Too many models, not enough lumens

[Xe54]

Hi:

I bought over $800 worth of flashlights about 18 months ago. Since then, there has been a proliferation of new models, yet what has happened with emitters? I don't even see R2 bins widespread. By this time I would have expected some significant increases in lumens. The lack of progress is disappointing. I guess it doesn't help that Lumileds basically fell off the map. Who is competitive with Cree XR-E these days?

The incredible number of makers and models is also overwhelming. Perhaps the depression will weed out some. I also worry about the quality of electrical engineering. I have a Dereelight pill that won't go anymore. I have yet to contact the maker for a replacement. I have also had wierdness with Fenix lights, with surges of light when switching modes. I wonder how many flashlight designers understand switching power supplies to the point of being able to model and prove stability of the control loop? Some of the electronic whines coming from the Dereelight made me wonder about this. Overdriving LEDs also strikes me as bad engineering.

Am I complaining? Not really. More like being a very picky buyer. I am willing to pay a lot for a good flashlight. I'm just wondering when there will be something that represents enough of a leap forward from what I already own to motivate me to buy something?

Good day!

 

[Marduke]

R2's are becoming more widespread. Warm tints and quad-die emitters such as the MC-E and P7 are also hot items these days.

 

[kramer5150]

R2's are becoming more widespread. Warm tints and quad-die emitters such as the MC-E and P7 are also hot items these days.

.... quite literally, they're practically too hot to be used to full capacity in smaller (6P) sized lights.

I agree with the OP overall though. I wish manufacturers put more R/D into efficiency, and brightness BIN progression.

 

[Benson]

Overdriving LEDs also strikes me as bad engineering.

Now that, I take issue with. Bad engineering can certainly lead to overdriving, but choosing to disregard the manufacturers specs isn't bad engineering if you do your own testing (since you can no longer rely on the manufacturer's testing).

To me, that seems like saying that turbocharging an engine sold as naturally aspirated is bad engineering. Certainly, many jobs of it are poorly engineered, but only because bad engineers do it. Nothing stops a good engineer from doing it, and doing it right.

 

[Seiko]

If a MC-E, p7 is not bright enough to get the job done or the XR-E, P4 wont throw as far as you want, you need to invest in a spotlight not a flashlight.

Id rather see battery technology take the engineering and R&D time and effort. There would be more practical applications outside of the relatively small flashlight market as well.


With the "next" generation of multi die emitters already out, in the 3-4000 lumen range http://www.kaidomain.com/ProductDetails.aspx?ProductId=6809
It might be a year or so before they are suited to flashlight size as they wont fit in a maglite head yet. So we know what is coming in LED's, but I sure wish batteries would make a big leap too.

 

[gajslk]

I also worry about the quality of electrical engineering. I have a Dereelight pill that won't go anymore. I have yet to contact the maker for a replacement. I have also had wierdness with Fenix lights, with surges of light when switching modes. I wonder how many flashlight designers understand switching power supplies to the point of being able to model and prove stability of the control loop? Some of the electronic whines coming from the Dereelight made me wonder about this. Overdriving LEDs also strikes me as bad engineering.

I suspect that most manufacturers don't have a EE on staff. As supporting evidence, I can offer a headlamp with copper battery contacts.

My guess is that an applications guy whips a design out for the manufacturer so that they can sell parts. I'd happily believe that the design may get tinkered with at some point by an ME(the one engineer who's actually on staff) or a production guy when a new LED or new model comes out. I do know that parts buyers usually don't get nearly as tense about specs as they do about price and availability. If there's no one around to yell at them, who knows what they'll buy and how the circuit will operate.

In other words, the electrical design, all one hour of it, probably wasn't too bad, but the odds that your particular flashlight will suck are pretty good ...

Feel better now?

Gordon

 

[old4570]

Some one needs to invest in R&D , build a good flashlight so everyone else can copy it , for a fraction of the cost .

I see the current problem as a driver problem .

