Car LED tail lights strobe rate too slow

[Pete C.]

The visibility is higher only if the pulse rate is lower
than the flicker fusion frequency - which is not the case
for the automotive applications being discussed.

Visibility is indeed higher since the peak emissions are higher. The
flicker also attracts attention better. I've seen LED lights on a number
of trucks where the flicker frequency had to be in the mid 20 Hz range.

Strobe lights operate at a pulse rate lower than the flicker
fusion frequency. If they did not, humans would see only a
continuous light output.

And your point is? They still operate with peak emissions that would
incinerate the strobe tube in seconds if they were continuous.

Pete C.

 

[Mark Kelepouris]

PWM is used in order to exceed the maximum continuous output available
from the LED. LEDs can be overdriven past their continuous operating
currents and generate higher peak light outputs (within some limits) by
operating them at a reduced duty cycle to hold the average current and
therefore chip heating within the continuous operating limits.

Pete C.

Hello all,
On subject of Led's and strobes, I intend to build a high brightness white
Led strobe, but due to the required duty cycle of 20 uS, the white Led wont
do because of 'phosphor persistence' . Unless someone knows of a super
bright white that doesnt have this problem, i'll have to use banks of the
brightest Red, green and blue. I have some questions on which transistor
would be best suited to drive strings/bunches of Leds at up to an amp, at
1khz, 20uS duty at 12vdc. With the low duty cycle I should be able to whack
200mA through a single Led (correct me) but at high freq. things would get
hot. My strobe will mainly be used between 0-100hz. So whats the best
choice of Led's and Tranny driver?

Thanks for any help,
Mark Kelepouris

 

[Roger Hamlett]

Hello all,
On subject of Led's and strobes, I intend to build a high brightness
white Led strobe, but due to the required duty cycle of 20 uS, the white
Led wont do because of 'phosphor persistence' . Unless someone knows of
a super bright white that doesnt have this problem, i'll have to use
banks of the brightest Red, green and blue. I have some questions on
which transistor would be best suited to drive strings/bunches of Leds
at up to an amp, at 1khz, 20uS duty at 12vdc. With the low duty cycle I
should be able to whack 200mA through a single Led (correct me) but at
high freq. things would get hot. My strobe will mainly be used between
0-100hz. So whats the best choice of Led's and Tranny driver?

There are currently two and a half white LED basic technologies on the
market!. Blue with phosphor, is the most common, then RGB, has multiple
LED's and produces 'white', by mixing the colours. There are also some new
'variants' of the phosphor technology, with slightly different
characteristics.
Get rid of the phosphor based units, and try units like:
http://www.theledlight.com/rgbdiodes.html
or for higher power:
http://www.optosource.com/lednewsoptosource.asp?newsid=15
Several of the ultra high output white LEDs from other manufacturers also
use this type of technology. Talk to an LED supplier, about RGB ultra
bright LEDs. They should be able to help.

Best Wishes

 

[TimR]

There may be some flicker or strobing present, I'm not sure, but I
suspect most of this is aliasing.

This phenomenon also exists on LED displays run on pure DC.

When you move your head, or chew potato chips, or play very low notes
on a bass trombone your eyeballs jiggle. Each time the eyes move you
lose vision for a small amount of time. This has the effect of
chopping the light into flashes even though the light may be steady.
It is not often noticable, but at night when the light source is small
it can be.

That's my theory, anyway. I'd go play my trombone now but it's all
full of potato chips.

 

[Pete C.]

The reason PWM dimming is used on LED taillights is because there is a
ratio requirement between the photometrics of the stop lamp and the
tail lamp of at least 5:1 at some of the test points. If you are going
to do the tail lamp by just dimming the stop, then to just barely
comply you have to be running at 20% of the stop current. With a
reasonable design margin, you proably want to have on the order of
10:1. Another consideration is that there is a maximum limit on tail
lamp output that could be exceeded even if you have the right ratio
between stop and tail which would necessitate using more than the 5:1
drop. If you read the manufacturers specs on the LEDs used in these
tail lamps, the performance of the led at the low currents is usually
not spec'd or is not gauranteed because the behavior of the diode is
significantly different in that region. So running the tail at a lower
current isn't an option. Most of the times, the PWM is a lot more
difficult/costly to implement than a simple current draw down because
the PWM creates EMI that has to be dealt with or you might find weird
hums on your radio when your taillights are on.



PWM really can't be used to increase the light output in these
applications as a previous poster implied, because you would need to
overdrive the chip significantly and for a long enough duty cycle that
the heat generated would negate any advantage and you might try to gain
from PWM.

