Why LEDs are Bad… and Great at the Same Time.

Posted: March 23, 2018 in General SSL Commentary, Uncategorized
Tags: ,

Before we get into this topic, let’s remember that history is filled to overflowing with cries of lighting technology destroying human health. When the incandescent lamp emerged, the gas light industry flooded the market with baloney about its harmful effects (including claims that electric lights caused women to get the vapors), in support of the healthful glow of gas lighting (gag). Fluorescent and HID technologies have been under attack from their very first introduction. It’s quite funny to see now the defense of HPS by LED detractors, as it has been the target for criticism for decades. My advise for the apoplectic LED detractors… take a very deep breath and get over it. Every light source carries with it a compromise, including daylight, with undesirable effects we must do our best to mitigate. I suggest to anyone interested in the processes of new lighting technology: An interesting read on the 60 years it took to get lighting off the gas. That said…

LED technology has brought us a great many advantages over prior source technologies. Luminous efficacy is the greatest of these, and can deliver unprecedented savings in energy. But, this is not the entire story, nor does efficacy cover some of the liabilities LEDs present, that are not always well addressed. In fact, there are problems with LEDs that can cause serious issues in developing lighting systems and applied lighting design, that are the next challenge to overcome in the pursuit of maximum lighting performance. The following is a summary of the issue and problems, and some basic details as to why they are issues worth resolving, sooner than later. Note that I am avoiding diving into the complex science behind these issues, as that adds little to the discussion, and confuses the underlying concerns raised. I do not provide optical design as a service, as I fully recognize that this is the work of experts in this field (many of whom are not doing a great job addressing the issues I raise here, BTW), with the tools to get the work done properly. This summary risks erring on the side of over-simplification here, for purposes of a more open discussion and exploration. To those offended by this approach, I apologize.

Let’s do the color mess first

Color, both gross CCT value and actual SPD content are important factors in light source selection. In this, LEDs utilizing a blue pump die of either 456nm or 405nm, with an overlay YAG derived phosphor, produce a peak spike from the blue pump, which matches the LED blue pump used. This has created a great deal of consternation and speculation about blue light hazard, etc… The freaking out about “blue light” has now landed in a paradox – with blue content being argued favorably for HCL (circadian effect), and opposed as a human hazard. Meanwhile, the reliance on characterizing LEDs by their generalized CCT, with brash generalizations about blue light content, has become ridiculous. To be clear, CCT is just a generalized descriptive value. Spectral Power is where the real information resides. Basing arguments around LEDs on CCT alone is like arguing about what car has more horsepower based on what color the seats are. It is absolute nonsense, and exposes those who persist on relying on CCT, as hacks.

The fact that many “experts” get it wrong is obviated by the AMA position document that includes a specific recommendation for 3000K as a desirable LED color outdoors, with no other qualification of SPD content, color characterization, or blue emissions balance. The AMA amplified its insanity by being so bold as to offer up its opinion on lighting and product design. This was then responded to by the IES,  demonstrating the heights of ridiculousness this issue has climbed to.

Then, we have the issue of HCL, where emission of blue light in the region of 459 to 484 and the metric of Circadian Stimulus (CS) is the hot topic of the day. Again, the pundits propose that great gains in human performance can be made by fiddling with CCT’s through the day. There is some objective basis for this, but in real application, this has become more about marketing than actual real world gains. The difference in CS performance between “3000K” luminaires, can be as great as 15%, while the difference between 3000K and 5000K can be anywhere from 2% to as much as 25%. You cannot establish any advantage of change in support human circadian function without focusing on the SPD of the luminaires, regardless of their CCT. It is, and has been proven, that applications delivering variable CS emissions can be accomplished with zero change in gross CCT value, just by changing illuminance levels. If you compare a 3000K and 4000K LEDs, at the exact same illuminance level, the difference in CS ranges from -5% to +15%. These differences are easily attained with dimming, which is a lot less messy than dual color lighting systems. Further, this assumes the difference of 11% is significant. Not only does circadian function depend on SPD content, and illuminance levels, there is a threshold minimum level that must be achieved to realize any real gains, or reduce melanopsin supression. An insight into this can be found in the WELL Building standards for Circadian Lighting In this, the maximum melanopic lux for sleep cycles (minimum supression) is 50. Meanwhile the Minimum required for “wake time” is 250 melanopic lux. That’s a ratio of 5:1. Other research points to a similar differential required to truly have an real impact. That’s a lot more than 11-15% that might be realized with dimming or fiddling with different LED sources. Further to this, when light is used theraputically to correct and establish circadian entrainment, the illuminance levels used are on the order of 10,000 lux. So, the question is, where, between the 11-15% the majority of HCL CCT shifting lighting systems produce and the 5:1 ratio suggested by the WELL standard, and 10,000 lux used in research and theraputic application, do we need to be? This is not an LED issue at all, it is an issue of reigning in the silliness that the HCL has come to, in order to establish some real standards to be applied, and limits as to who can apply them. Here is an interesting read on this subject for those addicted. NASA has applied this technology, as has Boeing. How much of what they have accomplished applies to terrestrial lighting has yet to be determined. The CIE has issued a statement on HCL, that I tend to agree with. We really don’t have enough information to apply SSL to purposefully manipulate the physiological responses of human occupants in commercial space, so should refrain from doing so, until we do. I’ve noted this in another article here.

