Good afternoon for those of you on the East Coast, and good morning for those of you on the west coast. And if you're dialing in from abroad, well, thank-you, and it's probably the middle of the night for you.

My name is Michael Myer. I'm with the Pacific Northwest National Laboratory. Welcome to today's webinar – "Understanding Technology Tradeoffs In High-Efficacy Luminaires," brought to you by the U.S. Department of Energy's Solid State Lighting program.

Our presenter today is one Ms. Naomi Miller. Naomi Miller straddles the line between design and engineering at the Pacific Northwest National Laboratory. She is located in our Portland, Oregon, office. She's called a senior scientist, but she uses her decades of experience in architectural lighting design in upstate New York as well as San Francisco to educate clients, customers, and practitioners, and help nudge the solid-state lighting industry toward better, more practical solutions.

Working on case studies and exploratory studies of solid-state lighting products, she is the tedious voice of experience. By bridging the gap between technology and application, she can promote the wise usabilities, working with industry to overcome the hurdles, and celebrate the opportunities. And now, Naomi Miller.

Thanks, Michael. Let's start with this top performers study. The reason for this top performers study occurred a couple of years ago when we started looking at the LED Lighting Facts Database. The database was started in 2009, and has listed over 70,000 LED luminaires since then as a way to help specifiers find products that are efficient and also have good physical characteristics, and color characteristics, and so forth.

In 2017, we started seeing products that were listed close to 200 lumens per watt. And if you look at the listing on the right, you can see that at the top of the listing, there are two of them coming in at 229 lumens per watt, and 200, almost 210, lumens per watt.

So we scratched our heads, and we said, really? Do these products really perform at these levels? And what are these products? They've got these cryptic 25 digit catalog numbers. Let's order some. Let's find out what these products have in common. And let's see how they really perform and appear compared to what the manufacturer's claims are.

So we identified five of these top efficacy performers from the database. And they range in efficacy from 140 – or 136, actually – up to 209 lumens per watt. And they were all low-bay kinds of applications. Plus we added two luminaires from familiar US manufacturers that had similar 200 lumen per watt claims.

We targeted a range of lumen output from 16,000 to 22,000 lumens. The intent was to match a 4-lamp T5HO industrial lighting product which might be used in these kinds of applications.

The target CCT was 5000k. Why? Because all the highest efficacy products on the listing were 5000k. As soon as you went down to warmer color temperatures, the lumens per watt dropped to below 200 lumens per watt. So we stayed up at 5000k.

We ordered three samples of every luminaire. Two samples to be tested and mounted, with a third sample that we used just in case there was a malfunction of the first two luminaires. We sent them first to Richland, where we have a NVLAP-certified laboratory for LM-79 testing. And then those luminaires were shipped to Portland, Oregon to a mock up laboratory for mounting in a movable ceiling.

We then contacted 23 lighting and energy efficiency experts, and we brought them in as observers to assess the characteristics and the luminaires. We then took measurements inside the mockup laboratory space to collect information on things like flicker performance, light levels, glare, et cetera.

So what did these products look like? Here's our collection of seven luminaires. The luminaires are mostly 2-by-4 format luminaires. So they're fairly long, with two exceptions. The third one 17-S3, is a round fixture that looks more like a traditional round metal halide, or high pressure sodium industrial luminaire. And the seventh one, 17-S7, is a 2-by-2 format luminaire, but with linear arrays of LEDs.

The others were primarily 2-by-4. The second fixture was something like 3 feet, or 30 inches, wide by 4 feet long. So you can see there that they all have quite different appearances.

So first, the testing. What happened? Here are the two samples of each of the luminaires, and how the laboratory measurements came out. So you can see the CRI values on the left. The tested power usage. The power factor lumen output. CCT. And then on the right, you see the laboratory-tested lumens per watt.

So what's circled there is the fourth luminaire – 17-S4 actually did come in at 200 lumens per watt, or very close to it. The others were also very high in efficacy. So we were kind of heartened by that. But let's look at the results in more detail.

