Text-Alternative Version: Successful Selection of LED Streetlight Luminaires Webcast

Below is the text-alternative version of the "Successful Selection of LED Streetlight Luminaires" webcast, held March 6, 2013.

Edward Smalley: Hello and welcome ladies and gentlemen. I'm Edward Smalley with Seattle City Light, Seattle's publically owned utility. As director of the Municipal Solid-State Street Lighting Consortium, I'd like to welcome you to today's webcast, "Successful Selection of LED Street Light Luminaires: Optimizing Illumination and Economic Performance." I'd also like to have a warm welcome to our Consortium members and also our friends in Canada who are joining us through LightSavers.

Today, we're happy to have with us, as speakers, Eric Haugaard of Cree Lighting and Chad Stalker of Philips Lumileds. Eric Haugaard is the director of product technology for Cree Lighting, a position he has held for the past seven years. Previously, Eric was engineering manager of the new product development for 11 years and mechanical design and product development engineer for seven years.

Over the past decade, Eric has presented lighting technology programs to diverse audiences throughout the world, including a strong focus on LED luminaire technologies. Eric holds a bachelor's degree of mechanical engineering from the University of Wisconsin-Parkside, with post-bachelor studies completed at NASA Research Center.

Chad Stalker is a regional marketing manager for Philips Lumileds. Mr. Stalker is responsible for working across the lighting industry with fixture manufacturers, lighting designers, and other market influencers to support the adoption of solid-state lighting technologies. He has been working in the LED lighting industry for over ten years with such companies as Luminus Devices, Osram, and Color Kinetics. So well, thank you both, Eric and Chad, for joining us. We will receive a number of questions during our presentation, and we'll go over those at the end. So for now, I'd like to get us started with our presentation. And to give you a little bit of background about the Consortium itself, the Consortium was created by the U.S. Department of Energy in March of 2010. It was designed to support the needs of end users as new developments occurred in solid-state lighting.

The vision of the Consortium is to accelerate the adoption of high-performance solid-state street and area lighting by leading the end-user collaboration in areas of performance – or evaluation, application, and standardization. The mission is really to increase knowledge and develop a structure for influencing standardization for LED roadway lighting throughout the country. As we start to look at the membership of the Consortium, what we see is that the Consortium is comprised of mostly municipalities in the U.S. – over 218 of them, but a strong presence of utilities – over 109 utilities, both, municipally-owned and investor-owned type utilities.

If we take a look at the map of the U.S. today and we look at the red dots on the map you're seeing now, what we really learn is this phenomenon of adoption of LED technology is happening throughout the country. There are basically just a few areas where LED lighting is – not of interest to utilities – yes?

Edward, excuse me. We do not see your slides transitioning.

Edward Smalley: I see. Ah, so apparently, my slides were not transitioning for you, so sorry about that.

I think you should be able to see those now, and what we're looking at now is a map of the country. That's the danger of going live on these, isn't it? So you may not have seen some of the earlier slides, but nonetheless, we really wanted to give you a flavor for how fast moving the LED phenomena is across the U.S., as many cities, whether it's Los Angeles, New York, Seattle, or even Boston, it's really occurring in many different parts of the country.

So with that, I would like us to get into our presentation for today, and right now, we'll transition over to Mr. Eric Haugaard who will help us get deeper into the subject of successful selection of LED streetlight luminaires. Eric?

Eric Haugaard: Thank you, Edward, and thanks to the Consortium, and a special thanks to all of you that are joining us today. Today's goal for our time spent together is going to be spent primarily just really defining a process for developing effective street lighting specifications that enable expectations to be met. And I always like to start off with maybe stating the obvious because evaluating the viability of LED solutions – any solution, for that matter, for illumination – is really justified based on two major criteria.

The first one, of course, is meeting the illumination performance requirements for the application and the second one being meeting the economic performance requirements, and both of these requirements have to be met without compromise. So you would never sacrifice a illumination performance requirement for the sake of economics, nor would you sacrifice economics for poor illumination performance.

So if we look back into the incumbent source technology, showing up on the right-hand part of the screen is some metal halide technology, and then we add a luminaire here. This is not a street light luminaire, of course. It's a very common glass refractor wall pack.

But if you look back into the way we had the opportunities to create different versions of a given product category, one of the things that was very confining was that we had difficulty, in many cases, controlling the light source. So your optical distribution choices ended up being few – as little as one, like the glass refractor wall pack; maybe as many as five for a well-filled-out outdoor luminaire.

The other thing we had to deal with was if we, in fact, chose the right optical distribution to set our application, and for some reason, we fell short in illumination performance. In certain areas of the application, we had to jump from a given wattage to the next higher wattage. And if you look at these buckets of wattages we had, typically, it was about a 30-60% jump in energy. So if you missed your objective by 5%, you maybe had to do jump up 25-65 percentage points more illumination to capture that objective.

Lumen maintenance was something that, of course, we've always been dealing with. All light sources depreciate over time. However, with the incumbent source technologies, lighting designers use rules depending on the lamp technology and the published results from the manufacturer.

So the idea of rated lamp life was that the luminaires would always require service, and the most obvious service-related element would be that you would change the lamp within the lifetime of the application for that particular luminaire. But realistically, the lamp life is really defining, then, service requirements.

So if we look at some of the, for example, a manufacturer's data for a given lamp – this happens to be a 400-watt pulse start metal halide. It's rated at 20,000 hours, which means that you would have a 50% mortality rate at 20,000 hours. It's not a very sensible value to actually use for servicing.