For those who want to evaluate the SC value for their lighting systems, this is a great little tool: http://www.lrc.rpi.edu/programs/lightHealth/index.aspAll you need is the SPD of the luminaire you are using.

Beyond this, I won’t get into any further discussion of the “blue light” issue of LEDs, as the vast majority of those wishing to have a dialog, are more interested in making their specific point, either that LEDs are killing us all, or that the issue does not exist, or that blue light will change our lives. All are equally wrong, as there is a lot more to this topic than is being “discussed.”

Another topic rarely mentioned, but of equal importance, is the role of red light and its absence in most LED product SPD’s. Red light is important to human health, specifically in the region 620nm to 750nm. LED products using down-converting phosphors simply do not generate a great deal of light in this region. Red light is absorbed and utilized by both visual and non-visual systems supporting cell functions not currently on anyone’s lighting radar. Just as it is with blue light, CCT of a product does not necessarily indicate good spectral characteristics in red light production. For example, one might suspect HPS lamps, due to their orange color, would produce high amounts of red light. In fact, they produce less than most white LEDs of any CCT. That means selecting that magic 3000K LED does not automatically produce a significant amount of desirable red light. This is a fun read on the effect of red light.

There is also a great deal of noise on the topic of wildlife friendly lighting, particularly concluding that a 3000K or lower CCTs, solves that issue.  In fact, wildlife foundation recommendations do not not share this view at all. In published data, such as: Florida Fish and Wildlife Report   any light in wavelengths shorter than 650nm are problematic. This excludes all HID sources, all HPS and LPS, and all LED sources using phosphor conversion over blue pump LEDs – not matter what their CCT. That means that all recommendations to use 3000K LEDs to “protect wildlife” are baloney. Lighting designed to have no effect (invisible for the most part) on wildlife will be narrow spectrum red-orange to red. Further, there are additional factors regarding wildlife friendly design, including seasonal timing, geography, proximity to what species, surrounding environmental conditions, etc.. There is no such thing as a uniform, single specification for “wildlife friendly” lighting (LED or not), but one thing is reasonably clear. 3000K LEDs do not produce a significant change in visibility for wildlife observers.

The same can actually be said in regard to astronomical observatory environments, where filtering ambient light is necessary. LPS has been favored in these applications due to its relatively narrow SPD in the region of 589nm. LEDs are broad spectrum sources, so are not suitable in this application, even at the ubiquitous magic 3000K CCT, since the blue light emission remains from the blue pump, far greater than what would be experienced with an LPS source. Again, not all 3000K LEDs are created equal, so whether any of them are suitable to the task cannot be determined by a CCT label.

Lastly, on the topic of color qualities. My personal view is that there will never be a good metric for defining color quality, that reflects the actual distortions produced by different light sources, as long as each product is characterized and compared to an ideal source of identical CCT. The idea that we can have a blue-ish white light of 6500K attain a 98Rf value, while a 2200K orange yellow source with an identical 98Rf value, is ludicrous. It makes the metric nonsense, and leads to non-sequitur comparisons. In my personal view, we should select a single idealized white light reference and compare all sources to that reference. I believe I made my views on this pretty clear in a LED10-What-Color-SSL 2016.

