Look closer at the results. With your LED on an aluminum star that's on aluminum, you're not going to get within 30% of the specs. Putting the LED directly on copper gets a lot closer to the spec. So my question is "Why wouldn't you want to get an 30% more lumens for the same power output?"

As I said earlier, if you're not going to do your research, use good vendors and won't do math, then you take a big risk when buying Chinese lights. If you're willing to put forth a little effort and do some simple math, you can get a LOT for what you pay for.

The actual lumens to dollars ratio with Chinese lights is very hard for US lights to match.

Due caution is indeed warranted if you are going to burn magnesium to make MgO. If you have never done it, it is an extremely intense fire that, once started, is almost impossible to put out until the metal is consumed. It burns very hot and with a very high UV flux.

I think you both misinterpreted my question. Looking at colleen c's picture, is each LED in the array a discreet LED with a discreet output or is the output measured for the whole chip? It is unclear to me, reading page 2 of the Cree data sheet that colleen c linked to, whether the output is measured for each LED in the array or for the whole device. At 700mA, they say that the lumen output for an XM-L T6 is 280 lumens. If the whole chip is the "LED", then would the total output at 700 mA be 280 lumens or if each LED is putting out 280 lumens at 700 mA, the total light output would be close to the 1800 lumens listed by some manufacturers.

Like I said above, I've run the 900 and 1000 lumen Magicshines next to the Crees I have. The Cree has a narrower beam but they are far brighter than the Magicshines. They seem too bright for the reflector to be the only reason for the output.

This is a very important, because you have people running around talking about how they can tell lumens by sight. Sure to some extent it's fair to say that a brighter light is putting out more lumens, but to then jump to the conclusion that someone is lying to them, usually the Chinese. On top of that, none of these folks have any idea of how the reported lumens were measured, and therefore don't know if any exaggeration is really going on.

None of which is to say that some sellers aren't deliberately and maliciously overstating their products; I'm sure there are. However, I think there is plenty of likelihood that testing variance, bench results as opposed to application results, and production variables could, any one of them, explain differences between claimed and actual results.

Again, I understand the position that producer statements should match up with consumer expectations in some way, and ideal world/optimized/theoretical max lumen outputs don't serve that goal well, so I'm all for industry standards of some sort, ideally real world/actual application testing.

Maybe the proof of widespread lying and deceptive practices is out there, but I don't think I've seen it yet, and it feels un-American to me to presume guilt, and doubly so without any evidence to the contrary.

Oh, and thanks, Colleen, for pointing me back to the MTBR tests; I thought all of the '12 shootout lights got the IS testing, but as you note, not all did, and of course, the ones I looked at the second time, weren't!

Just speaking for myself - because riding around with something like this as a supplement to every emitter doesn't exactly look like a worthwhile tradeoff to me: Total 'out the front' lumens are dependent on several other things besides the extent the emmitter is driven to. An efficient optical system and an effective beam pattern can make a difference of a lot more than 30%.

Since most people in this forum seem intent on 'trading up' every few years anyway - its probably incidental. Personally I don't have any such intentions and already drive with as much or more lumens than anyone else on the board. My emitters are only driven to 80%. I probably just deal with less heat issues.

I'm quite sure the Cree specs are talking about the whole LED... its all one chip, under a single bubble lens.

can you remove the reflector and lens entirely from your other Cree lights? I know you can on the magicshine style, just unscrew the bezel and the reflector lifts out. compare the brightness without the reflector assemblies and you'll have something more consistent to compare. XXX lumens in a 10 degree spot will be MUCH brighter than the same XXX lumens in a 30 degree spot (9 times as bright as its 9 times less area).

Thank you noglider, you got me feeling giddy

I'm almost for sure they list it as the total luminous flux per LED as a whole and not each individual emmiter. There is several reason for my belief.

One, Cree does not list the total amount of array in each emitter, otherwise they would have listed some total value as a sum.

Two, they list the XML as being about 100 lumens/watt. Given that fact, if the emiiter is driven at the 3amps mark, the chart show the relative flux at 325% making the total lumens at 910 lumens (3.25 * 280 lumens). Given the fact that the Vf of the LED is 3.2v and at 3amps, the total watt is 9.2 watts which is about 920 lumens going by the 100lumens/watt claim. That being said, makes me feel that the listing by cree is for the LED as a whole and for for each array.

