Nichia GS 40ma overdrive fade test

[JohnR66]

Normally I test the 5mm LEDs at 30ma drive, but since the GS is rated at 30ma continuous, I over drove it at 40ma for 192 hours (1 week). I also tested two other LEDs that passed the 30ma test just to see how they would hold up at 40ma.


Nichia NSPW500GS-K1 Lot 96CS14-c0W
No visible fading. Does not get as warm as other LEDs due its efficiency. Some don't like its angry blue oblong center. Seems to render colors better than most 5mm LEDs. According to jtr1962's graph it produces triple digit lumens/watt efficacy in the 1-17ma range and 80 l/w at 38ma (bin and rank unknown). I don't care for the "stopper" (Nichia term), the square protrusion on the leads that hinders surface mounting. It is no surprise why this is the king of the 5mm white LEDs. A fantastic LED when you need light at low power and cost.


Sylvania 3mm, hacked from xmas light string
This poor LED took some abuse. Its tiny form factor and the fact it was high off the breadboard means it could not dissipate heat and thus it got hot. It showed some fading. It probably would have passed this test had I given it some way to dissipate heat. It remains the best LED used in a light string I've found. Hopefully Sylvania uses it in this year's light strings as well.


Asia Engineer 5mm white
A super cheap ebay LED. It remains the only ebay 5mm white LED that did not fade at 30ma. It suffered a little here at 40ma. It is not very bright (I agree with the 13,000 mcd rating) and it has a cyan cast to the beam color. A somewhat uninspiring LED, but fairly stable. Unfortunate that the 55,000 mcd JELED and LOV (light of victory) ebay LEDs bowed out in the 30ma test. Is there nothing worthwhile on ebay?



Side report: Radio Shack 276-017 neutral/warm LED (4200K). Last Fall I installed the 18,000 mcd 15 Deg LED from Superbrightleds.com into my garage door opener button to replace the red one. It had faded and turned purplish after a few weeks. Next I installed the Rat shack LED and I'm happy to report that it is still bright after 9 months.

Driven LEDs left, control group right.
GS top, Sylvania mid, Asia Engineer bottom.
LEDs photographed through white plexi that sits on top of the LED's dome to capture light. It is not a true indication of beam width and shape.

 

[LEDAdd1ct]

Unfortunate that the 55,000 mcd JELED and LOV (light of victory) ebay LEDs bowed out in the 30ma test. Is there nothing worthwhile on ebay?

Sad, but not surprising. I wish there were some really bright 5mm on the 'Bay that weren't junk/didn't fade appreciably.

 

[Curt R]

What kind of instrument equipment was used in the fade testing. And what were the numerical results of the testing. The human eye standard response curve established in the 1920's required a change of 30% intensity to detect a visible change in light output, which is why most LED manufacturers output bins are established at a 30% low to high differential. Also as the human eye ages there is a requirement that the light output must double for every 13 years of aging to see the same object at the same distance with the same intensity and detail.

Curt

 

[JohnR66]

What kind of instrument equipment was used in the fade testing. And what were the numerical results of the testing. The human eye standard response curve established in the 1920's required a change of 30% intensity to detect a visible change in light output, which is why most LED manufacturers output bins are established at a 30% low to high differential. Also as the human eye ages there is a requirement that the light output must double for every 13 years of aging to see the same object at the same distance with the same intensity and detail.

Curt

That is correct. Looking at the LEDs through the white plexi or straight on, I can't see the change much.

The camera sensor is more sensitive to changes in light intensity than human vision. I check the difference in values of the grey scale converted image (256 levels). Difference of 3 or less I consider no change. A change 3-8 is a small change. A change of over 8 is a moderate change. I consider anything higher a failed LED because it will just get weaker over time and in a month or so be useless.

If you want charts and graphs, someone with the equipment can perform the test. I am quite satisfied with my test that if the LED survives a moderate overdrive, it should last quite a while when driven within its recommended current level. Many of the LEDs that faded in my other tests did so dramatically that it was quite apparent with my own eyes.