Question about XM-L efficiency at different voltages

Another posted asked why the difference outlined bloew in the Neutron 2xAA and LD20, but never received a satisfactory answer. Say we compare 2 settings of the XP-G Fenix LD20 to the XM-L Thrunite Neutron:

Neutron: 120 for 2.1 hours
LD20: 180 for 2.25 hours

The LD20 with the XP-G is clearly far more efficient. That said different circuits, etc. Maybe the Neutron design is just terrible.

So sayy I compare the Quark 2xAA XP-G and the same model in XM-L. A lot of different outputs which make direct comparison not 100% clear, but say one setting

XP-G: 4.0 lumens/5 days
XM-L: 2.7 lumens/3 days

The XP-G is clearly better, for say another setting:

XP-G 85 lumens/5.0 hours
XM-L 115 lumens/2.5 hours

The XP-G seems more efficient to me. At the very least - the XM-L doe *not* seem clearly more efficient. And in terms of usage - I would pretty strongly argue is worse. An extra 30 lumens on top of 85 (hardly noticeable) is not worth half the runtime.


But then say if I compare the Quark 2xCR123 XP-G and XM-L models - the XM-L is the very clear winner on all output levels.

So...is the XM-L quite a bit more efficient above only certain voltage levels? And below that, the XP-G is more efficient? Or am I confused?

XML is much more efficient at higher drive currents only. See the quark line as compared with the identical light with XML.

XPG has a bit of an edge in low. After that it is all over. At medium it has over 3x the output for half the runtime. Might not be preferable but it is more efficient.

At high similar runtime, double the output. At max >50% more output for the sameish runtime.

Keep in mind though that despite the massively higher lumen ratings, it probably wont look any brighter. but will have a bigger hotspot and MUCH bigger spill. You really don't get tunnel vision with XML. Up close and medium distances the spill fills most of your vision. Very cool LED for EDC type lights since they don't need to be throw monsters. Something XML won't be without a really big head.

Quark line:

123^2 XP-G
Moonlight: 0.2 lumens for 30 days, (1ma)
Low: 4 lumens for 5 days (10ma)
Medium: 22 lumens for 20 hours (50ma)
High: 85 lumens for 4.5 hours (250ma)
Max: 230 lumens for 1.8 hours (990ma)

XML:

Moonlight: 0.3 lumens for 25 days
Low: 3 lumens for 5 days
Medium: 65 lumens for 11 hours
High: 160 lumens for 4 hours
Max: 360 lumens for 1.7 hours

Are we talking about the .....Xeno XP-GR5 and XM-LT6

I think 4sevens may have changed either the cells used in testing or the point at which they end the test (50% or 10%) output. I don't have inside info but these can be done differently while claiming 'ANSI' standards and may be the reason for some differences between models which otherwise don't seem to make sense.

This is a normalized curve for the XP-G and the XM-L directly from Cree's published specs

And here is XP-G and XM-L comparison of lumens vs. watts based on the same Cree's source:



You may find more here: XM-L-vs.-XP-G-comparision-follow-up-to-discussions-about-XM-L-in-ZL-H51

Generally XM-L is more efficient but the difference is almost non existent at lower currents. The strange fluctuations in flashlights specs may exist because of:
-different driver efficiency
-different measurement methodology of lumens (ANSI, OTE, OTF)
-different measurement methodology of runtime (up to 50% of lumens from the beginning or something else...)
-used power source with different capacity/voltage (for example: AA white eneloop 1900mAh at 1.2V vs. black eneloop 2500 mAh at 1.2V vs. Energizer L91 3000mAh at 1.5V)
-different led bin quality
-because of power used by widgets like light localizator or variable ring.
-(not)intentional error in specs aka SOOA - specs out of.... ;-)
-different electrical resistance of the entire circuit.


PS. Sorry for the over sized image, It looks like something was changed in image embedding on cpf. Few weeks ago it was scaled down correctly.

yup, the XML needs a pretty big head to throw any distance. Among the smaller head lights, i think its the Olight i6 and the TK35 which have the smallest heads possible for an XML thrower. XML in the Quarks are definitely flooders with a bit of decent throw to maybe 100+ yards unlike the TK and i6 which can throw over 300 yards and 100 yards respectively.

