High-end Flashlight Makers Could Re-bin LEDs for Better Tint

[KeepingItLight]

One of the big reasons that tint bins have been established is so that a large array of similar LEDs does not show variations in tint. That works great when you have a group of LEDs working together. Take, for instance, the Nichia 219B. Twistedraven reports that the 219B SW45 is rated at 4200-4700K. Several of them together will blend quite nicely, even when there are small variations in tint between individual LEDs.

In many flashlights, however, there is only one LED. That is why tint is such a lottery. If you get a light where the tint is at the extreme edge of a bin, say 4200K in the case of the Nichia above, its tint will be noticeably different than one at the other end of the bin. At 4200K, the tint may be too "rosy" for some.

Does this indicate a weakness in the binning system? For applications in precisely tinted flashlights, should the tint bins be more narrow?

It might be nice if high-end flashlight makers would individually test each LED prior to insertion a light, and re-bin using a narrower range. Better still would be if LED makers would do the testing and offer "precise-tint" models. The Nichia bins cited above, for example, could be divided in half or even into thirds. The LEDs that did not make the grade could be sold, discarded, or used in lower-end models. If, say, five LEDs were excluded for every two that were acceptable, and extra testing expense were also added in, the cost of a high-end flashlight would still be in the ballpark where these things sell.

If this is unreasonable for a large manufacturer to attempt, would it be viable for a modder?

 

[GeoBruin]

Vinhnguyen54, perhaps the most prolific modder in this community, does offer this service. I'm not certain about the nomenclature he uses but for a cost, he will hand select the LEDs with the best and most consistent tint for use in his modded versions of factory lights.

 

[TEEJ]

One of the big reasons that tint bins have been established is so that a large array of similar LEDs does not show variations in tint. That works great when you have a group of LEDs working together. Take, for instance, the Nichia 219B. Twistedraven reports that the 219B SW45 is rated at 4200-4700K. Several of them together will blend quite nicely, even when there are small variations in tint between individual LEDs.

In many flashlights, however, there is only one LED. That is why tint is such a lottery. If you get a light where the tint is at the extreme edge of a bin, say 4200K in the case of the Nichia above, its tint will be noticeably different than one at the other end of the bin. At 4200K, the tint may be too "rosy" for some.

Does this indicate a weakness in the binning system? For applications in precisely tinted flashlights, should the tint bins be more narrow?

It might be nice if high-end flashlight makers would individually test each LED prior to insertion a light, and re-bin using a narrower range. Better still would be if LED makers would do the testing and offer "precise-tint" models. The Nichia bins cited above, for example, could be divided in half or even into thirds. The LEDs that did not make the grade could be sold, discarded, or used in lower-end models. If, say, five LEDs were excluded for every two that were acceptable, and extra testing expense were also added in, the cost of a high-end flashlight would still be in the ballpark where these things sell.

If this is unreasonable for a large manufacturer to attempt, would it be viable for a modder?

As mentioned, Vihn will do this for you....and its a lot of work.

Essentially, Bins give ranges, but, the manufacturing process is not so exact in mass production. So, no mass market maker can afford to spend the labor on this and be price competitive with hos competitors.

There is, statistically, zero demand for this...so, its not even remotely worth them doing it on a mass basis. The few thousand lights, total, that there might be flashaholic demand for are not enough to move a needle measuring overall demand.

It would be like a small number of people insisting that the knurling pattern on the body mate up perfectly with the knurling on the tail and head, etc....its just not on their radar.



If you want order a lot of spare LED, and then hand test them, and then hand sort them to get the tints you want, sure...but no one will do that for you w/o being paid for the hours of work involved.

 

[KeepingItLight]

If you want order a lot of spare LED, and then hand test them, and then hand sort them to get the tints you want, sure...but no one will do that for you w/o being paid for the hours of work involved.

Certainly you are right.

The only cost-effective way to implement something like this would be for the LED makers to do it. They already test each LED individually. They already have machinery to automate the sorting process. All they would have to do is occasionally make a low-volume run of "precise-tint" LEDs. Even if the price to flashlight makers were double that of an ordinary LED, it would be cheap.

 

[chuckhov]

Thanks,
-Chuck

 

[TEEJ]

Certainly you are right.

The only cost-effective way to implement something like this would be for the LED makers to do it. They already test each LED individually. They already have machinery to automate the sorting process. All they would have to do is occasionally make a low-volume run of "precise-tint" LEDs. Even if the price to flashlight makers were double that of an ordinary LED, it would be cheap.

