Text-Alternative Version: LED Replacements for Linear Fluorescent Lamps
Below is the text-alternative version of the "LED Replacements for Linear Fluorescent Lamps" webcast, held June 20, 2011.
Theresa Shoemaker: Welcome, ladies and gentlemen. I'm Terry Shoemaker with the Pacific Northwest National Laboratory, and I'd like to welcome you to today's webcast, "LED Replacements for Linear Fluorescent Lamps," brought to you by the U.S. Department of Energy's Solid-State Lighting Program. At this time, all participants are in a listen-only mode.
We're very happy to have as our speakers today Jason Tuenge and Eric Richman of the Pacific Northwest National Laboratory. Jason Tuenge joined PNNL in 2008 and has 11 years of experience in the lighting industry. He provides lighting engineering support for the U.S. Department of Energy activities within the Commercial Building Initiative and the market-based programs for SSL, Solid-State Lighting. Jason also manages the Technology Fact Sheets Program, producing educational materials on topics ranging from solid-state lighting itself to human factors and economics.
Eric Richman has been with the Pacific Northwest National Laboratory since 1986, working primarily in building energy efficiency and lighting technology application analysis. He is a part of the U.S. DOE Solid-State Lighting Program, working on the development of LED test methods and standards, performance program specifications, and field assessments of lighting applications. Eric is also involved in the energy codes and standards work with the 90.1 Lighting Subcommittee.
Today's webcast will discuss current Lighting Facts-listed products as well as products evaluated in the latest CALiPER Reports. Eric Richman will report on a recently completed GATEWAY Demonstration project, in which LED and fluorescent lamps were installed in a variety of recessed troffer luminaires for comparison in an office environment. The presentation will conclude with a discussion of specification/s listed in a newly updated Technology Fact Sheet
Jason, please begin.
Jason Tuenge: Thanks, Terry. So, in this presentation or in this webcast, we're going to be going over a number of subjects. This provides an overview on this slide. We're going to open with an explanation of why we're looking at this particular product category, and some may wonder why we're devoting so much attention to replacement lamps when sometimes — or oftentimes, luminaires can serve a more optimal solution — or provide a more optimal solution. We'll also look at the information provided by a number of the Department of Energy programs for Solid-State Lighting, including Lighting Facts, CALiPER, and the GATEWAY Demonstration Program. We'll also be looking at some suggested specifications that we've produced and are available online, and also put this in perspective and by looking at some other applications where we can put these products, such as parking structures, while also looking at some alternative technologies, such as luminaires, complete lighting fixtures, not just for replacement lamps.
So, why are we looking at LED replacements for linear fluorescent lamps? First, this product category is ubiquitous; you see them everywhere, in offices, in industrial facilities. And, basically, this is low-hanging fruit. There's a lot of money associated with this and a lot of interest in anything that might pose a superior solution to fluorescent in these applications. There's a lot of interest in LED technology. A lot of this is probably due to marketing, in general, publicity around the technology, not necessarily this product type or this particular product category. Also, when you're looking at replacing products, replacement lamps can be seen or often are perceived as something that's going to be cheaper and easier to replace than getting up there and replacing the whole fixture. There are a lot of this particular product type available out in the marketplace, and so this is well — is merit for consideration for this product category. Last, as we've mentioned in a number of our publications and previous webcasts, LED performance claims are often pretty lofty, either in terms of output or efficacy, efficiency or lifetime, and sometimes these claims are very vague, and when you get down to it, there's really no data to support the claims. And so, that also applies to some of these products, and we'll be looking at that as we move through these slides.
First up, we've got the Lighting Facts® program. This is a program that is voluntary. This is not something that manufacturers have to do. Those that agree to the Lighting Facts policies will show a logo on — as shown over on the right, as an example—on their product labeling, and the idea here is that you've got something that's analogous to a nutrition label that you'll see, say, on cereal boxes. So, this provides a very neat and tidy and consistent summary of the primary features for the product. How much light do you get from the thing? How much power does it consume? Then, you take the ratio of these two units and you get efficacy, which is another metric for efficiency; this gets more to the point. And then, also, you get two more metrics addressing color features: color rendering index which, basically, tells you, you know, does this thing render red and blue and green all equally well or, at least, relatively well, or does it do fairly poorly at possibly one of those? And then, correlated color temperature, and this gives you a sense of, is this thing warm in appearance or cool in appearance? Is it kind of orangish or bluish in appearance?
One thing that's important to point out is that just because you have a Lighting Facts label on your product doesn't mean that it's going to be a good product or that it's going to meet your particular needs. Just like in the old Calvin and Hobbes comic strip — I'm sure a few of those in attendance have read through a few of those — Calvin's favorite cereal was Chocolate-Frosted Sugar Bombs, and this was the sort of thing that my mom wouldn't let me eat when I was little, and for good reason, probably 99% sugar. So, in that case, it's got the label; that doesn't really mean a whole lot. At least, it tells you, though, what's in — what you're eating. The same thing goes for these light fixtures. So, it's good just to have the label so you can then compare two products, apples to apples.
Another thing that this Lighting Facts project serves to do is to collect all this data in a database. You can go to the Lighting Facts website and search through all these products and see what's actually out there, at least that's been submitted to the Lighting Facts project or program. And so, here, I've taken a screen capture of the website, and you can see that the search tool already selected the product type replacement lamps for linear fluorescents, T8, T5 or T12 tubes, and then we see our five parameters down the left side: output, wattage, efficacy, CRI, CCT. And, basically, you just hit "Go" and it'll go and find all the products. Currently, we're looking at about 102 of these products, as shown over on the right. Another thing that the Lighting Facts program does is provide snapshots or periodic reports summarizing what's listed on the site and what it might mean relative to other options out there, such as fluorescent. One more note is that, in order to be listed, manufacturers have to have their products tested per the Illuminating Engineering Society of North America's test procedure, LM-79. This is a standard test procedure, so people can't just test however they want, they need to test in a consistent manner, so again, you're comparing apples to apples. One thing to note is that you will see other labels out there that look similar to the Lighting Facts label — it's, actually, by the same name, through the FTC. And for more information on this, you can visit the Lighting Facts website, and it provides a comparison of the two very similar labels.
So, looking at this, the most recent snapshot from Lighting Facts, this was released on June 1st, I believe, and it's dated in May, and basically looks at three different product categories: A-lamps, so the traditional incandescent bulb, basically; reflector lamps, so the type of bulb you're more likely to put into a recessed downlight; and then, our linear fluorescent lamps that we're looking at today. And so, these pie charts give you a sense of the percent coverage or representation of the linear fluorescent lamps among all the Lighting Facts products that you'll find on the website.
