Need some friendly advice on choosing LEDs

Another-Nocturn

Newly Enlightened
Joined
Aug 1, 2010
Messages
5
Hey Gang; I am working on a new project and would like some friendly advice. You guys have worked and played with LED's far longer than I; so maybe you can help me shorten the learning curve and avoid a few mistakes along the way

I am designing a new product whose goal is to be placed into mass production. It's a specialized flashlight with both UV (for specialized work tasks) and also plain old white light (for general work tasks). The power source will be 3 standard D size cells, but could also be 4 AA cells (in a cartridge) if needed. In my testing, white-light output that is on par with a mini maglight should be enough to get the job done. Likewise, the light output from the 12-20 LED UV flashlights (the ones we have all seen on eBay) is also enough to do the job (although a little more UV output would be nice) The product will be designed with some level of "intelligence", so it wouldn't be difficult to also add a voltage regulation feature to "normalize" the output power from the cells.






My requirements are as follows:
  • The cost of the raw parts needs to be very low (when purchasing in volume)
  • The overall footprint of the circular LED PCB should be as small as possible, hopefully not much bigger than the diameter of a D cell.
  • If emitters other than the standard 5mm are used, their "reflectors" must also be accounted for in this size limitation
  • Device may be used up to an hour at a time, probably switching frequently between light sources (UV or White)
Using my CAD software, I've laid out an arrangement in 3D space which may work within the confines of my footprint. It consists of:
  • [16ea.] standard 5mm UV Led's, (These are typical UV diodes in the long wave-UV family)
  • [1ea.] 10mm White LED within the footprint of a D cell. (The single white 10mm LED contains 5 chips, draws 100ma and puts out 265KMCD in a 40degree beam)
5mm10mm.jpg



The output of this bundle seems to do the trick but i'd like to make sure Im not overlooking a better solution or walking into a problematic design.

Here are my questions:
Q1) Before I tell you my own doubts, what do you think of the plan as shown above?

Q2) I've read mixed reviews of those 10mm light cannons based on the fact that at full power, they overheat and die, is there a way to avoid that fate by reducing the loads; or are they really a dead-end design that should be avoided?

Q3) would it be better to explore the PCB-mount designs (like Cree) instead of through-hole LED's? If so, wouldn't those also require special drivers, reflectors and heatsinks?

Q4) I've seen Maglite retrofitted with Cree-style UV bulbs, but they are always very expensive and I think those bulbs are hard to find. It seems like this approach wouldn't be smart for a production model with a low price point…agree?


Again, I really appreciate all of your help; and fully plan to update this post as the design evolves
 
Welcome to CPF, Another-Nocturn :)

Very interesting project for your first post here. I think it probably belongs in our Homemade & Modified section for the time being, so I'll move it there.
 
I would say avoid the low power LED's at all costs.


Check out this fade test on cheap 5mm LED's: http://www.candlepowerforums.com/vb/showthread.php?t=264326

A Mini-Mag AA puts out about 10 lumens initially. 10 lumens is ridiculously easy to get with a modern emitter, and will require minimal heat-sinking. A neutral tint XP-E (not as efficient as the newer XP-G, but easier to focus in a small reflector) with a thermal pathway (ie: an aluminum plate connected to the body of the light) will be easy to manufacture, cheap, effective, and have better light quality than any 5/10mm LED. IT will also be a much better quality product.

A high power UV LED would also be much better than the 5mm ones, if you can figure out what wavelength you need. That LED could also replace the cluster of 5mm's, be put on the same heat-sink, and overall be cheaper to manufacture (~4 solder joints for the LED's, instead of 20+).

Also, why Dcells? Lithium AA, or even other battery types, such as CR123 would be much more compact and cheaper to manufacture (less materials cost). Even a 2AA format with a simple driver should cost less (in my mind, anyway) than a larger format light with multiple crappy LED's to be produced. Even if the two designs worked out to the same cost after tooling, my design ideas would lend to a better quality product that could be theoretically slightly more expensive for the consumer ( more profit in your pocket).

My $0.02
 
Awesome response John.
My electrical engineer had many of the same thoughts with regards to the battery choice and your right about the questionable quality of the plastic LEDs. I am indeed wary of creating a product that might end up being replicated, sold and shipped out of hong kong for 99 cents on eBay

My fears has always been what would happen to the price point if I started using the high end emitters. As much as you and I can appreciate the quality of a good tactical light, the average Impulsive Joe shopper may not be able to get over a price point that comes with it. I suppose this makes my job of making it LOOK like a premium design that much more critical.

In your reasoning, you explained that using good emitters was actually not all that expensive. I know it's not a fair barometer to gauge price by looking at single units sold at boutiques, but I need to be able to buy these things in volume for a few bucks tops. For me, volume means 5K to 10K.. *Do you think that is a reasonable goal? (I need 1 uv emitter at 390-400 and one warm or cool white. * I am now thinking about switching to four AA cells as my power source (even numbers are better for this particular project and I don't mind the extra capacity).

Definitely interested in additional feedback
 
Last edited:
A bit more on the "thinking like the buyers" topic.....

I stumbled across these two UV flashlights on Amazon. Both tout high output UV in the 395 area, and both undoubtedly deliver blinding output. (yes.... the obviousness of this pun almost hurts.... (Aha, a double dose!))

The first model is your run-of-the-mill ebay-special, sporting 51ea 5mm UV LED's packed into a standard Maglite-style casing. It sells for less than $20 Its got 50+ raving 5-star reviews from folks that are just blown away by the output and (cough) the build quality. most folks describe it as being "just as good as a maglite; which for many implies that its quality is top notch.
http://www.amazon.com/dp/B001Q70A0G/?tag=cpf0b6-20

The next light is also a high performance UV light, this time with 3 Cree-style 1Watt emitters. It sells for just under $70 and lacks any reviews whatsoever. When looking at it, the "Customers ultimately bought" section of the ad even acknowledges that 41% of the people who looked at this light actually ended up buying the 51-LED model listed above.
http://www.amazon.com/gp/product/B002ZBOSG4/?tag=cpf0b6-20


Now, I presume that the quality of the 2nd model is just as good, (probably much better) and we can all agree that the Emitters are probably an order of magnitude more consistant, predictable,and yielding a cleaner output. But clearly, the average joe's dont see the advantage or they simply decided that the lower-grade model was worth taking the risk at such a low cost.

I am definitely not being a troll here, but I thought this was a great illustration of the "stigma" I am up against while trying to make the decision between high end and low end light sources. If I can get higher end parts at a lower cost to make a pricepoint within 20% of the cheapy flashlights; il be in great shape. But a model that comes in 3X the cost with benefits that only elite hobiests understand will be pretty hard to move on a retail box environment.

Aaaaah weighing it out makes my brain hurt.... What is everyone elses take?
 
Top