Hi Guys,
I'm coming in way late on this discussion, but it's been a very busy couple o' months between some serious family illness, watching kids and house and work and hydroponics tomatoes contest and dogs and...
I'll try to weigh in a concisely as possible, but I've always been the wordy type, so try to bear with me.
I have always been an efficiency buff... when I first got into LEDs, it was because
a) they are more efficient (in the range/uses) that I would use them in, and
b) they finally had white leds...
It's been a good 20 years since I've had to do any electronic cct design, so I'm probably very rusty, but... one of the thing that has always galled me was the fact that passive current/voltage control has always been lossy/in-efficient. That resistive current limiting... can easily be less than 50% efficient. Even the semiconductor-based linear current limiting can be abysmally in-efficient. When I started looking at re-tooling to build my own, I realized that I wanted to get every drop of electrons that I could to the LED junction instead of having them burn off in friction/heat. Using 9 v batteries meant burning off electrons to control the amount of current heading into that PAL SMC LED... and you could feel it if you burned your finger on the LM318 chip while lighting the room w/ the cover off. 9v's don't have that great of current density to begin with -- less than two AA's, but they DO have a voltage higher than the 3.4-3.6v required to drive the white LED's into their "linear" conduction range. Two AA's might give a meager glow to a WLED, but no real "light" is shed on the matter. Step up to 3 AA's, and you've back in the "burn off the excess" to control the current, and even then, when the cells have gone from 1.5v each down to 1.2v each, you've exhausted your range and you're back to under-driving your LED w/ only 1/3 of your current capacity used. There was always the NiCds -- they put out "about" 1.2 volts (or 1.4, for the ones I used) so a three-stack would give you 3.6v (to 4.2v) -- just the right level for the LED to sip current at its leisure -- until the meager current density was drained enough so that the v level would drop into the under-driving level. Sigh. There HAD to be a better way.
(For those of you who are easily bored, you may wish to skip down a way... I do belabor a point below).
When I found the Maxim ccts, I was in heaven... step-up/ step-down, high efficiency (compared to the 40-60% of the passive/linear methods), etc. Single-cell, multi-cell, low-current/high-current... it was all there. The only problem being... these are absolutely miniature devices... not like the discrete components or even ICs I designed with. The ccts are cheap, too... $1.68-$3.?? for the majority of their products -- BUT those little "chips" are about 2/16's x 3/16's of an inch in size. How the heck does one "bread-board" those?
You don't. You need cct layout, and mechanical parts assembly. You couldn't even buy the protoboards w/out laying out $60-120 for a SINGLE cct. NatSemi and a couple of other companies have similar products, but again, if you want a prototype, you have to build it yourself or pay $$$ to get one to tinker with.
Why bother?
Well... 85%- 95% efficiency. SINGLE-cell input. HIGH current capacity output... What does that mean? Well, AA input down to .6v will provide 3.6v at up to .5 amps at varying efficiencies above 78%. Efficiency goes up w/ V(in), but still, 78% is not bad at all... what did that mean to me?
AA alk cell: 1.5v @ 2850 mAh (my memory... ) means 4275 mWh -- pretty nice considering I can buy them in bulk for $0.37 each... Those 9v batteries I use to throw at the PAL? 5355 mWh @ $1.50 each... and over 50% of that goes up the stack as heat...
With the Maxim ccts and an AA, if you "estimate" 78% efficiency, you have approx 3334 mWh available at the voltage of your choice... if you merely drive the WLED at it's rated current (20mA) you can get a _theoretical_ time of 46 hours... probably closer to 40, but that's a "full" brightness. Drive it at 40 mA for 20 hours? 60mA for 15 hours? All theory, as I don't have $60-$120 to "burn" for the proto board...
If you use the two cell input, efficiency is even higher. How hard do you want to drive 'em? ON-board current out can hit 500mA (3A for some ccts), and w/ external components, 5-10A. Granted, that takes your 40 hours down to a mere 1.8 hours (probably less), but you could drive a mess o' LEDs w/ that -- 25 at 20mA?
The best thing? It's constant. No sliding current capability... no voltage variations... CONSISTANT LIGHT OUTPUT.
You set the voltage and it provides as much current as you need, up to the capacity of the device -- and LEDs are current driven devices (IIRC). Two AA cells -- about the same size as one 9v, only now you have 7W available for your paltry 74 cents.
(Still w/ me? Let's pick up here...)
