Well holy chit... They really were using small car driving lamps. Hella Micro Xenon's were probably these:

https://www.rallylights.com/hella/Micro_DE.asp

(i've bought nearly all the auxiliary lamps for my car from www.rallylights.com)

I've thought about doing that with a set of smaller driving lamps i have laying around...

Then do what I do, run 2 bulbs with an A-B switch. If one burns out I can just flip it to the other. But I haven't had to yet because the bulb hasn't burned out on me. Redundancy is the key to having light. I have 2 MR16's wired up to a switch, a Cree xre headlight, and another 3 LED cheapie on the bar in case the rest fail.

I ride pitch black trails and can't afford to be without light.

Bang for the buck, the overvolted MR16 is the way to go...

I think I might have figured out what happened with my attempt at overvolting the Krypton star. The culprit is the fact that standard flashlight bulbs have short lifetimes around 25 hours or so. Compare that to quality MR-16s, MR-11s etc., which have lifetimes of 3000-5000 hours. Now, a say 20% overvoltage of the latter gives an acceptable lifetime of say 300-500 hrs, but do the same with the flashlight bulb, and one gets a measly 2.5 hrs.

Of course, bulb quality varies from individual bulb to bulb. So I'm guessing this is why my slightly undervolted magnum star is still working after two years.

Now, the interesting question is WHY do the MR lamps last so much longer than the PR-based bulbs.

Is PR halogen? I think halogen bulbs in general have longer life due to the regeneration of the tungsten filament by the tungsten-halide cycle.

True, halogen bulbs tend to last longer. But I checked with some halogen PR bulbs (aka HPRs), and their lifetimes are on the order of 20-50 hrs, not 2000-5000 hrs. So the mystery remains.

Uber lighting: 25 HID heads mounted to my handlebars.

Note: the following is a CONJECTURE, and I cannot say that it is correct. It is an educated guess only.
Here's a conjecture as to why MR bulbs last so much longer than flashlight-style HPR halogens. The HPRs have just one glass shell covering the filament. If it leaks due to age or whatever reason, say thermal cycling, the bulb is toast. MR style bulbs, OTOH, have a second cover that I am guessing is airtight too. So it would take TWO leaks before oxygen would enter the chamber with the filament and kill the bulb. This conjecture is supported at least by a comparison of rough lifetimes. HPR - 50 hours. 50 squared (assuming equal seal quality) = 2500, and incidentially, MR bulbs often have lifetimes around that value.

If leaks are the main killer of halogen bulbs, then this conjecture would make sense.

smartypants actually that's real good reasoning. I wonder if true?

I've read that Halogens are most stressed during turn on due to the inrush current. Anecdotal evidence here would seem to support that. I've also read about soft starts for MR16s/11s, has anyone tried that? I'm in the process of throwing together a DIY MR16 system and I think i'll incorporate a soft start.

Depends on how much time and money you want to spend to protect a $4 bulb. I like simplicity and I haven't had to replace too many halogens because they blew out.