XP-G R4 is better at 1.1A vs. 1.4A review...

Hi all,

I had 2 XP-G R4 emitters. Naibender put one in my L1 direct drive and the other in my E1b bezel with a 1.4A driver was put together by Krammer5150.

Here are the results in MrGmans work lab.

E1b bezel/ Aleph 17670/ Surefire clickie hosts:
250 OTF turn-on
210 5 sec
196 1 minute
193 2 minutes
190 3 minutes

Now the L1 which Gary said kept the 1.1A after the test was completed.
218 OTF turn-on
198 1 minute
194 2 minute


The turn-on lumens our eyes will never see, but the 5 second mark is what we can and do see. More importantly look at the 1 and 2 minute reading..

I now want my E1b bezel fitted with a 1.2A driver to test out. I now think 1.4A is too much current for the smaller hosts like what I have.

Here is my video.
http://www.youtube.com/watch?v=NHW_iJId_YY

bigC

I guess my thinking is a 1.2A driver is the best solution for sustained OTF lumens in these Surefire E-series bezels.

What KD 1.2A drivers fit these bezels?

Thanks,
Jose

Are the optics in both lights identical & if yes are the emitters focused the same with both lights?

darkzero said:

Are the optics in both lights identical & if yes are the emitters focused the same with both lights?

Click to expand...

Yes,

They both were installed and centered by Nailbender. They also share the same TIR optic Stock Surefire stuff.

bigC

Yeah that little 1.4A mod I did got warm pretty fast too.

let me know when the 1.2A boards get here and I'll get to work.

So this results are because of the little host/heatsink?
Could it be the other way round with a big host and good heatsinking?
I think I read somewhere here that you could push a XP-G up to 1.4amps safe.

A380 said:

So this results are because of the little host/heatsink?
Could it be the other way round with a big host and good heatsinking?
I think I read somewhere here that you could push a XP-G up to 1.4amps safe.

Click to expand...

Yes, My results are due to the small hosts.

The XP-G R4 can handle the 1.4A, but it didn't help where it counts in the OTF department. I don't know the paper specs on these XP-G R4's I am just reporting and experimenting on them in a real hosts with real OTF lumens as tested by MrGman in his IS Sphere.

If I were to put the XP-G R4 in a Mag D hosts and run it at 1.4A I am sure my OTF numbers would be at least 100 more lumens, but since most of us use tiny cr123, 17670, 17500 type of lights for EDC this I hope many more will benefit from learning.

Milky's Cremator Q5 variant tested in the same IS Sphere did 242 OTF with the 1.3A Acorn driver. I wonder if he uses a custom heatsink. I know I am using the same TIR optic as he is, but man even with a Q5 he is still smoking me.:shakehead

Obviously thermal droop is a big issue that doesn't get paid enough attention to around here at times. Whether 1.1A is actually brighter than 1.4A in a small light as a general rule isn't clear yet from this one test because of a number of important factors not controlled in this test:

- Thermal path differences in the construction of the light. We simply have no way to control for differences in the thermal resistance of the star to light body connection.
- Different thermal mass of the 2 lights. The L1 is .1 oz heavier than the E1B in a quick look at surefire.com. I'm assuming no active air flow or holding of the light during testing, so the flow of heat from the light to air/blood is small. The L1 will take more energy to heat to the same temp.
- Vf variance is uncontrolled. Even at the same drive current it can make for important swings in the amount of power consumed and heat generated.
- Driver efficiency. All the power lost in a driver is part of the heat that needs to be managed.

In this case the light that's brighter after 5 minutes also has probable smaller driver related losses (direct drive) and greater thermal mass. If there were no Vf or thermal path differences, the L1 should still have lower junction temp and higher output even at the same drive current.

