You should be able to run it on one pair, two pairs is a bonus. Try the pair of protected cells you have and leave the other two out…
Regarding protected cells I’ll say I didn’t give them much/any thought during the design process. Like I mentioned in my edit of post #19, if the user intends to run protected cells, different hardware can be used to accommodate them. The body tube depth from front lip to face of rear pcb is ~70.6mm. This means with a std protected cell at ~69mm, you’ll have 1.6mm remaining for a spring. This is reasonable as long as the right springs are used.
NOW, for those that are really eager to run prot. cells AND have beefy springs that are resistance modded, you can always add some spacers under the rear pcb and gain tons of room (relatively). The body tube with this setup no longer carries current so any material can be used. I’ll measure but the factory screws probably are long enough to give you ~2mm comfortably.
Any updates?
Yes. I’ve revised the design to enable mechanical lockout with aprox a 1/4 turn. I’m still working on learning eagle. I’ll post 2D drawings of the design here tonight. I may ask for help from the community to get this drawn up in Eagle for the sake of time. I just don’t have enough time to devote to this at the moment. :~
8000+ lumens? 
Something around that 
It’s at least 2x my K40M that’s running full throttle…
Would you say its more of a flooder than a thrower now comparatively?
Just did a ceiling bounce test with these two head to head, both on fresh cells. Light meter on the floor, sensor up, underneath a wooden table the lights were tail-standing on. 14lux startup and 12@30sfor the M6, 5 for the K40M (steady). That puts it around 9k+ lumens startup.
Thrower is such a subjective term. My personal descriptions of a thrower may not match yours, but last I checked it did in the 70K lux range at 30s. I don’t remember startup values, or if I’ve ever recorded them.