So, the summation snapshot is actually based on data gathered in February, so it's a little bit behind in terms of current products. But, as I'll show later, it's not that far behind. So, here, we can see that the previous snapshot was in September 2010, and you can see the growing population of these products on the Lighting Facts website. We've got, you know, basically, a doubling in the number, and we're also showing that in these reports, we also compare the LED products against whatever it is they're supposed to replace, in this case, fluorescent.
So, on this chart, this basically plots wattage along the bottom and light output along the left, you know, up at the left, and we've got LED products scattered about in blue, fluorescent products in the other colors, and it's broken down by T5, T8 or T12 lamps. A line — I'll try drawing here on the screen — a line from the bottom left extending up — that's not a very good line, sorry — extending up in a straight line long the page — there's a better line — you can draw a line such as this in any direction, and this line represents a line of constant efficacy or constant efficiency. So, what we've drawn in this case is — on this line here, we've got, basically, what's going to be required for fluorescent lamps as of sometime in 2012; I believe it is in July. Eighty-nine lumens per watt is going to be required for fluorescent lamps, not for LED, but it gives you a sense of where LED should be, just in terms of efficacy. And so, you can see that LED is, actually, fairly competitive in terms of efficacy, and that's just going to be — those continue to improve over time or can be expected to. As the technology itself improves, these products should also improve.
But what you can also see is that, while you have LED products saving you energy by being lower wattage, they're also lower output, so a lot of this energy savings is actually due to reductions in light levels or light output from the products, so not necessarily a good investment, something that you need to look closely on. The line that was shown there is from this point here, so basically, looking at the typical four-foot linear fluorescent lamp, T8, and most of your products are going to be under 4500 Kelvin for color temperature. But even if you're at a higher color temperature, you're still looking at 88 lumens per watt.
So, as I mentioned, the most recent snapshot was actually using February data from the Lighting Facts website. PNNL has taken some internal snapshots, basically, over time, and the most recent one was done May 31st. As I mentioned earlier, as of that date, we are looking at 102 of this product type; this is on the Lighting Facts website. Of these, 100 are of the type, basically, as I'd shown on the dropdown before, they're directly classified as T8, T5 or T12. There were of couple of others that just had T8 in their description, and they're just classified as "Other," so some manufacturers are submitting these products in different categories. But, for the most part, you're going to find them directly under linear T8, T5, and T12 tube category. Two products were excluded from consideration in this analysis because they were over 32 watts; in other words, they would not save energy.
Then, also, looking at the available information from the manufacturers for these products, found that these average about 50,000 hours for rated lifetime. This is just from the manufacturer literature, so there isn't necessarily supporting data for this, and it's worth pointing out that lifetime currently is not evaluated by the Lighting Facts program, so that data is not collected nor analyzed. That's something that is, actually, not really standardized at this point. It's best to get some clarification from your manufacturer as to where they get their rated lifetime from because we're still working on a standard for that, mainly TM21. I'll discuss that more later.
Also, interesting to note, that some of these products submitted to Lighting Facts are apparently submitted by consultants. Typically, when you'd hire a consultant, they're not also selling products. So, it's kind of interesting to see the border between consultant and manufacturer or salesperson blurring a little bit here, but you'll see that as you search through the database.
So, this gives a quick snapshot of available products on the Lighting Facts website. The average initial efficacy is around 80 lumens per watt, and output is around 1366. You do have some products that stand out from others. For example, you're getting — there are some products now listed at nearly 2500 initial lumens, and those are some real standouts. Average CRI, it's worth noting, were about 75, so this is getting pretty comparable to fluorescent, and average CCT is actually about 4500 Kelvin, and so this, again, is pretty comparable to fluorescent. It's not excessively blue, as sometimes is the case for these products that you'll find on the shelf.
Also worth noting, as I mentioned before, a couple of products were excluded because they were above 32 watts. Really, if we were more stringent, we would exclude products above 28 watts because fluorescent lamps are typically operated on — or, at least, T8 lamps are typically operated on 0.88 ballast factor ballasts, which then, effectively, underdrives them, and so they're really consuming more like 28 watts. So, to save energy, your T8 product — your LED product you're using to replace should probably be under 28 watts if you want to see some energy savings.
This diagram provides a quick snapshot of the available products on the Lighting Facts site. Here, we're showing relative efficacy with bubble diameter and breaking down by color into two different colors, so you've got above-average efficacy products and below-average efficacy products. And what you can see here is that there's no real clear distinction between, say, bluer-appearing LEDs in terms of efficacy versus the warmer or more orangish-appearing LEDs, nor in terms of Color Rendering Index. Efficacy is just kind of all over the map, it doesn't really — it's not really a function of CRI or CCT.
So, one thing we need to do to set up some of these future slides, in particular, looking at our CALiPER benchmarking, is to establish how we're going to evaluate fluorescent. The fact is that — or the way we're going to evaluate these replacements for these fluorescent lamps. The reality is that fluorescent lamps light in all directions whereas LEDs, at least, the replacements for these products tend to direct light in, really, just one direction. So, whereas the fluorescent lamp will direct light up into the fixture that it's installed in, LEDs will typically just send light straight out of the fixture without requiring ballast inside the fixture.
So, when you're looking at performance, you really want to see how they're going to perform in the product they're installed in. A great many of these products are installed in troffers, and so that's how we're going to be evaluating the performance of these products in the following slides. Most of these troffers break down into — or fall into one of two categories. These are all going to be 2 x 4 troffers, or most of them, and most of them are going to be either prismatic lens, like we see over on the left or parabolic louver, like we see over on the right. Prismatic is, actually — or, at least, was three times more popular in 2001, so it's, actually, fairly common.
As I mentioned before, we've got two terms to consider, efficiency and efficacy. Really, efficacy is the better metric to use in evaluating these products because it tells you how much light you're getting per watt. That's, really, what, ultimately, is going to be of importance here if you want to save energy and maintain your light output. LED, because they do tend to send the light directly out of the fixture, they do give you an efficiency advantage. Here, we've noted 12%, roughly, for prismatic versus 23% for parabolic, but this does not necessarily translate into higher efficacy. The LED product itself may not be very efficacious, and so the efficiency advantage may not overcome a lack of output. This provides a quick sense of what you can expect from LED products — or from fluorescent products and, basically, your benchmark for LED then in these applications, looking at about 70 lumens per watt for lens troffers.
