Comparing apples to apples, at last...

In the vast background noise written on the subject of LED vs other light sources comparison, this post is a little gem. James Alexander makes an excellent job at summarizing the issue with easy to understand day to day examples. Moreover, the cited numbers are among the most up-to-date at this point in time.

Although I disagree with him on the subjective matter of using 5 mm LEDs as a "real ligh" source, this is definitively worth reading.

Beyond the love of our own crap

effective design is driven by insight, strategy and purpose.

In this time of induced uber-individualism, our day to day has been invaded by herds of people who “trust their own style”. What a perfect risk avoiding position, as anyone can both be author and critic of its own creation! How equally rewarding for the would be guru which can make a living of celebrating any new fad of fashion, art or design.

Submerged by layers of styles, pressured by their peers, subjected to the average crowd demand, how could a vast majority of “creative people” resist morphing into “stylists” and reiterating whatever the latest fashion happens to be. And the rest maybe tempted to evolve into “creative geniuses”...

We all fall in love with our own creation. This has probably something to do with the lovingly instilled early suggestions that all our expressions of creativity must be deemed significant. Nevertheless, these expressions can still suck.

Coming up with design solutions effective at reaching a broad audience is in no way less difficult or creative than making work that is personal in nature. In fact, as it requires keeping personal perspectives in the distance, in an effort to understand the situation from a wider standpoint, it soon becomes much more challenging,

Clients and designers alike fall into the trap of bringing personal aesthetics or style to projects. As that has usually nothing to do with the task at hand, the result is an enormous quantity of pretty but ineffective “design” out there.

Like for every other business, I believe effective design must be focused on results. Although it can use a multitude of instruments, it must not be personal in nature, nor follow degradable styles, unless this is really necessary. Effective design is driven by insight, strategy and purpose. It is about taking away the stylish and fashionable stuff in favor of what is actually accomplished.

Heat is the great enemy of LEDs

Heat is the great enemy of Light emitting diodes (LEDs). This is ironic, since conventional bulbs produce light by heating a filament to such a high temperature that it glows. The most common way for LEDs to fail is by gradual decrease of light output and loss of efficiency. However, sudden failures can occur as well. All caused by excessive heat.

Driving power LEDs on constant current
LEDs use a different principle than incandescent or fluorescent sources to create light. LEDs are semi-conductors diodes that emit light when traversed by a current flow. LED diodes have polarity and, therefore, current only flows in one direction. A photo emission is taking place at the diode junction region when a DC low-voltage, constant current power is applied. Driving power LEDs is relatively simple as, unlike fluorescent or discharge lamps, they do not require an ignition voltage to start. Simply put, too little current and voltage will result in little or no light, and too much current and voltage can damage the light emitting junction of the LED. Consequently, to ensure a proper functioning of a LED light source we need some sort of power supply regulation.

When looking at a typical power LED forward voltage vs. forward current chart, we clearly see that, for a given junction temperature, a small variation of the forward voltage produces a large variation in the forward current. Conversely, as the junction temperature increases, the forward voltage across the LED drops as depicted on the forward voltage vs. junction temperature chart.

If we drive power LED light sources with a regulated constant voltage power supply, the forward current passing through the LED will increases as a result of a forward voltage drop, and in turn will generates additional heat in the junction. Ultimately, if nothing limits the current, the LED junction will fail by over heating.

Instead, by driving power LED light sources with a regulated constant current power supply, the light output and lifetime issues resulting from variation of the forward voltage can be eliminated.

Driving power LEDs for clean light
Luminous characteristics of power LEDs are specified for a specific forward current and a 25°C junction temperature. However most LED light sources are operated well above 25°C, and the “true” light output should be based on the anticipated operating junction temperatures.

As illustrated on the relative luminous flux vs. forward current chart, the light output of increases when the forward current increases. However the efficacy of the light source, expressed in lumens per watt, is adversely affected. Conversaly, the light output from LED light sources decrease with increasing junction temperature, as depicted on the relative luminous flux vs. junction temperature chart.