So most manufacturers will recommend that somewhere, far in advance of that, you do a group re-lamping, and the group re-lamping selection you made – in this particular case, it's 8000 of 20,000 hours or 40% of beam life – it means you're going to have 25% depreciation in just the lamp performance, and that is incorporated into a light-loss factor calculation for your initial lighting design. So we're always over illuminating the site to begin with to make up for lumen depreciation up to some – the lowest in-service value that requires maintenance.

So we certainly have some new opportunities with luminaire technology designed around LEDs. Some of the diagrams that we see here, the farthest to the left is showing an LED completely enveloped by an optic, and the beauty of this technology is that 100% of the luminous flux that emits from this particular LED package is controlled by that particular optic. And these are choices that we have today that were not enabled as eloquently as with the incumbent source technologies.

The other advantages we have – ways of incorporating many, many sources, very few LED luminaires for roadway lighting and area lighting have one source. They're multi-source optics, so we can balance the number of LEDs vs. power output to create light-engine configurations within a luminaire that really balance cost vs. energy.

So in places of the country where you have very, very expensive energy costs vs. not-so-expensive energy costs, your luminaire solutions, although deal in the same pattern of life, may have a different energy objective and cost objective. So this enables dramatic improvements in application-level performance optimization because of what I like to call "application fine-tuning."

So how do we optimize the process? Realizing this is not a roadway application – this is an area light application – if you were to see this application pre-summer of 2007, it was 100% high-pressure sodium. It could've been metal halide, also. It was designed to meet the IES recommendations for RP-20 for just over minimum illumination for minimum security, which is about 0.2 foot-candles min. with a uniformity ratio of something better than 20 to 1, which you're actually looking at on the high-pressure sodium side to the right. It's about a 20-to-1 uniformity.

And on the left, you're looking at far, far greater uniformity. However, on the left side of the application, we actually have half as much light as the right side of the application, and the reason is we have a pretty significant dump of light right where the poles are where the luminaires are mounted. So if I look at the specification sheet for that particular luminaire, if I direct your attention to the left part of the screen where there's a polar candela trace, if you look at the first arrow, that is really what the optical designer intended to do to direct most of what I typically call the "main beam."

If I look directly below the luminaire, I have about – oh, about a third of the intensity. However, when I look to the actual application at that same angle of 60 degrees, there's far less light at that point than there is directly below the luminaire, and the reason is we cannot envelope an optic system that controls 100% of the light. So some of that light is emitted uncontrolled, and it typically will manifest itself in too much light right at that point at the vertical error.

Why is it so much brighter with the intensity shown in the diagram being so much less? It's just a simple physical property called the "inverse square law," whereas if I give an intensity at a surface, if I half the distance, I will actually quadruple the intensity. So this is something we've dealt with incumbent source technology and primarily because we can't control all of it.

So when we talk to people that were interested in incorporating this technology back in those days, we said, "We're going to save you some energy," 53% in this particular case. However, if you really read it, looked into the details, there was that, actually, 61% light. And without doing a lighting design analysis, you would never be able to convince somebody this was a good idea – 61% light for 53% energy is not something that people get very excited about. But what we did was we designed the application for the IES standards, and rather than 0.2, this has to be 0.3, but nevertheless, it's above the minimum-security recommendations.

So we're going to talk about these two essential elements in a little more detail. First thing to incorporate, of course, is something that defines your required sustainable illumination performance. This particular application, again, an area lighting application, rather than going for 0.2 foot-candle minimum luminance or even a 0.5, which is what the IES recommends for enhanced security, a lot of retailers, in this case, want something a little bit more than that. And it's important to note that that has to be discussed upfront. Before you go into the specifics of luminaire positioning, what is the environment you want to create from an illumination-performance standpoint? And even the uniformity objective was better than the IES recommends – 10 to 1 vs. 20 to 1 or 15 to 1.

The other thing we have to discuss is the application duration, remembering that the lamp service requirements define the service duration. Now, if we're looking at near maintenance-free applications, where we have life cycle cost analysis incorporated for 10 or 12 years, one of the major value attributes is that maybe you don't touch the luminaire throughout that time. So it's very important you discuss that time period.

So we would simply, in this case, incorporate a lighting design process, and in the end, what's very important to the user is that, of course, you meet the luminaire _______ performance objective. But that's sort of done by inspection by the user by making sure they're happy with the illumination.

What's more important is having a lowest total cost of ownership. A great proportion of that, in many cases, is the actual energy savings. This particular case, 19 kilowatts of energy vs. 31 kilowatts of energy, and I'm mentioning this not specific to the luminaire because I think the composite of the energy, the actual kilowatts of energy at the site are much more important than talking about luminaire efficacy and other types of energy attributes. Certainly, another portion of this would be the avoided maintenance.

The second part, of course, is the economic side, defining what your economic performance is, and I have risk management there because you have to have ways of mitigating risk. We'll talk about that later, and we believe that a life-cycle cost analysis should always be incorporated when you're trying to do this, and the evaluation timeframe has to be a part of that because that's what life-cycle cost is all about. It's about a time-based equation for understanding that.

Product durability and performance requirements, we'll cover that a little bit later, but if you look at all these different, somewhat-arbitrary attributes with relative proportions, all of these things should be considered or not. I mean there's some that might be difficult environmental impact. Some people equate that different ways, but the point is whatever is an element that's associated directly with your cost over life needs to be incorporated.

So if we just retrace the steps here in a little bit more detail, number one, always, always establish what your illumination performance requirements are, and this is just sort of a short litany of what it could be – certainly, vertical and horizontal illumination, uniformity requirements, color quality, light trespass, and glare metrics. This would apply to, actually, all applications, not just to the application that we've shown in this example.

