LED Apparent Brightness?

[MikeMl]

I am planning to use a red LED as an On/Off indicator. The forward current will be half-wave rectified 60Hz Sine with the following values:

Peak=6.2mA
Ave= 1.97mA
rms= 3.1mA

What dc current would the same LED have to be driven with to get the same apparent brightness?

 

[alec_t]

Hmmm. I suspect that would depend on image retention, which could vary somewhat from one person to another?
You could easily run an experiment with two similar LEDs side by side, one driven by that AC and the other by an adjustable DC source. Repeat with the LEDs swapped over and average the two results to rule out LED mis-match.
If I had to guess, I'd go for the RMS value.

 

[crutschow]

........................
If I had to guess, I'd go for the RMS value.

Not a good guess. RMS current is only accurate with a resistive type load, where the load voltage is proportional to current . For a near constant voltage load, like an LED, average current is more accurate.

 

[audioguru]

A LED running from half-wave rectified 60Hz will flicker and drive you crazy. Each flicker lasts for less than 8ms and I learned from somewhere that our vision sees a light duration of less than 30ms as being dimmed which is why PWM is used as an efficient light dimmer.

 

[Sunnysky]

Apparent brightness depends on photopic and scoptic eye response and light or dark ambient. As eyes tend to store peak values at low rep rates, the peak is more noticeable than the average and RMS, so I would start with the RMS value.

The measure of comparison is ambiguity of threshold detection with ambient. i.e. what scaled down levels are noticeable if Off or ON. When DC matches AC perception of detection threshold, they are equivalent.

Peripheral motion sensitvity is ever higher , so a moving eye or a moving indicator ( Old Cadilac tail lights) can be seen to flicker even at 500Hz rates vs 100Hz for stationary.

Obviously 50 or 60 Hz flicker on LED's will be noticeable in both stationary and moving eye, but adequate for cheap indicators.

 

[tvtech]

Apparent brightness depends on photopic and scoptic eye response and light or dark ambient. As eyes tend to store peak values at low rep rates, the peak is more noticeable than the average and RMS, so I would start with the RMS value.

The measure of comparison is ambiguity of threshold detection with ambient. i.e. what scaled down levels are noticeable if Off or ON. When DC matches AC perception of detection threshold, they are equivalent.

Peripheral motion sensitvity is ever higher , so a moving eye or a moving indicator ( Old Cadilac tail lights) can be seen to flicker even at 500Hz rates vs 100Hz for stationary.

Obviously 50 or 60 Hz flicker on LED's will be noticeable in both stationary and moving eye, but adequate for cheap indicators.

Interesting observation. I am sure I can learn from this. Thanks for sharing.
Good higher language used. You seem glued upI mean clued up .

Faulty mouse again

 

[KeepItSimpleStupid]

Dunno if it's useful: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3057437/ but this shows the eye's response to red and blue light.

Going the other way, I had to do some intensity measurements of an mercury lamp which had a wierd curve, I did use an RMS meter and not an average responding one. I would think the number of photons would be proportional to RMS current. Human perceived intensity could be something else entirely.

 

[audioguru]

Dunno if it's useful: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3057437/ but this shows the eye's response to red and blue light.

Thanks for posting it.
It is amazing that when a bright light shines in ONE eye then the pupils of BOTH eyes contract. The reaction time of the pupils were very slow at 1 second to 4 seconds.

 

[k7elp60]

I am planning to use a red LED as an On/Off indicator. The forward current will be half-wave rectified 60Hz Sine with the following values:

Peak=6.2mA
Ave= 1.97mA
rms= 3.1mA

What dc current would the same LED have to be driven with to get the same apparent brightness?

Mike I don't know, but I tried to determine about the same thing with LED's. I have done some experimenting with a Lux(Lumens) meter that I have. It is a sensitivity from about 450 nanometers to 650 nanometers. The biggest problem has been the ambient lighting causing incorrect readings.

 

[MikeMl]

Mike I don't know, but I tried to determine about the same thing with LED's. I have done some experimenting with a Lux(Lumens) meter that I have. It is a sensitivity from about 450 nanometers to 650 nanometers. The biggest problem has been the ambient lighting causing incorrect readings.

Does your meter have a "weighting" setting that matches human eye sensitivity; not just power?

 

[k7elp60]

Does your meter have a "weighting" setting that matches human eye sensitivity; not just power?

No it does not.

 

[ronsimpson]

Ave= 1.97mA
rms= 3.1mA

The eye and the ear are not linear. Doubling the power makes a noticeable difference but you don't think 2X.
In the audio world 1db is 'just perceivable'.
With LEDs, if you get the power with in 20%, you probably can't see the difference.

Just noticeable difference = jnd

 

[audioguru]

I agree that our vision and hearing do not have linear sensitivity. I was going to say we can see in sunlight and in less than moonlight but the iris (pupil) in our eyes cause even more range.
The iris has been shown to adjust to brightness much slower than these 60Hz pulses.

 

[MrAl]

Hi,

I am not sure if measuring the response of the human eye will be something to consider here. It depends on the ambient lighting too. Most indicators are not run in total darkness but yes that is an option.

Also, RMS mainly applies for types of powers that are of the same energy type as is efficiency. If the power sources are different in nature then it depends on what it really is.
For example, to heat up a cup of coffee if it takes 500 watts for 60 seconds then it takes 1000 watts for 30 seconds, etc. In this case it is the average power which is found from the duty cycle.
So if you wanted to use an IR LED that was 1000 watts (a lot of LED's) then you would turn them on for 30 seconds too.
But a simple test is always better, and Alec outlined what looks like a good test for this. It's easy to pulse the LED with a micro controller if it is a lower power LED.