Therefore, when designing for specific light output, efficacy levels, wavelengths or color temperature, it is important to consider the effects of temperature and to maximize the thermal management of the application.

For a specific LED light source, the forward current may be chosen up to the maximum current recommended by the manufacturer. Driving LED light sources above that maximum may result in lower lumen maintenance or, with excessive currents, catastrophic failure.

High forward currents at elevated temperatures can cause diffusion of metal atoms from the electrodes into the junction’s active region, decreasing the radiative capacity through the creation of dislocations and point defects that produce heat instead of light. High-power LEDs are susceptible to current crowding, non homogenous distribution of the current density over the junction. This may lead to creation of hot spots in the junction, and increases the risk of thermal runaway.

When the epoxy resin used in packaging reaches its glass transition temperature, it starts to expand rapidly, causing mechanical stresses on the semiconductor and the bonded contact, weakening it or even tearing it off.

Higher junction temperatures resulting from increased power dissipation or changes in ambient temperature can have a significant effect on light output. Red and Amber AlInGaP phosphors are more sensitive to temperature effects than Blue and Green InGaN phosphors. Depending on the phosphor type, wavelengths can typically increase from 0.03 to 0.13nm/°C. White LEDs often use one or more phosphors. The phosphors tend to degrade with heat and age, losing efficiency and causing changes in the produced light color and slight shifts in color temperature. Similarly, some materials of the plastic package tend to yellow when subjected to heat, causing partial absorption, and therefore loss of efficiency, of the affected wavelengths.

Appreciating LED useful life time

With Light emitting diodes (LEDs), outright failure is very unlikely. Contrary to conventional lighting sources which typically fail suddenly or burnout, LEDs are solid state electronic components and as such gradually degrade. But because of their long expected lifetimes, conventional light sources’ life testing is impractical to estimate the useful life of LEDs.

Predicted Life Time
Useful life of conventional lighting sources is commonly expressed as the time to failure. This primary metric is based on the time it takes for 50% of lamps to fail. LEDs, being replaceable semiconductors, are using “mean time to failure” (MTTF) to express their failure rate. This is a well-defined statistical reliability metric commonly used in the electronics industry.
Power LED manufacturers typically predict high brightness LED MTTF to be on the order of 50.000 - 100,000 hours, provided LEDs “are properly packaged and used in accordance with manufacturers’ recommendations”.

Translating these durations in plain English, a LED light source would have an average life time between 5 and 12 years if left on all day. For “normal” general lighting usage, even when considering a 12 hours average daily usage, this would translate into an average life time comprised between 12 and 24 years. A very long life times indeed.

Average Lumen Maintenance
Even when operated within the manufacturers’ specifications, both conventional and LED light sources experience loss of light over time. This is known as lumen depreciation, and is typically expressed as lumen maintenance, i.e. the percentage of initial lumens remaining after a specified period of time.
If you have ever changed a light bulb, you have certainly noticed how bright the new bulb is compared to the older bulb, then you have seen the effects of lumen depreciation. Lumen depreciation in incandescent lamps mainly occurs by depletion of the filament over time and accumulation of evaporated tungsten particles on the bulb wall. In fluorescent lamps, it occurs by photochemical degradation of the phosphor coating and the glass tube, and by accumulation of light-absorbing deposits within the lamp over time.

LEDs also experience lumen depreciation, but many factors can influence light degradation, such as ambient temperature and humidity, drive current and thermal management. That said, the primary cause is heat generated at the LED junction. High junction temperatures accelerate degradation in lumen maintenance, but also result in a temporary reduction in luminous flux. Contrary to other light sources, LEDs do not emit heat as infrared radiation, so the heat must be removed from the component by conduction or convection. If a LED application has inadequate means of removing the heat, such as heat sinking, the temperature will rise and light output will decrease.

Let’s translate this into plain English. For general ambient lighting applications, the commonly admitted light output decline is set at 70% of initial lumens. If for example we are considering LEDs that deliver 70% lumen maintenance at a 50,000 hours rated life, it means we should expect to receive 70% of the initial lumens after 50,000 hours.