After that point, of course, we do have to define the variables, understanding exactly what the site light geometric target is and then the position constraints of the actual solutions to the light source is. Some of that is new vs. upgrade and, also, the service life or sometimes called the "application life" or, if I have it here, "years of near maintenance-free operation." I like to use the term "maintenance free" because some of the things that are – certainly, we're trying to avoid routine lamp replacement, but vandalism still occurs. Poles getting struck by automobiles still occur. Trees still keep growing and other things, of course, that are outside of the control of the – the people operating the site will control what will happen.

And so what is the economic assessment period? Well, certainly, that's up for debate, but we're seeing in the industry something around 10 years for the economic assessment – as much as 15-plus years in some cases. And part of this, of course, is in the lighting design process, you have to understand lumen maintenance, and Chad is going to be talking, in depth, about that.

I can't stress enough the value of lighting design. This particular architectural application has seven different exterior solutions incorporated, and I would say that the 19 kilowatts vs. the higher number, the 19 kilowatts without some pretty in-depth lighting design would be a bigger number. The economic impact would also be a bigger number.

So if you're looking for the most optimized illumination performance and that kinda performance, lighting design is the – really – the biggest value. The solutions can be wonderful, but if they're not incorporated properly, nothing really can override a deficiency in lighting design. And this requires that all the products that you're considering have very complete and comprehensive product-performance data.

And Chad, later on in the presentation, is going to go over, in detail, maybe even just the disqualification process. If you are trying to qualify certain luminaires and they're deficient in data, I believe it is grounds for disqualification.

So we've been talking a little bit about the architectural space where, in most cases – design-built, certainly, in most cases, and even in the upgrades – there's some design process incorporated. But this is a photograph looking out in the city of L.A. The city of L.A. is going through an upgrade of their street lighting. Roughly, about 140,000 of those luminaires happen to be just cobraheads.

It's not possible for us to take this application – we as an industry – and through an RFP process, otherwise, do a point-by-point for the surfaces being illuminated through their municipal roadway structure of 140,000 luminaires. So the idea is, here, that maybe you break this into certain categories – certainly, something more than one – and you create model applications that allow the industry to really, through a request-for-proposal process, give their solutions into a proposal and bid process.

So we're going to talk about essential data, and this is where I turn it over to Chad Stalker.

Chad Stalker: Eric, thank you very much. I'll make sure everybody can see my screen. I'm going to just switch. Can you guys see the presentation mode, now?

Yes.

Chad Stalker: All right. Perfect. Thank you. So Eric, thank you, and I want to thank everyone for taking the time to get on this and learn more about the whole roadway lighting criteria and everything that's going on.

So why do we want to talk about essential data? All right, first and foremost, LEDs are still new in the market, and people want to understand how this impacts the performance and characterization of LED luminaires. It's a different source. They want to understand it. Secondly, we want to really demonstrate how the markets align over this new technology with the needs and methods of the lighting market. At the end, hopefully, we'll show you that the characterization that goes into LED systems follows a lot of the same criteria that happened in traditional lighting, and that it's more and more aligned with what you, as people who are trying to specify and understand fixtures for installation and roadway, know and feel comfortable with when you're evaluating new products.

Let's go to the first slide. First and foremost, when people are looking at any lighting system, they're always looking to understand the photometric performance, and the takeaway here is that there's really no difference between what LED luminaire systems and how traditional luminaire systems are characterized and overall photometric level.

The methodology is actually covered under a standard called LM-79, and I'll get into it a little bit more in a second. But what it does is it looks at the SSL product, dominantly, as a bundled system and does that characterization. So in the end, you still get the standard IES data file that you would use to do your layout and design planning. And the thing to understand with this is it's still the same as any other IES file, where it's still the initial photometric performance, and what we're going to talk about next is some of the steps being taken to allow you to look at lamp lumen depreciation and lumen lamp light-loss factors.

Let's talk a little bit about LM-79 and what that is. Now, as I said before, it really bundles the performance of the fixture into a single report. It looks at the effects of thermal management. It looks at optical control. It looks at electrical power and efficiency, and the goal being is to provide a level of characterization that anyone specifying an LED fixture can look at and understand what the system looks like in an operating situation.

And there's an aspect to this that's called "in situ test measurement," which really looks at the system running in operation. So it's not just characterization of separate components. It's looking at it from a system.

Best thing I can reference here is – I've put a reference down at the bottom – is the DOE has a great paper. It's just a couple pages long but does a really good overview of what the LM-79 report is and how that's done. If anyone has any questions, really take some time to read through that. I think that'll address a lot as you're looking at this. But the takeaway there is that it's a level of characterization for fixtures that allows you to understand the system as a whole.

Now, Eric made a mention before about application fine-tuning. I wanted to talk about that a little bit more when you think about it from how you look at the characterization of the system, and he highlighted the fact that you have a lot of choices in optical distribution, so you can get good performance and targeted performance there. He also mentioned a point that I really want to highlight – is that the systems are scalable, whereas you're not stuck in kind of the big wattage change and light-level changes that you saw with traditional fixtures.

So from a design standpoint, this offers a lot of flexibility. But from a manufacturing standpoint, it offers a lot of skews, and that can actually raise a question when you're looking to purchase a product is, "How can anyone keep up with all this?"

And what's nice about that is that the MSSLC spec has actually looked at how to approach that. And they developed what's called the "product family testing," and it allows the manufacturers to do a level of characterization around the baseline products and look at iterations and actually not have to go through all the testing.

