Maker Pro
Maker Pro

LED lights for filmmaking

  • Thread starter Daniel Kelly \(AKA Jack\)
  • Start date
I

Ian Stirling

Jan 1, 1970
0
Touch an ordinary filament torch bulb while it is working, and then one of
those new high brightness white LEDs. Notice the difference. Filament lamps
are about 5% efficient, which means that for every 100W of light, you have
to dissipate nearly a couple of kilowatts of heat. I'm not sure what the
efficiency of LEDs is, but torches that use them run for months, so it must
be greater.

Actually not.
LEDs are cooler because they have the heatsinks designed so they are cooler.
If they are not cooler, they stop working.
LEDs work very well at low powers.
Bulbs don't.

The current generation of white LEDs is about as efficient as a 500W
conventional floodlight bulb.
 
R

Richard Crowley

Jan 1, 1970
0
I recently saw an advert for some LED lights for filmmaking. They looked
perfect - very efficient, flicker-free, dimable from 0-100% etc. The
problem is that they're extortionately priced. So now I want to make my
own
LED lights for use on film...

Has anyone tried this?

From
...
white LED's?

"Watson A.Name - wrote ...
 
R

Richard Crowley

Jan 1, 1970
0
Just recieved in email from Video Systems Newswire...
(Additional online content at:
http://videosystems.com/mag/video_litepanels_lighting_kits/)

Will LEDs Light Your Productions?
Dan Ochiva

In Barry Braverman's review of the LitePanels Lighting Kit in September's
Video Systems, the author praised the company's LED-based on-camera light as
an ideal solution for DV'ers. Braverman said the far more efficient LED
device bested tungsten-based gear in delivering a dimmable, daylight color
value along with higher output than traditional on-camera lights.

There are great benefits, it turns out, if you dump today's tungsten and
halogen lighting technology. Those lamps heat their light producing metals
and gases in such an inefficient way that some refer to them as heating
devices that happen to give off light.

Another strike against tungsten and halogen lights? They need heavy
capacitors and transformers to change AC to DC and filter out the AC pulses.
Such power supplies--which must provide continuous, always on wattage--are
heavy, bulky, and expensive.

But LEDs are different. Light emitting diodes are solid state electronic
devices. They use a pulsed, switching power supply similar to the one in
your computer.

More benefits come from combining an LED and a switched power supply. A
microprocessor, for example, can be used to precisely dim an LED.

LEDs also just keep on working. Some companies claim up to 100,000 hours of
useful life for their LEDs, albeit in a laboratory setting.

Those are some of the reasons that make LED-based systems such as LitePanels
such a production breakthrough. But don't think you can slap one together
after a trip to Radio Shack. It's trickier than that. For example, to
achieve the light's useful 80 footcandle (or lumen) output, a LitePanel is
studded with 140 small, carefully positioned LEDs, with a microprocessor
controller onboard too.

This new approach to lighting, says Braverman, enables shooters "to take
control of our craft, and the extraordinary LitePanels system can help us do
precisely that."

But if the benefits are so great, why can't we use LEDs for production
lighting to replace those hot, clunky lights we use today? Complexity is one
reason. Remember, it takes 140 tiny LEDs precisely fitted into a 6.75 x
2.25-in. form factor to create a LitePanel.

To make a system bright enough to light a set would only result in a pricey,
complex, and fragile light. Heat too becomes a problem with that many LEDs.
While LitePanels' smaller light array doesn't generate much heat, placing
thousands of closely spaced LEDs together would create enough heat to melt
the whole thing.

But a new generation of super bright LEDs will soon change that. One of the
leaders in developing and manufacturing these new LEDs is San Jose,
Calif.-based Lumileds. The company says its Luxeon V Star LEDs are the most
powerful LED light sources available.

How bright? The company says a single Luxeon V Star lights delivers 37
lumens at 1 amp, drawing only 3W of power. (Stats are from the Lumileds web
site.)

Considering that the current LitePanel needs 140 LEDs to deliver 80
lumens...well you can see there's quite a jump in capability. The Luxeon is
specified for AC or DC 12V operation, which means one bulb can be used in
existing low voltage halogen lighting systems.

