Using a PWM to extend run time

I have been thinking, is it possible to have this tiny circuit in leds such that it turns the LED on 10% of the time on and 90% of the time off? And to do that without anyone noticing flickering, it would have to switch at a extremely high rate. 1 ms on time, 9 ms off, 1 ms on, 9 ms off and so on. Is it possible? Maybe not with a PWM but some of our talented circuit designers here can do something about this?

Nerd, that *is* PWM you are referring to. The commercial flashlights I know that already use this are the Eternalights. Not so sure about the PALs, but maybe them too.

Unless the cycling is buffered by a cap, the LEDs will flicker annoyingly. Or the frequency should also be matched well to human eye's persistence of vision threshold.

wouldn't it just appear to be 10% as bright?

doubleganger, yes - sort of. But because the human eye is a strange beast, if you PWM at perhaps 75% duty cycle or more, you'll notice almost no difference. If you have an Eternalight, you'll see what I mean - their first dim mode is PWMming at about 60-66%, if I remember correctly and it is scarcely dimmer than full constant power.

Of course, by the time you hit 10%, you'd most definitely notice!

Maybe is should be done with incandescents instead because they still glow for a while after the power is cut off....

There is already at least one company making PWM circuits for incandescent AC lighting to save energy.

And the company is...? (Does it happen to be a torch light company?

Originally posted by Nerd:
And the company is...?

Click to expand...

<font size="2" face="Verdana, Arial">Hmmm - take your pick from a google search. Looks like this is even older hat that I thought!
Their products are probably being marketed under various guises - most ppl won't dig what PWM means.

Using PWM to extend battery life works but, at a price - the human eye is very slow and just averages out the brightness; that is why movies and tv works - persistance of vision

If you run 50% duty cycle, you will get 50% of the brightness and 2x the operating life; also if you run slower than about 30hz (depends from person to person) - the human eye will see it at "flicker"

I had a customer that wanted me to make a LED brighter than it was designed to operate using PWM.

He told me to run the LED current at 10x more than the LED was designed to do at 10% duty cycle;
it just gave normal brightness

I once made a very cool 8x8 array of TriColor (R/G/2B) LEDs for a demo unit for a suitcase; and the array would color-cycle the "wave" of colors - using programmable PWM using a Intel 8751, and it was wonderfull - too bad the company didn't pay me for my time

Unless you are using the LED for some special application such as digital communication, there is no advantage to try and make it last longer than just changing the operating current

I do not know technically if it is a "PWM", but The Princeton Tec Aurora Headlight, 3 LED's pulses on and off very quickly, and is not noticable unless you are reading.

Say your LED flashes at 60 time per second. Each duty cycle would be 16 millisecs. Let's say we on the LED for 2 millisec and off for 14 millisec, there would be a 60 hz rate. Would that suffice for persistance of vision? Or would it just reduce brightness?

you can't get something for nothing. turning the LED off for whatever interval, even if the eye cannot perceive it as being off, will be perceived as a dimming effect. the human eye will average out the light perceived.

from what i've read PWM dimming is the only way to go if you want good color. from personal experience non-PWM dimming (via potentiometer) will cause the color of the LED to shift, at least for white LEDs

the phospher of a white led would act as a sort of beam capacitor, as it glows shortly after power to it is stopped.

a 75hz flicker shouldn't be all that bad, and if you are using a white, you can probably get away with even less than that.

in multiled arrays, it would be interesting to see if different blinking properties were assign to different leds would result in a smoother output (one is off while the other is on, or maybe overlapping a little)

streamlight uses pvm in their new tailcaps for the (hp and xt?) they say it can increase runtime by 200%

Streamlight uses pvm? What's that? Are you refering to the PWM? If it can really increase run time by 200%, it's pretty good already. Wonder why didn't they do it for the Ultra Stinger as well.

Surefire should do it for their M6. 500 lumens for 40 mins....bwahahhaa.... I feel so evil already...

Cycling power to LEDs to conserve battery life has been done for a long time in notebook computers and cell phones. The period is short enough to not be seen and the duty cycle is whatever gets the desired result.

This hasn't been desirable in LED flashlights because they weren't that bright and battery life already met consumer expectations. It may become more common with 5W LEDs as battery life gets uncomfortably short.

yes, i meant pwm, i was typing very fast and about to leave.

it can't be done with the us because the tailcap acts as a vent for the gas from the battery.

The original "classic" Eternalight had a perceptible flicker rate that got more objectionable as you lowered the power. The contemporary Eternlites have raised the rate sufficiently to eliminate or greatly reduce that effect.

Brightnorm

What if you pulsed each LED so that one was on while the other was off?

Well; they are flashing the LEDs to lower the average current - you could call it PWM

If they are running their Transistor (BiPolar or FET) at Class-C operation - there should be very little loss in the control circuit

Originally posted by INRETECH:
Unless you are using the LED for some special application such as digital communication, there is no advantage to try and make it last longer than just changing the operating current

Click to expand...

<font size="2" face="Verdana, Arial">Question about this. I picked up somewhere that lowering the current to the LED causes the color to not be as white since you're not exciting the white phosphor as much. I can verify that my own current limited light gets bluer as you lower the current. The point of PWM being that when it's on, it's all the way on and the light color and quality are the same all the way down.

This doesn't make it last any longer, but it makes sense that the output would be better this way doesn't it?

I haven't experimented with this yet, but it's something I've been meaning to play with.

Thanks,
James