Pulsed drive of white LEDs?

[Don Klipstein]

In <875b2cdb-ea0d-4e22-a41a-4a12d7ce00f4@g23g2000yqh.googlegroups.com>,

Limiting factor is the die, not the phosphor. I've seen 200 nanosecond
rise times on white leds with standard yellow phosphors.

Thanks for the info - now I know phosphor persistence is pretty small,
as I suspected.

This data point appears to me to indicate that phosphor persistence
counts for something, since unphosphored LEDs usually have rated rise/fall
times of 30 nanoseconds.

- Don Klipstein ([email protected])

 

[Jon Kirwan]

Emergency vehicles already signal at intersections. The receivers sit
high on the wires or poles across an intersection and are IR-based.
They use optical baffling to narrow the acceptance angle downward
towards oncoming traffic at the right spot and optical wavelength
filtering to narrow the accepted IR color bandwidth, as well. They
use electronic receivers that are tuned with extremely narrow notches
so that only a very precise flicker rate of the emitter is accepted.
The sun doesn't yield much flicker in that band (circa 15Hz, if I
recall.)

I haven't been reading the thread and I apologize if I've duplicated
something already said. My thought though is for you to use a tube,
baffling and possibly some optics, a narrow-band wavelength filter
(thin film?), and a receiver/emitter pair that transmit/receive in an
extremely narrow bandpass. The combination appears to work,
elsewhere. I believe I've also seen similar things designed to count
traffic, replacing those hoses that lay across the road. Maybe
there's no time to screw with this stuff, though.

Jon

 

[ehsjr]

Has anyone tried driving ordinary white LEDs - the kind used in
cheap flashlights - with a pulsed current and done some kind of
evaluation of their performance? Any idea if they can cope with
frequencies of the order of a kHz?

I'm thinking of using them with photosensors with a pulsed drive
for discrimination against ambient light. It's for the drag
racing christmas tree system that I asked about a few days ago.
I've considered laser pointers and infrared devices, but a tight
schedule and possible problems with on-site alignment makes me
think that a pulsed white LED with a general-purpose photodiode
might be the best option.

I'm always open to alternative suggestions, but please remember
that quick procurement of anything but the most common
general-purpose devices is next to impossible from where I live.

If you can get laser pointers, your outdoor quick setup problem
can be easily solved. When building the equipment, install the
detector into a fairly large white surface. Set things up (at
home) such that the emitter and detector are perfectly aligned,
and are separated by the same distance they will be on race
day. Mount a laser pointer on the emitter assembly such that it
shines someplace on the detector white surface, but not on the
detector itself. Draw an X where the laser pointer dot is when
the emitter & detector are perfectly aligned. Make sure everything
is firmly mounted so that it cannot move when you bring the equipment
to the race. Then, when you are setting it up in the field for the
race, set it up so that the laser dot hits the X and it will be
properly aligned.

Your biggest problem may turn out to be getting coplanar level
spots at the race start line to place your emitters & detectors.

Ed

 

[pimpom]

Emergency vehicles already signal at intersections. The
receivers sit
high on the wires or poles across an intersection and are
IR-based.
They use optical baffling to narrow the acceptance angle
downward
towards oncoming traffic at the right spot and optical
wavelength
filtering to narrow the accepted IR color bandwidth, as well.
They
use electronic receivers that are tuned with extremely narrow
notches
so that only a very precise flicker rate of the emitter is
accepted.
The sun doesn't yield much flicker in that band (circa 15Hz, if
I
recall.)

I haven't been reading the thread and I apologize if I've
duplicated
something already said. My thought though is for you to use a
tube,
baffling and possibly some optics, a narrow-band wavelength
filter
(thin film?), and a receiver/emitter pair that transmit/receive
in an
extremely narrow bandpass. The combination appears to work,
elsewhere. I believe I've also seen similar things designed to
count
traffic, replacing those hoses that lay across the road. Maybe
there's no time to screw with this stuff, though.

Jon

You hit it with that last sentence. No time, *plus* procuring
things quickly from where I live is next to impossible. I have to
rely on what I already have, can get locally (which is of very
limited range) or can scrounge from discarded stuff.

 

[pimpom]

pimpom wrote:



To avoid cross detection, you can do this:

emitter1------->detector1

detector2<------emitter2

Each detector can "see" only its own emitter.

