Pulsing a luxeon

Amonra

Enlightened
Joined
Jan 18, 2005
Messages
779
Location
Malta
Has anyone ever tried pulsing a luxeon ( fast enough so that it is not visible to the naked eye ) ?

Im not an expert but wouldn't pulsing increase runtime?
I guess the LED would also run at a cooler temp, or even allow the increase of current without damaging the led thus increasing the brightness.

Also since small batteries do not allow large currents to pass through for an extended period of time due to internal resistance wouldn't pulsing rid us from that problem thus allowing the use of smaller batteries for larger currents?

i.e. a smaller, cooler/brighter, longer running flashlight.

Are my assumptions right or did i get something wrong ?
 

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
No, pulsing doesn't increase runtime beyond the time it is on. Add up all the pulses, and the runtime is the same.

The Luxeon itself is actually more efficient at lower currents. It can be up to 253% more efficient at lower currents vs. pulsing. Even for pulses as short as 1/1,000,000 of a second.

There is an old tale about pulsing being brighter to the eye when you pulse. It is one of those bogus tales. Sam talks a bit about it in Sam's FAQ, use google.

Use the search function here and enter PWM for the keyword. I recently did a bunch of testing on PWM dimming vs. current dimming.

Also, the higher current pulses cause significantly more loss in the battery.
 

OddOne

Enlightened
Joined
Mar 5, 2003
Messages
495
I built a PWM-drive circuit to fire pulses through some 1-watt Luxeons, and got an amp at a time into one with a duty cycle of 5%. VERY overdriven, so there's no guarantee whatsoever that a single pulse wouldn't turn the chip bonding wire into a fuse. Also, they do get hot, FAST, even when pulsed at a very low duty cycle.

They make good camera strobes with single pulses. /ubbthreads/images/graemlins/grin.gif

oO
 

Amonra

Enlightened
Joined
Jan 18, 2005
Messages
779
Location
Malta
ok forget the higher current bit.

lets say we pulse a 1W luxeon according to specs. i.e. approx. 350mA @ 3.5V at a frequency of 100Hz ( i dont know what would be best frequency for pulsing not to be visible for a LED having much faster lighting times than an incandescent but lets just say 100Hz )
wouldn't the 'actual' total running time of the LED be much shorter than the 'percieved' total running time?

for example:
pulsed - 1 second 'percieved' runtime = 0.5 second 'actual' runtime ( since the led is spending half the time 'on' and half the time 'off' )
continuous - 1 second 'precieved' runtime = 1 second 'actual' runtime

therefore i guess that since it is spending half the time 'off' it would run cooler right?

Also as pionted out in some threads the initial LED output i.e. first milliseconds of operation, is slightly higher than the rest. Now when pulsing wouldnt the initial LED output be repeated for every pulse thus resulting in a slightly higher output ?

I have made above assumptions through simple reasoning with no 'real' technical background therefore i am asking you if they are anywhere close to reality
 

IsaacHayes

Flashlight Enthusiast
Joined
Jan 30, 2003
Messages
5,876
Location
Missouri
PWM will however change the color tint I belive over steady on. You can somewhat control that...
 

raggie33

*the raggedier*
Joined
Aug 11, 2003
Messages
13,453
i belive my fav led lights the liomheart uses this technolgy and i love the results
 

greg_in_canada

Flashlight Enthusiast
Joined
Jun 7, 2004
Messages
1,146
Location
Saskatoon SK Canada
Your eye averages out the pulse of light. So the only way pulsing
will result in longer run time is if the efficiency of the LED is higher
at the pulse current than at the average current.

For example 100mA pulses 50% of the time equals an average of
50mA.

The PWM circuit will have some losses so you could lose some run
time from that too.

Greg
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
Your eye integrates the pulses into a relatively steady brightness equivalent to a non-pulsed light source of equal average brightness.

You make no mention of the duty cycle. Let's assume 50%. You are still better off efficiency wise to run at an average current of 175mA rather than 50% duty cycle of 350mA. LEDs are more efficient at lower currents, which would far offset the slight initial brightness advantage due to heating (which only lasts for about 20ms.