So, the CALiPER program is another program where we look at what these products are doing out in the marketplace and, basically, what we're doing here is actually sampling products, taking them off the shelf and testing them. This does include some long-term testing, and also includes testing of other technologies. We provide routine reports, so you see several of these a year, in addition to some periodic testing, periodic reports like, in terms of Benchmark Reports. This is an example of a slide from a forthcoming CALiPER Report, showing the improvement in efficacy over time. It's worth noting that quarter one, over on the right, is just the first quarter, so for the average for the year, we're likely to see continued improvement in the technology. Something you might hear is that CALiPER testing lags the market and, therefore, is not necessarily a good or a meaningful metric for the technology. The reality is that there — while the CALiPER testing has to lag, the market, to some extent, takes time to acquire products, and then test them, then report. When you look at the data, the results, it actually does a excellent job of keeping up with the market and giving you a sense of what you'd be looking at if you were to go out and buy products.
These two slides give you a pretty good sense of where the CALiPER testing falls relative to what's available out there. The little Xs scattered about show Lighting Facts' products available in January of 2011, and all the color-coded dots show — or, at least, the ones that are hollow, show where the LED products that were tested by CALiPER fell. And so, you look at the round — the blue round hollow circles and you see they actually fall right in the pack, including some of the higher performers, for this time frame. The most recent round of testing for CALiPER was, actually, in October of 2010, and is still looking pretty good relative to January, products in Lighting Facts. And then, looking at an update, you see that, while the population on Lighting Facts is expanded, we've got more of our Xs scattered about. The CALiPER testing still is holding pretty well. We've just got a few new outliers in terms of light output, pushing towards 2500 initial lumens.
So, then, looking at, basically, a summary of all the product testing done to date, you can get a sense of how these products are doing. You can see that through the rounds, efficacy has improved, such as, you know, for instance, looking at the louvered products. Over here, you can see that they've gone from being less efficacious than fluorescent to now being, in some cases, more efficacious. And then also, you can see where some of our numbers came from for efficiency differences. You can also see how these stack up relative to the standard [sic] criteria that I'd shown previously.
One thing that's interesting to note is that, as you increase wattage in these troffers, they tend to become less efficient, especially when you look at the fluorescent products, it's especially clear. Increasing wattage or increasing your ballast factor tends to actually drive down your efficiency because you're adding more heat into the system.
So, with that, I will hand it off to Eric, who will now discuss GATEWAY. Eric, you might still be muted.
Eric Richman: Oh, thank you, Jason. What we want to do is provide a little bit of information on a recent GATEWAY project involving LED replacement lamps, but first a slide on GATEWAY to give you a little idea of the program, for those who are not familiar. It is a, basically, a demonstration program aimed at LEDs, of course, also accommodating some benchmarks for comparison. The idea is to highlight appropriate applications of LED technology to provide information on where they actually work. But, of course, not everything is perfect, so the GATEWAY program does present both the positive and negative. The idea, again, is not just to showcase applications but to provide good, useful information on where they work, where they don't, what you might want to be aware of if you want to try one of these.
A typical GATEWAY project does involve an existing real facility demonstration. What we did this time was more of a test setup, but all of the GATEWAY programs follow the same format of looking for energy savings, maintain quality and quantity of light, and reasonable economics or payback, so those applied here as well. The particular project we're talking about is an evaluation of LED T8 replacement lamps. The goal, of course, was to highlight the characteristics of these products as a general group, evaluate what the important criteria are, and present some results, again, the idea being, information that's useful to users when they want to consider this application. The project itself, first, we identified some good-performing, high-performing products that kind of represented the best that are out there. Of course, we're limited in scope, so we only were able to pick a few of them. We then wanted to look at some fluorescent baselines that they might be replacing, measure their energy and light capabilities in a consistent format for direct comparison, and then present those results. Please note that this is not an evaluation of all LED T8 replacement products, it's just a sampling, again, the idea not to look at all T8 replacement lamps but to provide information on their characteristics and how they can be addressed.
So, these are the products that were picked, and this was the initial criteria, white light distribution; in other words, we didn't want a focus beam, we wanted something that was a general application for office lighting. We did want high efficacy, high lumen per watt, as much as we could get, to represent the best of the products out there; high lumen output, in addition to efficacy, again, because there is an issue with LEDs providing enough light; and then, finally, a fourth to kind of add was timely arrival because we were a little bit hampered by availability of getting products. For example, of the three that we chose, there was a fourth one we were considering, to add even a little bit higher light output, but we weren't able to test that. Note that this does provide some basic information on color characteristics and luminous flux.
There's also a CALiPER test number. Again, these were chosen from the CALiPER test data. We do have actual CALiPER test reports on two of the products. Another one, as you'll see in the notes, there is a CALiPER test report, but it's for a previous product. But if you want more information about these, you can go to the CALiPER website and look up those test reports. Baseline products for comparison, we looked at — there's a little bit of garbling there on the screen, I apologize for that — we looked a standard baseline of a 32-watt 735 lamp, which is considered to be a common application in today's retrofit market. We also looked at a T12, standard T12 option because there are, surprisingly, a lot of those around. We then wanted to compare with the three LED lamps. We also compared with a high 835 T8 lamp and a lower-watt or lower lumen output 25-watt T8 lamp, again, just for comparison, so we can provide comparisons for possible retrofits.
We also wanted to look at various luminaire formats because the output might be different, of course, because LEDs tend to be directional, and fluorescent lamps, of course, are omnidirectional. So, we looked at T8 and T12 in these four different configurations: prismatic lens, parabolic louver, a standard basket, and then a non-planar or prismatic high lumen fixture. So, these were all incorporated into the test. We weren't able to test every configuration in every luminaire, but we did our best to capture a sampling.
This is information on the test space. We utilized the mockup space at the Seattle Lighting Design Lab because they have a nice configurable space there. It was a 16 x 16 team with a nine-foot ceiling, standard reflectances, 30-inch work plane height was assumed, and the setup was with four troffers, spaced eight foot on center. Now, of course, in the real world, there are going to be all kinds of different configurations, but this was chosen as pretty typical and standard, the important point being this is a uniform set of test conditions for the different lamp types and kind of luminaire types that were chosen.
In terms of measurements, basically two grid measurements, the horizontal — see kind of the picture there on the left — and then a vertical seen on the right. The horizontal, it was actually measured in the lower left, upper right quadrant as representative of the entire space, and then for vertical, was on the, I believe it was the back wall, above desk type that — the typical places where you might be interested in vertical illuminance. Of course, it can go much beyond that, but we picked a typical situation. So, these are the measurement grids that were used for all of the different considerations. And, of course, we also took power measurements.
