Luxeon efficient driver

jrn

Newly Enlightened
Joined
Dec 23, 2004
Messages
5
Well, my skills in electronic are very low. I'm thinking about a way to drive a luxeon 1W very efficiently. I want to use a buck converter and a PWM circuit. I would like to have some advice from knowledgeable people on this forum.
For a 6V battery, with a simple 8-Ohm resistor the luxeon draws 350 mAmp. If the luxeon is ON for 5hours the total draws from the battery is 1750 mAmp-hour.

If I can have a buck converter that draws 200 mAmp, this reduces the total consumption to 1000 mAmp.

With a PWM circuit and a duty cycle of 10%, the total current would be 100 mAmp. Less than 10% of the first design and the human eye will not see any difference. This way I think the battery will last longer.

This can be done for a 12-volt battery too. The current draws from the battery can be 100 mAmp with the buck converter, and with a PWM the current drawn from the battery would be 50 mAmp-hour.

Once again I'm just thinking efficiency, I don't know if it is possible to build such circuit. Anyways theses values would be higher than because of power lost in the components. I don't know if the 10% duty cycle is not too low.

Any comments about this idea.
 

greg_in_canada

Flashlight Enthusiast
Joined
Jun 7, 2004
Messages
1,146
Location
Saskatoon SK Canada
Your eyes average out the pulses so if it is only on
10% of the time it will appear as bright as an LED
driven at 10% of the current (ignoring the effects
of LED efficiency differences at those 2 currents).

No free lunch I'm afraid.

Greg
 

jrn

Newly Enlightened
Joined
Dec 23, 2004
Messages
5
I dont get it.
if the PWM is 1kHz and Duty cycle 10%, so that means the led is ON 0.001 sec and OFF 0.009 sec. The human eye will not see the diffence since I think (I'm not sure about the exact value) the eye "keeps" the image for 0.12 sec.

Any way if 10% is too low, we can try 30% or 50%, the circuit would still be more efficient than drinving with constant current.

I don't want free lunch, just a cheap one.

JRN
 

teststrips

Enlightened
Joined
Dec 18, 2004
Messages
317
Location
USA, Pennsylvania
the photon freedom uses pwm to "dim" itself, it's very effective too. Good thought, but i think your theory on eyes "keeping" an image is unsupported
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
[ QUOTE ]
jrn said:
I dont get it.
if the PWM is 1kHz and Duty cycle 10%, so that means the led is ON 0.001 sec and OFF 0.009 sec. The human eye will not see the diffence since I think (I'm not sure about the exact value) the eye "keeps" the image for 0.12 sec.

Any way if 10% is too low, we can try 30% or 50%, the circuit would still be more efficient than drinving with constant current.

I don't want free lunch, just a cheap one.

JRN

[/ QUOTE ]

Sorry, this is incorrect, and has been discussed and proven sevaral times here.

Pure PWM dimming (i.e. not in a DC-DC converter) is never more efficient than constant current to achieve a specific light level. Your eyes integrate the individual pulses into continuous light, so 10% duty cycle = 1/10 as bright.

Some posts to explain:
http://www.candlepowerforums.com/ubbthreads/showthreaded.php?Cat=&Number=898927&page=&view=&sb=5&o=&vc=1

http://www.candlepowerforums.com/ubbthreads/showflat.php?Cat=&Number=836487&page=&view=&sb=5&o=&vc=1

http://www.candlepowerforums.com/ubbthreads/showflat.php?Cat=&Number=808893&page=&view=&sb=5&o=&vc=1

http://www.candlepowerforums.com/ubbthreads/showflat.php?Cat=&Number=657009&page=&view=&sb=5&o=&vc=1
 

HarryN

Flashlight Enthusiast
Joined
Jan 22, 2004
Messages
3,977
Location
Pleasanton (Bay Area), CA, USA
I will save you some time - visit the cpf r georges80 - he has a nice web site listed in his profile - and very nice boards already done for you.

Alternative include boards from the sandwiche shoppe, and leddynamics. You can pick from a host of boards from 3 different suppliers with this info, and all of them are pretty good, and are used by many.
 

CM

Flashlight Enthusiast
Joined
Sep 11, 2002
Messages
3,454
Location
Mesa, AZ
As said above, no free lunch. Best bet is to go buy a converter board from dat2zip. He's got buck, boost, buck/boost ready to go. I've tried building my own but the $$ saved was not worth it and you're better off buying a drop-in solution unless you need large quantities where rolling your own would make sense.
 

jrn

Newly Enlightened
Joined
Dec 23, 2004
Messages
5
OK guys, thank you for precison on the duty cycle and the PWM.

But what about the power conversion?

Can the buck-converter convert 1.2W from 12V*100mAmp to 3.6Volt*350mAmp?
 

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
No

Take the first two numbers and muliply them together, and then do the same to the second two, you'll find you are trying to take out more power than you are putting in.

So you'll need more input power.

Most converters are either voltage converters or output current converters. LEDs really like to be fed by current control.
 

jtr1962

Flashaholic
Joined
Nov 22, 2003
Messages
7,505
Location
Flushing, NY
You'll probably have about 125 mA from 12V if the converter is 85% efficient. As for PWM dimming, the only advantage is no shift in color. Other than that, it's not a good idea. It's less efficient than just lowering the current. And if done at too low a frequency, it creates flicker. Even if it's done at a high enough frequency to avoid perceived flicker, you'll still have a "stroboscopic" effect with some spinning objects. People used to complain about fluorescents flickering on magnetic ballasts. It seems makers of LED products are shooting themselves in the foot by using PWM and/or running them off rectified AC with no filtering. Soon you'll be hearing people complain that LEDs flicker, and then soon after that it'll become a perceived "drawback" of LEDs over incandescents. Let's nip this in the bud by not using PWM except when color shift absolutely cannot be tolerated.
 

jrn

Newly Enlightened
Joined
Dec 23, 2004
Messages
5
OK guys, let's be more precise :

P_out = P_in*eff <=> P_in = P_out/eff

Output
V=3.6
I=350 mAmp
P_out = 3.6*0.35 = 1.26 Watt

Assuming Eff=85%
P_in = 1.26/0.85 = 1.48Watt

So jtr1962 is right. I need to draw 123 mAmp (1.48/12) (~125 mAmp) from the battery.

The simplest way to drive the Luxeon is to put a 24 Ohm resistor in serie with the battery. This way P_in=12*0.35= 4.2 Watt. (Even if the LED is not drive at constant current)

To resolve the constant current problem, I saw a web site that recommends to use LM317 and R1=3.6 Ohm and R2=0. This way LM317 acts like a constant curent driver, but we still have P_in=4.2 Watt

1.48Watt vs 4.2Watt. I think the power saving is quite interesting.

So if PWM is not the best way to drive the led efficiently what is the best way?
 

UncleFester

Flashaholic*,
Joined
Apr 28, 2004
Messages
1,271
Location
Desert Hlls,AZ
jrn
What you need is to use a georges80 CC5W It's constant current and it is designed to run off 12V in a car. If you email George, he might be able to tell you how to provide a dimming function.
This setup will have everything you ask: efficient drive and current limiting dimming control.
 
Top