(Update) Two virtually identical builds, one pulls 0.66A less. What should I check?

Flashlight Modding and DIY Parts
1

Update: I measured the Vf of the LEDs on the highest mode.
Light 1: 3.81V at 5.16A
Light 2: 3.87V at 4.52A

So Light2 has higher Vf even at lower current - that’s bad, right?

The Cree data sheet doesn’t go above 3A, so according to djozz testing, one might expect to see 3.7Vf at 5A. So the Light2 LED which has 3.87Vf at 4.5A seems like an abnormally bad (below spec) example. Would you agree, or am I missing something?

OP:

Light 1:

  • Brinyte B158B
  • I dedomed the stock XM-L2 U4 1A on the thin (but still DTP copper) stock star, Arctic MX4 thermal paste between pill and star.
  • MTNelectronics 17mm DD FET+1 driver with spring bypass

Light 2:

  • Brinyte B158B
  • I dedomed an XM-L2 U4 1A on a Noctigon (sourced from MTNelectronics), Arctic MX4 thermal paste between pill and star.
  • MTNelectronics 17mm DD FET+1 driver with spring bypass

With the tailcap removed I measured with a 12awg jumper from the battery to body. I measured with a Uni-T UT210E clamp meter, HS1010A lux meter, and my recently built and fairly well calibrated integrating sphere:

Light 1:

  • 5.13A
  • 996 lumens
  • 226 kcd

Light 2:

  • 4.47A
  • 845 lumens
  • 175 kcd

I know there is variation within a flux bin of emitters, but the output difference seems to be related to the significant differences in current.

What can I check, for possible sources of these differences?

What should I check regarding my assembly of the two lights, before I start tearing the two pills apart to check the individual components?

I guess as a first step I could remove the LED stars and test them using a power supply and integrating sphere to confirm they are both making about the same lumens with a known current.

How can I test the driver at the component level? Hook the driver up to a battery on one end and a resistor on the other, then use the clamp meter to measure the driver output current?

What else should I check?

EDIT: Same cell swapped back and forth, results are consistent. The cell is not the difference.

2

Can you measure the voltage on the emitter while on?
If the voltage is the same it means light 2 has an LED with a higher forward voltage

If the voltage is less it means it has more resistance in the light maybe a cold solder joint

3

Same cell?

4

Yes

The less drawing LED may have some 10 mV more voltage as the battery sag less

You could try change tailcap and or driver between LEDs

5

Rbd, same cell swapped back and forth, results are consistent with the info in the OP so the cell is not the difference. Forgot to mention that.

I can definitely measure voltage across the LED while in operation, but will it tell me relevant info? Shouldn’t I expect the Vf to be different due to the difference in current?

The current testing was done without the tailcap, and during lumen and lux testing I swapped the tailcaps back and forth with no change. So the tailcap is not the issue.

I even swapped the lenses back and forth wondering if one lens had a blemish or reflection that I just could not see. There was no difference due to lenses.

6

You can check resistance in all parts. Between switch spring and tail contact ring. Between host and driver negative. Even between two sides of tube. Clean all electrical connected surfaces, sometimed it is better to use special spray for electronics or automotive electric connections.

7

Good point, I didn’t clean any contact surfaces. Maybe one is contaminated with grease or something. I’ll give that a try tonight before tearing out components.

8

Measuring the current and Vf of the LED will give you the result, that you can say its drawing less because the Vf is higher

If the Vf of both lights eould be same the one with less output should have less Vf than the other due to resistance in the driver or light

If the Vf of the one with less output is equal or higher you can blame the LED for drawing less current and cancel your teardown of the light for finding a bad resistance

A difference from 100-200mV in Vf of the LEDs reslt in a significant difference current on DD

9

Is it better to measure Vf at higher currents? (So I should measure Vf on Turbo mode, not Low mode?)

I will get out my darkest glasses to measure Vf while staring into a lit LED… :cowboy_hat_face:

10

Update:

I measured the Vf of the LEDs on the highest mode.
Light 1: 3.81V at 5.16A
Light 2: 3.87V at 4.52A

So Light2 has higher Vf even at lower current - that’s bad, right?

The Cree data sheet doesn’t go above 3A, so according to djozz testing, one might expect to see 3.7Vf at 5A. So the Light2 LED which has 3.87Vf at 4.5A seems like an abnormally bad (below spec) example. Would you agree, or am I missing something?

11

Yeah. That’s the culprit. Not necessarily a bad led but certainly higher Vf.

12

I did not realize that the Vf of the LED could have that much impact on the current draw (and subsequent output).

Since I’ve only got n=2 data points, is it possible that Light1 with LED Vf 3.81 at 5.16A is actually OVER-performing? Or based on those numbers I’ve measured is it more likely that LED2 is indeed under-performing?

Does the Vf not not impact the output binning of the LED at all? It seems like it should - I’ve got two U4 binned LEDs here, but one makes 1000Lm OTF, and one makes 850. That’s a much bigger % difference than the Cree bins specify.