Having seen the impact that lapping can have on CPU cooling the results with the 2010 seem to show what can be achieved with elbow grease. I can see now why people were getting excited about these.
Why is the XM-L U3 100 lumens brighter then the XM-L2 U2 even though the U2 has slightly higher amperage?
@30sec?
A bit useless IMHO.
What are the results on 5-10minutes?
Something useable. Unless You plan to use your light under 1 minute?
Tom - where did u find a ucl for the hd2010?
nice, thanks, I also grabbed two of the 320-800 assorted. That should be enough to do the 30 boards sitting here 
It’s not. The XM-L2 U2 is 387 lumens brighter than the XM-L U3. Sure makes my stock HD 2010 look sick. Lol. It ended up 770 lumens otf after 30 seconds and throws about 53k
you are correct, i misread the results
however, why does the post mod hd2010 have 1360 lumen @ 4.09A and the 7G9 have 1264 lumen at 4.2A?
What is a good soldering paste to mount led’s on a copper star? I have destroyed 1 star so far with no luck. :bigsmile:
I’m assuming those are tail current measurements, not current at the emitter, so assuming the measurements are accurate - driver efficiency, and the acceptable deviation of emitter performance (7% range for each bin?)
emitter current, good point, perhaps someone should test emitter current and voltage drops, a table of this would make a great sticky thread
flashlightlens.com - it's a little big, think it's thicker. You got usually 1mm to play with in width. Pretty sure I can't fully screw down the bezel as far as it used to, but its over the o-ring.
Bort - as far as the HD2010 doing better in lumens then the 7G9, amps helps but it's not the only thing that matters. We are talking minor differences here, and that HD2010 reflector does well for putting a lot of light out - good in throw, while the 7G9 targets throw more - tighter hot spot. So is my lightbox perfect? No, no chance, but could the HD2010 be a little better at putting out lumens in it's setup than the 7G9, certainly could be. There's also signifcant differences in batteries, LED emitter from one to another, etc. For example a binned 1C can be all over the place, more cool or more neutral sample to sample.
One difference is certainly the battery configuration - HD2010 running a 26650 (good one - KK) while the 7G9 uses 3 18650's in parallel - tested with Panasonic 2900's unprotected. Batteries make a big difference in measurements I've taken. You put 10 mins of runtime on a battery, brightness suffers quite a bit, even with regulated drivers.
great results, particularly the reduced output drop with the sinkpads. Possible lower Vf might help too, but everything else (host etc) should be equal. Although output after 5min or longer would be interesting, you’ll then be looking at the thermal properties of the host/ dropin etc, not the LED star.
Well, I used a paste that was marked to do aluminum, but it would do other metals as well - bit expensive because of it supposedly for soldering aluminum but I didn't have much luck with it, so, used the same stuff for the LED mounts - worked well.
Got it here: http://www.amazon.com/dp/B0058ED0LE/ref=pe_175190_21431760_3p_M3T1_ST1_dp_1
Where did u get the sinkpads?
nitro organized a group buy thing - nitro posted above. Not sure, but he may have more left. I wish I ordered more now 
, but 16 mm's would be great, but the 16's are not stock items at SinkPAD, the 20's were. Possible plan is to put another group buy together for 16's (or 14's), but include payment for the tooling costs to SinkPAD, so it will cost more but the more we buy, the less % is the tooling cost.
Thermal compound/paste is only good for bridging vewwy vewwy tiny gaps, such as microscopic irregularities in surface finish, down to the scale where you can't see any roughness with the naked eye. If you have a divot or pocket or wide gap, skip the paste, it won't do anything but make a huge mess later on when you take it apart again. Heat just doesn't transfer thru a big glob of paste.
I know a lot of you guys just like tweaking things for the sake of tweaking, but the fancy expensive exotic compounds aren't doing anything but lightening your wallets... just pick up a giant tube of white silicone-based paste from the electronics store. It'll last you thirty-seven forevers, and performance, when used correctly, will be within the margin of error for whatever equipment you're measuring with. Changes in ambient temp during the test will show more variation than between the most expensive and cheapest stuff.
Similarly, for filling gaps or pockets so you can correctly use thermal paste, or for permanently mounting stuff, plain ol' JB Weld is just as good as Arctic Alumina or any of the other fancy ones. And like the white thermal paste, you can get it in giant tubes that last forever ('Industro-Weld' - exact same JB Weld, just in bigger tubes). Skip the JB Quik - it's not as strong or as good at heat transfer. If you're curious, squirt out some of the black JB, and poke at it with a magnet... (hint: it's got lots and lots of metal in it, but it's not electrically conductive).
Interesting, comfychair. I’ve used superglue to bond stars to a heatsink before. apply a small amount, press with a lot of weight for about 10 seconds, and it’s on. I was surprised to see the star only be 4-5 degrees hotter than the heatsink. Worked quite well, and it’s temporary since the glue will not bond good to smooth aluminum. I would not use it in a real application unless in a pinch, Shock or lateral pressure seems to jar them loose sometimes.
I either use Fujik (probably a fancy name for overpriced silicone sealer) or thermal paste when I have something to hold the star down.
Has nothing to do with driver efficiency, I promise. They’re both linear drivers (AMC7135 Based) which means that tail current = emitter current.
PPtk
i see, do you know anything about voltage drops, as i am looking to order a driver very soon and would like the one with the least drop