Baterija,


That's all true about the testing methods, so I think the next step should be to install the emitter in a much bigger light with the same optic and see what happens. That will be the easiest way to verify droop as the reason for the results. I don't know much about small lights, but the p60 drop-ins I have from dereelight only dissipate heat into the reflector they screw into. There is no dedicated heatsink directing heat straight to the body of the light. That is an obvious sign to me that the led will get hot and stay hot. I am not surprised by the results.

luckybucket said:

Baterija,


That's all true about the testing methods, so I think the next step should be to install the emitter in a much bigger light with the same optic and see what happens. That will be the easiest way to verify droop as the reason for the results. I don't know much about small lights, but the p60 drop-ins I have from dereelight only dissipate heat into the reflector they screw into. There is no dedicated heatsink directing heat straight to the body of the light. That is an obvious sign to me that the led will get hot and stay hot. I am not surprised by the results.

Click to expand...

I have a SST-50 drop-in project right now, where the LED has already been tested at 5.68A at the tail and next I will test it at 2.5~2..8A at the tail to see which is a better way to go in a 6P type hosts, but thats another thread.lovecpf

Isn't there a luminous flux range as well? Theoretically you could have one LED from the extreme low end of the spectrum, and one from the extreme high end. The only way to tell for sure is to test both at the same Vf and currents and record lumens output.

Anto said:

Isn't there a luminous flux range as well? Theoretically you could have one LED from the extreme low end of the spectrum, and one from the extreme high end. The only way to tell for sure is to test both at the same Vf and currents and record lumens output.

Click to expand...

If I am not mistaken, CREE doesen't publish the Vf of these in the spec sheet. I know it makes a difference, but since I don't have anymore XP-G R4's I figured I can put a 1.2A driver and try the OTF test again and see where the numbers go.

Well a way to control for some of the variable items in this test would be using the L1 hooked to a bench power supply. It's direct drive so it would just be a matter of using the power supply to drive the different current levels (with the heat produced by any variable efficiencies being outside the light itself). That way you have the same emitter (same VF and flux), with the same thermal path from star to heatsinking, and the same thermal mass.

baterija said:

Well a way to control for some of the variable items in this test would be using the L1 hooked to a bench power supply. It's direct drive so it would just be a matter of using the power supply to drive the different current levels (with the heat produced by any variable efficiencies being outside the light itself). That way you have the same emitter (same VF and flux), with the same thermal path from star to heatsinking, and the same thermal mass.

Click to expand...

I should have asked MrGman to do that. I will have to convince him to try the L1 XP-G R4 DD with a 17500 AW cell. I think the current will stay put at 1.4~1.5A solid. I did try the L1 bezel on the AW 17670 cell, but the current was 2.8A at the tail and too risky to test. So, the next step with the L1 DD is to try the AW 17500 cell and the 1.2A driver for the E1b bezel.

Thank you all for the helpfull advice,
bigC

baterija said:

Well a way to control for some of the variable items in this test would be using the L1 hooked to a bench power supply.

Click to expand...

Exactly. This OP made a comparison of two different lights at different drive levels, which does not equal comparison of r4 @ two drive levels.

If the independent variable is going to be drive level all other variables must be kept the same. So two drop-in's for the same light, one @ 1100ma, one @ 1400ma, or swap drivers from the OP's original test and reshoot with the reversed drivers. This will give 2 data sets, the emitter's performance with-in a single host at different drive levels.

bigchelis said:

If I am not mistaken, CREE doesen't publish the Vf of these in the spec sheet. I know it makes a difference, but since I don't have anymore XP-G R4's I figured I can put a 1.2A driver and try the OTF test again and see where the numbers go.

Click to expand...

The Cree XP-G datasheet most definitely does have Vf data, including an If vs Vf curve, as well as data on typical Vf values for If = 350mA, 700mA, and 1000mA. What Cree does not do is bin their white XP-G emitters by Vf bin (unlike, for example, the SSC P4).

