Is a LED fast enough to strobe and is any color faster

[Seekonk]

I'm helping a friend rebuild a couple SUN Distributor machines.  They have a spinning neop flash tube that spins in an assembly that lights up the dial 0-360 when the plug would fire.  I just got the first machine to work acceptably, but it is likely that the other machines will need new flash tubes.  These are custom made and cost $250. A 10W LED would fit in the space just fine. The dial would be about 2 1/2 inches from the LED. Most any color would be acceptable.  Is a LED fast enough, it would have to be better than 3 uS.  White and some colors have dyes that might cause persistence problems.  I have a number of power LED but they are all at my camp.  Anyone with any similar experience. I'm incorporating a mini 328 to get real numbers out of this antique too.  I want to borrow this machine for Halloween.  Wear a white lab coat, some welding goggles and this spinning orange disk will scare the hell out of kids.

[DJ]

Use an RGB LED, and a position sensor. Have the LED switch from blue to red as it changes direction relative to the viewer.

Claim you have developed a quantum distributor which operates at the speed of light and apply for government funding.

[ConKbot]

LEDs would definitely be fast enough, and you're correct about phosphors being slower than a non-phosphor led. I'd look at a cree XR-E, in red or green, or a luxeon rebel color, in red or green also. 

[Leadfootin]

Very easy to test led's. Set up a modulated driver circuit for the led's in question and then use a fast photo transistor and a scope to see the output. IE a sig gen with a 10% duty cycle square wave to drive, bias receiver diode, photo trans or use a small silicon photo cell.

FWIW I have a similar unit and am contemplating such a project as well. The coil and power supply just don't seem to cut it any more. Note that for a point system you should use at least 100ma for a clean signal. Don't want to switch on the dirt on the contacts. It is easy to create a mag pick up circuit as well with only a few parts.

[mikeselectricstuff]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

[ConKbot]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

Similar to the phosphors used in a CRT. Though not meant to have persistance as a CRT would, there still is some. 
Grab an LED fashlight, a dark room, let your eyes adjust, cover the end of the flashlight, blink it on and off, and quickly remove your hand.  You can see a fading afterglow.

Would it make a difference? In this application? probably not, but avoiding white LEDs and using a color LED removes any doubt that it could affect it, and doesn't add any negatives for a 1-off situation like this.

[Seekonk]

Leadfootin, Have you replaced ALL the capacitors?  Besides the one at the flash tube there are four more under the power supply that are switched in and out as the speed changes.  Actually it is one big rectangular capacitor with four .3uF 600V packed in it.  Supply runs at about 520V.  I used X2 capacitors since I had a lot of them.  I modified a Harbor Freight freebie flashlight to strobe the inductor lobes but never got to try it.  Spent all the time getting the unit to just work in a cold shop.  It's 20 miles away and my flu has come back after being in the cold.  Be another week before I can retest.

[mikeselectricstuff]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

Similar to the phosphors used in a CRT. Though not meant to have persistance as a CRT would, there still is some. 
Grab an LED fashlight, a dark room, let your eyes adjust, cover the end of the flashlight, blink it on and off, and quickly remove your hand.  You can see a fading afterglow.

Some of that could be capacitance in the driver. White LEDs are so efficient that they will emit a visible glow on a microamp or two. Any afterglow will be orders of magnitude below the normal "on" level.
This LED flash claims durations down to 500nS
https://www.kickstarter.com/projects/vela/vela-one-the-worlds-first-high-speed-led-flash

[Zero999]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

Similar to the phosphors used in a CRT. Though not meant to have persistance as a CRT would, there still is some. 
Grab an LED fashlight, a dark room, let your eyes adjust, cover the end of the flashlight, blink it on and off, and quickly remove your hand.  You can see a fading afterglow.

Some of that could be capacitance in the driver. White LEDs are so efficient that they will emit a visible glow on a microamp or two. Any afterglow will be orders of magnitude below the normal "on" level.
This LED flash claims durations down to 500nS
https://www.kickstarter.com/projects/vela/vela-one-the-worlds-first-high-speed-led-flash

There is a significant afterglow and it isn't due to the leakage current.

I have seen it before too. The afterglow is normally yellow or orange but it may appear white as human vision is monochrome below a certain illumination level.

The LED filament bulb I have is the worst offender. The glow lasts a few seconds and is still faint after a minute. Again I'm pretty sure it isn't the capacitors because it's an orangish colour, as the blue LED dies turn of instantly.

[Psi]

Should be easy to test.
Shine a LED torch directly into another LED torch in a dark room then switch off and have a look at it.

[Yansi]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

Exactly. My friend tested the "delay myth" of white LEDs an has come to a conclusion, no light decay is present. As the power is cut off, so is the light.

Hero999: Of course that effect is due to capacitors in the LED bulbs. How could you think it is not? :-O

[grumpydoc]

I have seen it before too. The afterglow is normally yellow or orange but it may appear white as human vision is monochrome below a certain illumination level.

It is but it actually depends on ambient lighting.

Cone cells respond to colour but don't work in low light.

Rods actually "see" green-blue light but only work at all in low light. This is because the pigment which absorbs the light (rhodopsin) is destroyed in the process and has to be re-synthesised. This is relatively slow and the cell can't keep up during daylight.

So in daylight you see in colour but anything emitting below the threshold for detection will just appear black. At night you can see in low light but only in "black and white"

[dannyf]

Is a LED fast enough

Opto-couplers are typically spec'd to be within 10us (emitter output), and many times well within that limit. So I would say that LEDs are fast enough for your application.

