eBay 26,000mcd LEDs - output vs: I(f)

[jamaica]

I bought two lots of these 5mm white LEDs, mentioned in
another thread ( 26Kmcd on Ebay, anyone try these? Comparison pics).

They are very bright indeed at 20mA, but didn't seem equally
dazzling at lower currents, so I made some measurements
just to see. My findings follow.

I gauged relative outputs by applying a given current
to the LED under test, then measuring the photocurrent
produced in a solar cell illuminated by the LED.

Please note that the comparison to the Nichia LED is not
completely fair, as this setup did not catch all of the
light from either LED type. [In a setup designed to do
that, with a forward current of 20mA, the new LEDs
produced twice the photocurrent of the very best Nichias
(NSPW500BS, circa 2002-2003) I could find in my extensive
stash.]

NOTES:
<ul type="square"> [*]LED #1 = eBay 26,000mcd, from warden_jp2002
[*]LED #2 = Nichia NSPW500BS, from 2002-2003
[*]Luminous efficacy ('n') units are 'uA of photocurrent per mW of LED power'
[*]missing data points were filled in by linear interpolation,
redundant data points were combined by averaging.[/list]
<font class="small">Code:</font><hr /><pre> --------- LED #1 ---------- ------ LED #2 ------
i(LED) iPhoto (uA) iPhoto Ratio
Run #1 Run #2 Vf(1) n(1) Run #1 Vf(2) n(2) (n2/n1)
------ ------ ------ ----- ---- ------ ----- ---- -------
25 uA 0.29 0.28 2.346 4.9 3.33 2.510 53.1 10.9
50 0.78 0.82 2.401 6.7 7.00 2.551 54.9 8.24
100 2.53 2.49 2.447 10.3 14.3 2.604 54.9 5.35
250 13.1 11.8 2.494 20.0 35.5 2.693 52.7 2.64
500 42.8 38.2 2.560 31.6 69.1 2.779 49.7 1.57
750 79.0 40.7 101 47.6 1.17
1.0 mA 117 109 2.622 43.1 132 2.878 50.3 1.17
1.5 157 2.666 39.3 191 2.943 47.8 1.22
2.0 274 254 2.701 48.9 248 2.994 45.9 0.939
4.81 681 692 2.825 50.5 529 3.153 38.9 0.771
9.62 1320 1330 2.988 46.1 966 3.328 33.6 0.729
14.4 1870 1870 3.108 41.7 1340 3.449 29.9 0.717
</pre><hr />

I fully tested only one of the wardenjp2002 LEDs, but
then checked 5 samples at random from 2 lots to confirm
the fully tested unit was indeed closely representative
of the entire batch. It was. It seems wardenjp2002's
LEDs are closely matched.

In another test for longevity, I ran a single unit for ~300
hours, after which time the output was unchanged, +/-
a few percent.

Summary: the 26,000mcd LEDs are quite a bargain, but
inefficient at low currents. I suspect the dice used
are either leaky, or have been slightly damaged, neither
of which affects their usefulness when driven hard.

Jamaica

(edit 2/28/05: corrected a Vf data point)

 

[UK Owl]

Nice Info... but...

Is this based on the LED's as they are made (i.e. focussed) or are they comparisons made with the dome on the end of the LED's filed off to scatter the beam producing 'area lighting' and possibly giving a fairer comparison of the luminous efficiency of each.

The lower Vf also looks to be an advantage, especially for those using DC Converter circuits and chasing efficiency.

Further details including colour comparison, and how output degrades when driven at 30mA, 40mA and 50mA would be nice.

Maybe you (or some other posters) could offer some input.

 

[Lynx_Arc]

It looks like to me the nichias only start looking better below 2ma, which I consider not the range most people drive LEDs at. I agree that the lower Vf of the 26kmcd LEDs would make up for the lower output as the current drops. If you compare where the the nichia starts getting more efficient you will notice that at 2ma the 26kmcd is operating at 2.7v
and checking that Vf on the nichia chart gives you an output of 35.5 at .25ma input, which even though it is more efficient the output compares to 274, leading to approx 7.7 times as much light output at the same voltage. My guess is unless you are using tiny batteries you will probably get more useful runtime/light due to the lower vf than efficiencies at these minute current levels.

I would like to see a runtime graph vs output level comparison of these two LEDs using standard AA/AAA cells which most cheap lights tend to use.

 

[greenlight]

Nice to see some comparative #s.

 

[jamaica]

[ QUOTE ]
UK Owl said:
Is this based on the LED's as they are made (i.e. focussed)

[/ QUOTE ]
Yes, intact LEDs.

[ QUOTE ]
UK Owl said:or are they comparisons made with the dome on the end
of the LED's filed off to scatter the beam producing 'area lighting' and possibly giving
a fairer comparison of the luminous efficiency of each.

[/ QUOTE ]
Nope. As noted above I did separately compare the total photonic output
of both parts at 20mA drive, using optics to catch nearly all the photons. I
believe that setup covered the substance & spirit of your question, at least at
20mA drive. Total luminous output of the 26Kmcd units was exactly double the
Nichia's. This method was very fair, but I can't report overall efficiency
because I didn't measure the LEDs' forward voltages.

Sadly, the exact 26Kmcd LED appearing above met an untimely 200mA pulse and
is no longer feeling quite as perky, precluding further direct comparisons.

[ QUOTE ]
UK Owl said:The lower Vf also looks to be an advantage, especially for
those using DC Converter circuits and chasing efficiency.

Further details including colour comparison, and how output degrades when driven at 30mA,
40mA and 50mA would be nice.

[/ QUOTE ]
Sure! I can offer a little info along those lines:

The Nichia was decidedly whiter, while the eBay unit had a light bluish tint
with a pleasing even, diffuse pattern. The eBay unit throws a few degrees
tighter than the Nichia, delivering noticably more light to distant targets,
but less spill.

I didn't overdrive either LED, as I consider that punishment cruel and all
too usual!

Best,
Jamaica

 

[jamaica]

[ QUOTE ]
Lynx_Arc said:
It looks like to me the nichias only start looking better below 2ma, which I consider not the range most people drive LEDs at. I agree that the lower Vf of the 26kmcd LEDs would make up for the lower output as the current drops. If you compare where the the nichia starts getting more efficient you will notice that at 2ma the 26kmcd is operating at 2.7v
and checking that Vf on the nichia chart gives you an output of 35.5 at .25ma input, which even though it is more efficient the output compares to 274, leading to approx 7.7 times as much light output at the same voltage. My guess is unless you are using tiny batteries you will probably get more useful runtime/light due to the lower vf than efficiencies at these minute current levels.

[/ QUOTE ]

Yes, agreed, & that's a good point about the Vf. Lower Vf=longer run
in many lights, especially resistored & direct drive units. I've personally
swapped one the the 26Kmcd units into a keychain light. I switched the
2 x 2016 lithium stack to a single 2032 coin cell, and used it two days later
to hike out of the wilderness, a few miles in pitch dark!

Best,
Jamaica

 

[Lynx_Arc]

I would think regulated lights would most likely never dip down to the low levels where the nichia is more efficient, but only direct drive or resistored, or boosted lights.