So if we're comparing the results of the luminaire test versus the manufacturer's claim, the first thing I have to do is give you a caveat that says, when you look at the LED Lighting Facts Database, the luminaires that are listed aren't always based on the specific luminaire you're looking to order. Sometimes you may have a family grouping, and the efficacy and the other characteristics can be based on the best performing product in that family. That's one caveat.

The second is that because of that, we would go to the manufacturer's website, and we would look at the cut sheet, or the information, on the specific luminaire we are ordering which falls between that 16 and 22,000 lumen range.

The luminaires that were on the LED Lighting Facts Site were not necessarily in that range. So that's another reason why we are going to compare against what the manufacturer claimed on their website rather than what was claimed on the LED Lighting Facts site. So let's start.

Color performance was reported with CCT and CRI. Manufacturers often claimed either CRIs greater than 80, or a CRI greater than 70, rather than a specific value. Of those, all tests show that they performed in that same range that was claimed.

Of those with targeted values, like saying, oh, it has a CRI of 82, those test values differed by a maximum of 1 point. So very close. CCT varied by no more than 4.5% of the targeted, or the claimed, CCT value. So we gave the check mark to the first category.

Second, lumen output values. Well, lumen outputs, most lumen performance values, were higher than claimed by a maximum of 10%, but not all. One was as much as 8% lower, and a couple of them were lower than claimed. But since it was pretty close, within 10%, we're going to give it another check. The power draw varied by as much as 7% high – excuse me – 7% lower than claimed to 1% higher than claimed. So that was also pretty close.

The efficacy, if you take the change in lumen output value and the variation in power draw, it produced up to a 12% variation in claimed efficacy. Sometimes it was about 5% lower, and went up to about 12% higher than claimed. So there was more variation in the efficacy. But it's still in the right ballpark.

Power factors were all above 90%. And we were pleased to see that there really are luminaires performing at 200 lumens per watt. Ta-da!

Now, when we shipped the fixtures to Portland, Oregon, and had them installed in this movable ceiling, we contacted the local Illuminating Engineering Society section, and we recruited a number of specifiers from the IES. We also recruited a number of utility program energy-efficiency experts locally.

And we recruited a couple of facility managers from the Association of Professional Energy Managers. These are facility managers that are very familiar with lighting issues in their buildings. All were lighting knowledgeable, and we had one hour observation sessions scheduled over three days. Generally, we had no more than three or four people evaluating at a time.

And I would like to point out two things. Number one, the person leaning against the wall was one of our labmates, Tracy Beeson. She's one of the collaborators on this project. And she is also our resident flamingo.

And the other caveat is that, yeah, they were unpaid. Except we bribed them with lemon bars and chocolate chip cookies. So if you want to discount the study because of that, it's my fault.

Next, the observers were given a series of rules. The groups were given three minutes to move around the space and to imagine themselves in the shoes of the industrial employee and the building owner. And they were asked not to stare the luminaires, but were free to glance at them briefly if need be.

We asked them not to talk amongst each other in order to reduce bias – just in case the first person said, oh, man, is that awful! With the other people saying, oh, I have to rate this as awful, too. We didn't want that to happen.

We also randomized the presentation order of the luminaire pairs. The observers saw one pair of identical luminaires presented at a time. They were asked to comment on visual comfort, distribution on the work plane, shadows, color, appearance, and an overall dollar value. Here is a reflected ceiling plane showing the pairs of luminaires.

You can see that luminaires are spaced 14 feet on center, with one exception. We had one luminaire that was too wide to actually fit in on a 14 foot center so we squeezed it in on a 13 foot 6 inch center.

Here's the questionnaire that they completed. The written comments for each of the unidentified luminaires, except by a letter value – the written comments asked for them to comment on specific issues, or anything else that they wanted to mention. And then they were asked, given this type of luminaire, and on the assumption that luminaires cost $200 on average, what would you pay for this one?