Lumen depreciation has to be taken into account when appreciating the “useful” life of very long life components such as LEDs. In short, the “useful” life time of a LED results from a combination of the MTTF and the lumen degradation. From an application perspective, a catastrophic failure of the semiconductor or the lumen performance falling below 70% always boils down to a degraded service. Because of the LEDs very high MTTF, LED applications are more likely to falter because of lumen degradation, which in turn is highly dependent on the appropriate thermal management.

No retrofitting please!

Early attempts to apply LEDs to general lighting failed because LEDs did not meet the required luminous efficacy or color requirements. Technology has reached a point where using LEDs for general illumination is now a viable option. But, taking over where I left, I believe trying to repurpose existing technologies' lighting fixtures to house LEDs is inappropriate.

LEDs represent a disruptive innovation for the lighting industry. A disruptive innovation is technologically straightforward, using off-the-shelf components put together in a product architecture that is often simpler than previous approaches. These products are usually less capable in the traditional aspects of what is required in established markets, but feature different bundles of characteristics that were not considered important in the past. Applied to LED technology one may think of energy efficiency, resistance to vibration and unidirectional luminous flux, to cite a few.

Conventional approaches to developing power LEDs based general lighting often involve retrofitting existing fixtures to house the new technology. Many early attempts simply used traditional lighting standards and housings instead of investigating the challenges and benefits of LEDs. But LED technology breaks traditional rules, and it quickly become apparent that old thinking cannot be applied. A LED module may physically fit into an existing fixture’s housing, but that housing will not leverage the inherent qualities of power LEDs, mainly because:

• standard housings do not handle the challenges of LED thermal management, which is vastly different from those of incandescent or fluorescent lighting.
• optical design used in most traditional fixtures does not maximize the LEDs' efficacy.

Furthermore, power LEDs last a very long time, and the expectations for fixtures' life span are getting higher. Typical specifications for LED lighting fixtures tend toward more durable, longer-lasting products using higher quality materials than those commonly associated with other lighting sources. A state of the art LED module in a cheap fixture would defeats the purpose.

A disruptive innovation is not the same as a radical innovation. A radical innovation is just a major improvement along an existing performance dimension. A disruptive innovation creates a different performance dimension, one that is not particularly important to incumbent firms' most profitable customers. Let’s hope lighting industry experts accept the change and gain a better understanding of how to capitalize on that technology.

The incandescent frame of mind

There is one recurrent bias in many writings about the coming of age of LEDs for general lighting: the author always assumes the only allowed form factors for LEDs to become successful have to be similar to those of the other light sources technologies we know today. Or to put it more simply, that LEDs are only going to succeed in the “retrofit” market. I take as example this attempt at describing how LED based fixtures are now able to compete with traditional general lighting fixtures.

Beyond the fact that the post speaks a little too much about one single manufacturer to be exhaustive, and try to draw general conclusions based on this narrow view point, I believe it gives a wrong perception of the issues at stake by providing the wrong examples. This is particularly true of the cost based examples. But I am ready to assume this has been caused by jet lag and "red eyes"…

Let’s examine the "much work remains to be done to get the costs down" example. First of all it sounds based on that ubiquitous chart found all over the web comparing luminous efficacy of light sources that everyone makes its own. Let’s say it makes me suspicious when I see a comparison using a "800 lumens" figure as a basis for calculation… But it is more the peremptory conclusion that "cost" must be driven down that makes me uneasy.

After all it is nor fair nor possible to compare the price of an incandescent bulb with a “15-20 watt LED”. Doing so we are not comparing apples to apples, but rather commodities with luxury goods, and ultimately we only propagate marketing BS.

As a matter of fact, the "cost" of this 15-20 watt LED light source is a retail price. Knowing that at semi-gross prices this same light source is coming out at around $27 (without volume discount), the only valid conclusion one can draw from the example is that the retailer is taking a three fold markup. And this cannot be called the cost of LEDs.