When the CMG Infinity and then the Brinkman came out, I was very happy. They aren't efficient, but they make up for it by using inexpensive batteries... The CMG is v. small for it's day, the Brinkman v. bright. But the only way that there would be advancements would be through better power ccts or through better LEDs. We have both now. If you're running your LEDs off of a car battery, you don't care about the efficiency... any amount that you "burn" is negligible compared to the capacity of the battery. Similarly, if all you care about is bright, battery capacity isn't really that big of a deal... you're willing to throw weight and cost at the problem of intensity.
Me? I wanted it all... bright enough to see to do what I was doing, but efficient or cost-effective so that it didn't cost me $$ each time I had an outing... the CMG is plenty for 90% of my work -- because 90% of my use is running through the HOUSE late at night, or working up very close... in either case, I know the territory. But what if it could get better? That's what we're here to discuss... my druthers in the matter at hand?
Size: Smaller is better for me, because I am a casual user who needs something easy and convenient to carry. As with concealed carry, the .22 in your pocket beats the .45 at home in the dresser.
Battery selection: Important because it translates to $/use. Working by the side of the road late at night? Those extra couple of hours bought by the better cct are mighty handy.
Intensity: Important because it is related to $/use or task. The trade off again...
Circuitry: Wide-range step-up, OR, step-up/step-down cctry... 1-,2-,3-,4-,6-cell input? Doesn't really matter, because the efficiency is about the same, and the light output depends upon the V(out) design -- and the LED uses as much current as it wants/can at the V(out) level. Ostensibly, one could design a cct that would take anything from 3v - 18v in, and provide a rock-solid (well, w/in reason) 3.6-4.2 v out at just about any desired current level. Designer picks the right reference resistor and the cct locks on that that v. level.
Design... My FIRST WLED light was the cheapy "throw-aways" that TechAss used to make. Used to. They had design problems -- but what do you expect from a light that consists of an LED whose "legs" straddle a watch battery and is coated in a rubber/plastic? Extrememly light, reasonably cheap ($6.50?) and very convient design... this was my first "bite-light" -- with the benefit that I could also control it the same way. Walk the dogs at night? But the thing between your teeth and you have both hands free -- and it points the way your're looking. Small, light, flat. Bite, it lights. But once it's used up, your only recourse was to cut the rubber off and keep the expensive little LED.
My next was the PAL (in gold). Handy, but bulkier for the pocket. Could still function as a "mouth-held" light, only you couldn't really control it, and your jaws get tired after a while. Don't EVEN think about trying that w/ the Brinkman -- unless you're a mashocist.
The Ergo works, mainly because of the shape of the lanyard region -- fits reasonably nicely 'tween my front teeth, but then I drool, and the weight is well forward -- it gets tiring, so it's only a short term solution.
The CMG is similar... v. short term... hurts the teeth.
So where does that lead me? Design... a two-AA power source, side-by-side, should work well for no-handed operation -- even if I can't control it. I can turn it on, crawl under the hood at night, or squat beside the tire, and get to work.
So -- Put 'em all together and where does that leave us? Right where we're at... only I don' t know what cct the Arc folks are using that let's them hit (IIRC) .3 - .4 v and still get output. Whee-doggies!
One (or more) "Heads" -- has the power conversion cctry in it, w/ variety being expressed by the LED device, and,
More than one body... pick your need, design, etc.
Can I afford 'em? Er... at $60 per, not yet, but maybe one day...
I'd be happy enough w/ having an in-expensive "universal" head-unit, and just design a battery holder for it... which is basically what I would have gotten w/ the Maxim proto board, and what I plan to do w/ the Arc Light if I ever get it... twist its widdle head off, and make a converter to fit the CMG body... presto-chango, best of all worlds... real brightness, small size, inexpensive but higher capacity battery = Happy camper.
If the Luxeons make it to my table some day, I'll pick up there and be happy. If not, then there's always other "fun" projects... like the wall-charged, NiCd-driven, photo-transistor controlled LED power-outage lights... which we needed two weeks ago when Richmond lost electricity... If power goes out, or it it's un-plugged, then you have enough light to navigate a room by... for a little while (let me see.. 600 mA capacity / 20 mA requirements means... about 30 hours? Heck, even if it's just 10 hours -- figure winter dark-time -- then it can charge and dis-charge every day and still be of use. Then I can start dreaming of retro-fitting the motion-sensing units... heh heh heh)
Whew. Hope you guys lived through that... I won't be offended at all if you didn't.