I guess I'm still struggling with whether XM-L is more of an improvement, or more of a choice. If I look at my Quark AA2 R5 performance across it's operating range, would the XM-L version outperform it over that range and then on top of that offer an even longer range? Or, at range, would the XM-L be basically floodier while the R5 "spottier" without really much of a visual difference in brightness?

I am fairly certain that 4sevens did not use the ANSI/NEMA FL1 standard to measure the XPG R5 quarks (I am pretty sure they started using the standard with the S2 quarks). If they used the ANSI standard to measure the quark X, then we would be comparing apples to oranges.

tony22 said:

I guess I'm still struggling with whether XM-L is more of an improvement, or more of a choice. If I look at my Quark AA2 R5 performance across it's operating range, would the XM-L version outperform it over that range and then on top of that offer an even longer range? Or, at range, would the XM-L be basically floodier while the R5 "spottier" without really much of a visual difference in brightness?

Click to expand...

This should help, http://www.candlepowerforums.com/vb...d-Cat-XPG-vs.-Hound-Dog-XPG-vs.-Hound-Dog-XML!!!

The Wildcat is a flood monster triple XPG, the Hound Dog is a close to max driven XPG, and the Hound dog XML is a close to max driven XML. I'd say the XML is an improvement over the MCE, with regards to the XP-G it's a choice depending on whether you want flood and max lumens, or a tighter beam in a smaller package.

OK: My answer?

Going by the Cree graph provided: sub 200 lumens the the efficiency is rather close. That said, current circuits are probably optimized for the XP-G, so any slight bonus is overwhelmed in the circuits being not designed for this emitter.

So - XM-L is better, but for <200 lumens level...the efficiency bonus is small anyway, and too small to overcome the not-designed-for-this-emitter characteristic of current Quark lights and others?

tony22 said:

I guess I'm still struggling with whether XM-L is more of an improvement, or more of a choice. If I look at my Quark AA2 R5 performance across it's operating range, would the XM-L version outperform it over that range and then on top of that offer an even longer range? Or, at range, would the XM-L be basically floodier while the R5 "spottier" without really much of a visual difference in brightness?

Click to expand...

Its more of a choice, the R5 has a tighter hot spot even with medium-sized reflectors unlike XML's that need big reflector to get decent throw. R5's max out at about 400+ lumens, XML's run out to almost 1k lumens. I own both, when i need tight throw, R5, if i want big light with big throw, XML.

I also like the R5's whiter output unlike the greenish tint of the XML (boo!)

roadkill1109 said:

Its more of a choice, the R5 has a tighter hot spot even with medium-sized reflectors unlike XML's that need big reflector to get decent throw. R5's max out at about 400+ lumens, XML's run out to almost 1k lumens. I own both, when i need tight throw, R5, if i want big light with big throw, XML.

I also like the R5's whiter output unlike the greenish tint of the XML (boo!)

Click to expand...

D@mn you all. Now you're making me choose.

I choose to keep my R5 and buy an XM-L version! :devil:

One other aspect that the XML may have over the XP-G at lower lumen levels is better heat sinking ability due to the 2mm square die vs the 1mm square die of the XP-G. I suspect this may manifest itself in longer LED life, although I think Cree just gives a blanket 50k hours on all models. I'm sure this number though is based on absolute optimum lab settings vs real world use. Perhaps at sub 200 lumen levels the XM-L will have a better chance of making the 50k hours rating vs an XP-G driven for long periods on that range.

I struggled with the choice between the XM-L and XP-G in relation to runtime, brightness consistency, and total output... in that order.
I found that the lower Vf for the XM-L at lower levels are significantly different if you consider a simple CC driver (AMC7135 x 3).

Have a look at this graph to show at what voltage the 2 emitters begin to dim:

This shows that the XM-L will begin to dim much later than the XP-G at the same current.
It took a while for me to realize this but its been staring me in the face for months before I finally mapped it.

Please note the top curve uses 100% as 350ma for the XP-G and 700ma for the XM-L directly from the spec sheets.

When you compare STP's graph to mine above, you will begin to see this as I used Current and STP uses Watts. The divergence rate is greater when using wattage because it accounts for this Vf difference.

This is true in the use case where you don't have a boost or buck circuit and plan to run with a single Li-Ion cell at low current. Life changes when you step outside this window.

It's always good to throw a few lines on a few curves when in doubt