Instructables

I'd like to challenge everyone here that makes useful things.

I'd like to challenge everyone here that makes useful things.  I'm rebuilding an old light fixture and I'm focusing on efficiency.  

My challenge is to get the highest amount of lumens with the lowest number of watts. 

I'm stuck and I'd be willing to put a cash prize on this.

I'd like for everyone that gives this a try to stick to DC voltage.


hehehe, laser pointed at a diffuser. anyway what kind of bulb does that thing use?
00MrNotWrong00 (author)  nutsandbolts_644 years ago
I made an instructable http://www.instructables.com/id/LED-bedside-lamp/ Laser pointers actually use a lot of power , the one I use to play with my cat never lasts more than an hour or two. Check out my ible and if you want one of those bulbs pm me for a link.
00MrNotWrong00 (author) 4 years ago
That's what I'm talkin bout !!!!! Lots of light , little watts , sorry it took so long , life has been crazy . The ratio to beat is 420 lumins / 5 watts . That's 84 to 1. I know it can be done , and I still have that 100 bucks, my build cost a total of 50 dollars so if you do it for around the same price you will walk away with a really cool light and 50 bucks. I'm going to make an instructable for this as soon as I finish the one for my solar furnace .
35313_1483004429706_1071701142_1411988_8049922_n.jpg
00MrNotWrong00 (author)  00MrNotWrong004 years ago
I made an instructable http://www.instructables.com/id/LED-bedside-lamp/
00MrNotWrong00 (author) 4 years ago
I just got the 20 dollar light and O.M.G. it is so bright,  totally not a typo,  or a misprint.  It's not really useful at the moment but as soon as I get a few reflectors on it and a nice shade,  I'll be posting lots of pictures and adding an instructable on this topic.



Wow,  very bright .
Do you have any way to confirm how much power it's using?  Maybe an ammeter in series?
00MrNotWrong00 (author)  kelseymh4 years ago
I have a friend that has a fluke meter but right now I have it hooked up to two 6 volt lantern batteries,  I know they are limited to three amps.

How many amps is 5 watts @ 12 volts ?  I could put a fuse for that value in the line and if it does not burn out then we will know that it's over.

Man it's bright !
00MrNotWrong00 (author)  00MrNotWrong004 years ago
Well that's not going to work,  seeing as I don't think they make a fuse for  0.4166666666666666666666666666666667  guess we'll have to wait till I can get my hands on my friends fluke meter.
00MrNotWrong00 (author) 4 years ago
Well I think this comes out to 4.2 watts,  but like I said before,  I'm not the best with the math.ledsupply.com/creemce-w430.php

If anyone knows of anything with a better ratio please don't be shy.
kikiclint4 years ago
 So a 100 watt lightbulb puts out about 1600 lumens, and a 23 watt cfl should put out the same.  If you took cgi.ebay.com/5-x-3W-High-Power-White-star-LED-130Lm-3-watt-lamp-5pcs-/320523081829, 3 watts with 120-130 lumens for 1-3$ each, to get the same results you would need 12 leds at 35-36 watts total.  The 10 watt led is 700 lumens, and 20 watts is 1500.   Price is about a dollar a watt, and you can choose between the warm white and regular white.  They even sell bulbs in the 3 watts range pre made. 
00MrNotWrong00 (author)  kikiclint4 years ago
That's good,  I think that comes out to about 40 lumens per watt.

Witch brings me back to having thoughts that the 420 lumens for 5 watts has to be a typo or misprint or something funny about the way it was measured.

What dose that come out to ,  like 84 lumens per watt ?  That's a bit far out isn't it ?

I'm getting all getty waiting for the UPS guy.
00MrNotWrong00 (author) 4 years ago
I think this may be my answer www.amazon.com/BA15d-White-8-30V-AC60-Lumen/dp/B002BHSJSA
That's more lumens than a 40watt incandescent,  with a nice diffuser this could be what I'm looking for.

I went with the thinner wire because all the cars I've ever worked on,  work on 12 volts so I just figured,  the people that engineered these systems must have done it the most efficient  way possible.  I guess I'll be redoing that. 