That should indeed work except for two things. I had intended to
place the emitters between the two racers, battery powered
without any cables leading to the sides, two beams each emitting
outwards unidirectionally. The second factor is that an inward
beam could spill over to the other side and hit the receiver for
the other lane. That could be avoided by placing a baffle in the
center, but I thought it would be best to make the center piece
small, sturdy and low-profile.

Reflections could also be a problem. But I'm not rejecting the
idea out of hand. It might just be the best solution under the
circumstances.

 

[pimpom]

Would not red laser pointers work.

I'd already considered laser pointers but, as I explained
elsewhere, I anticipate problems with on-site alignment - four
narrow beams hitting tiny sensors at a distance of more than 10
ft over uneven ground. And once aligned, both emitters and
sensors would have to be fixed firmly enough against accidental
knocks and vibration.

 

[Adrian Tuddenham]

[...]

The beam won't carry any data. It's to be a simple
blocked/not-blocked switching action. I thought of using a
pulsed drive to eliminate the effect of ambient light. The
receiver will have a loosely tuned circuit (perhaps an RC twin-T)
or just a high-pass filter.

Assuming you have a cable between emitter and receiver I'd do two
things...

Square wave modulate the emitter and synchronous detect at the
receiver.

Then add a DC loop at the receiver to null out the ambient light.

Put a single tuned circuit in the receiving amplifier and feed its
output back to the light source. When the beam is unblocked, the whole
system will oscillate, but you don't need anything more complicated than
a diode to detect it.

There is only one tuned circuit , so temperature drift doesn't matter
and alignment isn't necessary.

I built a system like this to work over a reflected path across a wide
south-facing doorway at my local garage; the source and detector are in
the same box but carefully screened from each other. It has been
operating without problems for more than 10 years.

 

[pimpom]

Has anyone tried driving ordinary white LEDs - the kind used in
cheap flashlights - with a pulsed current and done some kind of
evaluation of their performance? Any idea if they can cope with
frequencies of the order of a kHz?

I'm thinking of using them with photosensors with a pulsed
drive
for discrimination against ambient light. It's for the drag
racing christmas tree system that I asked about a few days ago.
I've considered laser pointers and infrared devices, but a
tight
schedule and possible problems with on-site alignment makes me
think that a pulsed white LED with a general-purpose photodiode
might be the best option.

I'm always open to alternative suggestions, but please remember
that quick procurement of anything but the most common
general-purpose devices is next to impossible from where I
live.

Thanks for all your interest and helpful replies. Even the
suggestions I can't use directly were helpful in narrowing down
possible solutions.

I've tentatively decided to go for a pulsed IR system. What
particularly lit a light bulb in my head was when krw pointed out
that the emitter and receiver do not have to operate with narrow
beams. I had been stuck with the idea of collimating a thin laser
beam on to a tiny sensor.

I think it will be feasible to guard against false detection due
to reflections by limiting the angle of incidence with blackened
tubes at both ends. I don't think there will be a problem
aligning the path within a couple of degrees. A laser pointer
beam impinging on a sensor 2mm wide from 15 ft would require an
accuracy of something like 0.025º and that's simply not
practicable in the given circumstances.

The remaining problem is interference between the two parallel
paths, 7 inches apart, for each rider. Coded pulsing would
obviate the problem, but I don't think I'll have time to try that
out or get the parts. For the time being, I may just decide to
use only one beam for each rider to mark the starting line (and
detect false starts), and do without the less essential
pre-staging marker for now.

Refinements can be added for later events - the full two-beam
system, measurement of speed at regular intervals along the race
track, elapsed time, reaction time, etc.

 

[Adrian Tuddenham]

[...]

I'd already considered laser pointers but, as I explained
elsewhere, I anticipate problems with on-site alignment - four
narrow beams hitting tiny sensors at a distance of more than 10
ft over uneven ground. And once aligned, both emitters and
sensors would have to be fixed firmly enough against accidental
knocks and vibration.

It is much easier to keep all the electronics together in one place and
just fold the light path. The self-oscillating system I have described
in another post will work perfectly well over a double 10ft path in
bright sunlight (as long as it doesn't fall directly on the sensor).

Use a reflexor at the far side and keep the source and detector close
together in the same box on the near side. Test a few easily-obtainable
reflexors (such as number plate background material and the reflectors
for bicycles and the sides of lorries) to check that they work at your
chosen wavelength.

 

[Jon Kirwan]

Thanks for the info - now I know phosphor persistence is pretty small,
as I suspected.

This data point appears to me to indicate that phosphor persistence
counts for something, since unphosphored LEDs usually have rated rise/fall
times of 30 nanoseconds.

