Great, thanks! 
I am going to go with the:
Edmund AX32345 Grade 1
- LENS PCX 95MM DIA X 166MM FL
- 16mm thick and 3mm at the rim
Being a grade 1, it's crystal clear and coated, with no flaws. I will have to rethink the head. 3" PVC would work, well a 3" to 2" reducer, a 3" coupling and 3" pipe. That will reduce the size of the head. I also found a 95mm lens hood, but I would have to put 2 or 3 of them together because of the long FL. That would be very expensive, so PVC will probably still be the material I use.
Still if u like the beam and the light in the end you could dedicate some more time and create an metallic head. And it could help, driving one led at 10 amps need some serious mass and big surface to radiate heat.
Really curious to see how this will evolve.
So, I decided against PVC. I went to speedymetals online and bought 4" OD 6061 Aluminum pipe. That gives me 3.750" ID and the lens is 3.740" OD. I also bought a sheet of .050" 6061 Aluminum, to use inside the pipe, to let the lens rest on. I will form it to fit in the pipe and also on top, to sandwich the lens and hold it in place. The bottom, where it reduces to the OD of the Mag head, will be a piece of 1.750" 6061 Aluminum round stock. That fits in the pipe and I will drill and dremel the opening for the Mag head. It will have to be a press fit. Inside the Mag head will be a big heat sink, that will fit tight in the head and the LED on top. It should take a bunch of heat and it will be so heavy it will need a shoulder strap, LOL.
Argh, MRsDNF must be rubbing off!
Is it rutting season? 
I'm here watching with interest OL. I was led to believe that Texans did not need shoulder straps like some wimpy Aussies.
Photos of the Wavien Collar in the OP.
to bad we cant get a GB going on some collar's
Nice to see some closeups of that collar! Yeah, a group buy would be interesting, if they were priced about half of what they are Iād have snapped one up by now. Probably wouldnāt be the only one, either.
Cheers
I added more photos of the "comfychaid BLF collar", in the OP. I will also use it in the comparison tests.
looks like its gonna be a thrower. :bigsmile:
Thrower and a firestarter in one 
well i evaluate wavien collar inappropriately expensive in my book. lets make it clear, technically i find it appropriate to use in a thrower, but, the cost and the added complexity in the final design together with questionable total output in lumen have keep me away from using it in my projects.
that said i like to see it well implemented by other of course so here are my thought on how to use it. i was from some times going to make an separate thread on collars and ways to implement but since i have not find the time to do it i think its a good idea to contribute here.
for simplicity i quickly draw 3 main ways on how collar can be implemented:
(sorry for the low quality diagram but since i was uploading from my phone and have limited data plan i downgraded JPEG quality before upload. )
case 1. is the best way and the simple way to implement it. all the otherwise dead angle of emitted light from led that without collar is wasted, with collar is recycled being reflected back again in the die have a second chance to enter in usable angle of emitted light B. how much of the A is recycled to B this is very debatable. even that we for a moment assume that the quality of collar is perfect and focus is perfect(which in reality its not possible and always will have loses ) i think not all reflected light on the die will excite phosphor layer resulting in produced light THAT GO in the right direction. some will be wasted as heat and some will have to cycle again. if we assume that 2/3 of energy go in heat and 1/3 go in light as a rough estimation i could say ~30 % of A will go to B.
but pay attention, with different collars we could have different values for A and B! if we have for example A=B so 45 degree each in the best case scenario (perfect focus of collar, 100 reflective surface and so onā¦ ) we should see ~15% increase in in led luminance and so 15% increase in throw.
i think the blue emitted light will have more chances to excite phosphor layer and transformed in yellow so tint may shift a little from white to yellow.
BUT for a bigger A we theoretically should gain more in luminance ā¦.
so case 2. A is very big and B very small lets go to the extreme when A is 89 degree and B 1. we should gain ~32% in luminance and so in throw BUT this scenario have an big flaw. small angle B mean a very small surface in lens will contribute to throw so the total output will suffer from this. its clear the loses will be bigger that the gain from collar.
in this scenario i go to the extremes to emphasis the fact that the use of the collar more than anything else is an balance and the correct way to use it is very important in the end result.
implement it right we should see some gains in throw and tint shift, but, implement it wrong and we should see decrease in throw.
case 3 is more safe there we have some gains in throw. of course less than the fine tuned case 1.
Trying to improve it further
for some time as i said have thinking about how to implement collar.
here is my idea for an extreme thrower :
A go to the max, but since B is very small and we canāt lose precious surface on the main lens we can chose an lens with very big focal point placed far away from led, but since this its not an practical solution we can ad an second lens to defocus the beam. i do not know if this can be done and of course is all theory, but we can use or one small convex or concave lenses with very small focal points. this way we change angle of B to match the big surface of the main lens. making smaller the system we should use the type of main lens also with a big lens with small focal like in the draw.
hope what i said make sense, i have make like to many assumptions and for example if the amount of recycled light is bigger than 1/3 for whatever reason we should see more gain in throw as i predicted. :bigsmile:
01/27/15
I do love to get stuff in the mail. Photo added in the OP. The Aluminum came in.
Looking good OL. This is going to get interesting!
A little more is done
and is posted in the OP.
Hmm. Lookingā¦ heavy!
I believe he does like a bit of heavy metal going by the AC/DC songs in some of his threads.
There is a lot of work in that hole OL. Patience. I'll have to keep looking for it.
A couple more photos in the OP today. Also one of the dedomed SST-90 on the Maxtoch MT-G2 mcpcb.
01/31/15
The test collars:
Maglite Incandescent C/D reflector cut down to use as a collar, on the left.
Wavien Medium RLT Collar, on the right.
SST-90 led de-domed and mounted as it will be in the finished light.
Lens is in the tube and adjusted for the highest lux value at one meter.
Two different batteries. One complete test with each one.
Results
Battery One:
Bare led - 104,000 lux@ one meter
Mag collar - 102,300 lux@ one meter
Wavien collar - 104,500 lux@ one meter
Battery two:
Bare led - 102,900 lux@ one meter
Mag collar - 101,500 lux@ one meter
Wavien collar - 103,000 lux@ one meter
Battery two, with the collars raised 2mm, with shim stock:
Mag collar - 103,200 lux@ one meter
Wavien collar - 102,900 lux@ one meter
Those are the results. I have no opinion, not this time, no way.
Keep in mind that this is not a scientific test. This is a test performed by an old fart, in a garage, in East Texas. Form your own opinions from that... Or not...
Also posted in the OP.