That looks like a good bet.  It has an excellent output to power ratio (probably because they're using surface-mount LEDs without thick diffusers (which absorb a good chunk of the light).  This sort of commercial production is likely to be more reliable than a home-brew item, anyway.
00MrNotWrong00 (author)  kelseymh4 years ago
I've got more than twenty hours into my search now,  and I have not found anything better.  I'm starting to think that's some kind of misprint.  Even the expensive bulbs made to replace the incandescent bulbs that run on house voltage don't have that good a ratio.  I'll know in 7 more days.

I still need someone to look at the specs on those 10mm LEDs to see if I got that right.  I mess stuff like that up all the time,  it would be nice to see what someone else finds.
00MrNotWrong00 (author) 4 years ago
OK I still need someone better at math than me to have a look at these numbers.  I have "Kiteman's"  10mm diffused LEDs at 0.550 lumens @ 0.060 watts.  If that's wrong I need someone to get me reference material with the correct specs,  Those are the numbers I came up with.

If those numbers are right than it would take 6watts to get to 55 lumens and That's not as good as the 5watt 420 lumen bulb I post a link to earlier.

So someone please check my numbers and let me know if I'm right or wrong ?
kelseymh4 years ago
I will be very interested in the engineering solutions, if any are presented.  Here's the physics that limits the problem.  To maximize the visible light output for a fixed power, you need a narrow-band emitter. 

That rules out incandescent, since the blackbody spectrum (a good approximation) is extremely broad and peaks for any reasonable solid in the near IR.  Most of the power goes to heat, not visible light. 

So what's left?  Fluorescence and phosphorescence are relatively narrow bandwidth (actually a collection of transition lines), so you could find a material, probably an organic scintillator, with all of the lines in the visible.  The trouble is both of those processes are absorption-re-emission, they aren't driven directly by electricity.  So you run into efficiency limits -- you have to produce high energy (UV) photons, which are down-converted to visible, but with much less than 100% efficiency.

This leaves you with solid state systems that can take a voltage (energy gap) and convert it directly to visible light output.  That's LEDs, or organic LEDs, or quantum dots, or something similar.  There is still the efficiency question, but at least it's a single-stage process, not a cascade.  The efficiency is determined by the materials used for the solid state substrate and dopants.

So the bottom line is that your solution is going to be visible light LEDs.  Doing any optimization beyond choosing the right off-the-shelf components means going to a fabrication lab or facility, and doing some condensed-matter physics experiments.
Meh, my answer was shorter...
Pffft :-P  My answer wasn't really for the author, it was for all the users out there.  The "best" solution already exists, as you've shown.
00MrNotWrong00 (author)  kelseymh4 years ago
I think we are missing a few key words?  The first key word is "Useful"  a few LEDs taped to a battery isn't exactly useful.  As shown in that instructable ,  I like the concept and I'd love to see someone do something constructive with it,  but as is "Not Useful" .

The second key words would be "lamp" , Like something you might use to read in the dark,  or light a room.

As for what works best I've found that found that in fluorescent lights of the same wattage can differ drastically in brightness if they are labeled  "cool white" or "warm white"  and LEDs also very in brightness,  The LEDs used to replace mag-light flashlight bulbs are much brighter than the same rated LEDs you can get from radio shack. 

So if no one can put something useful together then I guess the LEDs taped to batteries will be the winner,  But I'd really like to see someone build something useful.
 I think you should spend some time reading the led sections on here, and waste $15 on a handfull of high power leds and some components for driving them.   LED's are quite versatile, and with some skill, can become far better than leds taped to a battery.  By the way, they do sell led lights ready to plug into 12 volts outlets premade on ebay.  That is where you will find the cheaper version of the mag light replacement bulb.  I would look at individual bulbs in the 5 watts or higher range for higher brightness. 
A few LEDs and a battery are a flashlight.  That definitely counts as a lamp, as I've read by flashlight many times, and those silly clip-on "book lights" are just flashlights with an overpriced handle :-)

Your third paragraph is right on the money.  Those differences are due to exactly the underlying physics I talked about.  With fluorescents (both tube and CFL), the difference in spectrum (what they call "cool" or "warm") is due to differing phosphor coatings, which have different output and efficiencies.  In the case of the really crappy cheap ones, they put color filter coatings on the glass, which reduces the total light output.  The different LEDs have different materials, sizes, etc., all of which affect the light output vs. current draw.