If those Ebay claim of 1800 lumens holds true, then the lights must require 18 watts (1800lumens/(100lumens per watt) of power to drive it. Doing the math for a 3.7 volt lithium ion battery means that the battery will need to be able to perform at 4.86 amps for a single cell flashlight and along the fact that the lights will need to dissipate that heat.

Disclaimer: not responsible for any math error .....now I got a headache

the fine print on those ebay '1800 lumens' ads say thats a theoretical maximum at peak voltage and peak current. the *charger* is 8.4V, the battery pack is more like 7.4V (2x3.7V liion cells, with 2 sets in parallel to achieve the claimed 4400mAH) in spite of the silly claim the battery pack is 8.4V. at $33 + $6 s&h, I can ignore some silly claims.

I tested mine for 3:05 on high before the battery indicator went to 'blinking red', if we assume I got 4 amp-hours out of the pack at 7.4V, the discharge rate *was* about 9.8 watts. now, the lamp got so hot I couldn't grab it, just juggle it on my fingertips, and I have rather heat resistant hands. I do need to measure the voltage before/after one of these tests, but I don't have a compatible connector to do it easily.

I think there's an indepth review of that '1,800 lumen' light here: https://forums.roadbikereview.com/com...mp-269236.html

Except that some of the numbers don't add up. The 18 watts and a 3 hour runtime would be consistant with an 1,800 lumen output - except than the reviewer states that the light output is more like 500 or 700 lumens and he still got a 3 hour runtime. So its possible that neither the stated capacity of the 8.4V battery pack nor the lumen output are accurate.

The heat issue is very real. Someone needs to hook that up to a coffee maker!

Here is another experiment that cyccommute can do fairly easily. If cyccommute have two battery pack, he can run the older MS900 and the XML light with this experiment. Find a room of about 10 x 10 ft. Anything close to that is fine, even the bathroom will do. Close the door and turn on both light on high. Use only the bike light in the room. Hold one light in each hand. Take one light and point it at the ceiling above your head so you don't get glare. At the same time make sure the other light is cover by placing it on your lap, table or whatever. Notice how much light intensity you have in the room. Now quickly swap lights meaning taking the one pointed at the ceiling and cover that one while taking the previous covered light and point it toward the ceiling above eye level. Do this several time within 5 minutes or less otherwise the lighthead will overheat.

I'm sure the XML will be brighter than the older MS900 P7. Just how much brighter? Twice as bright? 3 times as bright or just slightly? I think you will find that what difference you see out on the street using the two different light will be much more significant than what the difference in overall brightness in the room.

Using the room help evaluate better as to how much more lumens you are getting. The room is the control enviroment trapping all the lumens and allowing your eyes to perceive the total lumens given out by the light without too much influence by the reflector or optics. A Square Integrate Sphere....

my math says, 7.4 V (the actual voltage of a 2x2 18650 Liion pack) * 4.4AH == 31 watt hours. my test says the light was still running on 'damn bright' at 3:05 when the led went to blinking red, I haven't run it past that so I don't know how much residual power is left. IF they are putting the full 10 watts into the lamp, and if the battery was fully dead, thats 30 watt hours.

on the low setting, the housing runs barely warm to the touch when stationary in my 67F living room, and the light is about 1/3rd as bright (subjective, I know, but it looks like about 1.5 f/stops to this old photographer). the charge status light went blue to red(solid) at 5 hours on low (it did this at 1:45 on high), so I'm estimating anohter 4-5 hours before it goes blinking red....

Yeah - thats pretty much where the numbers are questionable. As per the reviewer:
I know there are different capacity 18650 batteries on the market, and would think that with a 6400mAh pack he would have gotten more than 3 hours. I can get 9hrs myself running a 10 watt light with a 6500mAh pack.

Old lamps are gone. The new lamps were so much brighter and so cheap that I got rid of the old lights this summer.

I also know what effect the reflector angle has on the beam from back when I ran halogens with 25 and 12 degree reflectors. I even had a 7 degree reflector which really concentrates the light. I was using the same halogen lamp...MR16...and driving them at 20% over nominal voltage. There was a difference in light but the difference wasn't as noticable as the difference between the 3 different emitters.