The brightness of Lumileds' Luxeon opens up new uses. In August, the company
announced that Sony chose Luxeon lamps for its top of the line Qualia 05 LCD
televisions. The lamps, installed in the Triluminous backlight, replace the
fluorescent lighting used by most LCD displays. The companies worked
together to develop a system suitable for the task, with results claimed to
deliver the highest color gamut. While conventional fluorescent/LCD
combinations reproduce only 65-75% of the NTSC color space, the Qualia 005
delivers a color gamut claimed to be 105% better than the NTSC color space.
(At press time, no one was available to explain that extra 5%.)

A smaller version of the lamp will even turn up in a new generation of
camera phones, due to roll out over these next few months. The two models
deliver either 40 or 80 lumens, said to be 12 times brighter than standard
mobile phone flashes.

Heat, however, becomes a problem when running this new generation of LEDs en
masse. "What heat?" you may ask. But these aren't the standard sort of LEDs;
those are meant to be looked at directly, which means they don't use much
power at all. These pop up everywhere, in signs, stoplights, and the front
panel of your stereo.

However, to light an object, LEDs need to be able to push through much more
power. While the lamp itself is only about 1mm in diameter, each uses
anywhere from 1 to 5 watts. Even a 1 watt LED, if not properly fitted into a
heat sink, gets hot.

In October, Westhampton, New Jersey-based Lamina Ceramics announced a
solution, one which will allow the creation of much larger light arrays that
could be useful in production.

Lamina doesn't make the high-power LEDs itself. Instead, this Sarnoff Labs
spin-off created a method to bond hundreds of LEDs to a single heat sink. It
works, to an extent. Their pioneering disk array design cuts some of the
output of the super bright LEDs, while sucking up much more energy than
expected. The cost is a shocker too: $4900, though that's not a bad price
for fixed installations used in theaters, for example, which will appreciate
the light's long lifetime.

That 5-in. diameter disk, though, delivers some 13,300 lumens. Wow.

You can get your hands on this new generation of LEDs, but not immediately.
LitePanels has a number of high-brightness lights nearing design completion.
This means that they're on schedule to release by NAB 2005.

The new on-camera/camcorder light will use only five high-output LEDs,
sourced from another company than Lumileds, says Steve Gillette, product
manager for the LitePanel.

"We're also planning a high-output IR (infra-red) model," says Gillette.
"For general production lighting, expect to see a 12-in. x 12-in. unit
that's only about 2-in. thick. You'll be able to interlock these to create
large grids, or mount them on a C-stand."

Gillette promises that the costs will be "affordable" and the heat bearable.

That's a relief. You might soon be able to fire that 'ol quartz lamp,
instead of firing it up.
 
T

Tim Shoppa

Jan 1, 1970
0
Richard Crowley said:
Just recieved in email from Video Systems Newswire...
(Additional online content at:
http://videosystems.com/mag/video_litepanels_lighting_kits/)
Another strike against tungsten and halogen lights? They need heavy
capacitors and transformers to change AC to DC and filter out the AC pulses.
Such power supplies--which must provide continuous, always on wattage--are
heavy, bulky, and expensive.

But LEDs are different. Light emitting diodes are solid state electronic
devices. They use a pulsed, switching power supply similar to the one in
your computer.

There's no fundamental reason why a switching power suppply couldn't
be used with incadescent lights. But they're missing what works so nicely
with LED's: PWM to control intensity with very little color shift
as intensity changes. And with RGB arrays and RGB sensors and feedback the
ability to make color shift from heating in the LED array a complete
non-issue.
More benefits come from combining an LED and a switched power supply. A
microprocessor, for example, can be used to precisely dim an LED.

Again, they're missing the fundamental point of PWM with LED sources.
After you take LED heating into account you need a feedback loop anyway
to regulate brightness to more than 30% precision (so you need more than just
a microprocessor, you need a sensor and a loop too) as the array heats
up and cools down.

Tim.
 
V

Victor Roberts

Jan 1, 1970
0
Actually not.
LEDs are cooler because they have the heatsinks designed so they are cooler.
If they are not cooler, they stop working.
LEDs work very well at low powers.
Bulbs don't.

The current generation of white LEDs is about as efficient as a 500W
conventional floodlight bulb.

They are actually not quite that good. If you use the minimum
guaranteed output instead of "typical" output and also correct for the
fact that the junction will be hotter than 25C even if the case is
held at 25C, (using the junction-to-case thermal resistance given by
the manufacturer) you get an efficacy of about 15 to 17 lm/W for the
best white LEDs available today.

