrance than the rest of the internal obstructions? So what can be done about this, other than moving them backwards?

 

[AilSnail]

I speculate that most alu flashlight threads have a very low contact area, and large amounts of voids. The same might apply to flat battery tube shoulders butting against an internal bulkhead in the head. I imagine that a more plastic material would conform to the recipient surface and mate better. Perhaps only one of the surfaces needs to be softer.. Ofcourse, the threads would have to be coarser, and perhaps "squared-off" (can't remember the term), like one sees in some plastic flashlights for instance.

Incidentally, titanium is quite a bit stronger than aluminium pr volume, and I believe it has a sort of bounce-back flexibility that al doesn't - I imagine you don't really need the same stiffness in a ti wall that you do in an alu wall.
This might make it possible with significantly thinner walls for the same impact/crush survivability. So now one might be able to either make the ti light stronger than the aluminium counterpart, or thinner, or perhaps one could combine it with some material that would take up what the ti lacks in thermal conductivity, while being as close as possible in the galvanic series, and also provide a soft mating surface between the head and the tube?

 

[Kiessling]

AilSnail ... the head center body (whrer the LED and converter sit) is hotter, and the tail and tip of the head are cooler.

As it works well right now I'll leave it that way and do an upgrade to the whole thing (or letting Don do the upgrade ) once we have new super-duper whiz-bang LEDs in the future ...

bernie

 

[AilSnail]

Ah, ok. You mean hotter and colder than the al counterparts, right? Makes sense. Doesn't seem like the threads are creating heat then at least.

 

[Kiessling]

Yepp ... that's what I wanted to say

 

[NewBie]

If I may add some snippets:

Bernie, :nana:

On heat acceptable to the hand, I don't know and would expect it to be a case to vary from individual to individual and also depend on the amount of moisture on the skin. The Ti light I tested was at 127 F and I had no discomfort in grabbing it and holding on to it. I have a prety good IR thermometer and gave up using it to measure the flashlights as it depends too much on understanding the nature of the surface it is reading and you need to make adjustments to your reading. It is always significantly lower that a thermocouple's reading. :shrug:

I believe that anodize is much more effective at radiating heat than the smoothor polished Al alternative. However, the anodize film is a thermal barrier in terms of conduction and in fact Ti is 4x more conductive than anodize! If you want to use your hand to cool an Al flashlight, you are better off with a raw light. If you want the light to cool via radiation of excess heat, the anodize is better. In general, since the boundary of the anodize is so thin, I suspect that the difference is moot! The environment as well as physical size and geometry of the light and its thermal path junctions all play a significant role here; not to forget the significance of the thermal load as it relates to the other factors mentioned!

Yup, the surface material emissivity is quite important. You can take bare AL and greatly improve it's emissivity by coating it with even plastic, resulting in a lower steady state temperature.

Then there are plastics one can obtain that are both thermally conductive and have very high emissivities.

Aluminum is even worse when it has a high polish....

 

[xochi]

I just read the McluxIII-PD FAQ and it mentioned something about not using the flashlight body or head for any kind of electrical path and was therefore completely Hard Coated.

I looked around and found these numbers
Titanium — Thermal Conductivity: 21.9 W/(m*K)
Aluminium — Thermal Conductivity: 237 W/(m*K)
Aluminum oxide--- 18 W/m·K
*Hard coat anodize - 0.5 - 1.0 W/(m*K)

I don't know how up to date these numbers are because they were established by a guy in 1987 who actually measured the thermal conductivity of an anodized hunk of aluminum verses a non anodized hunk. Apparently the average at the time was .7W/m*K and alot of factors were involved that made the thermal conductivity much lower than the oxide that grows.

In all honesty, I really know nothing about this stuff but it would seem that Hard coat anodize is about 40 times less thermally conductive than Titanium . If the PD is anodized all over then that means the ano is acting as a barrier to the heat getting to the aluminum. It should also insulate the aluminum on the outside and prevent heat from escaping.

I believe that it would be interesting to see the difference in heat buildup between a Bare Ti light, Bare Al light, Totally ano'd light and a light Ano'd only on the outside.

