Replacement LED?

On 3/21/2016 1:30 PM, DaveC wrote:


So it has a boost SMPS?

That's probably a downconversion phosphor. White LEDs are really blue
LEDs with a blob of epoxy on top, loaded with yag powder that converts
some of the blue to yellow.

I beg to differ with your diagnosis.  Unless the LED has magically
turned into a thermostat controlled flasher, it's not going to do
that.  More likely, the LED driver circuit is doing the flashing.
Based on zero detail about the actual flashlight, my wild guess(tm)
would be the big electrolytic that usually goes across the battery.
Broken or badly soldered connections on the driver PCB are also
likely.






If it's not too much trouble, could you disclose the maker and model
number of your flashlight?  Extra credit for providing a link to the
manufactures web site or China source link.  If there are no numbers
or sources, perhaps a photo of the assembled flashlight?




Yep.  You're looking at the wrong stuff.  Maybe something by Cree:
<http://www.cree.com/LED-Components-and-Modules/Products
Nope.  Ok, I give up.  What the photo looks like is a common dome
shaped LED (as in the Cree URL above) with a lens over it.  I did some
Googling looking for the lens and couldn't find it.





I have one that meets your requirements.  L3 Illumination L10:
<http://www.candlepowerforums.com/vb/showthread.php?376058-L3-Illumination-L10-%28XP-G2-R5-or-Nichia-219-1xAA%29-Review-RUNTIMES-BEAMS-VIDEO
Single cell, very bright, tolerably priced, small, adjustable
brightness, etc.  However, there's a catch.  It has 4 brightness
levels set by twisting the two sections of the body.  No problem
except that it makes it a two handed operation.  I would have
preferred a push button at the back end.  I've also measured the
brightness (using my highly creative and non-standard procedure) at
about 90-110 lumens (varies with temperature).



Again, I don't think it's the LED.  The driver board is a more likely
culprit.  Put it under a magnifier and see if you can find the broken
trace.  Put a light behind the PCB to make it easier to see the break.
If desperate, trace out the schematic and try to identify the LED
controller chip.  If it uses an MLCC capacitor, use a hot air gun to
reflow, not a soldering iron tip.

You might also get some help in CandlePowerForums:
<http://www.candlepowerforums.com

Good luck.

Nebo NU15J:

https://www.nebotools.com/prod_details.php?id31%&cid16%&subid94%&subsubcid=

I like this model for its good trade-off between brightness and battery life.  
Might be ?better?, but this fits my needs. (And it ft fits in my pocket  
along with my micro Swiss army knife...)

No controller chip. Looks like a boost converter (inductor, BJT, schottky  
diode, ceramic cap).

I scoped the voltage across the LED: it's constant 4v (with a very  
small--20mv?--rise and fall as it switches on and off). Haven?t yet  
measured current. Surely if there was a bad solder joint or other failure it  
would show up here.

Thanks for the referrals of other lights, but I?m not buying a replacement.  
I?m doing this for fun and to learn.

Cheers.

On 22/03/2016 05:56, DaveC wrote:


Desolder the original LED, carefully , as balance of probability it will  
be fine.
Jumper in any old high power LED and see if that flashes too.

Looks like the flashing LED is a common problem:
<http://www.candlepowerforums.com/vb/showthread.php?417142-Replacement-LED



Some day, someone will design a flashlight with an automatic PWM light
dimmer.  Shine the light at something bright, and the flashlight runs
at full brightness.  Shine it at something in the dark, and it goes to
fairly dim.



Yep, that's about it.  No controller.





Ok, I'm wrong.  The flashing is not caused by the non-existent
controller chip.




Like I mumbled previously, see if there's a dome type LED under a
plastic lens.  You'll probably need to unsolder the LED leads to do
this.  If they really are two parts, it should be possible to find a
replacement LED.

If you can't identify the LED by the power output, try measuring the
current drain with a new battery.  Multiply the current with your
measured 4V, and you should get the power in watts.  That should tell
you if you should be looking for a 1/2, 1, or 1.5v LED.

Jeff Liebermann sez:



That?s my post on canblepowerforums.com. (c:

lens. You'll probably need to unsolder the LED leads to do this. If they  
really are two parts, it should be possible to find a replacement LED.

Will do.


You mean 1/2, 1, or 1.5 *watt* LED, yes?

Thanks.

Oops(tm).




Oops 2.0(tm).  I was in a rush to get out of door.  Sorry.  Also, I
just noticed that there is no such thing as an individual 1.5w LED.

I think this is an under-used(?) 1W LED.



unit. That way can dissipate some W.

But that presents its own set of problems. Optics will need to be matched to  
the new LED.