--
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.
 
R

Richard Crowley

Jan 1, 1970
0
Tim Shoppa said:
There's no fundamental reason why a switching power suppply couldn't
be used with incadescent lights. But they're missing what works so nicely
with LED's: PWM to control intensity with very little color shift
as intensity changes. And with RGB arrays and RGB sensors and feedback
the
ability to make color shift from heating in the LED array a complete
non-issue.


Again, they're missing the fundamental point of PWM with LED sources.
After you take LED heating into account you need a feedback loop anyway
to regulate brightness to more than 30% precision (so you need more than
just
a microprocessor, you need a sensor and a loop too) as the array heats
up and cools down.

Agreed. They got a lot of the technical details wrong. Almost as if they
were parroting some of the phony marketing-speak of the vendors.
But I thought that the commercialization and especially lower-end
marketing of LED devices for illumination was the interesting part.
 
D

Don Klipstein

Jan 1, 1970
0
Touch an ordinary filament torch bulb while it is working, and then one of
those new high brightness white LEDs. Notice the difference. Filament lamps
are about 5% efficient, which means that for every 100W of light, you have
to dissipate nearly a couple of kilowatts of heat. I'm not sure what the
efficiency of LEDs is, but torches that use them run for months, so it must
be greater.

Filament lamps used for filming tend to be a lot more than 5% efficient
- try at least 10%. A USA "standard" 100 watt 120V lightbulb, known as an
A19, gets about 6.7%. High wattage halogens get a little more, and
photographic lamps with short life expectancy get as much as 12%.

Now for white LEDs: Really good ones get about 10%. I saw a couple
figures in Nichia's website consistent with about 12%.
As for LEDs of other colors: One manufacturer makes red ones that
achieve about 25 and orange ones that achieve about 20%%, otherwise 20%
for red and 10-15% for other colors is really good.

As for heat: Ever see how much heatsinking a 5 watt Luxeon needs?
One more thing: A majority of the energy going into a filament lamp
becomes infrared and usually escapes. With LEDs, nearly enough all energy
going in that fails to be converted to light (or wavelengths near those
the LED is intended to produce) becomes non-radiant heat that has to be
conducted out.

- Don Klipstein ([email protected])
 
I

Ian Stirling

Jan 1, 1970
0
In sci.engr.lighting Don Klipstein said:
As for heat: Ever see how much heatsinking a 5 watt Luxeon needs?
One more thing: A majority of the energy going into a filament lamp
becomes infrared and usually escapes. With LEDs, nearly enough all energy
going in that fails to be converted to light (or wavelengths near those
the LED is intended to produce) becomes non-radiant heat that has to be
conducted out.

Which is of course nice if you want a light with no infrared, but not
much else.
 
R

Roderick Stewart

Jan 1, 1970
0
Filament lamps used for filming tend to be a lot more than 5% efficient
- try at least 10%. A USA "standard" 100 watt 120V lightbulb, known as an
A19, gets about 6.7%. High wattage halogens get a little more, and
photographic lamps with short life expectancy get as much as 12%.

That's a lot? Even fluorescent lights are only about 40% efficient aren't
they? This is a great deal more than filament lamps, which is why they can
save money in the home, but the efficiency of light sources in general is
abyssmal, as it seems they all give out more heat than light.
Now for white LEDs: Really good ones get about 10%. I saw a couple
figures in Nichia's website consistent with about 12%.
As for LEDs of other colors: One manufacturer makes red ones that
achieve about 25 and orange ones that achieve about 20%%, otherwise 20%
for red and 10-15% for other colors is really good.

Thanks for this info. I had thought that LEDs were better than this, but
apparently not. There's certainly room for some serious improvement in this
area.

Rod.
 
T

Tony Morgan

Jan 1, 1970
0
In message <[email protected]>, Roderick
Thanks for this info. I had thought that LEDs were better than this,
but apparently not. There's certainly room for some serious improvement
in this area.

I've been quite puzzled at the (reported here) high powers of
current-technology LEDs

It's many years ago now, but the last time I encountered LEDs at the
sharp end was when the (then) high-powered LEDs first appeared. And to
use them we had to replace our front panels with 1/4" aluminium
precision drilled for a close LED fit to dissipate the heat. And this
was in an indication application, not that of illumination of a subject.
LEDs are, after all, semiconductors which are inherently prone to
destruction by heat at the junction. Worse, the polymer resin that
they're encapsulated in doesn't have very high thermal conductivity.
 