Enough about color already.

The real evil of LED technology is source brightness and intensity

LEDs are extremely bright light sources, with exceptionally small emissions surfaces. Setting aside photometric distribution for a moment, using a simplistic metric of lumens, a single high power 3mm LED die with phosphor plate, emitting 1000 lumens, emits 111 lumens per mm². As a comparison, a 2′ x 2′ fluorescent troffer with a surface area of 371,612mm² delivering 6000 lumens, distributes just 0.0162 lumens per mm². That means that the LED itself has a potential brightness 6,852 times greater than the fluorescent troffer, from a surface smaller than the plunger on a clicky ink pen. This condensation of luminous power creates massive challenges in controlling brightness and glare. Compounded this, is the fact that all LED light emission is directed forward, unlike arc lamps, which generate omni-directional light distribution. A 42″ T5 fluorescent tube generating 3000 lumens, delivers just .03 lm/mm², with less than 50% of that directed at any human occupant. A 250W clear MH lamp generating 22,000 lumens from its arc tube generates 3.26 lm/mm², with less than half of that directed at human eyes. The 1000 lumen 3mm LED produces 34 times the light per mm² that the 250W MH clear HID lamp, all directed into one photometric direction, resulting in a brightness differential of at least 68:1. That is a very bright little source, that if not properly controlled will produce undesirable effects on observers.

Looking beyond the simplistic lumens delivered description, consider that high power LEDs direct most of their light into a distribution pattern that is between 65° and 120° (FWHM), where HID and fluorescent sources generally produce light in a 360° distributed pattern, give or take. That means that for each lumen a raw LED product generates, it will appear roughly 3-6 times brighter when compared to an omni-directional source of similar total power. In other words, the aforementioned 1000 lumen LED has the effective brightness of a 3000 lumen source in delivered raw lumens per mm², packed into a surface area that is a microscopic fraction of any conventional light source. That means that the brightness apparent from an LED is significantly greater than a conventional lamp based technology. This is great for product package size, but an unmitigated disaster when it comes to controlling apparent brightness and disabling glare.

COB arrays, which are comprised of a collection or grid of LED die, covered by a larger area of phosphor, produce lower luminous brightness, but not by a massive amount. To make matters worse, COB arrays now generate such high  lumen outputs, in smaller and smaller packages, they too are approaching the 111 lumens per mm² power levels, making them just as much a challenge to control undesirable brightness. Add to this the issue of controlling the ‘etendue (radiant spread due to large emissions surface area) of these sources, which can create undesirable field brightness – and you will see that COB arrays present their own range of problems to work around.

Bottom line is this: If any LED is exposed to direct view by an observer, at normal viewing angles, a great deal of care must be invested in mitigating the intensity of light and the appearance of the emissive surface, particularly in the case of high power LEDs. Due to their extreme brightness and small relative surface area, they should never be exposed to direct view, especially at low viewing angles, even through optical components. This means that, for the most part, high power LEDs should never be viewed directly, at any angle above 40 degrees above nadir, unless from extremely long distances, or against very bright background luminance. This includes street and roadway lighting system applications.


There are four fundamental approaches to utilization of light energy from LED sources.

Approach One: This is fundamentally unchanged from prior technologies. This is to use of LEDs in groups, lines, grids, or arrays behind a diffuser, or directed into a diffusely reflective media (indirect). In this approach, the LEDs are concealed from direct view, so present no difference in the brightness characteristic  of the diffuser or reflector surface that would otherwise exist if using a T5 lamp or other source. The only caution here is that some LED products do push the envelope in delivering more lumens to reduce luminaire counts, creating too-bright a surface. But, this also applies to many fluorescent and HID products for the same reason. Others incorporate poor optical designs that simply blow light out in a Lambertian pattern with no directional control, aggravating an already too-bright surface presence from lack of consideration to emission in higher angles of distribution.