- Don Klipstein ([email protected])

Rare earth phosphors I use have taus on the order of milliseconds.

The term 'phosphor' refers to both fluorescence and phosphorescence,
now. Centuries ago, to any substance that seemed to emit light on its
own, without combustion taking place. (So a radium watch dial, if sent
backwards in time a few centuries, would probably be included by some
earlier texts as a phosphor.)

Some phosphors (usually those that do NOT exhibit a nice exponential
decay) can have decay tails that last well into minutes of time. One
of the earliest recorded phosphors were of the this very persistent,
phosphorescent kind. Some paintings in Japan or China used materials
that came from volcanic activity acting on seashells and sulfur (with
rare earth included, of course.) Although these have non-exponential
decays in the microseconds, they also have long emission tails lasting
well into seconds or even a minute or two.

Saying "phosphor persistence is pretty small" when talking about 200ns
periods of time cuts off at least half of the modern meaning of the
term, 'phosphor.'

Jon

 

[Baron]

Thanks for all your interest and helpful replies. Even the
suggestions I can't use directly were helpful in narrowing down
possible solutions.

I've tentatively decided to go for a pulsed IR system. What
particularly lit a light bulb in my head was when krw pointed out
that the emitter and receiver do not have to operate with narrow
beams. I had been stuck with the idea of collimating a thin laser
beam on to a tiny sensor.

I think it will be feasible to guard against false detection due
to reflections by limiting the angle of incidence with blackened
tubes at both ends. I don't think there will be a problem
aligning the path within a couple of degrees. A laser pointer
beam impinging on a sensor 2mm wide from 15 ft would require an
accuracy of something like 0.025º and that's simply not
practicable in the given circumstances.

Actually I don't think it would. I checked a couple of red laser
pointers I have here. Both produce spot diameters of around 6 or 7 mm
at 12 feet (the longest distance I checked). I also tried pointing one
at my car number plate. The reflected light was just a blinding blob at
that distance.

 

[Uwe Hercksen]

Has anyone tried driving ordinary white LEDs - the kind used in
cheap flashlights - with a pulsed current and done some kind of
evaluation of their performance? Any idea if they can cope with
frequencies of the order of a kHz?

Hello,

I drived red, green and blue LEDs with pulses of 10 to 100 microseconds,
no problem at all. But white LEDs may have a problem with the yellow
flourescent dye, so I would prefer single color LEDs for short defined
pulses.

Bye

 

[JosephKK]

Has anyone tried driving ordinary white LEDs - the kind used in
cheap flashlights - with a pulsed current and done some kind of
evaluation of their performance? Any idea if they can cope with
frequencies of the order of a kHz?

I'm thinking of using them with photosensors with a pulsed drive
for discrimination against ambient light. It's for the drag
racing christmas tree system that I asked about a few days ago.
I've considered laser pointers and infrared devices, but a tight
schedule and possible problems with on-site alignment makes me
think that a pulsed white LED with a general-purpose photodiode
might be the best option.

I'm always open to alternative suggestions, but please remember
that quick procurement of anything but the most common
general-purpose devices is next to impossible from where I live.

Modulate the LED and look for the modulated signal. Great aid to
{semi-)manual alignment. Plus it improves S/N a bunch. All
monochrome LEDs are good for MHz, whites may be good for many kHz,
depends on the phosphor.

 

[JosephKK]

Cigarrette - blow smoke so you can see the laser, line it up and fix. Two
minutes.

Or any other fine particulate source, like superfine flour in a
eductor type sprayer, or just kick up some dust.

 

[JosephKK]

I made a trial batch of street light controllers for my state's
electricity board nearly 20 years ago. Control was from
distribution points rather than at each lamp post. The problem
was that practically the whole state, including the capital, is
made up of small jagged hills. Daylight varies from point to
point and the controllers had to be set individually. They worked
fine when set properly, but the engineers left the setup to
unskilled workers who had no idea what they were doing. That
turned the experiment into a failure.

Design case point. What do the maintainers think? How will they act?
There may be more to the customer than the representatives you meet.

 

[JosephKK]

On Mon, 23 Nov 2009 10:17:54 +0000,

[...]

I'd already considered laser pointers but, as I explained
elsewhere, I anticipate problems with on-site alignment - four
narrow beams hitting tiny sensors at a distance of more than 10
ft over uneven ground. And once aligned, both emitters and
sensors would have to be fixed firmly enough against accidental
knocks and vibration.