D

Don Klipstein

Jan 1, 1970
0
That's a lot? Even fluorescent lights are only about 40% efficient aren't
they? This is a great deal more than filament lamps, which is why they can
save money in the home, but the efficiency of light sources in general is
abyssmal, as it seems they all give out more heat than light.

I think 40% is high side in a good case for fluorescents, and I have
some doubts even then. I think more like 30% for a really good one that
produces white or whitish light.
However, fluorescents do have one other advantage for "overall luminous
efficacy": Phosphors can be selected with spectra better matched to the
spectral response of the human eye than the spectra of daylight,
incandescent, or "black body" radiation are. As one example, the reddish
phosphor in triphosphor fluorescents produces an orange-red wavelength
that the eye is more sensitive to than to most other red wavelengths.
As a result, if a given fluorescent lamp is 3 times as efficient as a
given incandescent lamp at converting electricity to light, then its
"overal luminous efficacy" (lumens out per watt in) can be close to 4
times that of the incandescent.

Yes, light sources are maddeningly inefficient.

LEDs are advancing, but hardly anyone expects them to surpass
fluorescent lamps in efficiency of producing white light anytime soon.

I have heard of efficiency well over 50% for an underpowered low
pressure sodium lamp in an oven - assuming you don't count the
energy/power used to heat the oven.
Low pressure mercury (as in a fluorescent lamp minus the phosphor) can
reach/exceed 50% efficiency in producing the shortwave UV that excites
fluorescent lamp phosphors, but the phosphors have a major loss.

- Don Klipstein ([email protected])
 
D

Don Klipstein

Jan 1, 1970
0
In message <[email protected]>, Roderick


I've been quite puzzled at the (reported here) high powers of
current-technology LEDs

It's many years ago now, but the last time I encountered LEDs at the
sharp end was when the (then) high-powered LEDs first appeared. And to
use them we had to replace our front panels with 1/4" aluminium
precision drilled for a close LED fit to dissipate the heat. And this
was in an indication application, not that of illumination of a subject.

When was this?

Sometime around 1986, GaAlAsP red LEDs came onto the market. Those got
plenty bright, almost bright enough to see in areas illuminated by direct
sunlight, at currents low enough to take no special measures to dissipate
heat from the LEDs.
GaAlAsP has had an additional breakthrough and some evolution
quadrupling its output and efficiency over the years into the mid 1990's.
In the mid or late 1990's, InGaAsP gave us even brighter orange-red and
orange LEDs and yelllow ones as bright as GaAlAsP in the same traditional
indicator lamp packages. InGaN was put into use in the late 1990's to
give us similarly bright blue, green and white LEDs. LEDs in packages
that indicator lamp LEDs were made with since the 1970's have been used in
flashlights for a few years now.

Lumileds brought forth the first really popular high current
heatsinkable LEDs just a few years ago. These had a maximum current of
350 mA and wattage of 1 to 1.2 watts. There are now higher power versions
up to 5 watts.
Cree, Nichia and others are now making 350 mA LEDs.

- Don Klipstein ([email protected])
 
C

crzndog

Jan 1, 1970
0
I think that you should be comparing apples against apples (ie lumens/watt
against lumens/watt).

Here's an article which talks about the efficiency of flourescent bulbs and
LEDs.

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

Basically it makes a general assumption that the efficiency of an LED is
about twice that of a flourescent bulb. What it fails to talk about is the
angle of light projection. LED's typically have a projection of 20 degrees.
That simply means that most of the light power is focussed in a narrow beam
(spot lighting). If you need to have a larger angle you will need more LEDs
to cover the angles. This means that you'll need more power to get the
equivalent lighting.

The upshot is that LED's are more efficient "when and only when" they are
used in spot lighting situations. Even then they are only about a 2 times
more efficient.

Also, the heat will still have to be dissipated (albeit half as much as a
flourescent light source).
 
C

crzndog

Jan 1, 1970
0
Sorry, I forgot to mention another "percieved" issue with LED light sources.

