quantafoto wrote:
I've been thinking of building a camera flash powered by white LEDs. The LED's are typically rated in lumens, so I'm wondering how many lumens I'd need to use to get a brightness comparable to a standard flash? I know the terminology isn't exact, but i think you get what I mean.

Thanks!

Davey

I think you're looking it the wrong way. Think of LED like continuous light. check these out http://www.bhphotovideo.com/c/product/6 … 5600K.html

http://www.bhphotovideo.com/c/product/4 … t_for.html

http://www.bhphotovideo.com/c/product/3 … t_LED.html

quantafoto wrote:
I've been thinking of building a camera flash powered by white LEDs. The LED's are typically rated in lumens, so I'm wondering how many lumens I'd need to use to get a brightness comparable to a standard flash? I know the terminology isn't exact, but i think you get what I mean.

Thanks!

Davey

A "standard" flash will put out something like a million lumens, which is probably not the answer you were looking for. However it does this for a very short time, typically one millisecond or sometimes very much less. What you need to know is the total light output in "lumen seconds" (more correctly but rarely called Talbots.) In other words although the flash can output one megalumen, its total light output is typically 1000 lumen seconds or 1Kl-s. So it's not just the lumens you have to take into account, it's the flash duration, but if you multiply the lumens by the duration and get a figure of around 1000 you will be approximately in the right region.

PE Art Photography wrote:  in order to get the iso down to 100, and the fstop closed enough for 1 foot of DOF at 50mm  (6 feet away from model) or f8, you need about 3000 watts of continious lights

This calculation is ludicrous without a shutter speed.  ISO 100 is not necessary these days, though it might be desirable with some of the halfway MF digital backs.  ISO 400 is excellent for DSLRs and even 800 can be very good on the latest DSLRs.

Dan Lee Photo wrote:
Anyway, just DIY your own Xenon flash, thats what yuore looking for, you can use more than one xenon tube if you want more light output.

Be careful about this one: flash tubes need high DC voltage to operate - these high voltage power supplies can badly injure or kill you if not designed or handled correctly.

John Prentice wrote:
This is a project that I had worked on quite extensively with an electronics engineer. We tried several different ways of approaching the problem, but unfortunately with LED technology as it stands right now, they will only be good for a continuous lighting scenario. You simply cannot get enough power out of LEDs to be anywhere near cost effective, and the diodes we were using were over $80 apiece - The brightest that were on the market last year.

I won't go into too much detail, but there are some major things you're going to need to consider in pursuing this project. Firstly, heat. When dealing with high powered LEDs, you are going to need considerable heatsinks to prevent them from self destructing - multiply this by a large number of LEDs, and then consider that you're overloading the LEDs to squeeze every last bit of brightness out of them. It's just not feasible right now.

As an electronics engineer, it's a headache. What a flashtube does is hypothetically an explosion. It's an arc of electricity in a glass tube for an extremely short time. About the closest thing is a dirac-delta function, an 'infinite' amount of energy for an 'instant'.

Anything else 'non explosive' will take time to heat up, produce light, and then turn off. This can take many milliseconds to be effective. WAAY too long. And even then, because as previously mentioned you still have to produce approximately one megalumen. That's...a lot...and won't happen with LEDs for a long long time. They're already having trouble cooling 10,000mcd ones.

The best way to produce a 'flash' (an explosion, or arc) is with a high voltage and suddenly discharge it through something. If you drop a piece of metal on a large enough capacitor, it will vapourise and produce an almighty flash. Imagine what happens if you drop an LED (in the 2.7->3.7V operating range) onto a capacitor charged to 500V. Prepare to be showered in silicon/plastic/metal.

ImagesByEd wrote:
http://www.philipslumileds.com/pdfs/DR01.PDF

True, but 53 lux at 1m...they have MILES to go.

Liam G Martin wrote:

True, but 53 lux at 1m...they have MILES to go.

If you are talking about 1, then yes. They also have other offerings that are, I think, even more interesting. Another aspect is that an array of LEDs would not really have to match the output of more traditional devices because the higher ISO of digital cameras offers pretty decent noise performance. The trade off is very light weight systems, low power supply requirements (thinking of portable use here), ruggedness, small size, amazing recycle times, etc.

Managing Light wrote:

Be careful about this one: flash tubes need high DC voltage to operate - these high voltage power supplies can badly injure or kill you if not designed or handled correctly.

Current kills, not voltage.

Dan Lee Photo wrote:
Current kills, not voltage.

Actually, the mixture of both is the worst. High voltage is required to break the skin's natural zirconic effect/capacitance, and the current path to complete the circuit.

That's why charged capacitors are the worst. They can supply tens/hundreds of amps for a tiny period of time from a very high voltage.

Neil Snape wrote:
There are plenty of LED lights on the market. I have seen wonderful lights and accessories at PMA, Photokina etc.

You could build your own but be aware that the color temperature is a product of the range of light waves and spectrum produced by the LEDs.

The white LEDs are actually fluorescents. They are coated with a phosphor layer, just like the inside of fluorescent tubes, which glows when excited by the blue or UV from the LED. That (at present) is the only way of getting anything even resembling a continuous spectrum from LEDs. If you used LED light alone you would get a spectrum three or four very narrow spikes which would give peculiar colours. And since phosphors react very slowly, even if the LEDs themselves could be made fast, this would limit their usefulness for flash.

quantafoto wrote:
I knew we'd get side-tracked by conflicting units! :-)

So lets ask it another way. Forget the flash. Let's say I want to shoot a portrait from ~15 feet, 1/30 sec, f3.5 how many lumens would I need?

It seems if the little LED in my cellphone can work as a flash it can't be too far out of the ballpark to expect something practical from high intensity LEDs.