Here are the results. You can see up here, starting with the baseline 735 lamp, looked at a high lumen T8, looked at an older obsolete T12 technology, the three LEDs, and then finally, for comparison, a 25-watt lower wattage or lower lumen output T8. And you'll see the information here, in terms of price, and you can see there's clear cost — higher cost with the LEDs. Measured power, of course, the LEDs are, generally, lower, their definitional [sic] and manufacturer-listed information, also the calculated system efficacy, and here we see that the LED products are doing fairly well; they're certainly in the mix. Horizontal work plane illuminance, again, it's generally lower for the LEDs, and that's as expected, that's what we're seeing on the market, and that can potentially be an issue. Vertical illuminance, again, it tends to be lower on the LED side. One thing I will comment, they are lower, but if you look at strict IES guidelines, for example, they still meet those for typical vertical illuminance. If you look at other guidance, they're not quite as good, but they're in the running, so even though they are lower for vertical illuminance, they still do meet IES light level recommendations.
So, let's look at a few graphic representations here of the findings. The first one, again, down here, we have the luminaire wattage, and it's – this is mostly there to be able to categorize products. So, you'll see all the LEDs are in this area, and then we've got the T8 lamps in this area, two different types, and then the obsolete fluorescent products. Average work plane footcandles here in the left. The key here is that you do have higher footcandle levels in this test case for the fluorescent products, lower for the LED products. But, of course, you have lower wattage, of course, for the LEDs. One note here is that these here represent the basket-type LED, and those are not performing as well for LEDs, and likely, of course, due to the fact that the LEDs are directional in nature. This is something that could, of course, be explored further.
Another graphical representation here, looking at — let me go down here — looking at workplace uniformity. This is horizontal average to minimum illuminance with the two-lamp luminaires. With the fluorescent products, it's generally better, down here, in this region, the LEDs tend to be a little bit higher on their ratio, so not quite as uniform, although for the basket LEDs, it is, actually, slightly more uniform. We don't have comparable tests for the fluorescent in the basket, but there is quite a range here. Another look at vertical illuminance versus wattage for the different products, and again, as we saw in the table, the vertical illuminance for the LED products is not quite as good. According to IES minimum, it still met that, but it's not nearly as uniform as it is for the fluorescent products. And again, the LEDs are directional, and that can cause these kinds of issues.
Also, looking at the uniformity in vertical illuminance, and this isn't always critical, but, certainly, in office environments, we have a lot of people that you're dealing with, this can be important. And it's, generally, mixed, except for the parabolic fixture type where the LED is not doing as well, but for the other types, they are in the mix with the fluorescent products. Another thing we looked at, because of the issue of light output, if the LED lamp could be increased so it delivered the same footcandle levels, then they certainly seem viable. What you're looking at in this chart is the standard T8 performance line right here, and we then mathematically equated all of the other products, the other fluorescent products, the other LED products, to similar light output and then compared what their percentage energy savings would be. And as you can see, it's mixed. For the most part, they're doing pretty well, the LEDs. Their efficacy is pretty good. There would be good energy savings, but not so for the basket LED, these products her, there again, for directionality reasons, aren't doing quite as well. And, of course, the fluorescent one down here is, as expected, not performing well at all. But again, this isn't the real data. This is how we adjusted the data, just to give you an idea of what would happen if they provided the same footcandle level, and this is certainly possible.
We also wanted to do some economic comparison, of course, and we looked at life cycle cost, and we just wanted to show you what the inputs were, typical melded electricity rate. We used RS Means derived, relamping cost and labor cost, typical discount rates, hours of operation for the year and a 50,000-hour study period, and lamps price as were purchased. Now, the realization is that your economics may be different; your labor rates may be completely different, depending on your situation. Your study period may be different, depending on how you want to do the analysis. This is just one comparative analysis just to provide an idea of how these might compare. So, realize that, you know, your results may vary. The general takeaways here are that, simply because of the high initial cost, none of the LEDs will pay back over the 50,000-hour study period. Now, that doesn't mean that they won't in the future, even in the near future. It's just a matter of extremely high cost. The 25-watt T8 replacement was the best option in terms of lowest life cycle cost but, again, you have to look at the economics. The key here is that we believe it's important to look at these economics before you make a decision.
Another thing to consider is lumen depreciation. Again, this is a chart made up from data based on what the industry knows about these technologies. It's not real measured data for these products. But what we wanted to show is if you have these fluorescent products up here, for example, and then there's the 25-watter down here, they are going to have to be relamped during the 50,000-hour test period, so they get relamped here, here, and here, but that refreshes the light output, so at the end, over here on the right you'll see what the light output might be. The LEDs, of course, are expected to last 50,000 hours without relamping, and based on a standard lumen depreciation of 70% — again, this is just a standard industry number, this isn't necessarily what these products will perform like — you're going to end up with a different light level. And this looks like a 10 to 15 lumen difference or — I'm sorry, 10 to 15 footcandle difference at 50,000 hours. Now, this may not be critical for your application; it may be. It really depends. It's just one thing to keep in mind. And realize again, these numbers may not be totally accurate for the price you're looking at, but it's one more thing to be considered; that's the important part.
So, some application recommendations, if your space is currently overlighted, LEDs, because a lot of them are producing less light, they might be an effective replacement, but also consider a lower light output T8, because of the cost differences, may be the best option. Again, pro forma life cycle costs or, at least, a payback analysis, if you can get LED lamps at lower prices, they might be the economic choice. The key, again, is taking a look at that data and make a good choice. If your space is not overlighted, you're probably going to want to maintain the same light levels. You're going to want to make sure that LED lamp, if you want to go with that, provides enough light. And it looks like, likely, right now around a 2,000 lumen LED product would probably provide the same light output for most situations, but not every case is the same. You're going to have to look at that on a case-by-case basis.
A couple of other application notes, not all LED lamps and fixtures are wired the same or meant to operate the same. There's an issue with shunted sockets, for instance start versus rapid start, so you're going to have to be careful when you do an LED retrofit, you've got to follow the manufacturers recommendations for rewiring. Not all of these lamps are interchangeable; there are some with integral ballasts, external ballasts, you have to be careful with what you're putting in. The technology is new. Later, when it's settled down, they may all be interchangeable, but for now you need to pay attention to that.
In terms of rewiring issues, of course, you do have to typically disconnect the fluorescent ballast and rewire the sockets. You may want to consider labeling that fixture so someone doesn't inadvertently replace your LED with a fluorescent lamp and then it likely won't work because there's no ballast attached anymore. There is a lot of work being done in yield certification for replacement T8 lamps. Check with your manufacturer first; they may have already developed a kit that's UL certified or you can also check with UL and see what their latest is on that. And a final note, T8 LED replacement lamps aren't going to work real well on all fluorescent fixtures. We've seen that there are issues with the basket fixtures. So, the key here is, test something before you decide to replace everything, just so you know how it's going to operate for your situation.