Light tensity is another matter.

[Marco]

Led phosphors (except for remote phosphor) have to be designed for fast decay to avoid saturation. Papers on white LED optical communication state a bandwidth of a couple of MHz which would put the 90/10 decay time far below a microsecond.

[Mechatrommer]

Exactly. My friend tested the "delay myth" of white LEDs an has come to a conclusion, no light decay is present. As the power is cut off, so is the light.

every matter in the universe are capacitive. how do your friend conduct/specify the test and the what is the test setup limitation? because there is no such thing as absolute "instantaneous", just human perception percepting beyound there capability.

[Zero999]

Hero999: Of course that effect is due to capacitors in the LED bulbs. How could you think it is not? :-O

I know it's the phosphor. I've excited it externally with a black light and there is significant afterglow. It isn't just the LED filament. I've tested a bare power LED with the same result. I haven't tested standard 5mm LEDs. I don't know how transparent the epoxy is to UV.

[tggzzz]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

Exactly. My friend tested the "delay myth" of white LEDs an has come to a conclusion, no light decay is present. As the power is cut off, so is the light.

Not true.

Phosphors work by having electrons stimulated into a different energy state by absorbing photons at one frequency. Then, at some random time later the electron changes to a different energy state and a photon is emitted, optionally at a different frequency.

The question then becomes what are the statistics of the delay before the second photon is emitted. Depending on the specific phosphor and the frequency required in this application, the delay might or might not be significant.

So, what measurements did you make to justify your assertions?

Apart from that, LEDs are typically extremely fast, much more so than is implied by the speed of optocouplers, since there the speed is limited by the capacitances in the phototransistor.

[tggzzz]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

Similar to the phosphors used in a CRT. Though not meant to have persistance as a CRT would, there still is some. 
Grab an LED fashlight, a dark room, let your eyes adjust, cover the end of the flashlight, blink it on and off, and quickly remove your hand.  You can see a fading afterglow.

Would it make a difference? In this application? probably not, but avoiding white LEDs and using a color LED removes any doubt that it could affect it, and doesn't add any negatives for a 1-off situation like this.

People have developed different phosphors with persistences up to several seconds - classically for the radar displays, the first computer memory, and more recently for storage scopes.
http://en.wikipedia.org/wiki/Cathode_ray_tube#Phosphor_persistence

CRTs for TVs had phosphors with a persistence time chosen to minimise flicker, i.e. milliseconds not microseconds.

For this application, don't guess. Either measure the persistence, or use LEDs without phosphors.

[HKJ]

I do sometimes measure PWM on flashlights, here is a curve from a XM-L2 led where I measure light with a photodiode and voltage across the diode:



There is not any afterglow in that curve

Green is voltage across the diode.
Blue is voltage from photodiode, i.e. brightness.

[dannyf]

There is not any afterglow in that curve

But lots of afterglow in the human eyes,

[Marco]

i am very intrigued by how they managed such short pulse. boost up voltage to multiples in small cap and discharge it?

Doubt it, it's a pain to get nice variable width pulse (well you could use a PFN, but I don't think they have 1500 meter of coax in there). Probably a couple smaller (ie. faster switching) MOSFETs in parallel.

I wonder how much lumen they really get out of those LEDs ... if the phosphor is saturating the efficiency will drop like a brick.

[Seekonk]

I did a quick mock up with a 4W line operation (120VAC) white LED that I have. Normally it is a full wave bridge with a 560 ohm resistor.    I  removed their board wired it to 18V DC and 82 ohm resistor,all switched by a FET.  It is easy to over clock it and stop a computer fans motion at 1.6khz in a modestly lit room with the LED about 2 inches away.  With those short pulses you can stare directly into the LED and easily see the points of light.  It was a stretch for me to come up with 175V from parts I have.  I like the idea of the higher voltage LED as this has to be fed by brushes that aren't that substantial.  That article appears to have the yellow dye LED like I have and they allude to them being commercial varieties. That inspires me to try voltages 50% higher once a micro closely controlls the on time.  Willing to destroy a couple for science.

Since we are into an esoteric discussion, I'm wondering why they used neon.  Did the orange give a better contrast to the white arrow in room light? Does the eye have better persistence with that color? I really want to try it in red.   As a manager once told me....The optics of a project are everything.
Or was it just cheap and had a long life

[tom66]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

You'd be surprised. Typically red  and green phosphors have long delays. Blue phosphors are much faster. Delay can be on the order of 5ms for some phosphors used in plasma displays (which operate by a different principle, UV to light rather than blue to light, but the general operation is similar.)

[tggzzz]

I don't think phosphors would have any appreciable delay - how could they store enough energy to emit light anywhere near what they do when lit?

You'd be surprised. Typically red  and green phosphors have long delays. Blue phosphors are much faster. Delay can be on the order of 5ms for some phosphors used in plasma displays (which operate by a different principle, UV to light rather than blue to light, but the general operation is similar.)

Radar displays have phosphors with a persistance 1000 times longer than that; several seconds (not ms) in fact!

The delay is/was very useful with long-range primary radar, so that you could see how far the target had moved between successive sweeps.

[eliocor]

According to the Phosphor Handbook, the typical YAG:Ce phosphor used on white LEDs (combined with blue light chips) has a typical luminescence lifetime of about 10÷100ns (chapter 5.9)