First measure at start , then at 20 seconds , third at 5 minutes...both bodies have similar thermal resistance , the Yezl's is a bit bigger with a brass pill against a aluminiun one .
There's two measures because the second was made with the cells switched to null that issue.
The Lumens are inflated ...the ambient temperature here is 15 C , at 5 minutes boths heads became warm perhaps at 45 C...the P10 has mass added to the pill but I don't think that at 5 minutes this is an issue to count, maybe I'm wrong and this test must be done with the same body ...
At start 1100L at 20 seconds ,1040 L at 5 minutes 950 L
Apparently the P10's body dissipate the heat more efficiently since has almost not thermal sag . Anyway beside the at start output the flux is almost the same. At more current, say 3.5A , the results obviously should be worst .
This results could be predicted reading the datasheet carefully . Relative Flux vs. Junction Temperature isn't a bogus chapter which could be skipped simply ignoring it, is there for a reason
I was able to do a 10x7135 very easily once I bent the pins of the top chips down. Use small needle nose pliers and bend them to a 90 degree perpendicular. The drop down perfectly and almost touches the pins of the botom chip. Its a _MUCH _smaller (and easier) gap to solder bridge over. I don’t know how you would do a 3x or more stack though.
Sure np…. and for the record I absolutely could not do the mod without bending the tabs down. It was just too big of a gap to bridge with too many adjacent solder points. I would always end up bridging “sideways” contacts instead of straight down. Solder braid saved me though… good stuff to have around it can clean up even the most sloppy jobs.