Some people would feel uncomfortable, be thinking that there'd be small changes. But what's nice about the approach here is that the appendix outlines how – acceptable ways of approaching that, and that adds a lot of value when you're working with a fixture vendor, understanding when they're presenting different products. It's also a point of validation for the manufacturer. When you're talking with them, and if they're presenting a number of different products, it's a good validation to make sure that they're approaching it in the proper way.

So in some cases, you might look at this as if you can't get this information or the rationale from the fixture vendor, you need to dig in deeper in what they have for testing and requirements for the products. But most of the vendors that are working in those markets find reliable product, have thought this through, and already have their process and approach to it, and it's a very good thing for you to understand as you're talking with your vendors.

Now, the next part we get into is accounting for lumen depreciation, and this is one area that people, they look at it as — it's different than what's happened with traditional sources. And the benefit here and the thing that we're going to highlight is that the market has taken a very scientific approach to address this and align it with the way and the methodology of the market in the past.

So it's not changing and redefining how lumen maintenance is viewed at a system level. It's actually taking a scientific approach at the device level and bringing it through to allow you to do lumen maintenance and light-loss factor calculations at the system level with reliable and repeatable data.

So when we're talking about LED component lumen depreciation, there's actually a standard in the market – people have probably heard this a number of times – called the IESNA LM-80 standard. And this is basically a measurement methodology for the LED device lumen maintenance.

Now, what's nice about this is it's been adopted across the market and really allows any device manufacturer to consistently characterize the performance and lumen maintenance at the device level. It's repeatable. The methodology is well defined, and it allows a consistent level of reporting independent of the type of LED.

Now a couple things to highlight, this is not the characterization that you would typically look at for like lumen maintenance and lamp life of a traditional lamp. It is different. But this coupled with another methodology we're going to talk about in a second allows you to get an understanding of the performance of the system in an effective way that basically is in line with what you're traditionally used to.

So let's go to the next slide, and let's look at this from the standpoint of how it's done just so how you have an understanding about the process. And this, again, is independent of the LED type. What we're looking at is there are specific drive currents that are defined. There are specific temperature points that are monitored, and it's consistent between vendors. They all have to define a specific point, and that's coupled with the junction temperature of the dye, which we'll talk about in a second.

And what they do is long-term testing of the devices at a minimum of 6000 hours, where you're continuously looking and evaluating what the performance and the lumen maintenance is of the system. It does allow you to do interpolation, so you can look at other things, but you basically get a band of performance that you can look at and apply in your next steps when you're looking at the projected life of the product.

This is kind of what it looks like on the inside, what's going on when the LEDs are being tested. They're typically put into a chamber. There's heavy controls around temperature, both of the air and of the thermal management system to maintain a consistent temperature at the device level. And this is so that you're not applying any other variables other than time and drive current and keeping the temperature and just basically at temperature bands.

And that's from the standpoint of an LED manufacturer. Those are the critical things. With this information, you're able to pull together a lot of characterization on the LEDs.

Now, let's look at what happens at an LED level. So at the LED, it's being driven by a specific drive current, and again, you can set those at different levels. It's being maintained at a certain temperature level. But here's the one thing I mentioned before: People always think about junction temperature, but that's not what's being monitored.

What's being monitored is, typically, what's called a solder point temperature or a case temperature, and it's important to understand that in the fact that that TJ is actually a theoretical point looking at the thermal path between the test point and the dye. So you hear people talk about what I'm going to run the system at and how hot the LED can get. Where that's being monitored is typically at a test point along the case of the LED and not some point inside the architecture of the LED itself.

And we talked about this being a test that goes on for 6000 hours, multiple test points, also multiple devices being tested, so you have statistical validity for the information. And what's great about this is this is what you get for an LM-80 report. You get a lot of raw data that if you're having problems sleeping, it's a great reference. The key thing, what comes in front of this, is basically a summary report that summarizes the performance characterized over the period of time, and that information is what people then can use for doing more of the projected lumen maintenance of the system looking at farther times out into the future.

And this is where we get into another methodology, which is called TM-21. This is where, honestly, where I look at it as kinda the rubber hits the road. This is where you're able – they've defined a clear characterization method to take that raw data that all of the LED suppliers have put together for their products and look at what the performance – lumen maintenance performance – of those LEDs would be out into the future.

Now, remember we're still talking about the individual devices. We're still looking at those as single components. Overall, characterization and everything that goes into defining lumen maintenance of the system also has to take into play other things, which I'll talk about in a second, but what we're able to do now with this is develop a characterization method that people can look at consistently where the performance of the LEDs will be into the future.

And this was developed by players across the market. This was not something where people looked at – like someone came up and said, "This is now the methodology." It was a long-term development between these organizations to come up with something that would align with the market so that people doing the things like you're doing – evaluating products, making determinations on what's the right product to use in your application – can do it with an informed method and do it consistently, especially when you're comparing multiple products.

Now, a little bit about TM-21. So we talked about it's a test methodology that happens, that there's a minimum test duration of 6000 hours. But typically, the reported values basically derived from that are at different temperatures, but this can be done out however far you want. Now, there's a limitation on the projection. You can only look at it up to six times the tested amount. So an example I have here is, for example, if we have 10,000 hours of testing, that allows for 60,000 hours of reported lumen maintenance that you can do based on the calculation with TM-21.

Now, there's also the opportunity to project the values. We talked about interpolating. If we know that the system is being run at certain drive currents in between these areas, you can do a level of interpolation which allows you to do that. So again, it's kind of like the product family. They've thought of ways of dealing with, "What if I'm working in another area?"

There's also, typically, what you'll see is a number of manufacturers will talk about calculated values, and these are the expected lumen maintenance out in the future, kind of past the 6X level. And it's becoming more and more common in the market where manufacturers want to talk about what they can look at for calculated lumen maintenance – I mean – I'm sorry – reported lumen maintenance and also calculated.