Hmm...

Sounds about right.  If you can light it up, just measure the battery
current drain, and calculate or estimate the power dissipation.



If you can remove what I believe to be a lens, I think you'll find
that the actual LED is rather conventional and can be found in the
Cree catalog.  
<http://www.cree.com/LED-Components-and-Modules/Products
If you sort the above list by power output, there are only 3ea 1w
LED's listed.  Just find the right die size and good luck soldering
the tiny chip.  You can also dig throught the current flashlight
offerings and see which 1w chips are popular.



Learn by Destroying.... then buy a new flashlight.

[Jeff Lieberman]

I found these:

https://www.led-tech.de/en/0.5W-Power-LEDs_DB-86.pdf

Looks good, no?

Please deduct one point for spelling my name wrong.



No.  I thought you said your flashlight was bright and possibly 1
watt.  Those are common 5mm LED's with 0.5 watts maximum.  
<https://www.google.com/search?q=5mm+white+led+0.5+watts
These photos look very much like what I would expect to be hiding
under what I believe to be a lens of some sorts:
<https://www.google.com/search?q=5mm+white+led+0.5+watts&tbm=isch
This looks a bit closer:
<http://www.ebay.com/itm/161246348498

That data sheet is also slightly insane, specifying the luminous flux
as:
  Lumen typ.: 19 mcd
Lumens are measured in umm.... lumens, not millicandelas.  It is
possible to convert between lumens and mcd's using the viewing angle:
<http://www.rapidtables.com/calc/light/mcd-to-lumen-calculator.htm
Anyway, 19 lumens is not very bright but probably good enough for a
pocket flashlight.

And I tried so hard to get the ?Lei/Lie? bit right! (c; 2 steps forward,  
1 step back...


One of those images (located here):

http://tinyurl.com/j4pwvuw

looks exactly like the one in my light. It?s squat, with very (relatively)  
large yellow die(?) in the center. I don?t think there?s a separate lens.  
Maybe it?s molded to focus the light (integral lens)?
Wish I could find that LED other than @ Alibaba...

Ah-HAH! From that Alibaba image page: ?strawhat LED?. Search turns up  
similar-looking LEDs. Searching on the terms ?strawhat? and ?dimple?  
tells me that this form-factor is for radial distribution of the light, not  
throwing a beam. Which means the flashlight designer wanted the reflector to  
shape the beam, not the LED.

So I?m looking for one of these types.

Thanks.

(Why are my search skills so crap? Google keeps returning Manga images and  
cartoon faces...)

Hmm, that sounds like a good idea to put in my planned LED torch project,
which was going to have a manual dimmer anyway. I guess an LDR would be
suitable? Maybe mount it in a small tube to make it directional?

I guess I'll have to find an LDR and wave it around with my multimeter
attached. Ahh, the dignity of research. :)

On 22/03/2016 03:51, Jeff Liebermann wrote:

<snip>



Thermostat controlled flasher is probably the exact description. It  
happens when LEDs are driven straight from batteries too, and is  
probably the bond wire heating and cooling making intermittent contact  
at a few Hz.

I have a cheap LED torch which uses nine parallel connected white LEDs  
running directly from 3 AAs.  After a battery replacement, one is out,  
four flash at different rates, and four are on.


Cheers

Circuit:

http://imgur.com/YeELiHI

which is identical to the application note in the datasheet (scroll down to  
the ME2108A diagrams):

http://img.ozdisan.com/ETicaret_Dosya/431770_1622138.pdf

(easily translated via Google, but a bit more confusing...)

Measured current at the battery terminals is 275 mA (due to meter insertion  
loss it?s difficult to get an accurate current reading at the LED?s  
terminals--the LED dims). Voltage across the LED is a pretty-constant 4v.  
Estimating the efficiency of the converter at (roughly) 75 percent the LED is  
using 825 mW.

I guess I?m looking for a 1W replacement?

How?s my math?

Thanks.

On 3/22/2016 4:41 PM, DaveC wrote:


Didn't you say it used a *single* 1.5 volt AA cell?  I think that would  
make it 412 milliwatts not counting the conversion efficiency, so more  
likely a 1/2 watt LED.

1.5v boosted to 4v (measured) output from the converter. 275 mA (measured) at  
the battery terminals. I make that to be 1100 mW. If presume 75 percent  
efficiency, 825 mW.

No?

On 3/22/2016 10:00 PM, DaveC wrote:


Perhaps I am missing something.  The battery terminals will be at the  
battery voltage, no?  So why would you use 4 volts which is at the  
output?  It would be 4 volts times the LED current or the battery  
voltage times the input current.