In summary, there is good news. The efficacy of many of the T8 lamps is good and getting even better. The T8 LED lamp doesn't need to necessarily deliver equivalent illuminance, to deliver equivalent footcandles, because of the directionality and the type of fixture, it may be slightly different. Economic viability may arrive in the near future; it just depends on when the LED lamp costs will drop sufficiently. However, a few caveats. The light output [sic] is still lower than fluorescent for most products, so you need to make sure you're going to get the amount of light you need. Costs are still high in relation to fluorescent T8 LED, of course. The key, again, try a few before you commit to a whole project or, at least, look at some data where it's been installed before. Consider all your options. If you really want LEDs, you can get LEDs, but if you want to look at cost effectiveness and other considerations, you're going to want to do a little bit of analysis.
And with that, I'm going to turn it back over to Jason, who's going to talk about some specifications.
Jason Tuenge: Okay, thanks, Eric. So, there are two Fact Sheets currently available on the Department of Energy website related to this product category. One of them, basically, provides an overview of all the CALiPER testing done to date or, at least, as of May 2011, May of this year, and then another that provides suggested specifications for this product category, a little shorter, more to the point in terms of offering guidance. So, in the following slides, you know, we're going to be looking at some of these specifications that we have offered to date. First up is light output. And at this time, we're recommending a minimum of 2700 lumens for initial output, and this is based on the numbers, the figures listed below. Looking at a 0.87 ballast factor for fluorescent, note that when you install these LED products, for a number of these products, you actually remove the ballast from the circuit before installing your LED product. So, this is generally done just so you have a more predictable output from your LEDs; if they're being driven by a ballast, it can be a little more questionable what actually happens relative to LM-79 testing that might have been done on the LED product before.
So, looking at ballast factor, looking at lumen maintenance, 94% is assumed for fluorescent versus 70% for LED, since 70% is usually used as the point to define rated lifetime for LEDs, that's what we're looking at for lumen maintenance. And then, as I mentioned earlier, the average efficiency advantage for these LED products, once you install them in a troffer, is something on the order of 17%, and so we're effectively crediting LEDs and requiring less light from them due to the efficiency advantage in those applications. Also, surveyed a number of fluorescent lamps to get our roughly 2800 lumens initial output for those products. We also provide a recommended minimum rated useful life of 35,000 hours. As mentioned before, many of these products are going to be rated more like 50,000 hours, and these figures really just put you in the ballpark of fluorescent which, as you can see here, it really varies based on the type of ballast the fluorescent lamp is operated on and its cycling for how often it's switched on and off in its application.
Lumen maintenance is specified as being 70%. That's what we're going to expect from these things by the time they're replaced, and then as a checkpoint, something — we can't really test out to 50,000 hours, it's more realistic to test 6,000 hours or thereabouts, and we've got a requirement here for where the light output should be at 6,000 hours of operation. We require 94% of initial, in other words, only 6% light loss over the course of 6,000 hours of operation. This should be supported by test data for the product, tested as itself, not just the little LEDs that go into the replacement lamp, but the whole replacement lamp itself being tested this way. Also, recommend a minimum of three years for the product warranty to be proportionate to the rated lifetime. We've got criteria for color characteristics so that we're comparable to fluorescent, require a CRI of a minimum of 80. And then for color temperature, we don't require particular color temperatures or disallow particular temperatures that you'd commonly find, but rather, say, that if you're going to claim to be 2700 K, so that you look — like incandescent, then whatever we measure for CCT should be darn close to 2700 K, and the way to do this is to look to ANSI standards, and there's tolerances that are established there for CCT, as well as another metric called Duv.
The next slide shows, basically, the difference between these two metrics where, basically, you've got CCT essentially running — I'm going to draw some arrows here — you've got CCT running, basically, in these directions, and then you've got Duv running perpendicular. And, basically, what this does then is, once you establish tolerance for CCT, so you know that you're supposed to look orangish or bluish or somewhere in between, the Duv parameter which works in the, you know, a perpendicular direction, prevents your lamp from shifting from, you know, toward green or pink, which can actually happen if you don't contain Duv. So, it's also a good idea to review that characteristic, as well, and that is often provided in these reports, these test reports.
Electrical safety is a very tricky subject and, frankly, we're not going to give real specific guidance during this webcast on this subject. Really, what you need to do is work with your manufacturer, probably with your local inspector as well, to make sure that whatever is going to be required locally, in terms of safety, is being adhered to with your retrofit. And so, this is, really, going to vary from application to application; it depends what you're doing, what exactly you're replacing and how, and what kind of an environment and where you are in the country. So, it's not an easy thing to summarize on one page, and we could devote an entire webcast to this subject, really, but at this point, basically recommend checking with UL and/or CSA, and manufacturers and inspectors to make sure that you're going to be covered in terms of safety.
Intensity distribution, as Eric mentioned, is another parameter that should be looked at, but it's another one that's not easily summarized and, for this reason, we have basically skipped this parameter in the suggested specifications. I just pointed out that you should take a look at what's happening to your distribution. For instance, in a parabolic louvered troffer, you might have your beam shape changed dramatically and can have some ramifications in terms of work plane uniformity or shadows on walls, and things of that sort. You also want to pay attention to that. And, unfortunately, what you're going to need for your application is going to vary from application to application, so there's no one size fits all to what these things need to do.
We were targeting having an update for the specification series Fact Sheet ready before the webcast, and it is currently undergoing review, but the deadline for this has slipped a little bit — we're looking at probably in the coming weeks, probably, actually, now going to follow IES TM21, which should be released, from what we're hearing, pretty soon here. So, what we might do is just roll that into the updated specification Fact Sheet and give a more robust method for evaluating lifetime claims, which right now are a little fuzzy. We might also look at revising the criterion for minimum output to base it on 28-watt fluorescent lamps instead of the 32-watt, since this is something that's seen as being acceptable in terms of initial output and the number of applications. So, we're looking at some of these things, but they have not been pinned down just yet. And it's worth noting that, you know, if we do go to, say, to the 28-watt lamp as the new benchmark, then you should expect still less in terms of input power to your LED products in order to still save energy. They can't just get by with less output; they need to also have less input.
You want to make sure to do a good job with your economic analysis. It's not just a matter of looking at initial cost or one variable or maybe two. You really want to look at the big picture and see what's this going to mean for you on an operational basis when you replace these products, you know, how much is it going to cost to install these things, what's the upfront cost, access into the ceiling, disposing of products, so you pick up all the advantages and disadvantages of going with this type of a retrofit.