So the intent was not to have participants estimate the actual cost of a luminaire, but to give an overall value to this particular industrial luminaire if the type of luminaire cost would be $200 on average. This gave the observers a chance to say to themselves, if a typical LED luminaire is $200, would I be willing to pay more or less for this luminaire based on its performance in this space? So we gave them options of $100 up to $300 in $50 increments. And then we averaged all the numbers together for the final results.

Observer results. First, let's talk about appearance. Very few observers actually commented on the appearance of the luminaires themselves. Because let's face it, they are not exactly designed for aesthetic evaluation. But the few people that did respond commented that they liked the two fixtures that were smaller in form factor better.

And one included a round fixture that was – oh, I suppose it was 15 inches in diameter. And the other was a linear LED product, but it was a 2-by-2. So these are the two fixtures that received the top ratings for preferred appearance – that is, the luminaire itself. Not necessarily in its lighted form.

So here's the result of the questionnaire responses. And on the right hand side, you're going to see a graphical illustration of their responses. Green means a positive evaluation. Yellow is neutral, red is negative, and gray means there was no response from a particular observer on that issue.

So you can see in the visual comfort column, there's a lot of variation. Visual comfort – you notice this first luminaire, 17-S1, has a big chunk of green. So people, by and large, liked it. And those that didn't like it outright thought it was neutral in acceptability for visual comfort.

But let's now look at, say, 17-S3, and 17-S4, and 17-S5, and you can see that the negative comments about visual comfort dominated the stack of responses. So this is a very quick way to evaluate where you're getting big differences in responses among the different luminaires.

You can also look across on this particular plot. Because the first fixture 17-S1 – you can see that there's a lot of green here. And then if we look at 17-S3, you can see there's a lot of red there. Same thing with 17-S4. There's a lot of red there. So you can very quickly see which of these luminaires the observers liked best.

But we found that the visual comfort question dominated the responses, and corresponded to both positive ratings of the luminaire, and higher overall dollar value for the luminaire. And those luminaires with the poorest glare ratings, or visual comfort ratings, also got lowest overall value. So that was very consistent.

Let's talk about some of the in situ measurements. We took illuminance meters, and we measured horizontal illuminance underneath the fixtures, directly underneath each of the individual pairs of fixtures between the luminaires, and then at quarter points between them. And we averaged those values. And we also looked at the maximum to minimum values. So this column here is the average illuminance. And this column here is the average max to min value.

This upper section is based on light levels measured at the floor. This lower section here is measured at a 37 inch work plane height. Why 37 inches, you ask? Because we happen to have a rolling cart that was 37 inches tall, and very handy.

And it was the perfect height for assembling industrial tasks. So we thought, well, let's measure at 37 inches. Industrial tasks are often at standing desks of 36 or 37 inches high. So that represented a typical work plane height for industrial tasks.

Here you have the average illuminances underneath the luminaires, and the uniformity ratios. So let's talk about the uniformity ratios. Well, let's go back. First of all, these uniformity ratios for the floor level were very consistent. 1.1 up to 1.3, which is a very good uniformity.

The illuminance uniformity at a higher height were higher, as you would expect, because there's a shorter distance between the luminaire and the work plane height. And the range was from 1.3 minimum, to a maximum of 1.7, although most of the luminaires came in between 1.6 and 1.7.

The ones that came in at the higher maximum to minimum ratios were luminaire number 1, 17-S1, and luminaire 17-S2, the second one. And the reason these had the highest max to min ratios for work plane illuminance was that they both had linear baffles to limit glare and control distribution.

It may be hard to see here, but here are these linear baffles between each of the runs of LEDs. Now these are runs of LEDs. And this particular luminaire has a diffuser over those linear runs of LEDs. And then there's a linear baffle in-between.

This was the one luminaire where you could not see the bare LEDs. The one on the right also has these linear baffles, which cut down on the width of the distribution. But you could still see the individual LEDs if you were standing underneath the luminaire and looking up.

