At what frequency does the human eye notice flicker?

I think everyone is missing my point. Go back to the opener of this thread and read the first post again.

As others have stated, PWM is less efficient than producing the same (average) lumens by using a lower constant-current. So I’m not sure why you’d want to try to mimic a constant-current by some complex alternating PWM algorithm using multiple LEDs. All you’ll end up with is a more expensive and less efficient light, that will have lower overall output.

walkintothelight perhaps you are right. but if that were the case why have an array of LED’s at all? And within that array you say that there is no other manner, method, theory, strategy, idea, concept, or thought that could make better the array and how energy efficient it is or how much light is puts out? Have we actually become the GODS of LED efficiency?

Im not trying to be a smarta$$ for real… I appreciate everyone’s comments on this subject. I encourage it. I want to stimulate your thought.

Still nobody has even attempted to construct an equation as mentioned in the very first post. The model is real, the values are relevant, and the product I promise has significance.

“Complications” are not incorporated into anything without justification. I mean who would have thought that a computer chip would power our flashlights of today, twenty years ago…………

I need more input…….

You mean why use multiple LEDs? Simply: it allows for higher output, for very obvious reasons. Also, it could result in greater efficiency at a given (lower) output, because you’re driving each LED (at a constant current) less hard than you would a single-emitter light.

We already pointed out why your idea would not work in practice. But if you come up with another theoretical idea, by all means, state it, and maybe it really would work. Just be aware, like with perpetual energy machines, some may sound like a good idea on paper, but in reality can not work.

Nobody came up with an equation, because your idea doesn’t make sense.

There is a theoretical maximum efficiency you can achieve with light output. A 100% efficient green light source (most sensitive to our eyes, thus having the highest lumens), is somewhere around 650 lumens/watt. (I forget the exact number, or where I read it, but is was around that level.)

That has nothing to do with PWM or anything like that. It’s just basic physics.

If you come up with an idea, the first thing you have to ask yourself is if it violates known physics. If it does, you have to rewrite known physics, before your idea will be taken seriously.

That said, if I was going to come up with an idea to use PWM to increase efficiency of perceived light, I might do it something like a camera flash. A brief pulse of light would allow us to see quite a bit, even if the total energy used is fairly small (certainly much smaller than a constant source of light).

Perhaps there’s a point where you can increase the flash rate to a point where the human eye can no longer perceive it as flashing, while still using less energy than a constant light source would use.

It wouldn’t violate any physics, but might trick us into seeing more while using less power for the light source.

Or, it might be total B.S. I have no idea. I tend to think it’s B.S., though, as our retinas still have to turn light energy into electrical energy to send to our brains. Less light energy probably means less electrical signals triggered.

walkintothelight, Thank you for the insight. I will have re-read my own questions because the flash is exactly how this conversation was started I thought. Or that is at least the concept I was trying to convey. let me check

Well, just keep in mind that “500 lumens using PWM” looks an awful lot like “500 lumens using constant current”. So, at least at the PWM frequencies that are typically used, our brains are not fooled into seeing anything more.

I tend to think that by the time PWM is showing us more (by slowing the rate), the flickering would be getting very annoying. Kind of like a strobe in a night club.

As noted earlier, flicker fusion can’t be specified generally for use in an equation — it’s specific to any one individual/health/age
https://www.google.com/search?q=flicker+fusion+aging+variation

And PWM isn’t the theoretically pure square waves — even if that’s the intent, there’s some noise up and down. They can be sawtooth pulses, etc.

And each phosphor used in the emitter has its own persistence.

Specify those and build one, and you might be able to craft an equation — after empirically determining the values that come out of the circuit, and the phosphors, and the flicker rate perceived by the subject.

This is why cheap drivers use visible (low-rate) PWM, and why better drivers, if they use PWM, use a much higher frequency, to stay well out of the, er, gray zone where some people will see it.

Building a light with say three cheap low-rate-PWM drivers, each somehow coordinated with the others to keep their notches from lining up, might theoretically work. But egad.