It is much easier to keep all the electronics together in one place and
just fold the light path. The self-oscillating system I have described
in another post will work perfectly well over a double 10ft path in
bright sunlight (as long as it doesn't fall directly on the sensor).

Use a reflexor at the far side and keep the source and detector close
together in the same box on the near side. Test a few easily-obtainable
reflexors (such as number plate background material and the reflectors
for bicycles and the sides of lorries) to check that they work at your
chosen wavelength.

Ah yes, the three corner reflector trick. Reflects the incoming beam
parallel to and adjacent to itself. Then the detectors are adjacent
to the emitters and no wire to run and most alignment issues reduced
or eliminated.
Thanks Adrian, i had missed this.

 

[Don Klipstein]

I have what I consider to be "major experience".

My experience is in 2 parts, both disfavoring pulsing white LEDs in
attempt to increase ratio of visual perception of brightness to amount of
current or power delivered to the LEDs.

One is personal experimentation into this bit of LEDs appearing brighter
when pulsed. I found that human vision is a "good integrator" when pulse
rate is fast enough to achieve lack of human-perceptible flicker.

On that line, I also found that the legend of benefit from pulsing arose
from the benefit being applicable to a widely-used kind of LED having a
nonlinearity favoring efficiency and more-important-still ratio of
lumens to average-mA varying directly with instantaneous current
throughout at least most of the range of instantaneous current under
consideration.

I discuss this a bit more in a web page of mine:

http://members.misty.com/don/ledp.html

(slightly outdated due to excessive consideration to low-power LEDs at
the times when I wrote that long ago, enough long-ago for high-power LEDs
350 mA-plus to be either not-yet-in-existence or something quite new.)

Principles when translated to such higher currents of modern high power
LEDs still apply.

- Don Klipstein ([email protected])

 

[Don Klipstein]

[...]

I'd already considered laser pointers but, as I explained
elsewhere, I anticipate problems with on-site alignment - four
narrow beams hitting tiny sensors at a distance of more than 10
ft over uneven ground. And once aligned, both emitters and
sensors would have to be fixed firmly enough against accidental
knocks and vibration.

It is much easier to keep all the electronics together in one place and
just fold the light path. The self-oscillating system I have described
in another post will work perfectly well over a double 10ft path in
bright sunlight (as long as it doesn't fall directly on the sensor).

Use a reflexor at the far side and keep the source and detector close
together in the same box on the near side. Test a few easily-obtainable
reflexors (such as number plate background material and the reflectors
for bicycles and the sides of lorries) to check that they work at your
chosen wavelength.

Ah yes, the three corner reflector trick. Reflects the incoming beam
parallel to and adjacent to itself. Then the detectors are adjacent
to the emitters and no wire to run and most alignment issues reduced
or eliminated.
Thanks Adrian, i had missed this.

This physical efect has a name that I suspect is good enough to
web-search on:

"corner cube reflector"

It appears to me that plenty of bicycle reflectors including most other
than "aftermarket", some bicycle taillights, and some automotive
taillights use this principle.

In most vehicle-mounted retroreflectors (one more keyword), there is at
least one area of clustered small-sizede corner-cube inits with "cell
size" around 3 mm IIRC, and depending on (successfully) that a refractive
surface towards the incoming light source (boundary between air
and plastic) refracts the incoming and outgoing light rays equally so as
to maintain retroreflective property of the array of "corner cubes".

Many of these vehicle retroreflectors rely on arrays of small "corner
cubes", whose surfaces are angled sufficiently close to parallel to
incoming light rays desired to be retroflected, so as to achieve great
retroreflection from air-plastic boundary via "total internal reflection".
That principle of physics has its presence being a function of angle of a
ray being incident to a boundary from higher refractive index material
(such as plastic) to a lower refractive-index adjacent material (such as
air).

There are also retroreflective objects marketed to cyclists (and
elsewhere) involving very small glass spheres made of glass of whatwever
variant has refractive index that favors a significant and notable bit of
retroreflection. One thing that comes to my mind is "Scotchlite" (tm) by
3M.

- Don Klipstein ([email protected])

 

[JosephKK]

[...]
I'd already considered laser pointers but, as I explained
elsewhere, I anticipate problems with on-site alignment - four
narrow beams hitting tiny sensors at a distance of more than 10
ft over uneven ground. And once aligned, both emitters and
sensors would have to be fixed firmly enough against accidental
knocks and vibration.


It is much easier to keep all the electronics together in one place and
just fold the light path. The self-oscillating system I have described
in another post will work perfectly well over a double 10ft path in
bright sunlight (as long as it doesn't fall directly on the sensor).