As mentioned earlier, these suggested specifications are really focused on the major application of installation in recessed troffers and ceilings, mostly in offices, because they're so common. There are other applications, and for these other applications, it might make sense to have different criteria. For instance, you may not need as much light from your lamp if the light is typically lost upward. In a parking garage, some uplight is good, and this is an application where you do find strip lights, fluorescent strip lights, like the ones shown here at top right, installed in parking garages. In some cases, it might be okay to not have uplight on the ceiling and, indeed, many luminaires that are installed in these applications have no uplight, whether these are HID or LED or, you know, or fluorescent. But it often is a good idea to have some uplight on your ceiling, and so it's one thing to bear in mind. But in these applications, if you don't need the uplight, you may be able to get by with less output and, therefore, lower efficacy from your LEDs. So, the criterion really depends on your application, but what we've done here is target, really, the low-hanging fruit, which is troffers.
As mentioned earlier, LED luminaires offer an integrated solution whereas lamps, when you buy it on the shelf or when the manufacturer makes it, they don't necessarily know what kind of a fixture it's going to go into. Is it going to go into a lens prismatic troffer or an indirect fixture or a parabolic louver troffer or a strip light? You know, you really don't know. With luminaires, you know what you're getting from the product because it's already done, it's a complete unit. And so, you can go in — you can take a look at some of these products at our Next Generation Luminaires Design Competition site. It shows you some of the things that are going on on the luminaire side that can serve as competition to some of these products. Here, we're showing a 2 x 2 troffer as opposed to a 2 x 4.
We've also got some specifications for luminaires that are available through the Commercial Building Energy Alliances websites, and right now, we've got a parking structure spec that's already completed and a 2 x 2 troffer spec that's in development. You can find some more information on our website — well, I guess, first off, if you want to see this — reference this website at a later date, the slides in this webcast; you can go to this first link. If you want to register for Continuing Education Units, you can go to the second link and do so there, and then information on the various projects or programs that were discussed in this presentation are provided in these next three slides.
And with that, we will wrap up and begin addressing some of the many questions that have come in. We'll do our best to get to as many as we can but, obviously, there's only so much we can do in the time we've got. So, I guess, at this point, if Eric can sign back on, we can begin working through them.
Eric Richman: Okay, and I believe the first bunch of them, a lot of them start with issues that you talked about, so if you want to —
Jason Tuenge: Yes. So, first up, we've got a question in, asking about CRI as a Color Rendering Index versus another metric called CQS. And this is something that we've addressed in some previous communications and some of our Fact Sheets, it's called the Color Quality Scale, and was developed by the National Institute of Standards and Technology, or NIST, recently to serve as an alternative to CRI and potentially replace CRI at some point. The idea here is it's another metric for gauging how well a source renders its environment and brings out the color in its environment, and they're, really, from two different schools of thought. At this point, CQS has not been widely adopted yet, and so we're still helping to educate on the matter, but we don't have time to really get into it at length here. It does evaluate products differently, and perhaps better, depending on what they're being used for. But at this point, CRI remains the de facto standard, and so at this point, that's what we're using most everywhere on specifications and including for Lighting Facts labeling purposes.
Next question — actually, if you want to take this next one, Eric, on DLC?
Eric Richman: Yes, basically, like if you go to the DLC website, the way it will suggest for LEDs is to first look at ENERGY STAR® and see if it's qualified under ENERGY STAR. If it's not one of those categories, then it — there is, apparently — and I don't know all the details, but there is some qualification under DLC for those products not covered by ENERGY STAR. It also then goes on to suggest that you make sure that you have a product where you can see the Lighting Facts label and assess what the product quality is. That's about all the detail that I have, but if you go to the website and look for the LED category in DLC, it'll provide that information for you.
Jason Tuenge: Yes, so the DesignLights Consortium basically helps fill in some of these gaps, and we may see, at some point, a spec open up for this particular category.
So, if a product does not have a label to compare apples to apples, is that a red flag? That's the next question that came in.
And the reality is that you can't test every variation of a product, and testing is required for the Lighting Facts label. And so, you know, if you choose something that's 75 watts instead of 80 from a given manufacturer within a given product line, it might be that they tested 75 but not 80. It, really, comes down at that point to it, you know, being a judgment call, you know, can you trust the manufacturer, trust their data? Some of it comes, you know, just from what lab did they use, is it a reliable testing lab? But the reality is you can't really test every different variation of a product. It's nice if you only have two variations of a product, then you should probably get those both Lighting Facts listed, so that's not very difficult. But if you have thousands of variations that you offer, you really can't test all of them, so it is nice to have something from the manufacturer that's close to what you're looking for that's been tested and is Lighting Facts labeled. That does, then, give you a level of trust or assurance that testing has been done properly and if they extrapolate from that, then you can probably trust it, to some extent. So, the short answer is, no, it does not have to have the label, but it is very helpful if they've tested the product you're looking at or something very close to it.
How often is Lighting Facts' website updated?
From what I understand, it's, essentially, updated continually, that basically, as products come in, it takes time for the Lighting Facts team, which is DNR International, contractor to DOE, that as it comes in, they will review the products and eventually post them. It's a pretty quick turnaround, more like a matter of a week than months, so pretty quick, and, yes, it's done as they can get to it. So, it's pretty frequent.
A very generic question, lumens, directional lumens, with some question marks.
I think, basically, what this is getting at is doesn't it matter where the light's going, and the answer is absolutely, yes. But, as I mentioned, distribution is kind of kind to capture in terms of what you really need, so you do want to pay attention. You know, for instance, if you're taking a incandescent bulb and sticking it in the recessed can in the ceiling, you're going to lose a lot of light up there, the same thing for CFLs if they don't have a reflector built on them. So, yes, you want to pay attention to that. But at this point, Lighting Facts is just looking at total output, not specifically where the light is going.
Another question: Is slide number eight for lamps or fixtures?
Bear with me one second. Here, we've got slide eight. This is for lamps, and so — and then, that also holds for the next slide where we've got criteria from Department of Energy that go into effect — these are requirements that will go into effect in 2012, which apply to fluorescent, not LED. This is all for lamps, not for luminaires. You can see the numbers go down when you go to install them in luminaires. So, here we've got 89 lumens per watt required for lamps versus — I'll show you another table — once they're installed in luminaires, you're probably looking at more like 70 lumens a watt or 61 because you have some light loss in the fixture.
Next question: Are you only measuring light output in terms of lumens or are you measuring these items in terms of footcandles?
We're only looking at lumens because, otherwise, you have to assume where these things are going to get installed, you know, like what's the mounting height, what's the spacing between fixtures, what are the room surface reflectances, are you trying to light the floor or the work plane? Probably the work plane. Basically, as Eric pointed out, when you look at the output from the fixture, it definitely correlates to your light levels on the work plane, so ideally, you look at your levels on the work plane; you do a full analysis. But just, you know, for the purposes of Lighting Facts, which you can't know where these products are going, it's very useful to look just at the light output from the fixture.
I don't know if Eric has anything to add to that?
