what do I need to know about light emitting diodes?

rjo__

Howdy,

I'm headed in about 10 different directions at the moment with all of them converging on a central point:)
It's late and I should really wait til morning... but I just can help myself. Please bear with me.



Tonight's focus is LEDs.

I know that no two LEDs are exactly the same ... thank you Chip Gracey.

Here is an example issue...Let's say for argument, I have a global shutter camera capturing at 1000 fps. I know exactly when
the camera will be exposing the sensor... but the exposure lasts for only a tiny part of that one millisecond frame.

let's say that I am going to be using a series of different LEDs (possibly LED arrays) to expose successive frames.... sort of a poor man's spectroscope.
When we expose the frame, we want to be fairly certain that we have a determined total emission from the LED (or array) over a specified period.
Do I need to worry about how long it takes from the time that the target voltage is achieved to the time that the LED reaches some
stable state...or is this on the order of a nanosecond?

That's an example question. In real life that frame rate could go way up:)

The problem is that when you don't know... you sometime don't know what you don't know.

Any advice is deeply appreciated.

Where is the LED Bible or the best guide to research grade LEDs?

Thanks guys,

Rich

ErNa

I assume you are talking about colored leds, that is, there is no phosphor to convert UV to visible light. In this case the emission should be proportinals to the current. I have no idea why emission should not start much faster than a ms. But the spectrum of a LED is broad, so you might take into account laser diodes
however, the spectra of different colored leds should not overlap, so no need for laser

rjo__

thanks

kwinn

Not sure what the response time of modern led's is but the older optoisolators took about 50uS to respond. Not sure if that was due to the led or the phototransistor but you might want to check that out before you spend a lot of time, money, and effort on something that may not work as expected.

erco

ErNa wrote: »

I assume you are talking about colored leds...
however, the spectra of different colored leds should not overlap...

Please, please, please refrain, there is no need to use such terminology. They are called
"LEDs of color".

Rsadeika

Please, please, please refrain, there is no need to use such terminology.

Ha ha ha, I think, if their is some subtle joke about colored leds.

Ray

ErNa

ok, don't understand where the trick is, like in my blank induced morph.
Even google doesn't laugh

Heater.

It's a joke about Political Correctness and how some are trying to make many good old fashioned words and phrases taboo.
It's a scourge that will collapse civilized society.

RickB

I have also tried to find definitive data relating to turn on and turn off time for modern ultra-bright leds, both phosphor and non phosphor, but no luck.
The application is to replace the high priced nearly unobtainium flash tube in a yet to be modified General Radio 1531 Stroboscope.

ErNa

no political correct answer possible. But, can this happen: It's a scourge that will collapse civilized society. I fear no. But I also fear: It's a scourge that will collapse society. But back to the original question: what-do-i-need-to-know-about-light-emitting-diodes
One answer is: millions of light emitting diodes, arranged in an appropriate way and fired in time can influence people in an unbelievable way.

erco

Make sure to put your LEDs of color on the front of the main power bus.

jones

http://www.osram-os.com/Graphics/XPic5/00135349_0.pdf/High-Speed Switching of IR-LEDs.pdf

That's for IR LEDs. It looks like they were getting optical rise/fall times of the order of 10 ns with a pretty simple driver.

Phil Pilgrim (PhiPi)

You don't have to flash the LEDs one-at-a-time. This not only gives the LEDs more on time, but can multiply your apparent sensitivity. For example, for three-color RGB, flash RG, GB, and BR in sequence. Then,

R = RG + BR - GB
G = RG + GB - BR
B = GB + BR - RG

The same principle works as you add more LED colors to the mix.

BTW, the optical produce-grading biz is a harsh mistress. Been there, done that, never again! (Just in case that's what you have in mind. )

-Phil

Martin_H

I LOL-ed at Erco's joke and my wife and daughter asked what's so funny. I read it to them and they burst out laughing too.

Peter Jakacki

LEDs are diodes, and once they are passing current which is immediately, then they start producing light instantly, at least from a practical viewpoint. Remember too that IR LEDs are typically modulated at 38kHz or higher so they have no problem there with the speed. Other LEDS are using in LED laser printers, or used to be, and that needs to be modulated at a very high rate too.

However, to overcome the LED capacitance to help it switch faster it may pay to have the LEDs barely conducting in the "off" state.

BTW, we are talking about non-phosphor LEDs, since you need RGB and not white which is a phosphor type normally excited by a blue LED I think.

RickB

BTW, we are talking about non-phosphor LEDs, since you need RGB and not white which is a phosphor type normally excited by a blue LED I think.

Hmmm... I was under the impression that white leds were based on high power uv leds and phosphor. Wrong?

Heater.

You are not wrong.

So now I wonder how quickly does that phosphor respond?

Phil Pilgrim (PhiPi)

heater wrote:

So now I wonder how quickly does that phosphor respond?

Here's a graph I found online: <next post>

Phil Pilgrim (PhiPi)

Heater.

Cool.

Poses more questions than it answers. Why is red so much slower? Is that white slow due to it's phosphor response or because of the underlying blue LED?

Curious.

Phil Pilgrim (PhiPi)

I'm pretty sure it's the phosphor in the white that contributes to its slowness. Think of the phosphors in TV tubes and 'scope tubes, and their "persistence."

-Phil

Heater.

Interesting point. Some of those old phosphors would persist for a second or so.

What was old is new again.

Peter Jakacki

Phil Pilgrim (PhiPi) wrote: »

heater wrote:

So now I wonder how quickly does that phosphor respond?

Here's a graph I found online: <next post>

It would help to know how they were measuring it too. Do you a have a link?

Phosphor LEDs can be excited by either blue or UV LEDs but the area of interest is with non-phosphor LEDs since we would need discrete "LEDs of color" to perform spectral measurements.

btw. you can post an image into an earlier post if you simply draft a new post, insert image, grab the URL, and paste that into your earlier post, then ignore/discard the draft.

Phil Pilgrim (PhiPi)

Peter Jakacki wrote:

It would help to know how they were measuring it too. Do you a have a link?

Just found the original webpage:

http://www.jensign.com/Discovery/LEDFrequencyResponse/

The image I posted was from a link to a link in another forum.

-Phil

Peter Jakacki

Phil Pilgrim (PhiPi) wrote: »

Just found the original webpage:

http://www.jensign.com/Discovery/LEDFrequencyResponse/

Thanks Phil, that provides some more information about the test setup which looks about right too. One thing I noticed was that only shows a large signal pulsed response, that is, driving from whoa to go. It would be interesting to look at a biased drive where the LED is already conducting but barely visible in the off state.