Use a reflexor at the far side and keep the source and detector close
together in the same box on the near side. Test a few easily-obtainable
reflexors (such as number plate background material and the reflectors
for bicycles and the sides of lorries) to check that they work at your
chosen wavelength.

Ah yes, the three corner reflector trick. Reflects the incoming beam
parallel to and adjacent to itself. Then the detectors are adjacent
to the emitters and no wire to run and most alignment issues reduced
or eliminated.
Thanks Adrian, i had missed this.

This physical efect has a name that I suspect is good enough to
web-search on:

"corner cube reflector"

It appears to me that plenty of bicycle reflectors including most other
than "aftermarket", some bicycle taillights, and some automotive
taillights use this principle.

In most vehicle-mounted retroreflectors (one more keyword), there is at
least one area of clustered small-sizede corner-cube inits with "cell
size" around 3 mm IIRC, and depending on (successfully) that a refractive
surface towards the incoming light source (boundary between air
and plastic) refracts the incoming and outgoing light rays equally so as
to maintain retroreflective property of the array of "corner cubes".

Many of these vehicle retroreflectors rely on arrays of small "corner
cubes", whose surfaces are angled sufficiently close to parallel to
incoming light rays desired to be retroflected, so as to achieve great
retroreflection from air-plastic boundary via "total internal reflection".
That principle of physics has its presence being a function of angle of a
ray being incident to a boundary from higher refractive index material
(such as plastic) to a lower refractive-index adjacent material (such as
air).

There are also retroreflective objects marketed to cyclists (and
elsewhere) involving very small glass spheres made of glass of whatwever
variant has refractive index that favors a significant and notable bit of
retroreflection. One thing that comes to my mind is "Scotchlite" (tm) by
3M.

- Don Klipstein ([email protected])

Alas, that does not mean that any of this (with the possible exception
of bicycle reflectors) is currently readily available to pimpom (OP).

 

[baron]

JosephKK Inscribed thus:

[...]
I'd already considered laser pointers but, as I explained
elsewhere, I anticipate problems with on-site alignment - four
narrow beams hitting tiny sensors at a distance of more than 10
ft over uneven ground. And once aligned, both emitters and
sensors would have to be fixed firmly enough against accidental
knocks and vibration.


It is much easier to keep all the electronics together in one place
and
just fold the light path. The self-oscillating system I have
described in another post will work perfectly well over a double
10ft path in bright sunlight (as long as it doesn't fall directly on
the sensor).

Use a reflexor at the far side and keep the source and detector
close
together in the same box on the near side. Test a few
easily-obtainable reflexors (such as number plate background
material and the reflectors for bicycles and the sides of lorries)
to check that they work at your chosen wavelength.

Ah yes, the three corner reflector trick. Reflects the incoming beam
parallel to and adjacent to itself. Then the detectors are adjacent
to the emitters and no wire to run and most alignment issues reduced
or eliminated.
Thanks Adrian, i had missed this.

This physical efect has a name that I suspect is good enough to
web-search on:

"corner cube reflector"

It appears to me that plenty of bicycle reflectors including most
other
than "aftermarket", some bicycle taillights, and some automotive
taillights use this principle.

In most vehicle-mounted retroreflectors (one more keyword), there is
at
least one area of clustered small-sizede corner-cube inits with "cell
size" around 3 mm IIRC, and depending on (successfully) that a
refractive surface towards the incoming light source (boundary between
air and plastic) refracts the incoming and outgoing light rays equally
so as to maintain retroreflective property of the array of "corner
cubes".

Many of these vehicle retroreflectors rely on arrays of small
"corner
cubes", whose surfaces are angled sufficiently close to parallel to
incoming light rays desired to be retroflected, so as to achieve great
retroreflection from air-plastic boundary via "total internal
reflection". That principle of physics has its presence being a
function of angle of a ray being incident to a boundary from higher
refractive index material (such as plastic) to a lower
refractive-index adjacent material (such as air).

There are also retroreflective objects marketed to cyclists (and
elsewhere) involving very small glass spheres made of glass of
whatwever variant has refractive index that favors a significant and
notable bit of
retroreflection. One thing that comes to my mind is "Scotchlite" (tm)
by 3M.

- Don Klipstein ([email protected])

Alas, that does not mean that any of this (with the possible exception
of bicycle reflectors) is currently readily available to pimpom (OP).

There is always a reflective car number plate.