They do not normally test each LED...they do batch tests. The LED are on giant sheets.

Even lumen out put is only given as a range...hence the ~15% variation in output for the same LED too, and so forth.

 

[WalkIntoTheLight]

The only cost-effective way to implement something like this would be for the LED makers to do it. They already test each LED individually. They already have machinery to automate the sorting process. All they would have to do is occasionally make a low-volume run of "precise-tint" LEDs. Even if the price to flashlight makers were double that of an ordinary LED, it would be cheap.

I'm not sure if you're right about all that, but if so, I'd definitely pay the extra cost of the LED. Overall, LEDs are cheap, and wouldn't add much to the cost of the light.

I suspect, though, that if there was a market for premium LEDs, the manufacturers would already offer it. And, if they do already offer it, then the flashlight makers must not think there's any market for them to buy some of the premium LEDs and use in their lights. So, I think the aftermarket modding is pretty-much what we're stuck with.

Thankfully, many flashlight makers at least offer a neutral white version.

 

[KeepingItLight]

They do not normally test each LED...they do batch tests. The LED are on giant sheets.

Thanks for the correction. Does this make it impossible to implement a more precise binning?

Even lumen out put is only given as a range...hence the ~15% variation in output for the same LED too, and so forth.

For me, this is not an issue. Perception of brightness is on a near-logarithmic scale. A 15% variation is brightness is much more subtle than the variations in tint that exist at the extreme edges a tint bin.

 

[18650]

Certainly you are right. The only cost-effective way to implement something like this would be for the LED makers to do it. They already test each LED individually. They already have machinery to automate the sorting process. All they would have to do is occasionally make a low-volume run of "precise-tint" LEDs. Even if the price to flashlight makers were double that of an ordinary LED, it would be cheap.

Anyone buying these would be paying for it plus all of the LEDs that go into the garbage bin and all of the testing/QA.

 

[TEEJ]

Anyone buying these would be paying for it plus all of the LEDs that go into the garbage bin and all of the testing/QA.

Precisely.



I'm not saying that the difference in output OR tint really matters to anyone per se, or not...just that the data needed to sort them as desired doesn't exist in a practical sense, and obtaining it as needed would essentially be a process of hand testing and sorting, and, then, what to do with whatever doesn't make the cut.

Its easy to say "well, just use them in lower end flashlights" etc.

The problem is that LED are shipped, and loaded into mass production equipment , on giant rolls...and if you took them off to sort them, and had "discards", etc...they would no longer be attached to the means of loading them for mass production, etc.

So, again, its typically more hand sorting, and hand mounting, hand soldering, etc.

Handling is a big deal on a large scale.


I'll give an example - If a large chain store has say a refrigerator, and has a forklift put its blades through its box, so the BOX is ripped, but the fridge is fine...it costs so much to be re-packaged, then shipped back again, that the fridge is typically either dumped to a buyer of "lots" of damaged goods (Damaged at the store itself) or donated as a charitable contribution (Damaged at the distribution center), etc.

Imagine a production line boxing lights, or mounting LED....there's typically not really a good way to walk a light over to the line and try to insert it, or toss a loose LED in...it messes things up...and it costs money to stop and start a line, etc.


So, its just not on their radar due to the miniscule market for this sort of thing.

The vast majority of buyers do not actually know what tint is, let alone care what bin they get. If they had a "Hella Bright" 5 L incan, and get a 1,200 lumen ANYTHING, the "tint" is not what they gasp about...they are just thrilled that their 10,000 L Trustfire is insanely bright (Yes, its a 1,200 L light, advertised as a 10k lumen light, but they don't know what a lumen is either...and so forth...).

Remember, there are people who buy purple tinted headlights, on PURPOSE. (Not saying that's a good idea either, but, that's the basic drift of things out there)

 

[Mr. Tone]

Even if manufacturers did this there is still tint variation within the LED itself. The tint changes depending on angle. In flashlights, this is made more obvious by the use of reflectors. If you take a close look, you will see that the spill, corona, and hotspot are all a different tint, especially with Cree. Put the flashlight up against a white wall and shine it sideways down the wall and you should notice this effect pretty easily. The MT-G2 and Nichia 219A don't seem to exhibit this effect as much for some reason. I can't speak for the 219B as I don't have anything with that in it. The tint is one of the big reasons why I am so excited about the new Cree XP-L High Intensity LED. It is dome-less and will hopefully have a much more consistent tint throughout the angle pattern.