Really? I did a project a few years back using PWM for exactly that
purpose. Indeed the LED manufacturers data sheet included precise
specifications for such use as well.

Pete C.

 

[Boxman]

I would agree there might be specific applications for using PWM to
boost output, but for automotive tail lamps it didn't seem to benefit.
The LEDs used in the stop/tail lamps were predominantly the lumileds
snapled 75 and snapled 150's early on and they have since moved towards
the luxeons and K2's. Also used in stop/tail lamps were the Osram
power topleds. The snapled 150's usually had to be driven near their
maximum rated current (150 ma) to get the desired output for stop mode.
The composition of the red and yellow dies made it so that the light
output fall off due to junction temperture rise is really quite bad.
The LED loses about 60% of it's light output at the rated junction
temperature of 25°C when you drive it at any reasonable current. So
pulsing didn't seem to get any output back because the duty cycle had
to be so long, that there was still significant die heating.

Which LED's were you using the PWM to boost output with? I haven't had
a chance to try that.

 

[Victor Roberts]

Really? I did a project a few years back using PWM for exactly that
purpose. Indeed the LED manufacturers data sheet included precise
specifications for such use as well.

Pete C.

Which LED? What was the pulse rate?

--
Vic Roberts
http://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

This information is provided for educational purposes only.
It may not be used in any publication or posted on any Web
site without written permission.

 

[Douglas G. Cummins]

Victor Roberts wrote:

Visibility is indeed higher since the peak emissions are higher.

State your source. Note that this thread has been linked to
sci.engr.lighting. We lighting experts know what you're saying is
incorrect. If all you're doing is throwing back information that LED
sales people have given you, then you and they are mistaken.

The
flicker also attracts attention better. I've seen LED lights on a number
of trucks where the flicker frequency had to be in the mid 20 Hz range.

By what meter do you judge this? I see all manner of lights from
continuous sources to PWMs at 20 kHz or more. Once the pulse rate
increases beyond 5 Hz, it becomes very difficult to estimate a flash
rate by eye alone. I can't believe you're taking a light detector
hooked to an oscilloscope to determine a PWM pulse rate of an LED lamp
on a truck.

And your point is? They still operate with peak emissions that would
incinerate the strobe tube in seconds if they were continuous.

The point is that the eye is an integrator - it sees only the average
light output of sources pulsing at the frequencies we're discussing.
Same principle that allows people to see an image on the TV. Your
comment regarding strobe tubes incinerating is totally irrelevant to the
subject at hand. We're trying to educate you - you're trying to bully
your point through.

Peak intensity is only relevant for slow flash rates - below the '20 Hz'
you stated above. Warning lamps flash in the 1 - 4 Hz range where the
peak intensity does increase conspicuity (not visibility as is often
misused in this case as you are doing). Once all those flashes merge
into a continuous beam, conspicuity decreases.

 

[Pete C.]

State your source. Note that this thread has been linked to
sci.engr.lighting. We lighting experts know what you're saying is
incorrect. If all you're doing is throwing back information that LED
sales people have given you, then you and they are mistaken.

Drive higher = brighter = more visible

By what meter do you judge this? I see all manner of lights from
continuous sources to PWMs at 20 kHz or more. Once the pulse rate
increases beyond 5 Hz, it becomes very difficult to estimate a flash
rate by eye alone. I can't believe you're taking a light detector
hooked to an oscilloscope to determine a PWM pulse rate of an LED lamp
on a truck.

I've spent quite a bit of time in the film / video world so I'm a bit
sensitized to things in the 24-25-29.97 Hz range.

The point is that the eye is an integrator - it sees only the average
light output of sources pulsing at the frequencies we're discussing.
Same principle that allows people to see an image on the TV.

The phosphors in the CRT are what does the integrating for the most
part, or in newer technologies it's the drive logic that is holding the
pixel data.

Your
comment regarding strobe tubes incinerating is totally irrelevant to the
subject at hand. We're trying to educate you - you're trying to bully
your point through.

Care to explain exactly what's wrong with higher drive = higher output =
brighter / more visible instead of claiming I'm wrong with nothing to
back that claim?

Peak intensity is only relevant for slow flash rates - below the '20 Hz'
you stated above. Warning lamps flash in the 1 - 4 Hz range where the
peak intensity does increase conspicuity (not visibility as is often
misused in this case as you are doing). Once all those flashes merge
into a continuous beam, conspicuity decreases.

Peak intensity translates to visibility and low flash rates translate to
conspicuity. If it's brighter you can see it from further away and if
it's flashing slow enough to see the flash rate it grabs your attention.
They are not the same thing

Pete C.

 

[Pete C.]