Plenty of drivers for R2's , but the MC-E and SSC P7 drivers ??? , a lot of them seem to be very poor , though they seem to make good Cree drivers though .

 

[Closet_Flashaholic]

Over the past 2 years I have spent a lot (for me) on flashlights as well. I rode the "Cree" wave. The led technology was improving at a rate that encouraged me to do this. Now, however I believe things are definitely slowing down in terms of efficiency gains. I am not surprised by this, just disappointed. It might be a combination of the maturing of the die technology along with economic conditions reducing R&D budgets - may be less competition also has something to do with it. Eventually companies do have to make money and I view this period as Cree getting back some returns on their Q-bin Cree. Q5's are showing up everywhere in mainstream consumer flashlights which definitely helps Cree with these volurmes.

Ultimately I am still looking for a replacement for my Surefire M3 (MN11-225 lumens) in a Nightcore D10 form-factor (1xAA). Once that's reached, I think that's all that I need in terms of EDC. The technology was on-track to do this in the not-too-distant future, now it looks further away.

What I would like/want is something like a hand-held 0.1-1000 lumens, infinitely variable, adjustable focus running on AA batteries with 1000 lumens @ 1 hour. I think Nightcore or Fenix as the best bet at getting that to me, but based on the LED efficiency slow-down, it's probably not going to be for 18-24 months, at least and probably a lot longer.

 

[Jarl]

Overdriving LEDs also strikes me as bad engineering.

And yet you want more lumens :devil:

 

[kramer5150]

I almost feel like tapping Cree and Seoul on the shoulder... "Uhh OK guys, we all know you can push the ~800 Lumen envelope. Now lets get serious and refine things so we don't need a 3-D mag sized host to keep these things cool."

 

[Xe54]

Now that, I take issue with. Bad engineering can certainly lead to overdriving, but choosing to disregard the manufacturers specs isn't bad engineering if you do your own testing (since you can no longer rely on the manufacturer's testing).

To me, that seems like saying that turbocharging an engine sold as naturally aspirated is bad engineering. Certainly, many jobs of it are poorly engineered, but only because bad engineers do it. Nothing stops a good engineer from doing it, and doing it right.

What you say is true, depending on myriad conditions, of course. For electronics, it is generally the case that manufacturer's absolute maximum ratings are almost never able to be exceeded reliably. There are many reasons for this. You say that one can do testing. Sure, on the lot they have in hand. But process variations may make the next lot fail in the field under the overdrive that was tolerated on the lot used to qualify the design.

What to do, test every lot? Ok, tell me how you can guarantee LEDs will function for N hours at x% overdrive when you need to get product out the door NOW and you can't sit around for thousands of hours to test? Seriously, since the lifetimes of LEDs are expected to be very long, testing is very difficult. The failure probability is a highly nonlinear function of overdrive as well, with considerable statistical fluctuation. Therefore, consistent with what is generally true for most electronic components, it is simply bad engineering to push beyond max ratings.

If you want to see what the experts say about this, here is an interesting link:

http://www.analog.com/static/imported-files/rarely_asked_questions/RAQ_maximumRatings.pdf

But for hard data on what happens to overdriven LEDs (in fact what in general limits lifetimes). This focusses on epoxy encapsulated LEDs:

http://www.molalla.net/~leeper/5mmdeg.htm

What got me started on the anti-overdrive thing was testing some ARC AAA-P flashlights. I was simply astonished at how much overdrive they applied, especially with Li batteries. This prompted me to find the article linked above, and conclude that I don't think the highly revered ARC AAA is a good design. (That and several orders which came with mediocre machining made me finally give up on ARC.) Most people just don't notice because they don't run it for enough hours, and lost lumens won't be noticed anyway until is becomes severe.

Ok, so with your car analogy: The guy who supes up his engine isn't planning to mass produce and sell the mod. It is well known that suped engines blow up at inopportune times. This is almost expected. Just because "it works when I test it" doesn't mean it will keep working indefinitely. There is just no reliable extrapolation model, and that is the reason why it is bad engineering for marketed products.