And that's confused the market some, but I think it's important to understand that the reported values are basically statistically and data-based numbers, so they're based on actual test data, whereas calculated values go out beyond that. There's nothing to indicate that they aren't inaccurate. It's just the fact that the information is purely from a calculated level, and we're going to talk about where that has some value in a minute.

So far, we've talked about we've got a bunch of data that people can get around the LED system that aligns well with the traditional methodology for characterizing and evaluating products. So let's look at how this works into a system like when you're looking at writing your spec or you're evaluating product.

So we'll use the MSSLC model spec as an example, and we'll look at, specifically, 1.6. So this is looking at, "What's the information I'm going to be looking for?" So there's a first section, which is Section A, which is kind of a general content. If you look at the list, you'll sit there and go, "Yeah, all these things make sense." You'll look at, "Well, they want LM-79 photometrics. That makes sense. I understand that."

People typically get to Section C, and this is where things are a little bit different, and like I highlighted, the difference between LED and traditional product, all right, this is where things – people go, "Wait a minute. I'm not sure I understand what this is," and they're sitting there thinking about it, going, "Why am I looking at this vs. just give me the cut sheet on the source like you used to do with an HID or something?"

And because you want to get to the next step, we're going to start looking at light-loss factors. So why is this step in here? Well, it's there to do a level of validation on looking at what the lumen maintenance is on the system.

And when the standard first came out, there was different ways the manufacturers approached it. We're seeing more and more alignment in the market where the manufacturers have aligned better with characterizing what they do performance-wise in different environments and looking at the lumen maintenance. An example is here, where they look at what the lumen maintenance factors are depending on locale, depending on other kind of the environmental parameters.

And you can see a difference here where you look at looking out at calculated vs. projected, so you can get an understanding of what your lumen maintenance will be for when you're doing your design. As Eric highlighted, design is critical. As we talked about before, photometrics are absolute – basically initial photometric performance – but you need to look at what the performance of your system is going to be over the life. Information like this is now readily available to help you as you're working through this.

In the Appendix B of the MSSLC model spec, they even help identify different approaches and supported approaches for doing this. Probably the most common one that you'll see when you're working with your manufacturers is that they've looked at the lumen maintenance at the component level at the LED level and done everything we just talked about in the past, and they can provide that information to you.

It's very easy to ask them for the information. Information is readily available in the market as part of those systems. If you're not able to get that information or if they're not able to present it, I think, as Eric mentioned before and we've highlighted, these are areas where you need to sit down with a manufacturer and understand why they don't have it or get more detail onto it. It should be at least a red flag on, "Why can't I get this?" Beause you're about to make a decision long term on this technology.

Now, there's another option, which is Option 2, which is the luminaire performance, where they basically do a level of testing at the luminaire for an extended period. Some manufacturers will take that approach if they want to try to characterize the overall system but less and less valuable. It's – actually ties up a lot of product, and it's a long-term test, but it's there as an option. More so, it was critical maybe earlier when not all LED manufacturers had all the testing for their products. A fixture manufacturer could accelerate their time to market doing this.

So let's wrap up on essential data. First thing we talked about was the IES photometric file. That's pretty straightforward. There should be a check next to that, too, but that comes out pretty easily. That's easily available. There's methodologies for doing that. It's no different than what you used to get.

The LM-79 report is, from some perspective, you look at that as the additional validation on the system-level performance, and then lastly, the LM-80 report, which you work with the TM-21 approach, gives you derived values at a luminaire level, and you can look at that for lumen maintenance. Then you can move on in your evaluation and design to look at light-loss factors and everything.

So that essential data is available. It lines up with what the market needs. It lines up with methodology vs. maybe a number of years ago where different vendors had different approaches. Now, these methodologies and approaches are well defined and followed across the market. It is definitely clearly a way of validating the products that you're getting. If there's any questions from your suppliers that they cannot provide this, that should be raising a question in your mind.

Okay, I think I'll finish up. I'm going to pass this back to Edward and let him continue a little bit more in depth on the MSSLC spec.

Edward Smalley: Thank you, Chad. Appreciate that. And so as you can see, Chad mentioned several times the Consortium's Model Specification for LED Roadway Luminaires. And it really is a document that had its beginnings beginning with the Consortium in 2010, where we saw it was very important for us to develop a document that we could use as municipalities, as utilities, and as system owners that would stand the test of time. And so this document was delivered on October 15, 2011.

So that's the background, but what does it really do? What's the purpose? It's really, when you think about it, risk mitigation. Chad had talked about the LM-79 and the IES files. We can get those files, and we can do our modeling of our roadway with various different luminaires, and we can see that, through those documents and procedures, "Okay, this light will work on this roadway." But what we really wanted to make sure was that, over time, the luminaire would perform as advertised and do as we had expected on our roadways because, as pointed out, we are making a pretty large investment – in many cases, millions of dollars being invested in this technology as you go across the different cities. So it establishes a common framework.

We worked with numerous manufacturers – about a dozen different manufacturers to develop this document. We really wanted to make sure that if we utilize the document, the end user is likely to receive more than enough submittals from the different manufacturers of fixtures that will fit their application.

So we had a strong focus on streets and roadways and pedestrian ways as opposed to parks and area, and that's important here. There are a number of things the document was designed to do, but when you think about it, if you focus on retrofitting existing lights in place, it does a very, very good job with that and also in new installations, where your roadway is just newly being developed.

What it doesn't do, however, it doesn't focus on lamp and ballast retrofits, where you're gutting the fixture and replacing it with an LED driver and LED array. Now, that's not to say that there aren't good products on the market that can do that, but that's just not the design and focus of this tool.