Moving on to the “kitchen cans” example, we can repeat the same calculation. This time taking into account the additional elements necessary to create a retrofit light source, namely a constant current driver ($5), an anodized extruded heat sink ($2), an optic ($3), an outer fascia ($1) and fixtures ($1). We are now approaching the $40 for the bill of material. The resulting product will probably be assembled in China, adding another $5 to the cost. Once again the example only demonstrate the three fold margin applied by the retailer on the product.

However, I agree with the author that “only the most elite pocketbooks will open for LEDs” at this price. Like they did for incandescent bulbs when electricity lighting took over from gas lighting… But what a cumbersome way to say that most retailers are pricing LED retrofit light sources as luxury items! Hasn’t it always be the case with emerging technologies?

Nonetheless, to return where I started, I believe the author missed an important point. Power LEDs are a disruptive innovation and trying to mold them into the previous technologies form factors is flawed. LED general lighting will really take off when designers, writers and customers alike will step out of their “incandescent” frame of mind.

LED or Luminous Efficacy Demystified

Efficacy, power and costs of high power LEDs have now reached levels that make them attractive in general lighting applications.

The luminous efficacy of a light source is the ratio between the emitted luminous flux and the amount of the absorbed energy to transmit it. It is expressed in lumen/watt (lm/W), where the lumen is the measure unit of the luminous flux. However, the lumen is based on the subjective perception of light by the average human vision, corresponding to a particular curve inside the visible spectrum. To put it simply, a standard incandescent light emits radiation both inside and outside the visible spectrum. The radiations emitted in the infrared and in the ultraviolet do not contribute to our perception of brightness. A light source will have a higher luminous efficacy as much as it will be able to emit in a spectrum suitable for the human vision.

The LED technology has made significant progress on the emitted power front. Today standard high power LED are available in 1, 3 and 5 watt, and multi-chip emitters are becoming available that push LED power towards the 15 watt. But above all high power LED have ten times more efficacy than the incandescent sources. Every serious manufacturer offers high brightness LEDs with minimum luminous efficacies well above 80 lm/W. Nowadays 100 lm/W minimum luminous efficacies are quickly becoming the norm for white light.

To quickly illustrate the point, a 60 watt standard light bulb, with a source efficacy of 15 lm/W, produces a luminous flux of 60 x 15 = 900 lumens. A light source built with nine 1 watt high power LEDs, with a luminous efficacy of 100 lm/W will produce the same luminous flux, but will use only a power of 9 watt instead of the 60 watt of the standard light bulb.

For a more accurate comparison between different lighting systems we must take into account the entire system that produces the luminous flux:

• Source luminous efficacy (lm/W): it is the primary luminous characteristic of the light source, and varies according to the given technology.
• Electrical efficiency (%): it defines the incurred losses to adapt the standard electrical source to the need of the considered technology. For example, incandescent lamps are directly connected to the power line without any adjustment. This is not the case for other technologies such as the fluorescent lamps, which require ballasts with 60-70% efficiency. Similarly, an inverter for compact fluorescent lamps has an efficiency of 80-90%, while an AC/DC driver suitable for LEDs can have an efficiency higher than 90%.
• Fixture efficiency (%): standard incandescent and fluorescent light sources radiate in almost all directions and require reflectors and diffusers to shape the light beam for the required application. These systems have an efficiency which is usually estimated between 30 and 50%. By comparison, light emissions from LEDs are inherently directional, and 95% efficiency can be assumed.

Taking into account the entire chain, the efficacy of incandescent lamps will be as low as 7 lm/W, whereas fluorescent will achieve 38 lm/W and high power LEDs reach 76 lm/W. A better evaluation of the electrical power necessary to produce the 900 lumens would give 128 watts in the case of incandescent lamps, 23.8 watt for fluorescent fixtures and 11.8 watts for LEDs. A significant energy saving!

Evolving view points for an evolving technology

To me a great majority of blog posts talking about LEDs are frankly disappointing. Although some try to take an exhaustive approach at presenting the technology in terms that can be understood by anyone, their content often remains static and academic, often copied from the same source. As these blogs are likely to come on top of the search engines responses, the casual reader may end up drawing the wrong conclusions with regards to the usability of the LEDs technology. Obviously at this point in time it would be biased to present LEDs as “the” perfect, “do-it-all” technology for general lighting. But this is a fast evolving technology, and, as for computer hardware, Moore’s law is applicable: today’s “truth” may become obsolete within a year or two. As a result, one has to remain cautious when presenting LEDs technology strengths and weaknesses.