The efficiency improvment at lower currents is both due to LEDs being more efficient at lower current densities, and that the Vf of the LED drops with lower currents. Since power = V * I, 1/2 the current will equal less than 1/2 the total power since V drops as well. And since reducing current to an LED by 1/2 typically results in 60% light output (as opposed to 50% that you would expect for 1/2 current), running constant current at 1/2 the current is better than running PWM at 50% duty cycle. And due to less power dissipation, the LED will run cooler.
 

raggie33

*the raggedier*
Joined
Aug 11, 2003
Messages
13,453
man ya are smart i have no idea what ya mean. but i would guess that since some of the time the led receves no power it will be able to run longer per battery
 

beezaur

Flashlight Enthusiast
Joined
Apr 15, 2003
Messages
1,234
"They make good camera strobes with single pulses. . . ."

How short can you make a pulse and still have enough light?

Scott
 

Amonra

Enlightened
Joined
Jan 18, 2005
Messages
779
Location
Malta
My point is not about the efficiency of the LED since that remains basically the same.
My piont as Raggie stated is that the LED is actually running for only half the time therefore theoretically it should run twice as long on the same battery as in continuous mode, though that may not be true in reality due to various losses i would expect at least 35% - 40% gain in runtime.

in reality you would only be delivering a total 175mA in bursts of 350mA for half the time to the led instead of 350mA continuously but the eye will percieve it as if it would be running continuously at 350mA Right?
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
But you could just run it at 175mA and get the same increase in runtime, but have it be brighter and run cooler.
 

Beowulf

Newly Enlightened
Joined
Dec 17, 2004
Messages
5
[ QUOTE ]
Amonra said:
in reality you would only be delivering a total 175mA in bursts of 350mA for half the time to the led instead of 350mA continuously but the eye will percieve it as if it would be running continuously at 350mA Right?

[/ QUOTE ]

Wrong. If you only have the light on for 50% of the time, you get 50% of the output. Give that same LED 175ma continous, and you have more than 50%, because the LED is more efficient at lower current.
 

Amonra

Enlightened
Joined
Jan 18, 2005
Messages
779
Location
Malta
Has anyone ever tried this before or is this all just theoretical.

I have not tried it since i do not have experience in electronics and basically i dont know how to build the required circuit.
 

beezaur

Flashlight Enthusiast
Joined
Apr 15, 2003
Messages
1,234
Photons work this way, do they not? Twirl one at the end of a string on a low setting and you will see tha it blinks.

Anyway, I think you have to think of it as radiated visible energy per unit of time. If you are radiating 350 mW of actual visible light ( watts through the ether, not watts from the battery), it shouldn't matter whether that is accomplished in pulses or continuous. Your eye will "smear" the pulses in time.

But there should be a "rise time" or whatever, so your pulsing light output will not be a square wave but a rounded one. You should lose energy with each startup and shutdown of the diolde -- right? That would make pulsing less efficient, (average) watt for watt.

Scott
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
Yes, Newbie has a thread where he's tested the efficiency of PWM vs. constant current, and does brightness measurements for both.

The rise time is insignificant compared to the fact that at higher currents, the voltage drop (Vf) of the LED increases. Let's say that at 350mA, the Vf is 3.3V. During each pulse, the Luxeon will dissipate 1.16W of power. With 50% duty cycle, average power dissipated is 0.58W.

Now, run the same Luxeon with constant current at 175mA. Vf will be around 3.1V. Average power dissipation will be 0.54W. Not only that, but instead of generating 50% of the lumens from running the Luxeon at 350mA constantly, you are in fact generating about 60% of the lumens, or about 20% more.

So, you will get the exact same runtime (since average current draw is the same), but the LED will be dissipating slightly less average power, but will be producing more light.
 

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
[ QUOTE ]
greg_in_canada said:
Your eye averages out the pulse of light. So the only way pulsing
will result in longer run time is if the efficiency of the LED is higher
at the pulse current than at the average current.

For example 100mA pulses 50% of the time equals an average of
50mA.

The PWM circuit will have some losses so you could lose some run
time from that too.

Greg

[/ QUOTE ]


Sorry, this is not correct Greg.



Pulsing the emitter causes very significant efficiency loss in the emitter.

Details on this are posted in my thread:

Constant Current up to 253% more efficient vs. PWM


Mike Krames, one of the managers at the LumiLEDs Advanced Research area also confirmed this.

PWM dimming is very inefficient for the LED, and there is also a loss of efficiency in the battrey, especially at high pulse currents.
 

idleprocess

Flashaholic
Joined
Feb 29, 2004
Messages
7,197
Location
decamped
PWM is popular because it's relatively easy to implement, lets you setup a fixed current supply, and makes for relatively easy dimming with perceived "infinite adjustability" without the waste of a pot inline with the LED.
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
Its just as easy to build a constant current regulator that's infinitely adjustable, without a pot inline with the Luxeon.
 
Top