Due to the optical nature of the human eye, LED's appear to be brighter than
a broad beam light source (a bulb) because of the narrow beam nature of an
LED. What this does is saturate the rods in the back of the eyes and hence
the appear to be brighter. What you will notice though is that an LED will
leave a small spot on your retina (when you close your eyes you'll see the
negative image) whereas a bulb will leave a larger spot.

Basically the angular concentration of light is higher for an LED and as a
result it appears brighter.
 
B

Bruce Murphy

Jan 1, 1970
0
crzndog said:
Sorry, I forgot to mention another "percieved" issue with LED light sources.

Entirely appropriate quotes, as it happens.
Due to the optical nature of the human eye, LED's appear to be brighter than
a broad beam light source (a bulb) because of the narrow beam nature of an
LED. What this does is saturate the rods in the back of the eyes and hence
the appear to be brighter.

Psycho-optical mumbo-jumbo. How can an effect that *prevents you
perceiving the full brightness of something* cause it to appear
brighter?
What you will notice though is that an LED will
leave a small spot on your retina (when you close your eyes you'll see the
negative image) whereas a bulb will leave a larger spot.

An effect which does not require saturation.
Basically the angular concentration of light is higher for an LED and as a
result it appears brighter.

*raised eyebrow*

Unless you're extremely close to the light, this has nothing to do
with the narrow-beam nature of the LED. You may be talking about size
of the emitting area, but a lot of small bulbs have filaments similar
inside to a typical LED, even before you add the magnifying lens on
the front of most LEDs.

B>
 
D

Don Klipstein

Jan 1, 1970
0
I think that you should be comparing apples against apples (ie lumens/watt
against lumens/watt).

Here's an article which talks about the efficiency of flourescent bulbs and
LEDs.

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

Basically it makes a general assumption that the efficiency of an LED is
about twice that of a flourescent bulb.

I don't claim that, athough I believe someone else does or some other
people do.

Really good white LEDs are nowadays about twice as efficient as many
incandescents, but have efficiency (or more appropriately luminous
efficacy) still not exceeding that of short-life photographic lamps.
What it fails to talk about is the
angle of light projection. LED's typically have a projection of 20 degrees.

Or 15, or 30, or 45. Some have 60. Some have 110-120 or so.

Some are really narrow - like less than 15!
That simply means that most of the light power is focussed in a narrow beam
(spot lighting).

There are plenty of incandescents that also do this!
If you need to have a larger angle you will need more LEDs
to cover the angles. This means that you'll need more power to get the
equivalent lighting.

The upshot is that LED's are more efficient "when and only when" they are
used in spot lighting situations. Even then they are only about a 2 times
more efficient.

Also, the heat will still have to be dissipated (albeit half as much as a
flourescent light source).

Heat dissipation from LED high power light sources is often a major
problem, since they do not withstand high temperatures as well as
incandescents do. And incandescents can toss away about or somewhat more
than half their wasted power input as infrared, while with LEDs nearly all
energy that does not become desired output becomes heat that materializes
at the lamp(s)!

- Don Klipstein ([email protected])
 
D

Don Klipstein

Jan 1, 1970
0
Sorry, I forgot to mention another "percieved" issue with LED light sources.

Due to the optical nature of the human eye, LED's appear to be brighter than
a broad beam light source (a bulb) because of the narrow beam nature of an
LED.

Actually, the reason is small light source size.

Consider a smallish welding arc the size of a smaller-than-average pea
which may produce less light than a fluorescent fixture with four 4-foot
lamps ("bulbs").

What this does is saturate the rods in the back of the eyes and hence
the appear to be brighter.

Separate issue - in favor of bluer light sources including most white
LEDs over "warmer color" light sources.
Wavelengths from mid-blue to mid-green (roughly 460-530 nm) are good for
stimulating rods. Light with more of these wavelengths can have extra
perceived "illuminating power" compared to light that has less of these
wavelengths. Such an effect is greater in peripheral vision than in
central vision, and is greater in dimmer illumination conditions where the
function of the rods is more significant and less in brighter illumination
conditions.
This also makes cooler color compact fluorescents look better - but for
room lighting illumination at lower level with cooler color easily appears
"dreary gray".

- Don Klipstein ([email protected])
 
D

Dave Plowman (News)

Jan 1, 1970
0
LED's typically have a projection of 20 degrees.

Think that's rather old. I've been using 60 degree ones to replace
indicator lamps in switches that were forever blowing.
 
Top