For other concerns mentioned, it would be something of a middle ground between strobe and continuous lighting. Would be longer duration than a xenon flash but not on continuously so as to minimize battery drain and heat buildup. At most it would only be on for the time the shutter was open (say 1/30 sec.) If i got fancy, might have a 'preview' continuous on feature, or on continuous at 50% brightness and then 'strobed' at 100% for the shutter duration.

As far as brightness/cost, here's an LED assembly that produces 540 lumens for $28

http://www.luxeonstar.com/premounted-re … -p-647.php

It'll depend on the field of view of your camera, and the optics attached to your light source. Lumens measure luminous flux, rather than brightness. If you find your field of view, and angle of projection, you should be able to calculate how many lumens you will need, though. I would suggest, however, that you would be better off thinking in terms of energy than illuminance.

quantafoto wrote:
I knew we'd get side-tracked by conflicting units! :-)

So lets ask it another way. Forget the flash. Let's say I want to shoot a portrait from ~15 feet, 1/30 sec, f3.5 how many lumens would I need?

It seems if the little LED in my cellphone can work as a flash it can't be too far out of the ballpark to expect something practical from high intensity LEDs.

For other concerns mentioned, it would be something of a middle ground between strobe and continuous lighting. Would be longer duration than a xenon flash but not on continuously so as to minimize battery drain and heat buildup. At most it would only be on for the time the shutter was open (say 1/30 sec.) If i got fancy, might have a 'preview' continuous on feature, or on continuous at 50% brightness and then 'strobed' at 100% for the shutter duration.

If I get a chance, I'll try to take some actual test shots for you, with values.  As I say, this isn't theoretical for me, I have this stuff already built. 

You are right to think of it as a middle ground between xenon flash and continuous.  For instance, a 10ms pulse will give you a lot more light than a 1ms pulse, but is still relatively short as far as shutter speed is concerned.

ImagesByEd wrote:
The Canon 580 EX II manual gives a flash duration time (I'm assuming at full power) of 1.2 ms. LEDs can cycle at some pretty amazing rates; however, as another poster (correctly) pointed out, the white LEDs use phosphor which is quite a bit slower. That being said, Stocker-Yale's claims you can pulse the white LEDs at 5K/Sec before seeing any color shift – which, if I am not mistaken (I've had a few drinks), is .2 ms – If, in my admittedly buzzed state I am correct, it would appear that LEDS can easily outperform traditional technology (at least in terms of the 580EX II).


http://www.stockeryale.com/i/leds/lit/app001.htm



Note to self: Step away from the keyboard!

pulse width modulation is what you use for dimming white LEDs to prevent colour shift.

quantafoto wrote:
I've been thinking of building a camera flash powered by white LEDs. The LED's are typically rated in lumens, so I'm wondering how many lumens I'd need to use to get a brightness comparable to a standard flash? I know the terminology isn't exact, but i think you get what I mean.

Thanks!

Davey

A 640Ws Einstein studio puts out 28,000 lumen seconds with a flash duration of about 1/500 second.

This translates to an average light output of about 14,000,000 lumens during that 1/500 second.

If your LEDs put out 160 Lumens per Watt, then you need about 8750 Watts of LED power.  This is 8750 actual watts of power, not "equivalent" Watts.


Of course, if you want to match a lower power strobe, or one with a longer flash duration, you can use less LED power.

A good hot shoe mounted flash (like a Canon 580ex) puts out about a tenth of this; perhaps 1,500,000 lumens for about a thousandth of a second.

CTphotowork wrote:
It is possible. And LEDs would be fast enough. LEDs are semiconductor components and can turn on and off very fast. Think of fiber optics communications. Very fast switching of on and off.

These are big chips, with large parasitic capacitances: their turnon and turnoff times are much longer than that of the communications LEDs that are used for gigabit fiber systems. 

Further, it was mentioned by WMcK earlier that the white-light LEDs are using a blue- or UV-excited phosphor to create the white light.  The fluorescense risetimes and falltimes are slow.

CTphotowork wrote:
In fact, you can over drive the LEDs at much higher current to get more light out and it will be ok because it will not even reach it's critical temperature in that short burst.

This is generally true if the specifications for the LED chips allows very high peak-to-average current ratios.  Unfortunately, the specs for the white illumination LEDs that I've read (roughly 6 months ago) have virtually the same ratings for I(peak) and I(ave).  And phosphors tend to saturate at high illumination levels to further limit the peak-to-ave ratings.

As someone said earlier, we're not there yet.  And when we start to get close, flash manufacturers and companies that want to become flash manufacturers will be all over it like white on rice.

Managing Light wrote:

CTphotowork wrote:
It is possible. And LEDs would be fast enough. LEDs are semiconductor components and can turn on and off very fast. Think of fiber optics communications. Very fast switching of on and off.

These are big chips, with large parasitic capacitances: their turnon and turnoff times are much longer than that of the communications LEDs that are used for gigabit fiber systems. 

Further, it was mentioned by WMcK earlier that the white-light LEDs are using a blue- or UV-excited phosphor to create the white light.  The fluorescense risetimes and falltimes are slow.


This is generally true if the specifications for the LED chips allows very high peak-to-average current ratios.  Unfortunately, the specs for the white illumination LEDs that I've read (roughly 6 months ago) have virtually the same ratings for I(peak) and I(ave).  And phosphors tend to saturate at high illumination levels to further limit the peak-to-ave ratings.

As someone said earlier, we're not there yet.  And when we start to get close, flash manufacturers and companies that want to become flash manufacturers will be all over it like white on rice.

and by "earlier" we are talking 3 years ago. this was a 2010 thread.  not sure why it was revived without significant new technology to change things?

Paul Morgan Photography wrote:
Am I the only one who noticed this thread is from 2010???

clearly you did not notice the post directly above yours