I would agree there might be specific applications for using PWM to
boost output, but for automotive tail lamps it didn't seem to benefit.
The LEDs used in the stop/tail lamps were predominantly the lumileds
snapled 75 and snapled 150's early on and they have since moved towards
the luxeons and K2's. Also used in stop/tail lamps were the Osram
power topleds. The snapled 150's usually had to be driven near their
maximum rated current (150 ma) to get the desired output for stop mode.
The composition of the red and yellow dies made it so that the light
output fall off due to junction temperture rise is really quite bad.
The LED loses about 60% of it's light output at the rated junction
temperature of 25°C when you drive it at any reasonable current. So
pulsing didn't seem to get any output back because the duty cycle had
to be so long, that there was still significant die heating.

Which LED's were you using the PWM to boost output with? I haven't had
a chance to try that.

I don't recall which ones, it was a while back and someone else's
project I worked on. It was pre-Luxeon days as well.

Pete C.

 

[Pete C.]

Which LED? What was the pulse rate?

--
Vic Roberts
http://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

This information is provided for educational purposes only.
It may not be used in any publication or posted on any Web
site without written permission.

I don't recall, it was a while back and someone else's project I worked
on. It was pre-Luxeon days as well or we probably wouldn't have had to
bother.

Pete C.

 

[RHRRC]

snip <


Really? I did a project a few years back using PWM for exactly that
purpose. Indeed the LED manufacturers data sheet included precise
specifications for such use as well.

Pete C.

You must publish your results of your project 'from a few years back'
ASAP since your more modern work appears to contradict the body
understanding of held for over half a century on the human optical
responses and it is important that the truth of the matter be revealed
for the benefit of mankind.

It has long been suspected that the huge wealth of evidence
corroborating currently advocated principles of vision is totally wrong
and is, in fact, a conspiracy by every single one of those (other than
yourself) that have studied such phenomena in depth.

Of course it is ackowledged that waggling a leds supply around at speed
improves all led efficacies - the faster the better - as well as
improving perception.

The power companies (amongst others) suppress this information to try
to sell more electric.

 

[ian field]

State your source. Note that this thread has been linked to
sci.engr.lighting. We lighting experts know what you're saying is
incorrect. If all you're doing is throwing back information that LED
sales people have given you, then you and they are mistaken.


By what meter do you judge this? I see all manner of lights from
continuous sources to PWMs at 20 kHz or more. Once the pulse rate
increases beyond 5 Hz, it becomes very difficult to estimate a flash rate
by eye alone. I can't believe you're taking a light detector hooked to an
oscilloscope to determine a PWM pulse rate of an LED lamp on a truck.


The point is that the eye is an integrator - it sees only the average
light output of sources pulsing at the frequencies we're discussing. Same
principle that allows people to see an image on the TV. Your comment
regarding strobe tubes incinerating is totally irrelevant to the subject
at hand. We're trying to educate you - you're trying to bully your point
through.

Peak intensity is only relevant for slow flash rates - below the '20 Hz'
you stated above. Warning lamps flash in the 1 - 4 Hz range where the
peak intensity does increase conspicuity (not visibility as is often
misused in this case as you are doing). Once all those flashes merge into
a continuous beam, conspicuity decreases.

My favourite stroboscopic effect was when motorcycle shops used to sell
those novelty neon spark plug caps, if the bike goes by at a reasonably good
speed persistence of vision makes the flashes in the plug caps appear as a
line of red dashes. There's quite a difference between the flicker rate of a
fixed lighting installation and one mounted on a moving vehicle.

 

[Victor Roberts]

State your source. Note that this thread has been linked to
sci.engr.lighting. We lighting experts know what you're saying is
incorrect. If all you're doing is throwing back information that LED
sales people have given you, then you and they are mistaken.


By what meter do you judge this? I see all manner of lights from
continuous sources to PWMs at 20 kHz or more. Once the pulse rate
increases beyond 5 Hz, it becomes very difficult to estimate a flash
rate by eye alone. I can't believe you're taking a light detector
hooked to an oscilloscope to determine a PWM pulse rate of an LED lamp
on a truck.


The point is that the eye is an integrator - it sees only the average
light output of sources pulsing at the frequencies we're discussing.
Same principle that allows people to see an image on the TV. Your
comment regarding strobe tubes incinerating is totally irrelevant to the
subject at hand. We're trying to educate you - you're trying to bully
your point through.

Peak intensity is only relevant for slow flash rates - below the '20 Hz'
you stated above. Warning lamps flash in the 1 - 4 Hz range where the
peak intensity does increase conspicuity (not visibility as is often
misused in this case as you are doing). Once all those flashes merge
into a continuous beam, conspicuity decreases.

