Luxeon Vf shift over time

Candle Power Flashlight Forum

Help Support CPF:

Seems like a significant disparity in the two! Keep up the great investigation and thanks and happy holidays to you and yours! /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/thumbsup.gif
 
Good thing I over resistored my maggmods.
I was kinda sad to see them at just 650 milliamps on nimhs at 1st.
Now my test units are running 850 to 900 milliamps after extensive use at work.
My direct drive unit running on nicads only, has come up from 800 ma to 1100ma on the same batteries. /ubbthreads/images/graemlins/str.gif
 
Final data for the TX1J tests:

tx1jVfdropfinal.png


The Vf bouncing up and down on #1 is due to it being right at the 1/2 way mark between those two voltage measurements, and slight temperature variations in the garage causing 0.01V fluctuations.

and data:

<font class="small">Code:</font><hr /><pre>
#1 #2
current Vf Vf delta (Vf) Vf Vf delta (Vf)
(mA) 0 hrs 100 hrs 0 hrs 100 hrs
130 3.01 2.95 0.06 3.01 2.94 0.07
310 3.21 3.13 0.08 3.22 3.12 0.10
620 3.42 3.31 0.11 3.44 3.32 0.12
910 3.56 3.44 0.12 3.58 3.45 0.13
1540 3.80 3.67 0.13 3.82 3.68 0.14
</pre><hr />
 
I wonder if one can infer from all of this information that paying a premium for low Vf LEDs is a waste of money ?

This data "indicates" that the Vf will gradually come down to roughly the same point, (+/- 0.1 - 0.2 V) and will become more consistent over time from LED to LED. At least that is how it looks to me.
 
The two tenths of an amp may be significant for those doing direct drive or using resistors. That translates to roughly 200ma more current. It probably boils down to the intended use....

If I use a current limited converter, the light will function the same whether using an L or a J. The light will probably run a bit longer on the J, and will be a bit cooler.

So if I have a small light where heat might be a problem, or if I have run time issues, the lower Vf will still win out.

Of course, I could be wrong /ubbthreads/images/graemlins/smile.gif

Daniel
 
[ QUOTE ]
HarryN said:
I wonder if one can infer from all of this information that paying a premium for low Vf LEDs is a waste of money ?

This data "indicates" that the Vf will gradually come down to roughly the same point, (+/- 0.1 - 0.2 V) and will become more consistent over time from LED to LED. At least that is how it looks to me.

[/ QUOTE ]

Well, maybe. I really wish I had a light meter to check my u2. I swear it has gotten brighter after I ran it through 20-30 sets of batteries.
 
[ QUOTE ]
turbodog said:
[ QUOTE ]
HarryN said:
I wonder if one can infer from all of this information that paying a premium for low Vf LEDs is a waste of money ?

This data "indicates" that the Vf will gradually come down to roughly the same point, (+/- 0.1 - 0.2 V) and will become more consistent over time from LED to LED. At least that is how it looks to me.

[/ QUOTE ]

Well, maybe. I really wish I had a light meter to check my u2. I swear it has gotten brighter after I ran it through 20-30 sets of batteries.

[/ QUOTE ]


You may be seeing a different effect from the Vf shift in this thread. I have seen presentations at LED conferences which indicated that Luxeons actually go through a brightness increase during the first few thousand hours of use, then it drops to roughly the same level again and is stable.

I have asked some of the "life time test" presenters about this information, because it often is not shown. Their answer - "The output does increase, but we deleted the first few thousand hours from the data, because we were interested in finding the failure points, not the actual LED variation information."

30 sets of batteries - I guess that puts you in the several hundred hours of use catagory. - fairly heavy user. /ubbthreads/images/graemlins/smile.gif
 
[ QUOTE ]
HarryN said:
You may be seeing a different effect from the Vf shift in this thread. I have seen presentations at LED conferences which indicated that Luxeons actually go through a brightness increase during the first few thousand hours of use, then it drops to roughly the same level again and is stable.


[/ QUOTE ]

Well, if it takes a few thousand hours of increasing before coming back again, then I guess most of us won't ever see our LED's get dimmer than where they started when we bought them, only brighter. At an hour a night, every night, it'll take over 8 years to begin the journey back to where we first began when we bought them. Good news indeed /ubbthreads/images/graemlins/smile.gif.