In trying to find info about the thermal conductivity of hard coat, some companies claimed Hard Coat increases thermal conductivity while other platers claimed it reduced thermal conductivity. Go figure!

 

[McGizmo]

xochi,

I recall a couple years ago running into the poor conductivity of Al oxide and it was around the time I was playing with diamond dust in a bog for mounting the Luxeons. I asked around and was told that the film thickness of the oxide, being so thin in the case of the anodize film, significantly reduced the impact of its poor conductivity. :shrug:

Many use Arctic Alumina for a bond as it is low in thermal resistance but provides electrical isolation for the slug. It is low in thermal resistance but consists of aluminum oxide particles in an epoxy binder?!?!?

I don't know how to go about modeling these lights in a legitimate thermodynamic framework and ir is certainly due to my ignorance and stupidity when it comes to a working knowledge of thermodynamics!! :green:

I do know that heat does migrate and seemingly well through the aluminum parts whether there are HA barriers or not. If HA were such a great insulator, I would expect to see it used in insulation applications. I think fire bricks contain Al oxide but they are a heck of a lot thicker than a HA film.

My feeling is that when one doesn't know where the limits are, it is best not to try to push them. In the case of the LED's, there is a lot of emperical as well as theoretical justification in not driving them hard to start with. Heat management is less of an issue if you are not generating copious amounts to start with.

An interesting and related observation:

I recently built a couple 27LT lights using the new BBx2 driver and XX1T 5W LED's. In one case, the host is a proto 27LT Ti. On the bench, I measured 850 mA on high going to the LED. The Ti light gets very warm very quickly right at the overlap of the tube over the head and in the plane where the heat sink lies. With your bare hand, you can easily feel a real thermal diffrential when you move your hand away from this surface hot spot. The thermal load is not well balanced or distributed in this particular light. Since the location of the hot spot is also where you are likely to hold the light, does the ultimate steady state benefit any? Heck if I know! My gut tells me that overdriving a 5W LED in a Ti host is not the smartest thing to do to start with.

 

[Kiessling]

[smartass mode on]Overdriving a LuxV is never a wise idea anyhow. :nana: [smartass mode off]

The behaviour you are describing with this Ti-27LT is exactly how my McLuxIII-T is behaving btw.

bernie

 

[xochi]

I was only able to read the first page since I didn't have 25.00 dollars worth of interest in the topic but I guess that one of the important points was this:

"Since the principal constituent of these coatings is aluminum oxide, it is commonly assumed that the thermal conductivity is somewhat less than that of aluminum oxide (about 30 W/m/K). This value of thermal conductivity is correct for polycrystalline aluminum oxide (corundum). The aluminum oxide contained in anodized coatings is, however, quasi-amorphous and, since the coatings also contain very large amounts of other compounds, in is not surprising that the thermal conductivity is very much lower than that of polycrystalline aluminum oxide. As a rule of thumb the thermal conductivity of an amorphous dielectric material , such as aluminum oxide, is usually about an order of magnitude lower than the thermal conductivity of the crystalline form of the same material."

As Don said , though, since they are so thin , probabally not much insulation effect. I did also see that the porousness of hardcoat can very a great deal depending on technique and the less porous the lower thermal conductivity.

Here is the link to the article (first page with option to buy the other 4) if anyone is interested http://www.aiaa.org/content.cfm?pageid=406&gTable=mtgpaper&gID=94859

Kind of neat that it amounts to high tech, accelerated rusting.

 

[xochi]

I guess the great thing is that Don's lights are designed with all this stuff in mind and the Ti lights that he designs utilize appropriate materials and drive levels to insure reliability while maintaining excellent usability. The systemic thought involved in the design of Don's Ti light serves to help legitimize their value (in my own mind at least).

A titanium coffin , complete with LED viewing window really has no value to me, actually, I'd feel kind of foolish for spending so much for one. Thanks for doin it right Don.

Crap, I'm starting to sound like a McGizmo Fanboy.

 

[AlecGold]

Sorry to start up an old thread, but this is mighty interesting stuff.

I've done some things with warmth, heat and metals in my time and got some education in it as well.