I've measured the following Vf variations (apologies for the non-constant drive currents used -- the measurements were not originally done to test Vf for a given drive current level):

8mm MCPCB XP-G R4 #1, 3.3V@950mA
8mm MCPCB XP-G R4 #2, 3.5V@960mA
8mm MCPCB XP-G R4 #3, 3.3V@950mA
8mm MCPCB XP-G R4 #4, 3.6V@910mA
8mm MCPCB XP-G R4 #5, 3.4V@1047mA

10mm MCPCB XP-G R4 #6, 3.1V@1042mA
10mm MCPCB XP-G R4 #7, 3.3V@1196mA

It is true that the Cree flux bins are fairly narrow (e.g., R4 is 130-139 lumens at 350mA). But if you don't hold your variables (such as the specific emitter) constant, then I would postulate that flux bin variations and Vf variations combined can easily lead you astray in your conclusions about lumens vs drive current. Just look at XP-G #4 vs #6 above. A Vf of 3.6V at 910mA vs 3.1V at 1042mA. If #4 had been representative of the XP-G you drove at 1.4A and #6 representative of the XP-G you drove at 1.1A, you can already see that #4 could have a Vf of perhaps 3.7V at 1.1A (~4.1W), and about 3.9V at 1.4A (~5.5W). Compare to #6: maybe a tick above 3.1V at 1.1A (~3.4W).

The low Vf XP-G #6 could run almost 3/4 watts cooler at 1.1A vs the high Vf XP-G #4, and #4 could run quite hot at 1.4A (5.5W). Then if XP-G #6 were also at the high end of the flux bin while XP-G #4 were at the low end, you'd get even more skew.

Unless you also quantify the Vf of your XP-G, even if you put the direct drive head on a bench supply to help keep your variables constant, your conclusions may not be broadly applicable to other situations if you happen to have drawn an overly high Vf emitter.

Justin Case said:

The Cree XP-G datasheet most definitely does have Vf data, including an If vs Vf curve, as well as data on typical Vf values for If = 350mA, 700mA, and 1000mA.

I've measured the following Vf variations (apologies for the non-constant drive currents used -- the measurements were not originally done to test Vf for a given drive current level):

8mm MCPCB XP-G R4 #1, 3.3V@950mA
8mm MCPCB XP-G R4 #2, 3.5V@960mA
8mm MCPCB XP-G R4 #3, 3.3V@950mA
8mm MCPCB XP-G R4 #4, 3.6V@910mA
8mm MCPCB XP-G R4 #5, 3.4V@1047mA

10mm MCPCB XP-G R4 #6, 3.1V@1042mA
10mm MCPCB XP-G R4 #7, 3.3V@1196mA

It is true that the Cree flux bins are fairly narrow (e.g., R4 is 130-139 lumens at 350mA). But if you don't hold your variables (such as the specific emitter) constant, then I would postulate that flux bin variations and Vf variations can easily lead you astray in your conclusions about lumens vs drive current. Just look at XP-G #4 vs #6 above. A Vf of 3.6V at 910mA vs 3.1V at 1042mA. If #4 had been representative of the XP-G you drove at 1.4A and #6 representative of the XP-G you drove at 1.1A, you can already see that #4 could have a Vf of perhaps 3.7V at 1.1A (~4.1W), and about 3.9V at 1.4A (~5.5W). Compare to #6: maybe a tick above 3.1V at 1.1A (~3.4W).

The low Vf XP-G #6 could run almost 3/4 watts cooler at 1.1A vs the high Vf XP-G #4, and #4 could run quite hot at 1.4A (5.5W). Then if XP-G #6 were also at the high end of the flux bin while XP-G #4 were at the low end, you'd get even more skew.

Unless you also quantify the Vf of your XP-G, even if you put the direct drive head on a bench supply to help keep your variables constant, your conclusions may not be broadly applicable to other situations if you happen to have drawn an overly high Vf emitter.

Click to expand...

Wow, A ton to digest and thank you.

Justin Case,
I purchased these from KD to attempt to drive the XP-G R4 in the E1b at 1.2A with 2 cells. Do you know anything about these:
http://www.kaidomain.com/ProductDetails.aspx?ProductId=9912

6V~9V Buck Current Regulated 3.0V~9.5V 1.2A Flashlight PCB SKU: S008848

Thank you,
bigC