No hair splitting. Does that or does not do that, simple.
If you've worked with brass you know how easy it can oxidize and how it smells.
Fujik (BAB 900) is 0.88W/m.K rated. Still way better than air at 25C 0.024W/m.K
I'm sorry, I forgot to include the sarcasm sign. Here it is:
Now that we agree about brass oxidizing and me having explained what I wanted to say, mind if we move on? ;)
Wow, that bad.. thank god I never used it to glue PCBs. Thats almost as good as toothpaste. :D But thanks for providing that info.
I mentioned it a least a couple of times.
These are the thermal conductivity ratings for all “Fujik” branded thermal pastes:
BAB 900 – 0.88w/m.k (The common one that everyone buys from DX)
BAB 950 – 0.75w/m.k
BAB 660 – 1.10w/m.k
BAB 680 – 0.90w/m.k
BAB 690 – 1.10w/m.k
Toothpaste also tastes much better than Fujik does.
You just need to buy the right brand, “Coolmint” toothpaste has far better thermal conductivity than fujik.
Tested it in my mouth, definitely cooler… 
I actually use Weleda toothpaste. :D
I think electrical resistance of host and pill is negligible. Contact resistance where two materials meet is a lot higher. And think about all the springs, switches etc. Also, don't forget the 0.1V overhead needed for the AMC to work properly.
It's not negligible, spring can be made of several materials, you can have a better electrical conductivity with a better material, that is the point where yo udo not have to increase mass.
Copper is 100% then brass is 28% or phosphor bronze 15%. Silver(pure) is actually rated 105%.
7135 are linear chips, there are other drivers that use sense resistors like buck or boost drivers. Several companies built them, AMC is only the part where you talk about the manufacturer not about the product, you can have AMC7150 and many other AMCXXX.
EDIT: NightCrawl edited his post after my post.
The simple days of dropping a cardboard wrapped alkaline into a tin host with a dull glow orange bulb are gone. 
Hail the orange bulb!
Long live the memory of ancient times, one single 2D MagLight on top of the fridge to serve the household. (Spare bulb intact in the tailcap)
Our lives are so complicated nowadays… 6061, 6063, 7075, T6 or not T6, copper or brass? xpe, xpg or xml? cree or luminus or nichia or….? IMR, INR, ICR, LiNiMnCo?… where did we go astray?
This of course is said in jest.
I modded a few lights recently that had beautiful design, perfectly executed drivers, one of them double stacked with plastic clips plugging the switch and emitters into the board. All this thought put into an exceptional light. But about 1/16” of aluminum under the pcb’s holding 4 XM-L U3 emitters and a cut-out almost a third of the way around the base of the star apparently for nothing more than the wires to come through. They even have anodizing under the pcb!! Where does the heat go?
Sure looks good though! That last one, a name brand light retailing for some $150 and more.
Don’t hold back, name and shame! 
There is no denying that for anyone who mod, especially with hand tools, copper is the way to go. Its just easier to work with. That is why everyone use it. Me included...
There is no denying that MRsDNF makes the holy grail of copper pills that everyone drools over.. :love:
But this is not about modding in general, its mostly about the C8 and HD2010.
One argument some use is that copper should or might help with emitter life, but there is absolutely no reason for that. If you want longer emitter life you want to run the emitter at a lower junction temperature. Copper mcpcb makes the LED run at a lower temperature, which is why there is increase in output.
A copper pill in those lights have shown zero flux increase in the 4A range compared to aluminium pill. If there is zero gain, that means the emitter is running at the same junction temperature as when using aluminium pill. Which means no increase in emitter life when using a copper pill..
Just for reference, picture showing output vs temperature.
Junction temperature is basically measured under the LED die (as far as I have seen). Under the LED die is the built in heat sink in the emitter, then there is solder attaching the emitter to the mcpcb. On top of the mcpcb is normally a dialectic layer, then the baseplate of the mcpcb, then thermal paste, then pill, then flashlight body.
Once the dialectic layer is gone and mcpcb base is copper (Sinkpad/Noctigon upgrade) People should worry more about getting the mcpcb attached to the pill as good as possible instead of focusing on copper pills which basically give you zero gain in those lights (or most lights in general).
Nightcrawl have refereed to tests that showed no gain with copper pill even over brass. Tom E have tested copper C8 pill in the 4A+ range and there was no gain in that case either.
People could claim/hope that you just need higher current in order to see the advantage but we are talking about a C8 and HD2010 here, both single cell lights. Even with resistance mods and high current 26650 cells you will not stay above 4A for a long amount of time. And for people who normally use protected 18650 batteries (Sanyo/Panasonic), peak output will normally be around/below 3,7A to a single XM-L2. And it will decline down towards 3A fairly soon...
Im all for copper, im all for optimizing light output, but in terms of performance, which is the main reason most buys these. Based on what have been documented so far, its not worth it from a performance point of view..