I don't think the thermal transfer through the body is the big issue here. I think surface/air is a much, much bigger issue here. Aluminium lights are often HAIII, which is an insulator. Therefore I think the thermal properties of AL are more reduced as Ti lights are often "naked".
I think that in the end the Al is still better, speaking in terms of thermal management, but Ti isn't half as bad as many people believe.
I always learned that dissipating heath though mass wasn't a big problem in metals, but getting it into the air is the problem. Hence the form factor for CPU-cooling ribs, which are more about surface-air contact and still maintaining a form of airflow throughout it's design than about mass.

You can dissipate heath very fast with thermal paste and enough mass, but you would still need to get it out of the metal and into the air at some time, have very short runtimes or have an enormous mass to put heath into.

If you look at the Notebooks from Apple, you will see a heath pipe build in. These heath pipes are simple in form, but highly efficient in transporting heath. They are made from copper, filled with some fluid and used to enlarge the area that heaths up, without using air alone.

Last point I would like to make is the maximum time HDS EDC's are able to perform at maximum. For a while I had an U85. If not held very tight in my hand, or placed in a cup of water, the light would shut off after as little as 4 or 5 minutes! The aluminium wasn't even that warm, but the electronics got warmed up very fast until 40Celcius. At that point the electronics kicked in and put it back to around 60 lumens. When you would hold it in you're hand from the start and let it warm up more slowly, I could let it stand on it's own for 7 or 8 minutes on maximum, after the warming up period.
Strange, because when I let it stand after burning already 9 or 10 minutes on maximum (tightly held in my hand), the difference between the air and the surface was smaller than when I started with the light on it's tail.
But because more mass was warmed up, more surface was available to transport warmth into the air!

 

[McGizmo]

Alec,

I am in way over my head on any of this thermal stuff but I too agree that steady state temp of the LED is much a function of the whole system and this includes the ambient environment the light is situated in and its thermal connectivity to the environment. Newbie did some interesting studies with black surfaces VS high polished and showed how in the absence of other conductive paths (actual contact or a stream of air), the escape of the heat through radiation was significant and a black anodized surface was better at reducing the core temp than a "naked" and polished surface. Once you grab the light or set it in a puddle of water, the relief via radiation takes back seat to conduction and other thermal properties of the light and its material composition and form override its ability to radiate heat.

Doug S has been around lately and it would be cool if he could add some snippits of wisdom to this thread. I think there are any number of significant factors involved in the thermal state of a LED in a light, in use and a simple model or view may well overlook key considerations. The big picture has major impact on the little pictures it is comprised of. :shrug:

 

[tron3]

I'm no expert on this subject but I would solve the heat problem a couple of ways. Keep in mind, I didn't read the other posts first.

1. The inside of the light is steel or aluminum to be a heat sink, with a bonded outter Ti shell for strength and beauty.

2. Alloy the Ti, if possible, into better heat conducting metals. Maybe not as strong, but still rather pretty, better heat sink and lower cost.

Just some thoughts.

 

[AlecGold]

1. The inside of the light is steel or aluminum to be a heat sink, with a bonded outter i shell for strength and beauty.

2. Alloy the Ti, if possible, into better heat conducting metals. Maybe not as strong, but still rather pretty, better heat sink and lower cost.

Just some thoughts.

Hhhmmm, wel, there is a heatsink in the form of the MCPCB and it rests against the outer shell, perhaps there is even some arctic silver in between, but the important thing is it is clamped with a bolt and a tiny nut, ensuring there is enough pressure for adequate transfer of heath.
So I would say there is a bond between the inside alumnium and the Ti-outer.

2. IIRC it is Titanium, but not pure, it is Grade 5 or as it is called as well: 6AL4V which indicates there is aluminium and vanadium in it. There is 6% aluminium and 4% vanadium. The Aluminium and Vanadium makes it easier to machine it.
More aluminium in the titanium is very well possible, but would degrade the corrosion resistance and also the strenght of the alloy. Also, it would take a whole lot of aluminium to increase the conductivity of the alloy!
Pitty i don't have the big alloy book anymore. I used to have a extensive reference work for all kinds of metals and alloys.
And it doesn't sound as sexy as 90% titanium