Approach Two: Apply an array of LEDs within a unifying optic that blends the light from multiple sources into a larger surface area component that is then controlled with reflectors or refractors to produce a desired photometric distribution. This can be used to mitigate the issue of viewing LEDs directly, creates a smoother beam pattern, with less apparent glare and brightness. It’s a sort of hybrid concept, where LEDs are just lumen pumps for a larger optic, that in many ways acts like a conventional light bulb. A similar approach is in the use of LEDs inserted into structures that reverse their direction, emitting light into a larger reflector system above them to gain the desire output. This is somewhat like the old days of silver-bowl incandescent lamps, with similar effectiveness in creating glare reduction.

Approach Three: Use of a single stand alone LED, whether a single packaged mid or high power LED, which contains a single or tightly arranged LED grid in a case, or a COB, which includes an array of blue pump LEDs under a phosphor dome. Optics for these products are conventional lens, reflector, or TIR optical configuration, which generate output similar to a halogen lamp in appearance. This is typical of down-lighting, track heads, and other similar point source lighting. Glare and brightness can be readily controlled with proper shielding, reflector design, cover lenses, etc. While this appears to be relatively innocuous, due to its familiarity, some caution must be taken here. LEDs have created opportunities to shrink product packaging, and light source openings. This has led to introduction of small aperture luminaires delivering light output of much larger conventional sources. When combined with the intensity of LEDs in general (as noted above), the end product can create discomfort glare. It might seem neat to have a 1200 lumen down-light with a 2.5″ aperture, but without careful optical design (and in spite of it), that aperture is generating .3 lm/mm², or 10 times the surface brightness of a 6000 lumen 2 x 2 troffer. That is not going to be comfortable, or desirable, no matter how neat it might seem on paper.

Approach Four: The one that causes the lion’s share of problems with brightness and disability glare, is the use of arrays of individual high power LEDs covered with individual optics aimed into zones, to create a light pattern as a combination system.  The first use of this approach included retrofit lamps, using multiple LEDs to produce the desired light energy, each behind discrete TIR optics. This is now very common in street lighting, where the optical distribution of the luminaire is created by aiming individual or groups of LEDs into the desired zones, to produce a beam shape, such as Type 2 or Type 3 in appearance. This is also where conventional photometric measurement and design practices have run into trouble.

The issue with discrete distribution arrays described in Approach Four, is that in addition to the simplistic lumen brightness noted, there is very concentrated beam shaping included, that focuses light energy very tightly, amplifying the brightness effects inherent with the approach. This is manifest in optical components with sharply defined distributions that focus a great deal of energy onto a narrow target. The end result is an amplification of brightness effects that makes this approach extremely problematic. To observers, the end result is a source (luminaire) that aims individual LEDs directly at them, anywhere in the lighted field. Sort of like having a hundred gamer’s on a 30 foot platform with laser tag rifles aimed at the eyes of passers-by. This maximizes the dazzle, visual noise, and objectionable glare of the system, even if the cutoff angles reduce light in the surrounding field. This is, by far, the most hateful optical approach, and the cause of more complaints than any other approach. That this approach is used in tunnels, where the strobe effect of individual luminaires is layered onto the dazzling mess of LED arrays, is truly remarkable.

The LED array optic fail – Or, how to glare bombs are made

The majority of photometric tests, and most designs, use what is termed far-field photometry. This assumes all sources originate from a single virtual point in space, while capturing measurements of all light falling onto a meter at specific solid angles to create a model of the luminaire, source, or an optic’s distribution. For multi-sourced products (LEDs), there is no separation of individual sources, they are simply blended together, as the meter has zero discrete fidelity. This works just fine when the sources are large, the optics more generally focused, and the brightness of the system somewhat uniform for any given distribution angle. When the luminaire utilizes arrays of discrete sources, aimed into specific zones, or worse, at a specific small target, far-field photometry fails to produce any data indicating glare or brightness issues. It is blind to all source brightness patterns.

The problem is further exacerbated by the issue of contrast between discrete individual sources. For each aimed/targeted LED optic, there is a space at the source end (luminaire) that is not emitting light at all. The contrast ratio between the sources and spaces between can be essentially infinite, meaning high brightness to black. The end result is the presence of a great many individual spots of light, many aimed directly at the observer, presenting decidedly uncomfortable glare due to the extremes in contrast between very small extremely intense light sources and the surround. Conventional photometric tests and hand held light meters are inadequate to expose this issue. In stead, due to the way far-field testing is accomplished, these pips and dazzling glare effects are simply buried, averaged out at the meter and supporting software, combined into a single candela number.