1 18650 is 3.7V, so two in series are 7.4V. you charge them with a 8.4V charge voltage, and they might read 8.2V or whatever right off the charger, but that's just 'surface charge', and as soon as you hook even a small load up to them, they will be 7.4V.

the individual 18650's I've seen have all been in the 2400 to 3000 mAH range. 2 in series and two of those in parallel is how you get 7.4V(8.4V) and 6000mAH or whatever.
examples: https://www.batteryjunction.com/18650.html

consider that I can't even buy 4 of those batteries for what this light cost, hah!

You have also mentioned the heat issue in the past so I tested mine tonight on the way home. An hour of running on high...I only ever use high...at 50 F ambient temperature, the lights were stone cold. No heat whatsoever.

I have melted a hole in the carpet while doing static run tests in the past. Air flow is very important.

You're right about that set up in the first link having a lot of thermal mass. That was the setup for the XM-L on star, and it still suffered from a lot of thermal sag. Did you click on the second link? I see much less thermal mass. This could actually go in a flashlight or bike light.



It's this setup that produced the spectacular results, including a 12% improvement in output at 3 amps, and seemingly no thermal limitations.


Now maybe you see why I like the idea of making my own copper pill. As far as upgrading, that's just an emitter swap away. Optics don't change much, although getting a lens with good anti reflective coating would be an excellent idea. Batteries are an ongoing upgrade because they lose capacity over time and better batteries should continually become available.

I would love for budget lights to have copper pills, but since copper is expensive, I don't see that happening. Are there any US manufactured lights with copper pills? I'd assume they would already have anti reflective coated lenses and high quality batteries.

I don't know if you noticed, but I didn't mention anything about OTF performance, just getting the emitter itself to perform to spec. Thank you for expanding the subject though. The emitter is just one small part of a lighting system.

Yeah - missed that that copper pill wasn't also imbedded in another large thermal heatsink. Completely agree that, combined with a few other other goodies - would result in an exceptional light package. Nice to see you here! Anyone suffeciently obcessed to go to those lengths is OK in my books.

Personally I'm unaware of any light manufacturer currently using copper in anything except custom made flashlights. I'm guessing it could be because the increased cost and weight are far more tangible to a customer than a slight effeciency gain. Particularly in a business where there are no hard standards for measuring lumen output and the bicycle industry is, for the most part - totally obcessed with 'small and light'. And dare I mention 'cheap'?

Even the guy on CPF who did the Betty board upgrade (you've probably seen it) apparently only used an aluminum core board. https://www.candlepowerforums.com/vb/...GRADE-BOARD*** I did note that the Betty was and still is so designed to the limits of 'small and light' that any upgrades (including any factory upgrades) are seriously complicated by heat issues to the extent that major components have to be replaced to keep things from melting. I'm willing to live with a little extra weight myself - already have a stove thank you!

Don't know how you feel about collimators - have you looked at them? https://www.rpcphotonics.com/collimators.asp Those are becomming increasingly common on better lights and particularly in multi emitter lights - are more efficient and smaller than complex reflectors.

I've also already moved to a higher voltage system myself (12VDC) because it lets the wiring carry less amperage to do the same job. I find the wiring used in most bicycle applications is pretty marginal anyway and succeptable to breaking as well as unable to support large demands. Automobile manufacturers are already contemplating a move to 42V for similar reasons.

Guess what I'd really like to see is a no-holds barred collaboration by board members here to define what they think is an ideal light or lighting system. I say light or lighting system simply because I picked a modular spproach. The bottom line is industry has always been pretty much the same - one you've defined an ideal product - its much easier to find ways to make it less expensive than it is to start with an inexpensive product and find ways to make it better quality.

There are light manufacturer employees that monitor or participate in these forums and - given enough feedback - someone would likely end up building the thing for us anyway.

Serfas is on the right track, and deserves recognition for this:

"Serfas has changed the game on high powered lighting by testing its lights in house with a 20” N.I.S.T traceable standard intergrating sphere. As the name infers, The Serfas “True” lights promise that you will always get what you pay for."

www.serfas.com

www.nist.gov

Agree and its nice to see someone at least starting to do SOMETHING. Thing is - both NIST and ANSI are both American and since we're really looking at a global economy - ISO standards would make more sense in the long run.