So right now, I want to turn it back over to Eric, and he'll start talking a bit more about the concepts and how we can move forward with that. Eric?

Eric Haugaard: Thank you, Edward. So let's discuss some of the details of the concept. It's quite simple actually. It's a concept derived from the objective to enable, as closely as possible, application-specific performance evaluations for maintained illumination performance at the application level rather than matching lumens.

So a lot of the people in the market have a tendency to look at their inventory and say, "Gee, we're really interested in upgrading from maybe a high-pressure sodium technology. I'm going to send you the specification sheet on my luminaire," or even, "I have a 250-watt high-pressure sodium or a 150-watt high-pressure sodium. Send me the best replacement."

And the answer, when we get in the details, is it really depends on a lot of things because we have not tried, as an industry, in LED roadway luminaire manufacturing to replicate the performance of the incumbent source technology products. We've really tried to improve upon it and embrace and incorporate all the different things that LED technology can do better and differently.

The actual templates that I'll be showing are designed to be – this was mentioned – customized by each adopter to meet their unique requirements. And there's actually two methods being outlined by the Consortium doc. The one that we prefer that everyone use is the one that is application dependent, and it requires some application-specific details to be developed and allowing the applications to cover a range of variables, and we'll go into some of those details.

The part we're not going to be covering is the application-independent portion that's also contained in the specification doc. It is really more product-related features and actually is kind of more lumen and IES type. And of course, if you do that, there is almost absolutely a compromise in illumination quality and economic value.

So I've broken up parts of the template into a few different slides here, located in Appendix A, and this is the application-dependent greatly preferred methodology. The first portion I'm showing here really is just the site-specific parameters, so the roadway geometry in this particular case and the position of the luminaire over the roadway. And it's important to note that this is probably best utilized if you have a model specification or a model application.

So if you're breaking your municipality, for instance, into five different roadway configurations, roadway types, plus the luminaire configurations for the positions over the roadway, maybe this would be just one of those five. But it has all the critical information that a lighting design process would require.

Edward Smalley: Hey, Eric?

Eric Haugaard: The next – yes, sir?

Edward Smalley: Eric, this is Edward. When you say breaking your municipality into different roadway types, would you consider like residential-type roadways, collector-type roadways, major arterial-type roadways, those types of applications?

Eric Haugaard: Sure. Absolutely. And then within those applications, once you categorize them as broadly as that, you can then get specific as to which ones have illumination on both sides of the road vs. one side of the road, which applications have a boulevard included with illumination through the boulevard vs. maybe applications with a boulevard with luminaires located at the curbs at the residential side.

So it could be, really, whatever the varieties dictate, but also I think you have to look a little bit more forward and look at your existing inventory management for that infrastructure, and then really determine, "Do I want to expand that inventory?" And of course, the advantage to that might be that I can get more application fine-tuning, as we put it, but then, of course, your inventory gets bigger, and it's difficult to manage the inventory.

So I'm glad you brought it up because it's really a discussion that should take place from a visionary standpoint. When you're doing your first roadway application and you know that it's going to just probably happen more and more, plan for the future because inventory management is a critical part of this whole selection process.

In this particular case, we're looking at the actual application performance from a photometric standpoint. So what are we really trying to achieve? A good resource for this would be the IES recommendations – could be populated. And again, I think if there are not really too many non-applicable categories here. Some kind of override others but needs to be completed in as much completeness as possible.

And then ending up here with the last four that I'm going to highlight – the luminaire functionality and, somewhat, the durability requirements – for instance, input power, this particular example shows a nominal maximum input power of 103 watts. Why would you want to cap it at some given wattage? Well, you could have an electrical power infrastructure objective that you want to make sure that you do not exceed for a variety of reasons.

Or this could be driven by a utility requirement from a rebate standpoint, for instance, where if you have a 206-watt luminaire existing and you need to cut at least 50% energy to initialize the rebate or qualify for the rebate, then, of course, those are typically pretty substantial part of the life-cycle cost equation. You want to make sure you achieve that. The correlated color temperature and all the rest of the variables are certainly things that could be a constraint on your system that you do not want to conflict with.

How do we address application variability? This, again, is some stuff directly out of one of the formal presentations from the MSSLC spec and just showing here that this is somewhat of a typical residential – municipal residential roadway. You do not have a lot of commonality when you look at some of the spacing differences, including the intersections. And of course, we just can't simply place luminaires on poles where we have driveway approaches and fire-safety type conditions, trees, etc. So it really just stresses the fact that we have to have some type of model that somewhat normalizes the range of this variability, and it's going to take some work to do that.

So just further stressing this again, taking a look at the city of L.A., for example, there's an impossibility here to try to create every single location into a lighting design process. So I'm going to follow up here with some examples.

So the first thing I'd like to identify is that the IESNA has a standard that they like to promote, and it's very widely used. It's called the RP-8 Lighting Recommended Practices for Roadway Lighting, and so imagine that we have a specification that we have created for an RFP, and we're going to try to utilize the RP-8.

So directly from the RP-8, for my example, I'm using a collector, medium conflict area, and if you look at the excerpt from the table in RP-8, I have highlighted "collector medium conflict" – wouldn't have to be medium; could be low or high. If you look at the selected pavement classification, there is an illuminance value of 0.9 that is highlighted and a uniformity ratio of 4.0. It could be lower than that but cannot be any higher than that.

So 0.9 foot-candle average and a 4-to-1 average to min. or less, so it would have to be above 0.9 or above. And then really, the next part is, "What is my application constraints? What's my geometric target?" Essentially the position of the luminaire over the roadway.

