NCR18650 - about 60-61 mins
Solarforce V2 2400 - 50 mins
TF Flames 2500 - 45-47 min (i measured about 3.9A just before it cuts off). I've cycled 8 times for these 2 cells, so pretty indicative of longer term performance.
Xtar 2600 - 53-54 mins.
Draw ranges from 2.5A fresh to just under 4A. Light is regulated so it starts off low current and the amps then gradually rises. So the tests are pretty high draw for a LiCo.
Incoming - Xtar 2600
Which version is the NCR 18650? Is that the Panasonic 2900's?
I have the same light and im using the TF Flames in it now, But I think with the great regulation the light has and how it boosts so well when the cells start dropping off the extra run time of the 2900's should be perfect for it.
Im saving my Panasonic 2400's and HiMax 2600 for single 18650 lights. The power curve of the Xtar and HiMax are almost identical iirc and it drops off so fast at the end of the charge and will stay regulated most of its run as its voltage is more linear that it may not be the best for the Sky Ray 3xT6, But thats just my theory.
But i forward look forward to your tests. Where is the regulation drop off point on your's? 3.3v like its been reported?
Yep, they are Panasonic 2900s. NCR is nickle something, which the 2900s are and presumably 3100s too. Panasonic 2400s are not NCR, they are CGR as in CGR18650.
I learn 10 new things a day here! I thought NCR had to do with a company name.
XTAR 2600 - 53-54 min.
I can get you pics of the bottom of the driver! lol
Sorry but im not popping it apart for curiosity reasons, Im still to much of a noob to feel safe doing stuff like that unnecessarily. Sorry tho i wished i could.
I think i asked this before, But has it ever been proven the Xtar 2600 and HiMax 2600 are both the same Sanyo cell? My HiMax's will be here soon as they just left NY a day or so ago.
Same with my Panasonic 2900's and those will be going into my Sky Ray 3800, And the HiMax will power my single cell XM-L 3A lights.
2100- Do you have any of Reverand Jim's Panasonic 2400 unprotected cells? Id like to know the run times on those also. If you dont i can test them and post the results if you can verify the lights protection will kick in and shut down before i drain the cells too far? Also what exact standards are you using like starting voltages on the cells and anything else?
I really love those 2400's especially for what i paid for them. Here is the pdf of the cells specs: http://www.rosebatteries.com/pdfs/Panasonic%20CGR18650D.pdf
Starting voltage is about 4.20-4.21V thereabouts (the TF Flames I could only get them to 4.19V). Basically for the initial 4.21 - 4.15 the drop is very fast, no big issue if it is out by 0.02V at the high range unless one is extremely anal to the exact last minute (is there any point anyway?). Cut off is 3.6V thereabouts. I am not in CONUS, so no CGR18650 2400.
BTW Dereelight is selling the Panasonic CGR18650 2800s.
I looked at it and frankly there's not much to see. It's dual stack but they sanded off the surface of many of the chips (prevent copying?). Looks kind of like this but the bottom layer is tucked in really close so can't see what's on it:
http://www.dealextreme.com/p/12v-sst-90-5-mode-memory-circuit-board-for-flashlights-50026?r=43033281
Thats a good price on the Panasonic 2800, But with the 2900 and 3100 just a few bucks more its hard to buy lesser capacity at this point.
Did you use any active cooling like a fan pointed at the light? I ask as it makes such a long high amp burn easier on the emitters, But they still aren't driven very hard. But for long life of the light it would help.
But also it can skew the results as cooler conditions could make it pull higher amperage or it could go down for all i know and that could make a pretty specific large difference in the results. Ill do it with no cooling as i don't think you did and check back for your answer as i should have a while before heat affects this light much.
(Edit) Well it will be a while as none of my cells are matched nor are any at 4.2v so i have two in the Pila now. Ill be up late most likely so im gonna try and do this if they charge fast enough.
My battery tests are run under extremely controlled conditions. Nothing gets hot above 30 deg C, drivers/batteries etc....each and every run is the same. I know temperature can cause electronics and battery discharges to be somewhat erratic esp for different batts. These differences can cause slight difference in discharge power and of course the voltage cut off point. It is quite repeatable, as you can see...all the tests are run at least 3X.
The Panny and Solarforce are crazy. I have 2 pairs (4 batts) of Solarforces and they charge up and cut off at exactly the same within 0.01V. I have absolute confidence in running 3S or even 4S for them. 3S light coming....3xT4. So in a real world scenario if you are not gunning the light for > 10 mins at a time, there is absolutely no issue. There is probably more danger setting off a 100-shot fireworks cake.
Actually I think that CGR18650S 2800 is 11.90 per cell, so its more expensive than HKequipment's NCR18650 2900s at 19.90 per paid. 
Lol, your very first batch is higher quality still than even mine. Mine came w/ 1.3 o-rings (part of second o-ring stuck to first), and I have no idea if they did that intentionally or not becuase the lens is loose with only 1. Certainly not water tight.
Maybe....that is why when news came out here, within 1-2hrs I already purchased from manafont. LOL! Too bad they had a long weekend so it was delayed a bit. Mine is water tight till the bottom of a big pail, never tried a swimming pool, and there really is no need anyway.
Yes thats true, But the Caliies 3100 at $13.95 is higher, But since they are located in the USA and i can have them in my hands in 3 days ill gladly pay the extra money.
I already have 4 of the HKE Panny 2900's on the way and should be here soon as they are heading to my local USPS distribution center.
Im also having a hard time pushing these cells too 4.2v on my pila charger so i may have to let them sit overnight as it will eventually take them to 4.2v on the nose if left long enough. If that don't do it ill pull out the old trusty TR-001 charger that has no issues getting the cells high enough.
Ill just use a fan as im not putting my light in a sink full of water as with my luck ill just have another paper weight that needs new electronics. So my readings will just be generally related as im not even using the same exact light as you as no 2 lights are exactly the same no matter how well made.
Mine came the exact same way... 2 o-rings at the lens on top of each other. The lens also was loose. You can either correct this problem by buying a thicker lens (this one works, http://www.dealextreme.com/p/52-8mm-replacement-glass-lens-for-flashlights-5-pack-25231 )
OR
Use the O-ring from the head of a D sized maglight. It fits perfect on the skyray and the lens is very watertight and no longer loose. I did this to mine, and the lens doesn't move anymore (even if I try to make it move I can't)
You should do just fine with a 5-10kph fan breeze at night. It is just that over here it's always very warm, you place i can see now its 23-26 deg C in ohio.