Whats the big deal with Bin Codes?

Ive read around on this forum for a while and am supprized that there seems to be such a difference between model #'s. Is there realy a big difference between say... two 5w leds of different codes? Did somone make a table listing the good bin #'s? How do you know which ones are good and which ones arent?

Bin Codes Explained

that web page is what i reference when i'm determining the bin codes.. a lot of people consume brain cells with mezmorizing such, but i'm of the type that it's best to remember where to look than memorize.

The basics are this:.. there is a wide range of variance on the LEDs off the assembly line... so just like most things of the sort, (pun oops).. like say CPUs or RAM... they are tested, and grouped or sorted (hence the pun) into 'bins' based on how bright they are, what color they are and what voltage it takes to operate them.

When people brag about or clamor for a special bin.. typically it means one of three things..

1) Very bright... the same model emitter can vary by more than twice the brightness...so a 'bright' bin means with the same power, more lumens.
2) Very white... some emitters are very green or yellow or sickly.. others are perfectly white.. the lux3s are soo white in general they split up white in to several 'almost white' bins
3) Very low operating voltage (this can be bad or good depending on your design).... if you want to direct-drive.. having a high Vf can be beneficial, because you can operate the light w/o using a resistor or drive circuitry.. in addition.. some circuit designs need the Vf to be above the Vbat... and a high Vf will help.. in other designs.. Lowest Vf wins... because you can use a battery with less voltage to operate it full-power.

there ya go, hope it helps.

-awr

Another thing to keep in mind is that every bin is a range of values so ,even within bins , there can be a wide range of tint,efficiency, vf. Bin codes should also be taken with a grain of salt because leds can change quite a bit through the early minutes of life (I have heard) so if the testers don't give the led a real chance to see what it can do then the code may be way off.

Take bin codes with a grain of salt because even if the bin sounds good and the led produces a nice light in one part of it's beam , leds can have big bubbles in them that essentially makes them useless for use in a custom flashlight.

I've got a short FAQ entry on this with some links you may find useful. Hope this helps /ubbthreads/images/graemlins/smile.gif

Now Ive Noticed that the Bin #'s get better as they progress down the alphabet.
U... 87-113
V... 113-147
W... 147-192
X... 192-249

Now when I was an engineer I used to make plans for certain items. And if I were designing the same thing with a slight improvement I may give a model number to corespond with it. For example: I start making bulb model A, then I go back and re-do my designs untill I come out with new improved model #B, then C, and D, and so on and so forth.

SO I'm wondering is the fact that these model #'s seem to be improving with each sucessive bin #, does that mean that each bin # is actually a slightly improved design as well? Are these bulbs improving over time?

Yes, Newer LED's are generally better than the old ones. Just like computers keep getting faster and faster.

I believe the performance improvements are more a result of refinement of the manufacturing process, not the design itself.

Well, "performance improvements" is not quite the right term. "Improvement in average performance" is a better one. Emitters are binned by the batch, not individually. The emitters in a batch (for the most part) perform similarly. Over time, the process is improved and the average performance will increase, but any given batch may still only meet the U bin spec. There should be fewer U batches and more X batches in the future, though (hopefully!).

Hallis' analogy to computer chips raises interesting points. As things are going now I sure wouldn't plunk down 150.00 bucks for an led flashlight and try to justify the expense by saying I'll never need another flashlight. Also, being extremely anal about bin codes can really set yourself up for dissapointment because in six months the 'run of the mill' led might be better than the premium of today.

My 180.00 arc4 rev1 was only bought last year but it's already feeling long in the tooth and I'll be happy to replace it with a lionheart (as long as the LH get's a convenient, quick access carry method). Well , sorry to ramble off topic....

Da shoat ansuh - the approximately 100cm diameter thin silicon wafers from which LED chips are punched are organically "grown" and thus Mother Nature has a large part to play, and having some degree of random crystalline structure formation to affect the process means there is still a bit of black art going on here.

No LED mfr. deliberately makes lower-ranked products, but out of every batch there will be good performers, regular ones, and poor ones. Gotta move the poor ones somehow or that is money lost so they get shipped to the mass production users. Gotta get maximum revenue from the good ones hence the "bidding wars".

I myself have sent out countless bribes of Kona coffee and Macadamia nuts... /ubbthreads/images/graemlins/naughty.gif

Similar to computer CPU manufacturing. generally CPU's from the center of the wafer will be of slightly better quality and tolerence then the ones on the outer edges for the very reasons that Mr. Bulk just explained. The CPU manufactuere's colution is to underclock the less tolerant CPU's and sell them as say a 2.6ghz instead of a 3.2ghz like the ones from the center.

very true hallis my cpu is a amd xp 2100 ,but has been running at over 3400 speeds for very long time 24/7 they could of sold it as a 3400. well let me say they could of sold it as a 2800 .i had to raise vcore to get 3400 speeds. if i had better memeort i could do better.

Right... CPUs are a good comparison. Over time, faster CPUs in the same family become more available as the process improves and more of the resulting chips meet the higher specs, even though the chip itself hasn't been redesigned.

Hmmm, I wonder if the chip-farming process is the same for the two? If so then Lumileds should talk to the computer industry!

I dont think Lumenleds is going to be needing a .13 micron manufacturing process /ubbthreads/images/graemlins/wink.gif

oh oh! my 1700+ has been running at 2400MHZ (3200+ PR rating? not sure they don't make a T-bred that fast lol)...Thanks to it's great bin code, I mean, stepping! lol

Ialso, I was reading an article that computer processors could be grown also! of course, they are extreamly slow and don't just "plug in to work" and I wouldnet expect any instruction algorithms to be built in either heh...but it was interesting!

my JIUHB Tbred-B 1700+ would do 2014mhz on air, My motherboard didnt support multiplier change so that was all FSB. /ubbthreads/images/graemlins/smile.gif Good CPU's

Wow! My Athlon 1200 has been running at 1140 24x7 for 3 years. It was last rebooted Sept 12th when we lost power. Before that was in April (when we lost power).

Sometimes they are pessimistic and sometimes they are optimistic. Mine is unstable at 1200. A program would fail about once a week.

I have to wonder what causes the "sweet spot" in the silicon slice for LEDs. I sure wish they could duplicate it.

Daniel