In general lighting applications LEDs have advantages and disadvantages when compared with other light sources such as incandescent or fluorescent lamps. When looking at the positive side, the most significant advantages are fast turn-on, lower heat generation, lower power consumption, higher operating life, and high resistance to shock or vibration.

On the negative side, many blog have not been updated and retain obsolete information which may induce the casual reader into drawing hasty conclusions. Amongst the recurring limitations these blogs describe are the narrow viewing angle, and the need for electronic driver circuits to operate.

Starting by the later, LEDs need to be driven properly to ensure optimal performance and long life. An effective driver is key in obtaining all the benefits of LEDs. If early driver's implementations made of discrete components were not cost effective, this is not the case anymore. Pushed by the fast adoption of LEDs in the automotive and aeronautical industries, today almost every integrated circuit manufacturer proposes a vast array of LED drivers to suit almost every aspect of general lighting requirements. Furthermore, the leading manufacturers’ constant current sources, which until recently were only able to drive a limited number (usually 3 to 6) of power LEDs have recently been superseded by new affordable sources with 3 to 4 times more capacity.

Let’s now look at the viewing angle. First of all, light emissions from LEDs are inherently directional, thus reducing the need for reflectors and diffusers that can trap light. As a result, general lighting LEDs fixtures can potentially deliver light more efficiently to an intended location, leading to potentially higher application efficiency than other light sources in certain lighting applications. By comparison, fluorescent and incandescent lamps emit light in all directions. In their case, much of the light produced by the lamp is lost within the lighting fixture, reabsorbed by the lamp, or escapes from the fixture in a direction that is not useful for the application. For many fixture types, it is not uncommon for 40-50% of the total light to be lost before it exits the fixture.
Now, if the early generations of power LEDs were exhibiting narrow light emission angles (from 30º to 50º), the latest generations emit light at much larger angles, between 120º and 160º. This cannot be qualified as “narrow” anymore, and opens up new general lighting applications to the use of power LEDs.

Once again, as in too many cases of emerging technologies, the information for the public tends to remain far behind the actual technical advance...

Display of power

Yesterday as I was reviewing various power LEDs’ technical datasheets, trying to extract useful parameters and feed them to my lighting design software, it struck me how manufacturers use such diverse ways to express the same characteristics of their products.

Luminous efficacy, expressed in lumen/W, is one of the most important parameters for lighting design. But when it comes to high power LEDs, it can be wrongly interpreted. It is probably the result of using a simplified description for these LEDs calling them 1W LEDs.

In effect, the actual usable power in the LED is equal to the driving current multiplied by the forward voltage. The latter changes significantly between LED families of the same or different manufacturers, and may vary according to application usage between 3V and 3.6V.

For LEDs driven at 350mA, the usable power will vary from 1.05W to 1.26W, inducing a 20% variation of the emitted light flux, which is certainly not to be ignored. By the way, this is the power value to be used when comparing luminous efficacies, as it includes the small amount of power wasted as heat in the LED.

Did you say white LED?

The LEDs (Light Emitting Diode) are semi-conductor diodes that emit light when traversed by a current flow. A photo emission is taking place at the diode PN junction region, and the total quantity of emitted photons, therefore the light intensity, is proportional to the current intensity that passes through it.

The spectrum of the emitted light is primarily defined by the type of materials used to build the diode's PN junction, although it also depends on the current’s intensity and on the junction temperature.