 

[TEEJ]

Even if manufacturers did this there is still tint variation within the LED itself. The tint changes depending on angle. In flashlights, this is made more obvious by the use of reflectors. If you take a close look, you will see that the spill, corona, and hotspot are all a different tint, especially with Cree. Put the flashlight up against a white wall and shine it sideways down the wall and you should notice this effect pretty easily. The MT-G2 and Nichia 219A don't seem to exhibit this effect as much for some reason. I can't speak for the 219B as I don't have anything with that in it. The tint is one of the big reasons why I am so excited about the new Cree XP-L High Intensity LED. It is dome-less and will hopefully have a much more consistent tint throughout the angle pattern.

Excellent point(s), but, the sorting is assumed to be FOR a specific light though, in the example posted by the OP...so, that aspect would at least be included.

What heat sinking, and other factors would also come into play.

That all would have also impacted the lumens and throw, etc.

Its a messy concept to wade into, hence, no one wading into it on a mass production basis.


So, if you wanted that service, a guy like Vinh who does it daily, is a great source for example. The other option is, again, buy a crap load of LED, test them, sort them, and mount the ones you like.

 

[ForrestChump]

If this is unreasonable for a large manufacturer to attempt, would it be viable for a modder?

Modder most definitely. This is not lucrative enough for any manufacturer to offer as a service.

Manufactures are simply not set up for it.....less is best, sell lights, make them better, sell more lights. ect.

 

[Mr. Tone]

Excellent point(s), but, the sorting is assumed to be FOR a specific light though, in the example posted by the OP...so, that aspect would at least be included.

You are right, but my point is which part of the beam would they specify the tint is when all three parts are a different CCT? Does that make sense? The testing would need to say that the hotspot is 4000K, the corona is 4300K, and the spill is 5000K etc. since this would be for us picky tint people. Otherwise, we would still be unsatisfied and say the tint sucks

 

[KeepingItLight]

Everyone from Selfbuilt on down has described flashlight tint as a "lottery." The evidence is clear. As far as flashlights are concerned, the binning process used by LED makers is not doing the job.

In early June, this led me to suggest that flashlight manufacturers could take up the slack by testing each LED that is destined for flashlight use. As a know-nothing beginner, I did not see any problem with the idea. I thought that the cost of testing, along with the cost of discarded LEDs, would not be prohibitive. In response, a number of people who know a whole lot more about flashlights than I do explained why flashlight manufacturers could not—in practice—perform such testing. They recommended I head for the modders.

Only two months later, the game has changed. A few days ago, Zebralight announced its intention to individually test the tint of every LED going into its new SC5Fc and SC5Fd models.

This welcome event has reinvigorated my interest in the issue. Hopefully other flashlight manufacturers will follow Zebralight's lead.

 

[Mr. Tone]

Dang, Zebralight is doing something very enticing there. Also, almost 400 OTF from 1 AA Eneloop is extremely impressive.

 

[twistedraven]

Not to mention it's 400 otf lumens from 1 AA and also with a good color temperature and CRI. Zebralight stepping it up! Can't wait to see 18650 version.

 

[fnj]

Anyone buying these would be paying for it plus all of the LEDs that go into the garbage bin and all of the testing/QA.

Yes, obviously the selected LEDs would cost more. So? They would be a premium part. As has been pointed out, the LED manufacturers do not test every LED; just samples. For these premium parts, you would scoop a bunch of LEDs out of the most promising bin, test a whole batch of them, and take the winners and package them as the premium selected-tint LEDs until you have enough. Obviously you would take the non-winners and not throw them away, but simply throw them back in the bin they came from.

 

[fnj]

The SC5Fc and SC5Fd are frosted lens units - that's pretty much required because of the multi-segment EasyWhite LEDs they are using - but it does mean they are pure flood with no throw whatsoever. I have a previous generation F and the flood is very nice where you don't have to see very far, but it does mean you need way more lumens to get the same lighting effect on your target.

 

[StorminMatt]

Not to mention it's 400 otf lumens from 1 AA and also with a good color temperature and CRI. Zebralight stepping it up! Can't wait to see 18650 version.

Kind of odd that Zebralight would use these emitters in the SC5 before other lights, especially since the 400 lumen output is so limited when it comes to usability. I would think they would come out with it in an 18650 light first.