But these two luminaires, number 1 and number 2, produced the highest max to minimum ratios. But the other luminaires weren't that much different. 1.7 versus 1.6 maximum to minimum. Now let's talk about flicker measurements. This is Naomi Johnson Miller the 14th, and she has to talk about flicker at all opportunities.

Several observers commented on flicker from the luminaires. So we went back and we measured flicker from the seven luminaires. And found that the observers were right. Luminaire number 6, or S-6, produced 45% flicker at 120 Hertz, which exceeds the IEEE standard P1789 for the low-risk level.

Now the other luminaires all met that low risk level. They produced from 0.6% up to 33% flicker. 33% flicker could be a problem, but not at 1500 Hertz. So there's a sliding scale there.

And actually, what I want to point out to you is the flicker index, which I am starting to believe is a better metric combined with flicker frequency then percent flicker. The fixture that flickered exhibited a flicker index value – flicker index is basically the percentage of light output over a single waveform, a single period of waveform – of 13. This is 0.139, which is approximately 0.14.

That means that 14% of the light output over that period of time is emitted above the average compared to the total light output. So that's a value that I'm starting to look at in combination with the frequency that may give you a good indicator of whether a product is actually flickering to a noticeable degree or not.

You can compare that to a magnetically ballasted fluorescent fixture which has a flicker index of 0.09. And 0.09, that's a point at which a lot of people can start to see flicker and be bothered by it at 120 Hertz. So 0.139 is higher than that. So that does indicate that a number of people would be able to identify that that fixture is flickering at 120 Hertz.

Now what was also interesting was a lot of these luminaires were operating at very high frequencies. I was surprised. So as soon as you get the frequency up to between 1,100 and 2000 Hertz, flicker is much less noticeable – unless it's more extreme. And in all of our cases, flicker for those products was very low as measured by the flicker index.

Let's talk about glare. Now there is no way to measure glare directly from a luminaire at this point in time. But we decided to try measuring the luminance of the LEDs, or the diffusing tube that covered one of the luminaire’s LEDs.

Unfortunately, with our luminance meter, we were unable to capture just the light from the tiny LED itself. In all cases, we were capturing more area than the LED itself, including that reflector area behind the LED.

So because there was always some background area included in the captured angle, that additional surrounding area essentially diluted the measured luminance. So the actual luminances of the LEDs would be higher than shown here.

But let's take a close up view here of some of these luminaires. The top rated luminaire, which is number 1, had a luminance of 40,000 candelas per square meter. Now since we were measuring a diffuser and that filled the full capture angle of the luminance meter, that one is an accurate number.

So that is a very good indication of how bright that luminaire was. Compare that to a T5HO fluorescent lamp, which is about 25,000 to 30,000 candelas per square meter. So you can see this was higher than a T5HO lamp.

The luminances of the visible LED packages were very high. And the highest ones, which were measured at above 400,000 candelas per square meter, corresponded to the highest rankings of low visual comfort, at least at the mounting height of 11 foot 3'', where our fixtures were installed.

Let's look a little more closely. So here's a close up of that table. And we've added a close up of the LEDs themselves. And you can see for this fixture that was the highest rated fixture has the diffuser over the LEDs, so you can't see the LEDs themselves.

All the others had exposed LEDs. And some of them, as I said, were very high. Above 400,000 candelas per square meter.

Interestingly enough, this one had considerably lower LED luminances. So they may have been a mid power LED, for example. And they compensated by the lower light output by having a lot more of them. So in this luminaire, we counted 2,862 LEDs producing the light from this luminaire.

But because there were more of them, that may have been more comfortable than the ones that were brighter, where you had fewer LEDs. So that may be something that we should consider in the future when we're evaluating glare.

So the fixtures that were the lowest rated in terms of comfort, and ultimately overall value, were the third fixture and the fourth fixture that were both very high in luminance.

So let's look at the observer ranking. The preferred luminaire out of the seven was number 1. The second preferred was number 2. Now I don't know if you remember, but this is the luminaire that actually had linear baffles.

