I think it does matter if you’re going to run the light at 7A and more without melting the solder off the wires and emitter.
I think it doesn't matter since I've gone higher than that and none of what you fear actually happened. Show me some data to back this up, other folks have gone looking for it and not found anything yet.
Falling lumens argues with you Comfy. Your Convoy S6 SSSP Triple XP-G2 at 5.33A vs. my L2m Triple XP-G2 at 5.23A. Yours has a pill made from aluminum. Mine has a 3 ounce copper pill. Yours starts at 1580.1 lumens, mine starts at 1542.15. In 30 seconds, yours is making 1442.1 lumens, mine is doing 1486.93. Yours is running from a Samsung 18650 20R with 2000mAh capacity, mine is running from an AW IMR18350 with 700 mAh capacity.
The copper allows higher lumens for longer run times, even from a smaller weaker cell. In just 30 seconds, the L2m made up the 37.95 lumen gap AND passed your SSSP by 44.8, a mean difference of 82.75 lumens in only 30 seconds.
Data.
Shall I run em both for 5 minutes and take down the numbers? Measure the heat output at the head? Do you need that much more data?
Most of lightmeters have a +/-5% tolerance on the lower ranges and up to +/-10% on the higher ranges, just a thing to consider. The accuracy is not on voltmeter level for example.
If the meter is off, it’s off for both of em.
Wow! 5 minutes on high, that's longer than I've ever dared go with the tiny triple! I've never dared go much past the two minute mark. That thing drops like a rock after the 60-90 second mark. Lots of heat in that small head. No way around that unless you want to go bigger or heavier, which is exactly the opposite of what this light is all about.
That said, I really don't believe you would see as big a difference in the big 2 x 26650 host going to copper. It is a MUCH larger pill, extremely thick with lots of contact area (tons of threads) with the head, a world apart from the tiny SSSP pill and head.
Oh, I think the S6 did a very good job, being limited to an aluminum heat sink. They were both so hot I could barely pull the tail cap off after 5 minutes. But look, the L2m did it with a considerably smaller lesser cell! Copper, my friend! The C8 trifecta proves it out, my Cu M8 proves it out. Whether at 3A or 10.5A, copper does your body good, pass it on… 
Looking to buy: One chunk of sweet copper.
Bucket is the Cu go-to guy. Fast, excellent workmanship, and reasonably priced. He da man!
OK, not doing lights for anybody but me from now on. Dale, until your chart shows battery voltage alongside the light output, you're only seeing half the picture and pretending like the other half doesn't exist. Have fun guys.
I didn’t say the light you build was bad in any way, it just has limitations due to it’s size and whether or not copper would make any difference at all I simply do not know. Yeah, I could put my light engine in an L2P and run it with an 18650 20R so they’d be apples to apples across the board. Not much point in setting up to read voltage or amperage or anything else, as they are flashlights, handheld tools to chase darkness away and that’s how I approach em. When the power goes off, how long is it going to do what it does? All that matters to me, doesn’t even matter which one does it better or longer, I’ll just grab another when those cells die. No biggie.
Your end of the S6 is gorgeous, nothing wrong there at all. Already told you that.
For those who were thinking about grabbing the new 2x26650 host here is a little write up I did; It shows how well the dry driver and MT-G2 fits in it
Build Using New Host from Mountain Electronics
Well I don't understand at all then. You claim your data proves something, then say it doesn't matter. The light output is only the end result of temperature, input power, and output power. It tells you less than nothing about the pill material.
Some people are always going to believe that copper is a substance infused with pure magic that causes heat to evaporate away into a parallel universe, I guess I'll have to work on accepting that I can't fix irrational beliefs.
I only know what I read in the forum, and I assumed that copper is superior in delaying heat sag.
Copper is a better conductor of heat than aluminum but aluminum radiates heat faster than copper due to its lower density.
Other words copper pulls heat off faster than aluminum but aluminum radiates heat to ambient temperature faster than copper.
It converts heat into that dark energy they can’t ever seem to locate? :bigsmile:
Much the same way that while solder has not very good thermal properties it still is able to work well because the layer of solder is so thin, the pill-to-head/body contact area is so much greater than the area where the heat comes from (LED's center pad) that even a relatively poor material is able to work not much different than a really good material.
Why do copper MCPCBs work better than aluminum MCPCBs? Is it because they're made of copper? Or because the dielectric layer can be eliminated since copper can be readily soldered to using non-exotic solders & flux? If it were possible to solder a LED direct to an aluminum board without a dielectric layer, I'm pretty sure there would be no performance advantage to using copper and copper boards wouldn't exist because there'd be no reason for anyone to invent them. Again, compare the contact area between the LED center pad and the MCPCB, and the contact area between the MCPCB and the pill.
SinkPAD makes both aluminum and copper direct thermal path stars. I have a couple of the aluminum ones, though never did a head to head. I believe their site has test results. I thought recalling it was pretty clear the copper was superior, just the aluminum was offered for reduced cost and weight, but still the SinkPAD aluminum was better than the 10 cent aluminum stars in most flashlights.
q = k A dT / s (1)
where
q = heat transfer (W, J/s, Btu/s)
A = heat transfer area (m2, ft2)
k = thermal conductivity of the material (W/m.K or W/m oC, Btu/(hr oF ft2/ft))
dT = temperature difference across the material (K or oC, oF)
s = material thickness (m, ft)