I've noticed that manufactures of LED lighting devices really like to
boast about actual LED lifespans; 10,000 and 100,000 for two different
flashlights I own. 10k hours comes to about 417 continuous days or 1.14
years; and 100k hours is almost 4,167 days or 11.42 years.
Especially in the case off the 100k hour claims, thats a good while.
Got me curious... Has anyone here actually experienced an 'in normal
service' LED failure? (For the record, I haven't...)
(Who as of today is dash over 506.8k hours old; or 21.12k days.)
On Sun, 06 Nov 2011 05:59:57 -0600, Ignoramus14302
None of my LED flashlights has lost any LEDs yet, but I seldom use
them, and even then, seldom use them for more than a few minutes at a
time. Even the LED book light I used for 5 hours once hasn't given me
any problems. Hmm, all of my LED lights are discrete bulbed, not.
It's hard drives and CFLs which aren't living up to their MTBFs.
Most of my CFL failures came from Feit. The 23W (100w) dimmable ULA
bulbs I got from an eBay vendor last years longer. I bought a few GV
25W (100w) CFLs from Walmart a couple years ago and haven't needed to
change any yet. And I haven't lost any of the 17W (60w) 4100K or 5000K
Satco bulbs yet this first year. In home automation, even the
dimmable bulbs flash every couple seconds when off if they're on
standard lamp modules. I had to switch to relayed appliance modules.
They're a buck more apiece and make a noticeable racket when switched.
GE incandescent bulbs from Walmart didn't last a year.
"GE, we bring good things to life, for a very short while."
My MadeInChina electric tea kettle switch started dying within the
first several months. I wish I had a contact burnisher pen. Have you
seen what they sell for today? Crikey!
Live Simply. Speak Kindly. Care Deeply. Love Generously.
The LEDs will make it 100K, it's the support circuitry around them.
I had a batch of landscape flood MR-16 replacements (the clustered
5mm diode style, not a single Cree) all go partially to totally out
within a year, but the maker blamed it on power surges - the HOA was
using the original 500W magnetic transformer.
(If your design sucks because you didn't put any voltage regulation on
board to save a dime, blame it on another vendor...)
Gave us replacement lamps for free, but we had to swap out the
transformer for a proper LED Driver power supply - all in all, a wash.
The lamp end of a Compact Fluorescent Lamp isn't a problem, they have
the same 12,000 hour or so life as any other fluorescent lamp - It's
the ballast section that almost always goes first. Surface mounted
board with the cheapest components they can get away with using, and
the MTBF of the components is the MTBF of the lamp.
And if you think Feit is bad, Lights Of America is worse. I've taken
apart several of their fixtures that have failed, and traced out the
circuits - Power to the 26W lamp channeled through a 1/8W resistor...
You really do get what you pay for - and at that, only if you inspect
before purchase, and really insist. Take a Sharpie or a Paint Marker
and always date your lamps and ballasts, and take them back and scream
at the vendor or maker if they don't make the grade. Once in a long
while they'll actually give you another one.
And if you scream loud enough word might get back to the Bean Counters
who specified floor sweepings for components, and the Engineering
staff who designed it down to a price under orders.
If you have a choice, go for separate lamps and ballasts in the
fixture. The ballasts tend to last a lot longer than the lamps then
because they get better cooling away from the lamp, especially good
ones like the Fulham Workhorse line.
I'd still like to get Good Old Magnetic Ballasts for 24/7/365
applications like hallway drum fixtures, but that's not happening
anymore. And even certain major makers (Lithonia) are trying to use
cheap imported electronic ballasts that won't take constant running
usage anymore. Two to four years, and *pop* goes the ballast.
Incandescents are only going to get worse, because the US factories
are long closed and they are all relegated to third world manufacture.
It *is* possible to get quality still, but the US Based maker has to
ride their foreign QA crews very vigilantly to keep it up. And that
takes away from the profits...
--<< Bruce >>--
On Sun, 06 Nov 2011 08:11:53 -0800, "Bruce L. Bergman (munged human
Right, and boy do those little boards stink when they melt and let the
magic smoke out. Whew!
Lights of America cannot possibly, under any circumstances, meet or
exceed the UL listing codes. No way, no how, never in a million
years. I tossed three of their pieces of shit away (back to Home
Depot) back in the '80s and warned their manager that they were liable
to get some lawsuits from fires caused by the blasted things if they
didn't change brands. I've never considered another LOA since then
and constantly warn people away from them.
Feit sent free replacements, but they failed, too. It wasn't worth
Yeah, I do try to give the idiots feedback, but most of the time, they
don't want to hear it. So much for another mfgr...
Separable lamps and ballasts in CFLs? Nevahoiduvit.