Now remember, I’m some guy on the Internets and know nothing about this stuff, so you may find better advice from someone else.

I think I have read somewhere that in this situation the light with the PWM will look brighter.

Source?

Because, I’ve never seen PWM described that way. I think if it really did fool our eyes to give increased sensitivity, it would be much more popular and marketed heavily. And, efficiency ratings (on lower levels where PWM is used), would be much better than they are when tested. Granted, testing equipment wouldn’t be fooled, but I’ve never noticed that a 100 lumen PWM light lasts any longer than a 100 lumen constant-current light (it’s usually the opposite).

Okay I really like the conversation this has brought about. Everyone seems to have a valid point in some regard or another.

My point is this. If a flash of light lasts a known amount of time and is quantifiable in terms of brightness, temperature, current draw and is repeatable at lets say 500Mhz frequency, does the millisecond timing provide enough lapse for the LED to recover to cycle EXACTLY the same. YES because we haven’t put a limit on the millisecond recovery value.

Again, follow me on this: We all agree that the luminous output of an LED begins to diminish within milliseconds after being energized and reaching its apex output. That is the holy grail of LED performance. That single measure of time when an LED gives all it can. It is at that point the most efficient.

The performance curve.

Now explain to me that if an LED is on and off an equal amount of time during lets say a 1 minute frame at a rate 30 times per second that after 1 minute an equal amount of energy has been consumed from this arrangement as opposed to an LED that was provided constant current during the same. It may not use exactly half as much energy but indeed significantly less.

But this isnt about energy, this is about achieving maximum luminous output.

energy efficiency is a bonus.

Btw, constant current indeed operates in pulses. All electrical current does. Hertz….its how we measure it.

I’m only replying to this part.

WHAT?

No.

So tell us what hertz your AC batteries operate at.

Again, I don’t think you’re going to see that in a real experiment. In reality, PWM has demonstrated itself to be less efficient than constant current. I don’t think there’s a special frequency that changes that. You haven’t given any plausible explanation why.

No, direct current does not operate in pulses. Alternating current is probably what you’re thinking of.

Earlier in your OP it was about minimal temperature and maximum efficiency.
But now that doesn’t matter because it’s about max luminous flux. Right?

And suddenly your flashlights operate on alternating current.
Are you trolling?

Did you even read that PWM is less efficient than Constant current in the quote I posted by Wight?

It doesn’t sound like he’s trolling. He’s putting in too much work for that.

I think he sounds more like someone trying to come up with a perpetual motion/energy design. That is, there’s no way to make it work, but some people won’t let that dissuade them.

I think(?) I understand what dehc111 is getting at.

Hypotheticals:

Yes we know that an LED is not at it’s most efficient when being pulsed at 6amps 50% cycle. 3amp CC is more efficient.

But let’s say the peak efficiency is at 2amps.

We could pulse 2amps at 50% cycle (net draw of 1amp) and in a perfect world it would be more efficient than running a constant current at 1amp.

The problem being that the pulse are not instant on/offs they are ramped with the LED not being as efficient on those ramps.

Am I getting both sides to an extent?

I agree that I think this is a fundamentally flawed venture.

You could easily calculate that if you want 6000 lumens, and your emitter does 800 lumens at it’s most efficient point, how many emitters and how much power you would need. No problem. But you still want to go CC, not pulsed.

The problem is that if LEDs have an efficiency peak, it’s at an extremely low current (at least for Cree). Take any led from http://pct.cree.com/dt/index.html , I’m pretty sure that for all tested currents the efficiency lm/W decreases as the current increases. Also the efficiency lm/A (Current), which would be useful if you use a linear driver, seems to decrease monotonically as the current increases. I believe that most if not all leds follow the same pattern.

Of no practical use, but … interesting

Also a very misleading title. All they’re doing is using heat from the environment to power the LED. It’s not greater than 100% efficiency.