This is what makes the optimization difficult, and why you need some sensible constraints (e.g., minimum power usage or voltage) to make it solvable.
00MrNotWrong00 (author) 4 years ago
OK I need someone better at math than me to have a look at these numbers.  I have "Kitemans"  10mm diffused LEDs at 0.550 lumens @ 0.060 watts.  If that's wrong I need someone to get me reference material with the correct specs,  Those are the numbers I came up with.

If those numbers are right than it would take 6watts to get to 55 lumens and That's not as good as the 5watt 420 lumen bulb I post a link to earlier.

So someone please check my numbers and let me know if I'm right or wrong ?
00MrNotWrong00 (author) 4 years ago
I think we are missing a few key words?  The first key word is "Useful"  a few LEDs taped to a battery isn't exactly useful.  As shown in that instructable ,  I like the concept and I'd love to see someone do something constructive with it,  but as is "Not Useful" .

The second key words would be "lamp" , Like something you might use to read in the dark,  or light a room.

As for what works best I've found that found that in fluorescent lights of the same wattage can differ drastically in brightness if they are labeled  "cool white" or "warm white"  and LEDs also very in brightness,  The LEDs used to replace mag-light flashlight bulbs are much brighter than the same rated LEDs you can get from radio shack. 

So if no one can put something useful together then I guess the LEDs taped to batteries will be the winner,  But I'd really like to see someone build something useful.
From what I have seen of "high intensity LEDs" (I have a tiny lamp that has ONE in it, and uses a coin cell battery, and it is nearly bright enough to ready by,  two or three would suffice for a small reading lamp, if directed.  10-30 for a wider spread of light.  The specs for these LEDs normally come with purchase or a spec sheet can be gotten fairly easily.  
00MrNotWrong00 (author)  Goodhart4 years ago
I was thinking about something like this  www.extremestyling.com/2006-FORD-F150-Use-5560-Watts/M/B000LL12GO.htm
But as you can see It's above 50watts ,  and that's all I have to work with.
Yes, but if you don't need quite that intensity, one could construct one that uses a bit less power.....if you DO need that intensity, then, as Kelsey mentions, you become limited by present day physics and options at hand.
 
Since you moved your reply up to the top level (not unreasonable), here's my followup moved as well.

A few LEDs and a battery are a flashlight.  That definitely counts as a lamp, as I've read by flashlight many times, and those silly clip-on "book lights" are just flashlights with an overpriced handle :-)

Your third paragraph is right on the money.  Those differences are due to exactly the underlying physics I talked about.  With fluorescents (both tube and CFL), the difference in spectrum (what they call "cool" or "warm") is due to differing phosphor coatings, which have different output and efficiencies.  In the case of the really crappy cheap ones, they put color filter coatings on the glass, which reduces the total light output.  The different LEDs have different materials, sizes, etc., all of which affect the light output vs. current draw.

This is what makes the optimization difficult, and why you need some sensible constraints (e.g., minimum power usage or voltage) to make it solvable.
NachoMahma4 years ago
> I've replaced the wire in my lamp with a thinner wire to reduce power loss from the resistance of the wire.
.  Using the same material, thinner wire has a higher resistance and more voltage drop. You need the largest wire that is practical. 16ga zip cord works great for anything under 10A.
.
.
.  A high-efficiency power supply and high-efficiency LEDs (number determined by how much light you want) are probably your best bet.
.
.
.  You can buy LED replacements for standard 110/220VAC E26/E27 base incandescents, but they are expensive (~$40).
Kiteman4 years ago
00MrNotWrong00 (author)  Kiteman4 years ago
I like what you have there,  but I'm looking for a useful amount of light,  like what a lamp would output.  I guess I should specify a little more,  I'd like to see in the range of 5 to 10 watts.  I'm stuck at 13w,  it's a very useful amount of light,  but there has to be a way to get that much for like 5watts. 

and I'd like to see it in lamp format.  I've replaced the wire in my lamp with a thinner wire to reduce power loss from the resistance of the wire. 

Also the cash prize will be small,  like 100 bucks,  so try to do it under that price range so you can at least recap your expenses. 
But that isn't what you said.  You were very clear "maximize the number of lumens while minimizing the number of watts."  Kiteman has demonstrated exactly that solution.

If you want to set a minimum input power, or a minimum light output, then you should do that in the text of your "contest."  Changing conditions after someone presents a solution is not considered very appropriate.