Now add to this a whole gambit of human visual factors, illuminance levels in general and human visual responses to that dynamic, and the effect of dark backgrounds behind street and roadway lighting, and one can easily see why conventional design approaches are not going to deliver a desirable result. For this reason, virtually every street lighting luminaire on the market today produces undesirable optical qualities, excessive dazzle effects, glare, and discomforting visual noise. Add to this the combination of poor CCT color choices (read bad SPD characteristics), over-lighting of streets and roadways, lack of shielding, and too-great spacing between luminaires, and you have a toxic brew of epic proportions.

Now, all that said, who is now making most of the decisions about what street light system is to be applied? Lighting professionals? nope. Optical design professionals? nope. Engineers expert on the dynamics of optics and visual perception/human visual performance? nope. Who IS making the decisions are municipalities, utilities, public hearing groups, crackpots with opinions all over the map (with minimal information and zero understanding of the factors involved), and recommendations by organizations like the AMA? In other words, the pressure on manufacturers producing street lighting is all focused on cost, simple maintenance, low energy use, and in a few cases, CCT options, specifically the ubiquitous 3000K magic bullet every committee now seems to believe is the magic bullet fix for all their street lighting woes.

Moving to near field photometry

Where conventional photometry uses light meters to collect data at points around a product, from a distance, to generate a generalized photometric file, this protocol fails to capture the finite light patterning that comes from extremely bright discrete light sources aimed at specific target angles. By replacing the light meters with imaging meters/cameras that capture pictures of the light pattern that is directed at any angle in 3D space – we can begin to reveal where undesirable brightness patterns exist that far-field photometry is blind to. This paper describes the differences between the two photometry models very well. Take a look at this paper to see how near field worksThis is also an interesting read on this topic.  There are numerous technical papers on this topic, from characterization of LEDs and their unique emission, to lab protocols and how software is used to create a usable photometric file for lighting calculations – while presenting data for evaluation of intensity distribution with more fidelity.

Many LEDs also generate complex 3D Lambertian emission (sides and top as separate emission surfaces) difficult to integrate into design models without highly sophisticated 3D ray data. By creating a 3D modeling using near field photometry, optical designs can be refined to improve distribution characteristics, while luminaire testing will produce visualization and evaluation of observer effects, including contrast ratios in the luminous field, as well as to backgrounds beyond the luminaire in application. In the design of optical systems, near field models (ray trace files) capture the unique 3D character of an LED package for optical design, to improve both efficiency, as well as realizing glare and brightness control.

Lighting design software needs to step up and offer an observer view of the lighted field, including the luminaires present, with reporting and rendering of visual appearance, brightness ratios and patterns, glare indices, etc., similar to what might be seen when using near field photometry. This can only be done if near-field photometric data is available to work with, as current photometric data simply has none of the fidelity required for this purpose.

None of this is easy to do, and very few actually have the capacity for it. This will need to change in order to break out of the cycle of poor product designs we are trapped in today – specifically for designs that employ individual LED/optic packages visible to observers.

Near field photometry is equally applicable in the design of all LED products, as it exposes issues of beam pattern uniformity, color uniformity/distribution, beam and field striations/behavior – while revealing brightness patterning that conventional far-field photometry simply cannot support.

Building a better LED outdoor luminaire that will even satisfy LED haters

The underlying issue with LEDs is not the technology, or the sources themselves (other than a few nit-picking details about SPDs, the lack of red content, and control of the core blue/violet pump emissions) but how they are used. Further, conventional photometric design approaches and measurement protocols simply do not create the data necessary to evaluated and develop LED luminaires that utilize arrays of discrete LEDs topped with individual optics.

There are a couple of solutions to this.

The first solution involves abandonment of the discrete LED model all together and using Approach Two described above. Creating arrays of LEDs to be placed under a larger unifying optic, that distributes a more uniform, smoother pattern, while presenting a larger surface area, is one solution. This can be accomplished using any number of optical designs, from large single lenses to modular lens segments, with a few specific design criteria; 1.)  Erase/conceal the presence of any single LED anywhere in the system, 2.) Reduce distributed light power to less than 1 lm/mm², and 3.) Increase surface area of optics aimed at the highest angles to the largest areas of the visible optic, to spread the highest brightness over the greatest area possible. This can also be done using concealed LED sources and reflectors or refractors with various shielding and glare baffling components. While efficiency will suffer somewhat, the end product can be made superior to both HID technology and discrete LED solutions.