Btw, very recently I ordered a complete Convoy L4 (on sale), cost me about the same as a copper pill. Based on what I have seen on pictures, the L4 does not have a pill. It seems to have direct thermal path to the head. Its even got tapped holes and screws for the mcpcb. Might be a worthy alternative/host for people who are wondering if they should buy a copper pill for a C8 light..
Picture borrowed from this revew.
if I wasn’t on a purchase embargo, I’d have a l4 on the way to me, and I do agree, eliminating the pill entirely is likely to be a good step up rather than changing the pill material, as said above, there are two junctions you have to overcome, star to pill then pill to host, eliminate the pill, you eliminate one of those junctions, lap the star to that host, and your pretty close to perfect without resorting to a copper pill-less head, when someone starts making one of those, then I’ll jump on the band wagon, but I largely think copper pills are a flashlight go faster stripe, nothing more.
If it makes people feel better then fine, thats the whole point of modding in the end.
Bold part is at the end of the post, just feel like saying that my edit does not matter at all before I go into the rest.
It is, for the flashlight host. Spring, of course not. Metal springs = bad. Gold plated metal springs = good. Bridged spring with copper braid = very good.
You said:
That clearly refers to the host, not any specific part of it. And I said "I think electrical resistance of host and pill is negligible." Now, whats wrong about that? You say voltage drop.
Ok, time for meth math.
Lets take a specific resistance of ~0.0265 Ohm*mm²/m for aluminium and ~0.8 Ohm*mm²/m for titanium (stainless steel is a little better).
Now, for a flashlight of 20cm length and lets just say 3mm² thats ~0.00177 for aluminium and~0.053 for titanium. Thats a drop of ~0.009V for aluminium and 0.265V for titanium. But thats at 5A. And 3mm². Realistically, a flashlight has at least 1mm wall thickness. A flashlight that takes 16340s has an ID of at least 16mm. So about 50mm² to conduct electricity. Now, in most flashlights the power does not have to travel 20cm to reach the driver, in a typical 18650 flashlight, lets assume 10cm.
That brings us down to 0.000053 Ohm or 0.000265V drop for aluminium and 0.0016 Ohm or 0.008V drop for titanium. Any idea how little that is? It is insignificant. Even more if you refer to such a tiny thing like the pill.
What matters more, is contact resistance. Cell to spring, spring to clicky, clicky to host, host to driver.. and so on.
About your point with the AMC I mentioned.. you do realize that people buy the copper pills to drive a C8 or HD2010 (single cell lights!) at about 5A with a 17mm driver.. and mentioned AMC stacking. Well, what driver are they probably gonna use? Wow, hard question. No, seriously. There are only few drivers that even allow 5A to the LED on a single cell, so its DD or an AMC-based driver for pretty much everyone. And still, for MOST drivers you need at least 0.1V overhead.
Now, about my post being edited.. I added exactly ONE word in the second sentence: "contact" before "resistance". Woah, that totally changed the meaning of my post!
Geez, you gotta be kidding me.
P.S.: I prefer to work with brass, its a lot easier to machine than copper. :P
BST?
Bingo! Sky win’s the door prize! Why did they design the light so well, only to overlook the basics? Anodization between the star and head? Such a thin layer of aluminum directly under the star? Yes, indeed, there is thicker aluminum to one side of the star, but not under the emitter. I’ve seen thermal graphics that indicate the thermal path to be a cone shape, much like the cone of light coming out of the emitter. So right off the top, there is a bottleneck as the heat is forced to find a way to the outer perimeter where the fins can disperse it to the air. I guess it works, but I wouldn’t think it’s anywhere near optimal.
The light looks fantastic, the driver is a work of art, love it and greatly appreciate how it came to me. But can I improve it? Hmmmm, putting Noctigons with XM-L2’s in the last one I saw…no other changes…netted a 700 lumen OTF gain. Copper played a big part in that, no doubt.
Knowing you, you will definitely find a way to improve it. Is there space to add copper under the platform to increase the thermal path to the body?
I had a BST once and was planning on modding it with copper stars, xml2, DrJones driver in salve configuration. However, the end cost for the project would have been too high for my to justify it. If I had the money though I’d do it. There is just something about the BST design and form factor that I really like.
EDIT: Sorry for the off topic post…
why don’t we petition cree to make the input connections to the emitters bigger that must be a bottle neck right there….
[sarcasm mode off]
NightCrawl, no need to be sarcastic and attacking. I wasn't attacking you. I am "people" too and personally I do not have interest in 7135 drivers right now. If you take the "AMC" vs "7135" as an attack for you or me specifying that you edited your post you are paranoid. I found that after you edited your post mine was not looking that good.
Attacking? No, not really, you just threw in some fancy data and made claims without anything to support them.
I dont know what driver you will use to get 5A to an XML2 in a single cell light, but it will most likely also need 0.1V overhead.
And.. paranoid? You said I edited my post and implied that this edit made your post look silly. But my only edit was to add ONE word, so stop making stuff up?
But I'm glad that math convinced you to join my opinion on electrical resistance (of host and pill) being negligible.