Calculating Candela Power for LED

[fsonnichsen]

TL;DR Summary

How is power specified

I am looking for the power in watts, total, put out by and LED. This for example, if I enclosed the complete device in an integrating sphere, what power output would be measured in watts. Unfortunately a lot of the vendors report power in Candela. I could not get a response on this, but I am now clear what is being meant in relation to an LED for example:

Suppose that the LED shows and output of 4.5 Candela and a half point angle of he beam at 8 degrees.

Does this mean that the 4.5 is spread through all directions for the LED (4*pi steradiains) as though it were measured in an integrating sphere----or----does this mean 4.5 within the lightbeam of the LED--e.g. as though measured with a light meter with sufficient diameter sensor.

Thanks all
Fritz

 

[alan123hk]

Candela is a unit of luminous intensity, not power. The input power to an LED is equal to its voltage multiplied by its current. The power efficiency of LEDs is very low, only a small fraction of the input power is converted into luminous power.

The candela of an LED generally varies with angle. The candela value given in the datasheet is maximum value. For details, please refer to the light intensity distribution diagrams at different angles in the manufacturer's specification.

If you want to accurately measure the total output lumens of LEDs and the distribution of emitted light intensity at all angles in space, you need to use a LED integrating sphere spectrometer system.

 

[alan123hk]

An example of measuring an LED with an integrating sphere gives you all the information you need.

 

[fsonnichsen]

Thanks Alan. I am versed in the many ways that electrical power, photon-flux, luminosity etc is expressed, and viewed from theory they are all quite simple- but my problem is knowing how the vendors are using the terms.

As you state-the Candela is a function of angle-and it is not clear what the vendor means when they just give a value, almost always without tables of data so I can do my own estimation. Usually I am capturing as much light from all angles as is reasonable for entry into a measurement device--the divergence of the beam of course can be difficult to guess.

I have the spheres, spectros etc to take lab measurements, but I am perusing datasheets for particular devices and need to make decisions without the lab work at first. I always like total optical power output for estimations (e.g. integrated) as usually the geometry of the LED and my optical constructs handle the geometrical concerns after approximation from datasheets.

Ultimately this value feeds the sensitivity of the spectro or PD that I am using for a design. Lumens, Luminosity, Candela etc have their purposes but I find them least useful for the kind of work I do. With the subsuming of the LED market by illumination use I see more of these type of specs.

Thanks again
Fritz

 

[anorlunda]

but my problem is knowing how the vendors are using the terms.

Didn't this give you the answer in post#2 ?

he candela value given in the datasheet is maximum value.

 

[fsonnichsen]

Yes--of course to get the overall power one would want the details of the curve---in other words I can draw several curves of differing power, all with the same max peak height
tnx
fritz

 

[tech99]

Wiki quotes 1/683 Watts per steradian. I wish the illumination people would just use the same terms as antenna engineers (Radiated Power in Watts, Efficiency, Directive Gain, Isotropically Effective Radiated Power, Power Flux Density etc), just include a factor for the eye response.

 

[alan123hk]

In fact, the general Led specifications have provided enough information for us to evaluate the actual luminous power output, so that we can roughly calculate the power conversion efficiency of the led. Those who are interested can try to calculate it by themselves.

We may find that the power conversion efficiency of LEDs may be only a few percent or even lower (I'm just guessing, I calculated it many years ago, but I can't remember the calculated value), but most people don't care about the actual power conversion efficiency, because the most important thing is the luminosity we feel with the naked eye.

https://docs.rs-online.com/02eb/0900766b814f533e.pdf
http://hyperphysics.phy-astr.gsu.edu/hbase/vision/efficacy.html#c1

 

[fsonnichsen]

I absolutely agree with tech99. In fact, what we need to know about light- the number of photons, the wavelengths and the directions can be used by a proficient electrical engineer or physicist to calculate all the measurable values of what I believe is more than 50+ specialized and often nebulous definitions for light.

Lumens in particular will evaporate with the human race but the photons will go on. They aren't even well defined--I think there are 2 (or more?) standard curves for lumens, in an IT document hidden behind a paywall where it is useless. About 20 years ago I built optical transmitters in the ocean-the best wavelength for these is around 520-540nm - the optimum transmission wavelength for seawater. It should come as no surprise that the optimum wavelength seen by humans is about 540nm as well. We did, afterall, crawl out of the water millions of years ago and most sea creatures today sense and transmit this wavelength as well.

Consequently the maximum of the "lumen curve" is at 540nm. As tech99 states this circles around the 1/683 W/sr value-but it is a finite curve-not a laser line-hence for accurate work a table or equation is needed. The "foot candle" and "candela" are right there with the lumen in quaint history-the names alone give one a hint that the derive from centuries ago in what is sometimes called the "oldest science".

Alan-the LED tables sometimes tell what we need to know--often more they do not. I am comparing LEDs for spectroscopy and need good values to compare their efficiency on a photodiode or CCD. And while I can convert-when I am looking a 50 tables of LEDs trying to find he one that I need--the whole art of consistency in engineering goes out the window.

In addition many important definitions are never made clear in the table-angle vs half-angle, 90% vs 50% knee in the output curve etc.

I think the move to lumens is representative of the major portion of the LED market being in human lighting. Spectroscopy and physics probably only concern a small number of us. The sad reality is that a major driving force of engineering physics here in the US at least--the use of commercial items--is always an afterthought and tenuous at best. In fact with the present IC shortages I am dealing with this becomes very clear.

Commercial LEDs are rarely in the wavelengths we need. In many of my designs I have to resort to laser-diodes or notch filters-----all very expensive when we are trying to make low cost instruments for science.

Thank you both of you for your responses!
Fritz