If Approach Four is the chosen course, using discrete optics can also include pattern concealing features, glowing background effects to reduce contrast at the source, or use of a greater number of lower powered LEDs to produce the desired distribution pattern with the lowest individual LED/Optic brightness possible. This is a far more complex task than the unifying optical approach, as optical modeling in 3D is critical in determining the end result. To date, I am unaware of anyone who has succeeded at developing a product using the discrete LED model. Virtually all of these products create undesirable glare, visual noise, dazzling effects, etc…


The need to stop pushing for brighter numbers and focus on better luminance,  illuminance, and visual comfort

The single greatest issue with LED technology today is the escalation of brightness from smaller and smaller sources in the name of lumens per watt efficacy. While many products are doing just fine using the current processes and protocols, such as down-lights (given they are not over-delivering), track heads, cove, indirect architectural products, and decorative features, a few others are not doing as well. LED retrofit lamps (using individual optics over high brightness LEDs) and street lighting seem to have the greatest issues of controlling intensity, glare and dazzling that reduces visual performance and increases discomfort.

I am not a fan of individual Optic/LED packages arranged in arrays for any application where exposed to occupants. This includes track heads, down-lights (especially those where the arrays are barely recessed into the ceiling), linear strips, outdoor lighting, or retrofit lamps. LEDs are simply too intense, even when inside a reflector or optic, to be viewed directly.

In my opinion; While the safety standards, particularly IEC EN 60825-1 and S009/E:2002 do not consider LEDs hazardous, or simply exclude them outright – the issue of discomfort glare, annoying pixelated brilliance (dazzle), direct brightness, all effect visual performance and comfort, and demand reconsideration.  I cannot personally understand any manufacturer putting their product up in an application, can look at the glaring and blinding arrays of LEDs, and feeling these are acceptable products to market, I really do not care that they produces an acceptable photometric file, especially when that file uses an obsolete test method. The glare bombs we are exposed outdoors and in parking garages or retail spaces, are simply not acceptable, and are causing more issues with observer complaints about LEDs than any other factor in the deployment of SSL. Just for the record, I have had the same complaint decades ago with HID sourced luminaires creating unnecessary glare, brightness, and lack of cutoff control, so this is not an exclusive SSL issue. However, LEDs in direct arrays make these conditions chronically bad.

Another failure in street lighting is the lack of quality design influence and participation, leaving decision making to utilities and city council meetings, that simply do not know what they are doing. It is bad enough to base entire systems on retrofitting existing pole products, but when this is done with a budget-first approach, there is little chance advanced design considerations will play a role in the process. City councils addressing angry residents, reacting by making calls to change CCT as a solution to what is an optical issue, are ridiculous. Finally, the lighting standards community, which includes the vaunted IES Street Lighting Committees who have spent millions of hours making recommendations, have apparently dropped the ball when it comes to SSL deployment, and have failed to establish appropriate recommended practices and metric protocols to address the issues making LED public lighting so aggressively unsatisfying.

The issues of LED deployment are numerous. Photometric testing being centered on photopic vision, while street and area lighting exist in the regions of mesopic and scotopic vision, coupled with total confusion of CCT and SPD, assumptions about the connection of CCT with visual performance, over-statement about blue light hazards to humans and the messy combination of the 3000K solution as friendly to nature, are just the tip of the iceberg.  Next is the use of random “case studies” to promote, evaluate and communicate solutions, combined with “opinion” surveys to generate data points without proper qualification of test processes. These all make the issue of SSL roadway and street lighting a particularly confused disaster in the making. Based on this, there are now numerous “experts” coming out of the woodwork, deriding the lighting industry as a whole, and criticizing SSL technology itself as the culprit – through social media and trade conferences. Due to the relatively slow pace of professional engagement of this issue, it is easy to see that SSL street and roadway lighting will continue to be deployed as it is, with little real improvement in the near term.