I assume this is not for the LV.

Chris
 
So, does this mean that if I take my new Nuwai Q3 and connect it to an external power supply (to save batteries) at about 700mA for about 10 or 15 hours then my SYAK bin Q3 could become an SYAJ Q3 or better!!

Streak
 
[ QUOTE ]
Streak said:
So, does this mean that if I take my new Nuwai Q3 and connect it to an external power supply (to save batteries) at about 700mA for about 10 or 15 hours then my SYAK bin Q3 could become an SYAJ Q3 or better!!

Streak

[/ QUOTE ]

Quite possibly. And while I only have 2 samples showing this, running it at 1A may acellerate the process and cause a larger drop. Safely running it at 1A would require you remove the luxeon from the light, attach it to a large heatsink, and run it from a constant-current power supply.
 
Thanks for that Evan.
Do you use the basic LM317 with a single resistor as your constant current supply?
If so then how do you deal with the really small resistor values and fractions of an ohm to change the current from say 750mA to 1000mA?
If not then is there a better (simple as well) circuit that I could use?

Streak
 
Streak,

I use an LM350, which is just a higher current version of the LM317 (exact same pinouts/usage).

Here's what I do - Instead of finding really small resistor values, I have a bank of 5 switches that will switch in/out 5 resistors (one resistor for each switch), that are wired in parallel. The resistor values I have are 47 ohms (~30 mA), 10 ohms (~130mA), 3.9 ohms (~310mA), 2 ohms (~620mA), 1.4 ohms (~920mA). The 1.4 ohm was constructed from 3 0.47 ohm resistors in series.

For instance, to get about 440mA, I switch on the second and third switches, putting the 10 and 3.9 ohm resistors in parallel to get about 2.8 ohms, for 440-450mA. If I turn all switches on, I get all resistors in parallel, for 0.64 ohms, just under 2 amps.

This gives me a good number of different currents, to drive all manner of Luxeon LEDs, as well as regular (5mm) LEDs. And, if I keep an eye on things, I can charge NiCD/NiMH batteries with it as well.

I feed it with a 24V, 3A switching power supply. There is a large heat sink with a fan bolted to the '350, since it will dissipate up to 48W (max current, short circuit, the '350 is dissipating all the power).

If this doesn't make sense, I can draw a circuit schematic for you.
 
Thanks, makes sense.

You probably want to test multiple LED's in parallel hence the high input voltage I would probably use an input of 12v which is easily available and have less heat generated by the LM350.

Thanks for all of the research.

Streak
 
Last bit of data to display:


vfdropsum1.png


vfdropsum2.png



The titles should be self explanatory. The test current graph is less informative, but the one comparing Vf drop at 620mA to initial Vf at 620mA shows common ground for all tested Luxeons (I could just as easily show this for each current level). You can see a general trend forming. TV1K #1 and #2 are outliers because of the higher test current.
 
Okay, considering this, and some folks using 4.2V Lithium Ion batteries with a J bin emitter, in DD or combo DD/PWM, I wonder how close they are to smoked emitters at 100 hours or less runtime....
 
[ QUOTE ]
NewBie said:
Okay, considering this, and some folks using 4.2V Lithium Ion batteries with a J bin emitter, in DD or combo DD/PWM, I wonder how close they are to smoked emitters at 100 hours or less runtime....

[/ QUOTE ]
If I were to guess, I would say that some will be on the wrong side of "close" /ubbthreads/images/graemlins/banghead.gif
 
[ QUOTE ]
Doug S said:
[ QUOTE ]
NewBie said:
Okay, considering this, and some folks using 4.2V Lithium Ion batteries with a J bin emitter, in DD or combo DD/PWM, I wonder how close they are to smoked emitters at 100 hours or less runtime....

[/ QUOTE ]
If I were to guess, I would say that some will be on the wrong side of "close" /ubbthreads/images/graemlins/banghead.gif

[/ QUOTE ]


Well, guess it is their pocketbook, oh well.
 
You must log in or register to reply here.

Latest posts

Back
Top