Thanks Douglas. Your explanation is better than mine.

--
Vic Roberts
http://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

This information is provided for educational purposes only.
It may not be used in any publication or posted on any Web
site without written permission.

 

[Mark]

There may be some flicker or strobing present, I'm not sure, but I
suspect most of this is aliasing.

This phenomenon also exists on LED displays run on pure DC.

When you move your head, or chew potato chips, or play very low notes
on a bass trombone your eyeballs jiggle. Each time the eyes move you
lose vision for a small amount of time. This has the effect of
chopping the light into flashes even though the light may be steady.
It is not often noticable, but at night when the light source is small
it can be.

That's my theory, anyway. I'd go play my trombone now but it's all
full of potato chips.

No, I do not see this with incandescent lights and I do not see it when
the LED tail lights are on full bright when the brakes are on. The
effect is seen only on the LED tail-lights when they are being dimmed
(brakes off but tail light on) using a slow PWM.

Mark

 

[Don Klipstein]

Visibility is indeed higher since the peak emissions are higher.

Are you claiming that with pulse rate higher than the flicker fusion
frequency, the visibility is affected by the peak even when the average is
the same and the lamp appears to be glowing steadily?

The flicker also attracts attention better. I've seen LED lights on a
number of trucks where the flicker frequency had to be in the mid 20 Hz
range.

That part I agree with, except the pulse frequency for a lamp visibly
strobing only when moving being that low. I have played around with these
things enough, and it takes about 50 Hz or a little more for a pulsing
lamp to appear non-flickering. Sometimes a lamp will visibly flicker
at 60 Hz even if it is not moving.

- Don Klipstein ([email protected])

 

[Don Klipstein]

Drive higher = brighter = more visible


I've spent quite a bit of time in the film / video world so I'm a bit
sensitized to things in the 24-25-29.97 Hz range.


The phosphors in the CRT are what does the integrating for the most
part, or in newer technologies it's the drive logic that is holding the
pixel data.

Not true - ever roll your eyes while looking at a CRT monitor, a color
TV or a B&W TV?

Care to explain exactly what's wrong with higher drive = higher output =
brighter / more visible instead of claiming I'm wrong with nothing to
back that claim?

Peak intensity translates to visibility and low flash rates translate to
conspicuity. If it's brighter you can see it from further away and if
it's flashing slow enough to see the flash rate it grabs your attention.
They are not the same thing

Then why do LED tail/brake lights in "brake" mode and LED traffic lights
either operate continuously or with peak not a whole lot above the
average?

- Don Klipstein ([email protected])

 

[Don Klipstein]

Hello all,
On subject of Led's and strobes, I intend to build a high brightness white
Led strobe, but due to the required duty cycle of 20 uS, the white Led wont
do because of 'phosphor persistence' .

See if it is actually that bad. The amount of energy the phosphor over
an LED chip stores is so low and it emits at such a high rate (when the
LED is glowing at least) that I suspect it won't glow much for more than a
few microseconds after the LED turns off.

- Don Klipstein ([email protected])

 

[Don Klipstein]

I would agree there might be specific applications for using PWM to
boost output, but for automotive tail lamps it didn't seem to benefit.
The LEDs used in the stop/tail lamps were predominantly the lumileds
snapled 75 and snapled 150's early on and they have since moved towards
the luxeons and K2's. Also used in stop/tail lamps were the Osram
power topleds. The snapled 150's usually had to be driven near their
maximum rated current (150 ma) to get the desired output for stop mode.
The composition of the red and yellow dies made it so that the light
output fall off due to junction temperture rise is really quite bad.
The LED loses about 60% of it's light output at the rated junction
temperature of 25°C when you drive it at any reasonable current. So
pulsing didn't seem to get any output back because the duty cycle had
to be so long, that there was still significant die heating.

Which LED's were you using the PWM to boost output with? I haven't had
a chance to try that.

It's (or was back then) common practice with the digital displays of the
1980's, since the LEDs there were often run with just a couple milliamps
per die, and the particular dice in most of those had a nonlinearity where
they were much more efficient at 40 or 50 mA than at 2 mA or whatever.

- Don Klipstein ([email protected])

 

[Pete C.]

Not true - ever roll your eyes while looking at a CRT monitor, a color
TV or a B&W TV?

Plenty of times. As I said "for the most part". Phosphor persistence has
a lot to do with it. And of course now the pixel data is latched anyway
on the LCDs and whatnot.

Then why do LED tail/brake lights in "brake" mode and LED traffic lights
either operate continuously or with peak not a whole lot above the
average?

Dunno, I don't have any on hand to test or verify.

Pete C.