My own Personal experience running a 64623...I have once run one for the full runtime around 8 mins constantly when I needed to get out of a dark place. ( I think I had a break of about 30 seconds half way into the run, but It didnt cool down alot)
The reflector survived fine...As far as I can tell it was / is a 2" litho reflector ( I didn't make that 623...but it looks like my other new and unused Litho reflectors, although this must have been drilled out wider...or bought wider.)
Its my favorites, as its quite thick...and seems to be able to take the heat well.
That runtime "test" was done in the Summer...although England summer isn't exactly hot...but its not like it was snowing..if you get the idea.
However...Reflector coatings do "bubble". I have messed about with 250W osrams, and even being realy carefull, 10- 20 second bursts...I have very very very slightly bubbled a FM2" reflector.
I have seen destroyed photos reflectors using the same lamp ( 64657) ...and even heard of 64458's melting FM3" reflectors....which is weird, as I run a 64633...which tops out at the same temp as the 458 ( thanks again for the charts lux)...but I guess My friends 64458 was pushed longer, and harder than I push the 633 ( and the 458's I have used in the past.)
I think the 64623...is about the limmit that these reflectors can take for a few minutes runtime....and to look at it another way...at 10A on the 64623...thats about all the runtime you get in an 2D 13 cell package.
I have run a 6C 4 x emoli 26700 set up with the 64622, and 64633...but only for a few minutes at a time...4-5 mins tops, and that reflector survived OK aswell ( FM2") but its always hard to tell what has more V lamp over that and the 2D 623/633
I'm Still looking into reflector alternatives...With regards to not peeling..not sure if they will work though....its a bit of an uphill struggle ATM.