I did a pulse vs continuous current test a while back (long while now). If found that at about 1/2 the rated current level the LED did not change much in brightness but the color did shift. That's because many white LED's have a shift in color tint with current level, and this is more controlled by the peak current level. Red LED's should not have this problem though.

Another smaller point to consider is that the light output of the LED is not linear with current. It's almost linear but there is after all a curve for current vs light output. Luckily at lower currents it is more linear than at higher currents and also higher efficacy.

 

[KeepItSimpleStupid]

I think what Mr. Al is saying is valid. Way back when, when I set up an EDAX (Energy Dispersive Analysis of X-Rays) system in a dark room, a yellow monitor was really hard on the eyes. Yellow was good for florescent lighting and green was good for dark rooms. I have an eye sensitivity table somewhere.

Nowdays, no one seems to care about ergonomics. I'm getting tired of computer Yes/No questions having to be answered with OK/Cancel.

 

[audioguru]

A few minutes ago I put a blue LED on my soldering iron. Its transformer has a 6VAC winding for an incandescent light bulb that is not available anymore. The blue LED lights with 3.2V at 15mA.
I used a 270 ohm series resistor and a 1N4148 diode anti-parallel to the LED.

It is bright enough to be an indicator in the daytime with plenty of sunlight in the room.

It flickers (60Hz) only when I move my eyes.

Last night I saw something outside, "What the heck is that, I said to myself??"
It was a man riding a 3-wheel long and low pedal bicycle that was covered with blue LEDs. COOL.
I guess the LEDs were powered by DC from a battery because they did not flicker.

 

[audioguru]

A few minutes ago I put a blue LED on my soldering iron. Its transformer has a 6VAC winding for an incandescent light bulb that is not available anymore. The blue LED lights with 3.2V at 15mA.
I used a 270 ohm series resistor and a 1N4148 diode anti-parallel to the LED.

It is bright enough to be an indicator in the daytime with plenty of sunlight in the room.

It flickers (60Hz) only when I move my eyes.

Last night I saw something outside, "What the heck is that, I said to myself??"
It was a man riding a 3-wheel long and low pedal bicycle that was covered with blue LEDs. COOL.
I guess the LEDs were powered by DC from a battery because they did not flicker.

 

[k7elp60]

No it does not.

Today I remembered I started to build a gadget that would do the measuring. I found a photo cell or a photo diode and I mounted it in black project box so it was
looking up near the top. I forward biased the detector and had the Vf available to some connections on the outside of the box. I also had drilled a hole in the top of the box for T1 size LED's just so the dome would enter the hole and the flange would keep the LED from going all the way in. That is as far as I got. I never finished the project, but the idea was to keep the spacing between the LED and the detector constant, and vary the forward current of the LED and see if I could measure different light intensities because of LED current, color and different mcd values. Mike I wonder if this type of test fixture might be of help to you?
Ned

 

[MrAl]

Hi,

That sounds pretty good really and worth building, but often i prefer a human eye to a photo detector of any sort for problems that may involve the oddities of the human eye response. But even without that requirement it's not too hard to build a comparative light detector which allows quick comparison of light sources using the human eye.

The construction is quite simple, and i'll explain the basic construction. You can do it better if you like, but this is the least that works good.

First, find a white candle at the store, probably the kind used for table lighting like around 1 inch diameter. They are usually something like 10 inches long with a skinny end and one fat end. Somewhere along the length the diameter is fairly constant.

Find the place where the diameter is fairly constant over about 1 inch of length, then make a clean cut with a sharp razor blade or maybe hobby saw. That gives you two pieces. The flats of the two pieces should be parallel.

Next, cut about 1/2 inch off of each piece off the ends that were just cut away from each other. That gives you two round pieces of white candle 1/2 inch long each. Discard the rest.

Next, cut a piece of aluminum foil about 3 inches long and about 1.5 inches wide. It's a rectangle shape. Then, fold it over so that the two most shiny surfaces are facing out and the two dull sides in. This forms a piece about 1.5 inches square.

Next, apply some glue to the two dull sides and press the foil together so that it forms one piece two layers thick, then glue one candle piece to each shiny side, wait for the glue to dry.

Congrats! You now have a comparative light detector

After the glue has dried thoroughly, hold the device up in front of you so that you are looking at it from the side. You'll see one candle piece on the left and one on the right.
Shine one light source directly onto the round surface of the left side, and shine the second light source to compare on the round surface of the right side, while still looking at it from the side. Be sure that there is enough light spread to cover the entire round surface of each piece.

What you will see if both sources are the same is the light on the left piece will look equal to the light on the right piece, from the side. If one source is dimmer then you'll see a slight difference.

To use this with the pulsed vs constant current test just shine one on the left and one on the right, and adjust the constant current until the two intensities look equal. When they look equal, you have the correct current needed to match the pulsed light source.

The key to this simple device is the human eye and brain find it hard to remember actual light intensities, but are incredibly good at detecting light differences when there is something in the field of view to compare it to. So even a small difference will be detectable, and what else is key is it is after all the human eye that is doing the comparison not an abstract instrument.

 

[ronsimpson]

When measuring light from LEDs there are two different types of measurements.
1)Total light output. (in all directions) I can't remember the name of the instrument. It is round and the LED goes inside. I have one. Have made several. The shape of the LED lens makes no difference. (viewing angle)
2)Light that hits a small spot in front of the LED. This is very dependent of the lens in the LED. Very dependent on the shape of the LED. The LED might be slightly out of alignment. It is hard to get repeatable readings.

Because you are interested in what the eye sees, the second measurement is probably what you want to measure.
**broken link removed**