LED manufacturing uses diverse technological process variations that lead to the production of different families of light color, such as:

• Gallium arsenide (GaAs) for light from infrared to red (650 nm);
• Gallium arsenide and phosphate (GaAsP) for light from red to yellow (630-590 nm);
• Gallium phosphate (GaP) for wavelength from blue to green (565 nm);
• Gallium nitride (GaN) for blue light (430nm);
• Indium and Gallium nitride (InGaN) for light from deep blue to ultraviolet (390 – 360nm);

The white power LEDs are usually based on blue LED chips of the more recently introduced InGaN family. The white color is produced using a blue light chip covered by one or more semi-transparent layers of phosphors. Using the light’s complementary color combination, appropriately chosen phosphors layers combined with the base blue light allow the creation of various white lights.

As a result, today’s white power LEDs generally behave like the InGaN products' family, with some variations, due to the presence of the phosphors.

Is the general lighting industry ready for LEDs?

Reading this post I was wondering if today’s lighting industry is ready to design and produce power LED lighting addressing general users’ needs.

LED technology is something we have been hearing more and more about from television, magazines and news reports. However, very few people understand why this light source is such a powerful alternative. The advantages of LED light sources over traditional lighting sources are numerous but can be simplified into energy efficiency, longevity, power and versatility. Despite these advantages, creating power LEDs based lighting fixtures is challenging because the technology and the associated constraints require a different knowledge from the traditional general lighting know how. To highlight just a few differences, a power LED based lighting fixture generally involves:

• a number of electronic components which are not usually found in today’s standard lighting fixtures,
• a thermal management system to ensure the proper functioning temperature of the semi conductor which produces the light,
• an optical system to propagate and adapt the light for the specific purpose of the lighting fixture.

Beyond these technical differences, I see two major human reasons behind the slow adoption of LED technology in general lighting:

• the rather feeble interest shown by designers of general lighting fixture for the technology
• the heavy investments made by the industry in trying to make existing technologies more efficient

First of all, the components of power LED fixtures come in different form factors and present different usages patterns from what many lighting professionals, including designers, are using in their day to day work. In summary, using power LEDs requires a different type of knowledge, which in turn implies a certain learning curve.

Besides, I am under the impression that many interior lighting designers behave as stylists, “kooky kids” who like to do fun, pointless things, being more preoccupied by the actual “look” than by the “function” of the lighting fixture. When a given style becomes “hot”, legions of designers imitate it until it is everywhere, and as a result, most of today’s interior lighting designs boil down to flurries of boring light shades and diffusers stuck around inadequate incandescent, halogen or fluorescent sources.

As to the manufacturers, their weak interest make me believe that they re-iterate a well known incumbent behavior: they dismiss a disruptive technology to “protect” investments made in the previous technology. Furthermore, LED manufacturers only address part of the overall market, forgetting to properly educate other sectors of interest. This is particularly obvious of power LED manufacturing, where a vast majority of today’s production is destined to the automotive industry. Obviously, with the ever increasing amount of electronic in our cars, automotive engineers do not seem afraid of an additional “pile of semiconductor chips” and certainly know how to put them to good use.
Unfortunately, the same does not apply the general lighting industry! There, in view of the substantial investments made over time in studying, designing and using the current range of halogen and fluorescent lighting, everything looks like the adoption of power LED technology will take another century.

Finally, there is a split within the general lighting industry regarding the form LED luminaires should take. Some manufacturers are focusing on producing LED light sources that will accommodate existing lighting infrastructure and resemble existing decorative and task lighting fixtures. In my view doing so is rather short sighted, and wastes the potential of a promising technology. A better approach is, in my opinion, to design new light source formats that will deliver lighting in unexpected and innovative ways that could not be accomplished with previous lighting technologies. The fact that LEDs can be incorporated into just about any material, including fabrics, means that lighting could be made to emanate from the very surfaces and shapes of objects, from architectural elements to furniture.

At the risk of sounding overbearing, I believe some of these attitudes by the industry's actors delays a particularly attractive and efficient technology and limit the opportunity of participating to higher level discussions. In particular, although it is unlikely that incandescent lamps will be banned outright, the lighting efficacy standards under consideration all over the world would set minimums so high that most incandescent lamps would become ineligible for use by 2012 at the earliest. In this context, the lack of traditional sources efficient enough to meet the proposed efficacy standards will leave a huge gap in the market, that could be partially filled by LED lighting fixtures and luminaires.