Eric Richman: No, I think that covers it.
Jason Tuenge: Yes. Was the T12 lamps symbol on slide nine actually more efficient than the T5 or T8?
T12 — and so, basically — okay, here, you've got the red — I assume this is slide eight, not slide nine, as referenced in the question. It might look like the T12 are looking the best because they fall furthest to the top right, but again, the line of efficacy, you know, running diagonally, that's a constant — that's a line of constant efficacy, and so you can see they actually aren't doing that well. They're below 89 lumens a watt whereas most of our T8 lamps are above 89 lumens per watt, so they're actually not doing that well. You just got to make sure that you're reading the graph properly, that basically, what we're seeing here is that these things produce more light and consume more wattage than T8s. It might seem at first that they're outliers, you know, the highest performers when, in fact, they're actually not.
All right, so if we look at the next batch of questions here. Who's CRI standard are we using?
It's CIE's standard, and I can't remember the particular number. Eric might be able to remember. But, yes, basically, CRI is defined by CIE, and, traditionally, it just looks at eight different materials being illuminated, all of them being pastels in color. There are other materials that are used in the same standard, but that's not part of calculating CRI. So, yes, just looking at pastel colors as defined by CIE.
So, a comment on slide 12, basically saying that you would expect more correlation between CRI and/or CCT when looking at efficacy.
And this is true, when you look — and I'll go back to slide 12, sorry — that's really why I showed the slide, is that you would expect to see a stronger correlation between CCT in efficacy and CRI in efficacy, and you do see some here, but it's not a very strong correlation. Basically, what you're looking at here is the variety and quality of products that are out there. So, if you look at any given manufacturer, you'll probably see pretty good correlation between, say, CCT and efficacy, basically, as they get bluer, they get more efficacious, so that you may not like the appearance of them, but they're more efficient. But then — so, that's within any given product line offered by one manufacturer.
When you look at all the manufacturers that have submitted, what you see is that correlation disappears because you've got all kinds of stuff out there. You've got good products, bad products. You know, somebody's low CCT products, a warm appearance product, may be more efficacious than somebody else's high CCT, so a bluish-looking product. The opposite of what you'd expect, but it's explained by the difference in quality. So, it's just good to be aware that they're not necessarily correlated when you go pick out these products.
Let's see, what does LER stand for?
You correct me if I'm wrong, Eric, but it's Luminaire Efficiency Rating, or Efficacy Rating?
Eric Richman: Yes.
Jason Tuenge: Luminaire Efficacy Rating, and so, it's basically, it's just luminaire efficacy is, really, all it is. It's defined in a NEMA standard, and there are some exceptions, but that's, basically, all it is. For most products, it's just simply luminaire efficacy. In some cases, they only look within certain zonal, like a zonal range, you know, so a cone of light, but for the most part, it's just luminaire efficacy.
Okay, and if you want to take this one, Eric —
Eric Richman: Sure.
Jason Tuenge: LM-79 versus 80?
Eric Richman: There's a question here about, Could you please take a minute to describe LM-79 versus LM-80 testing?
These are the two preeminent testing methods right now for LEDs, the two biggies, so to speak. LM-79 is a product capacity test. It's a one-time test of a luminaire, of an LED luminaire that outputs the light output of that luminaire, as well as its color characteristics, and it also measures its electrical draw. So, from that, you can identify color; you can also identify the efficacy of that LED luminaire. And LM-79 is the standard test that everybody uses when you rate an LED luminaire product. They need to be tested as a complete unit because of the heat issues, so that's a little bit different than a typical relative test with a fluorescent product. But that's the standard test for LED products. LM-80 is a separate test. All it does is relate to the industry how to do lumen depreciation testing for an LED chip or module, not a luminaire but a chip or module. With that LM-80 testing data, that gives you an idea of how that LED product is going to perform over time. What LM-80 doesn't do is tell you how long it's going to last or when it's going to degrade to a certain point because it only requires a minimum of 6,000 hours of testing.
Jason mentioned the TN21. TM21 - you can call it a companion to LM-80. It's the second back half, which then will take LM-80 data and try to estimate how long that LED chip or module lasts or how it will degrade over time. So, those are the differences.
I think the next one's up to you, Jason, if you'd like —
Jason Tuenge: Yes, it's a quick one on ballast factor, just clarifying — asking for clarification.
You know, ballast factor for fluorescent lamps is, basically, a matter of how hard you're driving the lamp. And for fluorescent lamps, typically, as you change your ballast factor, the efficacy holds relatively constant. So, I think this is, really, referencing the chart I'd shown earlier, which — give me one second and I'll pull it up — this one, and so, basically, showing that even though efficacy holds pretty much constant for fluorescent lamps, if you change the ballast factor, how hard you're driving the lamps, that's for the lamps. When you're installing a fixture, it is then producing more heat in the fixture, which fluorescent doesn't like heat, at least not T8 lamps; T5 might like it, but not T8. And so, that's something to keep in mind. It's something that's been borne out by testing by CALiPER, demonstrating that this does, indeed, happen. So, you do get an impact on output, whereas I believe the wattage was not affected and, therefore — yes, so the wattage is not affected and so, therefore — actually, no, sorry — basically, the output was affected by the temperature, therefore, your efficiency dropped down.
Let's see here. Next question, okay, and we're taking these ones now real time. Power factor in THD? Actually, you want to take that one, Eric?
Eric Richman: Sure. There, of course, are issues with power factor in THD for any building system. Are they worse for LED T8 replacements? I don't know, I don't have all that catalogued in my head, but if you go to the CALiPER data, some of that, I believe, has been tested for some CALiPER products so you can see what that is. But another item here is that it really depends on the system and whether that's going to be an issue or not. Sometimes a system already has a power factor or harmonic distribution signature, so to speak, and the lighting system may not affect that. So, yes, it is important and, yes, it's good to get good quality products, but you really need to look at what your requirements are. And this is just another thing you need to look at when you look at the characteristics of a lamp that's being provided: ask the manufacturer for that data, and then determine if that's going to be a problem. But, overall, are they worse than others? I really can't say that. I don't think I've seen that, myself.
Jason Tuenge: The next question, so basically asking, if the baseline values of fluorescent efficacy in lamp life use compare to LEDs being updated with the current advance by fluorescent manufacturers? There are fluorescent lamps available now with lamp life well above 50,000 hours.