Fulham = ChaCHING, oui?
Lithonia is starting to feel a lot like LOA. I haven't liked the last
couple of fixtures of theirs I've put up recently.
Aren't 'they' somehow phasing out incans as early as 2012? Oh, that's
China. LoCal does it by 2018. http://goo.gl/LIRc
Live Simply. Speak Kindly. Care Deeply. Love Generously.
On Sun, 06 Nov 2011 08:11:53 -0800, Bruce L. Bergman (munged human
Trust me, the engineering staff already knows. And getting screamed at
by bean counters for process failures caused by the bean counters'
unwillingness to appropriately staff a design project doesn't make the
next product come out better, it just shortens the service life of the
I can't understand the attitude of folks who drive BMW's and Mercedes,
and who won't listen when an engineer tells them that "quality costs
Control system and signal processing consulting
"Bruce L. Bergman (munged human readable)"
WHich is really a meaningless observation, as you well know.
Depending on the other factors involved. A 27 W lamp could be drawing
.01 A, or 100A depending on the voltage, or a 1/8 W resistor could
easily handle 100A at that rating or .001A all depending on its
But on the other hand CRAP is CRAP, true enough.
Oh yeah - I'd sit down and do the math with the voltage and wattage
and resistance/capacitance ratings on those components - but they're
blown to heck already, so you can't read it.
And the factory in Nankang isn't going to share the wiring diagrams
and parts pick lists and layout diagrams with you or me - you're
probably a competitor looking to copy their crap even cheaper. This
is why they came up with Un-marked discrete components and proprietary
User Part Numbers for the IC's - to make it harder to reverse engineer
Suffice it to say the components usually don't blow up if the circuits
hadn't been "Muntzed" (over-simplified - Google it...) and weren't
being overdriven or over-voltaged in the first place.
I may not be an electronics engineer, figuring out the block diagrams
and finding all the blowed up parts and changing them is about my
limit. But I can see it when they do it - they tend to leave the same
circuit board design and component location masks with a lot of vacant
holes. "We don't need any over-voltage protection there, eliminate
those four MOV's on the power lines, and the three on the Input
--<< Bruce >>--
I have, several. But on boat interior lighting and all appeared to be
over voltage problems. But if you are talking about flashlights they
likely will go the distance as every one I have encountered is
substantially under powered. A 3 watt LED, for example, running at a
measured 1.75 watts, using the flashlight case as a heatsink.
On Sun, 6 Nov 2011 08:10:26 -0500, "Stormin Mormon"
I'm not familiar with a LED Mag but every light that I have looked at
- quite a few as I use them for bicycle lamps - two things seem to be
universal. (1) the output (1 watt, 2 watt, etc) marked on the case has
no relationship to actual output and (2) the actual power, referencing
the LED, not the marking on the case, is always very conservative.
However, I think that this is not so much a matter of the LED but on
the relative ease of fitting batteries. i.e., how can we fit the
batteries in the case.
In bulk LEDs are cheap and I suspect that the cost difference between
fitting 1,2,or 3 watt LEDs is insignificant in reference to the retail
price of the light. I see 1 watt White LEDs quoted for $0.30 each, for
example, and it is very likely that a negotiated price would be
Of course, heat is the enemy of a LED but normally flashlight LEDs
consist of the LED fitted to a circular aluminum heat sink which
should suffice. I have never actually tested 3 watt LEDs using only
the supplied heat sink but I have fitted them to reading lamps and
drove them with a 2.75 watt constant current driver and they didn't
get hot and they lasted at least two years. I can't say how much
longer they might have lasted as I sold the boat.
The big problem is that most LED lamps and flashlights are not built with
premium, top-name, first-run diodes, but mostly with rejects, out-of-
current-spec devices, and 'cosmetic rejects'. It takes a big search to
find a currently made, reasonably-priced LED _anything_.
I bought a 4 Watt (3 C-cell) flashlight with a Sylvania 4W LED. It's a
hell of a good light, and damned-near burns grass at ten yards.
It has a limited lifetime warrantee -- as follows. It is warranted for
the lifetime of the original owner except it is _not_ warranted against
the following: Accidental damage, water leakage (It's a "waterproof"
flashlight), or LED failures.
Ray-O-Vac, "premium", Hecho in Chine
So, don't believe ANY of the lifetime claims on LEDs in consumer crap
from overseas. All the LEDs are production culls. I've had numerous
ones fail in the multi-LED flashlights, which I use as equipment
illuminators and inspection lights.
there is a good article in Spectrum a few months back about LED drivers,
led lighting and poisoning of the substrate at high currents - if you
actually want to understand, it would be worth searching for.
otherwise, suffice it to say that running a chip near its limits
shortens the lifespan
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.