And then some other details that are important – the ambient temperature environment shown here at ten Celsius. Why do we need that? Well, I think Chad covered that in great detail. He showed you a table that showed a map that characterizes the average nighttime ambient temperature, and applications in Northern Wisconsin vs. South Florida will have different lumen maintenance given the exact same conditions with the exact same application conditions using the exact same luminaire because the warmer conditions will accelerate that lumen depreciation.

Then there's a discussion on the required hours of operations – 50,000 hours, approximately 12 years, and then we typically have to make a choice on what our correlated color temperature is. And in this particular case, we're just selecting 4000. So at some point, through a design iteration process, somebody believed they've optimized by selecting a given luminaire.

This is just a specification sheet on a luminaire. It really doesn't have much bearing on our discussion except for when you plug in the photometric data for that particular luminaire with the application variables all considered, we're just highlighting here that we have met the average of 0.9 or greater. In this particular case, it's 0.91, and the average-to-min. uniformity is well within the 4-to-1. It's actually just shy of exactly 3-to-1.

Also important to mention here is that the lumen maintenance factors for this particular luminaire were incorporated, so this is not the Day 1 performance. This would be the, in this particular case, the 50,000-hour value. Also important to stress is this 0.91. I wouldn't say we always get this close from a fine-tuning standpoint, but this is pretty well optimized if you're only roughly about a percent over the bogie.

So what about alternate evaluations? We have the proposed product, and it's defined as an IES Type II Short. So for this given manufacturer, their IES II Short distribution is going to be available in a variety of different outputs.

What if I wanted to build a different IES Type II Short? Same distribution type, but maybe I wanted to do it for a little bit – a different economic set. Well, they're both IES II Short. What's the difference? Well, I can build this luminaire, so that adds 50,000 hours. They produced about the same amount of light, but somehow – maybe by balancing a number of LEDs in the system with the optical distribution, again, being exactly the same – I can create a luminaire for 139 watts vs. 119 watts.

And typically, what you would see is for that equal performance, photometrically, and for that savings in wattage, there's going to be a difference in cost. So it would be 139-watt power consumption luminaire yielding the same results, would typically be less expensive than the 119 watts. So the question is, for that 20 watt of extra savings, which one provides the best value? And although there's a composite value that incorporates a lot of different variables, one of the most important ones, of course, is energy cost.

So just for example, this is something that we collected off of Wikipedia. I can't speak for the accuracy, but just identifying the source there. Just from a simple payback standpoint, if the luminaires are $100.00 difference in cost, maybe at 43 cents a kilowatt-hour you would do that. Perhaps, at 6 cents a kilowatt-hour, yielding $60.00 savings over time, you wouldn't do that. Now, what this does not incorporate is other economic variables that would be affected by rebate structures and things like that, but it's important to note that this is a very, very important part of it, and it speaks to the fact that we could do economic fine-tuning by sort of balancing the – sliding the fulcrum that has, on both sides, the cost of the luminaire vs. the energy savings.

Second step in the process is kind of what I call some of the pass-fail functions, and I tend to call this "functionality and durability requirements." They don't directly affect photometric performance, but the electrical input range, most of these luminaires today are available in multi-range power supplies that will take anything from 120 volts to 277 volts, and that's just one of the ranges.

Maximum power consumption – of course, the infrastructure rebates and things like that, things like power factor – THD – and we can go on and on with some of those.

Mechanical issues also – the same environmental factors that affect incumbent source technology products are still there. So if you're concerned about different levels of vibration and shock resistance – mechanical shock resistance – and corrosion resistance, there are ways you can characterize that and certainly nothing better than backing all of that up with a good warranty.

So I'm going to go through a short litany here of some of these examples here, but basically, this particular slide just shows that you've got some mounting constraints, for instance, you need. You got some constraints on the actual mounting structure, the pole, the mast arm. Do you want it UL-listed? Does it have to have anti-vibration for a normal applications resistance or for bridge and overpass specific to the application? Does the driver have all the things you want? Is the electrical power characterization FCC compliant, etc.?

I won't point out the second bullet point. Internal surge protection – 10 kV is what the Consortium specification shows as one of the high limits, and I will just refer you to the MSSLC model specification Appendix D for that. I think it's well on its way.

As was mentioned earlier, by the way, the specification is designed to be upgraded over time, so that's one of the categories that I believe will be refined over time. Even things like passive thermal management, people typically don't want active cooling, such as fans incorporated, because it's another failure point. International Dark-Sky Association is something that's typically what municipalities are looking toward and then things like your warranty and your finish warranty.

I show this slide as "nice-haves." These may not be pass-fail metrics, and it's something that would disqualify otherwise of a manufacturer if they don't have this. But if you have a quality system, doesn't have an ISO-9001 facility, maybe that's not something you desire – things like tool-less entry; recyclability; things like adjustable output power, which allows you to manage inventory differently, so one single luminaire can be used in a variety of different locations, one piece of inventory; leveling indicators; quick-disconnect power supplies; and even people that are looking for a manufacturer that has maybe some years of experience.

There's a lot of people entering this business that are not luminaire manufacturers that are startups, and not saying any of those are bad. Matter of fact, I think a lot of those will turn out to be excellent solutions, but it does ask the question, particularly on the warranty side, "Do you have something to back it up with?"

This is the slide that may get a little contentious. Do not specific product performance variables that directly or indirectly affect illumination performance. What does that mean? Well, for example, we have a tendency to look at spec sheet variables – that's a luminaire – total luminaire efficacy or lumens per watt. Why would you want to or not want to do that?