So, as I mentioned, the baseline values for lamp life, for fluorescent, they really vary based on what ballast you're using to run the lamps and how they're being cycled, and so these ratings, you can see, are quite high sometimes, but those are for optimal situations where maybe the lamp is cycled every 12 hours; in other words, it's not on occupancy sensor, and where it's on a program start ballast when, in fact, instant start ballast is most common. So, it really depends on what it's being operated on and how it's being operated on that particular ballast type. It also depends on which lamp you purchase. And so, some of these standard lamps, they're just going to be rated for 24,000 hours on standard cycling and on instant start ballast, whereas you can find some that are rated above 60,000 hours. So it really depends, it varies on the lamp type and the ballast, and how it's being operated. But this gives you, basically, a sense of what you should expect from these LED products if they're going to be replacing fluorescent and not requiring increased maintenance.
So, next question: Do graphed lumen outputs of fluorescent and LEDs represent L70 and end-of-life lumens for their respective technologies or are they initial?
And we're pretty careful to show that these are always — pretty much always going to be initial lumens that are shown in graphs and in Lighting Facts on the label, those are all initial values. The reality is that L70 is not a one size fits all metric; it's not, you know, requiring that your product maintain 70% of initial output. It may not be okay everywhere, you know. You may want to have L90, you know, so only 10% reduction in output at end of life for your application, or if you have a decorative product, you may want — it may be okay to have it drop to 50%. So, it really depends on the application. L70's pretty common, basically, the de facto standard. And so, we do include that when we establish things like minimum initial output as a parameter criterion, but typically, when you look at the charts and graphs and quick snapshots, you're looking at initial lumens because it's not really clear what you're going to need in terms of lumen maintenance.
All right, so next question: LED downlight, wide or narrow affects picture design to achieve proper distribution? — with a question.
So, I think, basically, what this is getting at, is just asking whether the distribution coming out of an LED product will affect the distribution coming out of the luminaire and, therefore, distribution in the space, and the answer is, yes, you know, so you just want to make sure that you're pairing your lamp well with your luminaire and your application. And, really, the best thing to do when you're evaluating distribution for these products is to do a mockup. I mean, if you're looking at installing, say, 1,000 or 100 of these products in your facility, it's probably a good idea to test out one room and make sure that it actually looks good and that it's not going to cause glare problems or leave the walls completely dark if you want light on them. So, yes, ultimately, it's best to do some kind of a mockup before making a big investment, also making sure that things continue to operate for more than a day.
I guess, the next question, on to Eric.
Eric Richman: Yes, the question's about LED lamps running on different voltage formats and all requires some fixture field modification. How do you deal with the UL issues?
The issue being, that when you rewire or change a luminaire, that essentially negates its UL listing. What you UL is working on is retrofit kits and other types of retrofitting situations that will provide UL recertification, essentially. What you really need to do is to talk to your manufacturer first and see if they've got a UL certified kit, what they're providing. If they don't, you can go to UL and ask them what their requirements might be. It may change, depending on the situation, but you're right, you do have to get some kind of UL recertification. You need to check with them on all the details.
Jason Tuenge: And we're getting pretty tight on time now. I think we're down to about four minutes, maybe five, so just a couple more questions now. The next one, Why is the selection of CCT so high? This is unacceptable for the NAFTA market. Better comparison for Asian market.
And it is true — to respond to this, it is true that color preference will vary from region to region, but, at the same time, it's also worth pointing out that, quote, unquote, "daylight spectrum products are pretty popular here," and that would put you in the range of 5,000K to 6500K, so pretty blue stuff. And there are, clearly, people that like those products because they're offered by the big manufacturers. You can find them, you know, on the store shelves. So, again, what we've done with the CCT criteria in the suggested specifications is not say that you should use something of a certain CCT. but saying that, you know, if you're to use anything in that range of 2700K to 6500K, going from basically warm white to cool or even cold white, daylight, then you should be within certain tolerances, so what we provide there are tolerances. It gives you a way of evaluating a product's test data versus claims. So, if somebody claims 2700K, but is actually measured at 4300, that's not a good thing. It's going to be a lot cooler looking than you probably want for your application. So, that's really what that guidance is there for. And again, that was actually developed by NEMA and the NFLG, and we're just referencing their criteria. It's been very useful.
Eric Richman: I can …
Jason Tuenge: You want to touch on that one?
Eric Richman: The question, really, is about a bottom line for DOE. It says: Previously, DOE recommended against this technology, LED replacement lamps. Do the new findings change that?
Well, yes, in general, what the findings — DOE going through CALiPER program, et cetera, the findings were that the LED replacement lamps really weren't that good early on. The efficacy wasn't that high, and the light output was extremely low, and that made them not a good choice for a direct retrofit replacement. The technology has gotten a lot better, and while not all LED products are as efficacious as they could be or we would like to see them, a lot of them certainly don't have the high light output, they're getting a lot better. So, the findings can't say, yes, they're ready to go, start using them. What we have shown, we hope, is that here are the things you really need to look at. This is what you probably need to have to make this a successful project, so please take a look at these duty evaluation, get the right data, and pick a product that will work. So, yes, before they weren't so good; they're getting better. We hope they're going to continue to get better.
I think there's a couple of other questions in the next —
Theresa Shoemaker: Eric? Excuse me. I think we have time for one more question.
Eric Richman: Okay. A question here about — we have a couple of questions, actually about testing products that had 50,000-hour LED life, why didn't we test — one question was: Do LED products really have this long of a life? Another question was how come we didn't test products with 100,000 or 150,000 life? It turns out two of the three products we tested specifically said they have a 50,000-hour life. Now, we all know that's just a kind of a generic number the industry is working on. There's very limited data that supports that, although manufacturers do test some to come up with those numbers. We just picked a few products to test. We picked 50,000 hours, as a standard number. As I mentioned, that will greatly affect your economics, and so you will have to take that under consideration carefully. Who knows how they're actually going to last? Unless the manufacturer can show you 50,000 hour or 100,000 hours' worth of data, you're not ultimately going to know. Of course, no one really has that much data. That's why we're kind of all waiting on TM21 to come up with that.
So, I think a big point from this is that a lot of these items are variable. Lumen degradation is variable, lifetime is variable. You have to take those under consideration when you're doing comparisons.
Jason Tuenge: And at this point, because evaluation of lifetime is so preliminary, you know, we're still working on a standard, which we should have pretty soon, from the sounds of it.
Eric Richman: Yes.
Jason Tuenge: Because of that, you really can't look at rated life as a quality metric and say that, oh, well, this one's rated 100,000 and this one's rated 50; therefore, the 100,000's better. Yes, don't do that. At this point, it's hard to say. It might be that one is better, but that's not something that's cut and dry and reliable.
And I think that's about it. I think we're out of time.
Eric Richman: Yes.
Theresa Shoemaker: Perfect timing. Well, thank you, Eric and Jason. And thank you for everybody participating in today's webcast that was brought to you by the U.S. Department of Energy. You may all now disconnect. Have a nice day.