Do lighting designers and the professional design community deserve the hits they are taking on this subject? Absolutely. There has been too little work done by lighting focused professionals to resolve the issues in hand, and the community as a whole is taking far too long to resolve the issues experienced on the streets and roadways of communities across the country. If the lighting community were really doing its job, the current state of street lighting would not be such a glaring failure (pun intended) for human observers.

Are LEDs Bad?

In a word… NO! That is a ridiculous assertion.

The LED is just a light source technology, with a characteristic that can be made to do pretty much anything we demand of it. Something we cannot say of incandescent or gas discharge technologies. LEDs are now available with full SPD’s, including reds, and limited blue content, not to mention selection of either 405nm or 456nm as the core blue pump, or simply using 7 distinct colors of LED to generate any shade of white one might desire. We can get LEDs in narrow spectrum’s that will enable human vision for navigation and safety that are virtually invisible to animals in the wild. We can create optics with minimal glare, controlled brightness, that also control the brilliance of LED light sources. It is also possible to add controls over when lights are on, to what intensity, and of what SPD to suit changing needs during the night to mitigate impact on natural surroundings, and humans trying to get a good nights sleep. The technology is all good, if we choose to make it good.

However, like all other technologies before it, LED technology does have a few liabilities that must be considered. That’s how lighting has been for more than a century, nothing has changed here. From buildings burned to the ground from gas lighting failures, electrocutions that lead to strict code regulation of the electrical era, physical burns and fires from halogen lamps, etc… etc…. every light source has to be considered and carefully applied.

What is truly bad about LEDs is that they generate a lot of noise, marketing claims, consternation, argument, false logic, silly opinions, short cut solutions, and strange contradictions, that detract from our using them to their best capacity. If we put our heads down as a solid-state lighting community, we could be well on our way to producing light that is healthful, friendly to neighbors and nature, long lived, and energy efficient. That means ending the constant bantering about how they are a danger to humanity, as this is clearly and most definitely… you guessed it… baloney.

On a positive note… By the time all the harping about LEDs is settled, carbon nano tubes, OLED, and other as-yet to be released technologies will emerge – firing up an all new spate of arguments and controversy. It’s the way lighting has always been, and the one thing that will remain the same – even if everything else changes within it.






  1. First of all, thank you for taking the time and effort to write such a elaborate narrative about LED-based lighting sources. The lighting industry could take away a good deal of valuable information out of your article.

    I’ll start with your claim, “The LED is just a light source technology, with a characteristic that can be made to do pretty much anything we demand of it”. Your second claim, “LEDs are extremely bright light sources, with exceptionally small emissions surfaces. Setting aside photometric distribution for a moment, using a simplistic metric of lumens, a single high power 3mm LED die with phosphor plate, emitting 1000 lumens, emits 111 lumens per mm².” Your third claim, “The 1000 lumen 3mm LED produces 34 times the light per mm² that the 250W MH clear HID lamp, all directed into one photometric direction, resulting in a brightness differential of at least 68:1. That is a very bright little source, that if not properly controlled will produce undesirable effects on observers.”

    … While LEDs are extremely bright light sources, the brightness differential you approximated (i.e., 68:1) is actually off by orders or magnitude. The number 68 would go up tremendously along center optic axis… This (is the) fundamental reason… why LEDs have such tremendous brightness and consequently why it produces blinding glare. While perceived glare can be improved by raising the background light to high levels, the amount of light that can fall into people’s eyes when viewed directly at an HBLED LED chip, module, or luminaire from close distances are still extremely high and very damaging.

    Finally… the techniques you mention to handle LED illuminance, luminance, and glare are only somewhat effective. One needs to understand the source cause of LED glare and find the right technique to manipulate light from the LED chip itself; and the LED chip itself has to be carefully designed to produce the type of light that one can do lighting design with.

    I urge the LED, lighting, and the optics industry to read your article and my response and take the issue of LED lighting very seriously because enough roll-outs have already taken place without people recognizing just how damaging current LED street, car, and high-bay lamps and luminaires are.

    Dr. M. Nisa Khan
    IEM LED Lighting Technologies

    • kwillmorth says:

      Cool. Not sure where all this leads, and did not submit this for grading. This was a summation written in simple enough terms to get tge points across, and was stated as such in the intro. I removed the SPAM portion of your response, as I do with everyone. Good luck with your work.

  2. Paul says:

    Fantastic write up!