Well, I think the example that we showed earlier, we can adjust the lumens per watt by adjusting the cost of the luminaire. If, in fact, I go through a life cycle cost analysis and I select what is the most economical solution that yields the required illumination performance and then somebody says, "Well, we have a limit on efficacy of 70 lumens per watt, and yours is 68 lumens per watt," my option might be to throw some more money at it.

Maybe I'll just use the $100.00, for example, and I can do that by adjusting the LED system to be more efficient at a cost. However, maybe I go to 72 lumens per watt. The luminaire costs $100.00 more and for that over the life of the project, I get a $50.00 savings.

So I just caution everyone that putting efficacy limits on luminaires does not have to be done if you drive everything to the application and look for the lowest total cost of ownership because where energy is free vs. energy is expensive, you're going to have different solutions. You cannot, even across the U.S., do this, and the Big Island of Hawaii, where it's 43 cents a kilowatt hour vs. some parts of the where it's 4 cents a kilowatt hour, you're going to have different solutions, and those specifications will conflict with one another if, in fact, you do that.

The other thing to remember is the total lumen efficacy is the composite of a lot of different variables and absolutely do not specify optical efficiencies, thermal efficiencies, electrical efficiencies. And I bring this up because we, as manufacturers, see that level of granularity and, as Chad said earlier, keep it all bundled, keep it bundled at the best piece of data that we have to work with – to start with – is the actually the photometric file. That has all these things incorporated.

Specific IES types – again emphasizing if you have a specific IES type for your roadway applications, for instance, do not assume that that exact same IES-type classification is going to be the best. As was mentioned, IES types are more the gray area than the black and white. Every one of those type classifications can be right on the cusp of another completely different type classification by just barely changing the performance.

What would an example of that be? Well, here's a Type II Medium from a given manufacturer. Here's a Type II Medium from a given manufacturer, and here's a Type II Medium from a given manufacturer. I think in this particular case the most noticeable difference is the illumination behind the pole is significantly different. If I go back to the initial slide and forward to the last slide in this deck, the roadside illumination may be significantly about the same, but for applications where you want or don't want sidewalk illuminance, the applications where you want or don't want significant house-side illuminance, these are going to be very, very different.

There are many economic variables that are outside the scope of this spec and really any good spec, I think. Energy cost is certainly one of them, although it's a very significant variable. So the message here, I think, is that if you're paying for energy a certain way, maybe explore with your energy provider different ways of buying energy, and I won't go into the details of that. I'm not an energy professional, but I know there are ways you can work with the energy providers to buy energy differently that makes it mutually beneficial.

Maintenance costs, I believe, are the most – probably most misunderstood portion and therefore the highest degree of variability in a life cycle cost analysis. I can certainly cite cases where we've seen plus or minus 300% differences for similar applications on total maintenance costs for the people that are maintaining their own lighting infrastructure.

There's a lot of other system operation options, too. Public-private partnerships are becoming more and more prevalent – maybe not as much in the U.S. but certainly outside the U.S., and there's always different ways of doing financing. And again, not being an expert on that, I just encourage people to take a look at the way they can actually get these financial options brought into scope.

The MSSLC also has a retrofit financial tool that can help with some of the financial part of this. The link is highlighted, and this detailed analysis provides various different ways that give you economic analysis yielding net-present value internal rate of return, even things like estimates on annual greenhouse reductions.

So, just as we move toward a wrap-up here, I think the important takeaways for all applications – and I know we're focusing on municipal street lighting – but really, this applies to any application where you're considering LEDs is application-based performance evaluations really give you the best opportunity to optimize your best illumination and economic performance. And it requires, of course, complete and comprehensive data, and then the lighting design has to come along with it. You have to incorporate this at the application level as specifically as possible, and then also realizing that for large opportunities, maybe you have to characterize that with model applications that actually, as accurately as possible, have categories for those different applications.

And as Chad mentioned, don't be ashamed to disqualify solutions that have incomplete data. Sometimes the best way to qualify your luminaire supplier base is to have a disqualification process at the front end of it. So we've seen evidence that people look at all these different luminaire solutions, and they try to qualify them. But in essence, there's probably a way you can sort of disqualify the ones that will never make it to the end – as early in the process as possible.

Always use life cycle cost analysis approaches whenever evaluating your illumination solutions. Of course, it does not have to apply LED, but that's what we're talking about today. Any solution really has to go through this, and the goal, of course, is sustainable illumination performance and sustainable economic performance. And with that, I will turn it over to Edward again.

Edward Smalley: Thanks, Eric and Chad, both of you. So we've really done a good job today. Both of you, I appreciate what you've done so far. I wanted to just give a brief overview, and I won't go through any more slides here – just one or two. But some bottom lines and some takeaways, as Eric has just given us, we really want to do a pilot evaluation of your process.

You want to really go through the process of evaluating and modeling your different fixtures, and then what we recommend at the Consortium – and we have a brief process up on our website that you can follow – is we recommend that you do a pilot evaluation, that you get your customers, your citizen opinions. You survey your citizens and ask them what they think.

And then after you've gone through this entire process, you modify your specification, whether you've adopted the Consortium specification or modified your own previously, but that you modify it, taking into account the different opinions that were formed by your community, but also the lessons learned that your engineers have learned, through the process. And that way, you're likely to get a product, as you go for full deployment, that's not only going to satisfy the engineering checkboxes that we all like to do but that your community would accept, as well, because there's nothing worse than putting up a million-dollar project, and your community really doesn't like it.

So I wanted to kind of just end there with that note and basically really thank you both – to Chad Stalker and Eric Haugaard. We appreciate your input here. Until then, we thank you, again, and have a good day.