Thermal properties of titanium?

In typical rookie fashion, I'm building a dyno-driven 3-cree setup for my winter/distance bike. I'm going to use Martin's circuit #10 (with the Hi/Lo switch), and build the circuitry into the same housing as the emitter/pcb/optic.

One question I've been unable to answer: housing material.
I've got access to lots of Grade-9 titanium tubing in appropriate diameters, with wall thicknesses running between .8 and 2mm. Building the optic/emitter/pcb and circuitry into this stuff will be very simple. What I don't know is whether or not the tubing will dissipate sufficient heat, even assuming that I build a good thermal pathway between the emitters and the housing.
My thought is to attach the PCB to a 10-15mm thick aluminum disc with arctic silver, and then anchor the disc to the inside of the housing with more thermal compound and machine screws. This ought to be a really good thermal path to the housing. The light unit will wind up being ~6" long, and the rear cover will be another aluminum disc, sealed with silicon this time.
Will this allow the housing to dissipate enough heat? Or is the wall thickness of the tubing too thin? There will be no standlight included, so Moving=lighting. As well, I see no reason not to seal and insulate the circuitry with tool dip or latex.
Thoughts?
Sarcasm?

Eamon

The thermal conductivity is about 1/10 that of aluminum. If the Al heat sink slug gets hot it will expand volume wise more that the titanium tubing. So the Al slug will heat up, expand radially, the Ti wall will not be able to dissapate the heat, the slug will be in compression with the wall and cause ring stress on the tubing with the epoxy in between. Probably crack you epoxy if it is brittle. I'd stick with aluminum. Included copper for comparison.

Al = 136 Btu / (hr - ft -°F)
Ti (pure) = 12.6 Btu / (hr - ft -°F)
Cu = 231 Btu / (hr - ft -°F)

And the coef of expansion is quite different also.
Al = 13.1 (in/in/°F x 10-6)
Ti (pure) = 4.7 (in/in/°F x 10-6)
Cu = 9.8 (in/in/°F x 10-6)

Got it.
It would be really cool, but not very functional.

thank you

I have been very successful with a sandwich of aluminum core and Ti surround. The aluminum is pressed in and the wall thickness is great enough that uneven expansion has not caused problems.

However this is for a hand held light and the Ti heats up nicely and is dissipated by the hand holding the light.
I have no idea how you get rid of the heat just with air flow.
I think this is a case where you just have to experiment.

At about 500 ma provided by the dyno, I don't think you have to worry very much about the heat. What worries me more though is the length and weight of your light, it will be subjected to a lot of vibrations. A heat sink half the size of what you are suggesting will likely be more than enough.

The following thread may help.
http://candlepowerforums.com/vb/showthread.php?t=102758