Even though you could see the bare LEDs, there were linear baffles that helped minimize the view of the LEDs above a certain viewing angle. Above – I don't remember, but it was something like 60 degrees. That was the second rated.

You could also look at the estimated value that was assigned. These two luminaires got the highest estimated value as well. The luminaire that got the worst rating was the one that had – the round one – it had a very high LED luminance. And it also got the lowest estimated value. And the one that rated number six out of seven, also 400,000 candelas per square meter, it also got a low dollar value assigned to it.

So let's look at the installation and see where these are. The best rated fixture was the one with the diffuser. That's this luminaire here, and this luminaire here. And you can see from this view that the baffles are actually cutting off the view of the bright diffuser. You can see it from this angle, but not from this angle.

Same thing is true with the second rated fixture. This is a viewing angle at which you can no longer see the LEDs because the baffles are blocking the view. But you can see it from this angle. This is that second rated luminaire, which is still very high in brightness when you're looking at it below a 60 degree viewing angle per meter.

The least preferred luminaires were these two. So this was the sixth rank, and this one was the seventh rank. So let's go on to the next, and let's talk about the limitations of this CALiPER study.

The space where we evaluated these luminaires had a limitation in ceiling height because of some mechanical issues of 11 foot 3''. We had planned to mount them at 15 feet, but we were only able to get them 11 foot 3'' above the floor.

Now 11 foot 3'' is still a tall height, and it can still be a fairly normal mounting height in some industrial applications. But we also are aware that many of these industrial fixtures get mounted at 20 feet, or even 25 feet, above the floor.

The horizontal illuminances that were measured between the pairs of luminaires ranged from 811 lux up to over 1,000 lux. That's not outrageous. That's very normal, especially when you're talking about high detail tasks, or small industrial tasks. But the lower mounting height may have exaggerated the glare from the luminaire for the observers.

The volunteer observers that we brought in were lighting knowledgeable. So we weren't bringing in people off the streets. They had experience in industrial lighting. And although we took care to reduce bias and order effects through the protocol that we use, this was not a rigorously designed human factors study. However, we always think it's important to get feedback from industry experts when we're looking at luminaires, because they have important insight and important experience.

So let's summarize the results here. And what you're seeing on the right hand side is one of our observers who is enjoying the glare from one of the least preferred luminaires that were set up in the high bay space. The luminances of the exposed LED packages were very high and corresponded to rankings of low visual comfort.

The highest efficacy product that was close to 200 lumens per watt had luminances of 400,000 candelas per square meter, and received the rating of sixth out of seven. So it was very efficacious, but not the preferred product.

The lowest efficacy product that we mounted was the fixture that had the diffusers. And that came in at only 136 lumens per watt. However, because it had the diffuser, you could not see the bare LEDs. And because the luminance was so much lower than all the others, it received the highest ranking of the group.

The results on glare support the development and the adoption of a glare metric that incorporates luminaire-luminance distribution. Now, what does that mean? If you're going to look at two luminaires, and one has a completely diffuse plastic or glass across the aperture, and it's even, and you can take a luminance value, and it's the same luminance value all the way across, that is the very even luminance distribution that the unified glare rating metric that we often use for glare prediction uses.

But it's unrealistic, especially for LEDs. Six out of seven of these LEDs did not have diffusion. You could see the bare LEDs themselves. So the distribution of high luminance and low luminance varied widely across the aperture.

Now the good news is that the CIE has a joint task force that is currently focused on modifying the UGR, Unified Glare Rating metric – to accommodate such nonuniform luminaire apertures, and this study supports that work.

More results. We looked at whether we would have sustained some efficacy reductions if we had ordered warmer CCT options – warmer than 5000k. 5000k is obviously a very blue-white color. And we did find, from the manufacturer's literature on websites, that it would have resulted in a 13% to 17% drop in lumens per watt for the luminaires where we had information. Not all of the websites had this information. So that would be going from, say, 5,000k down to a 4,000k or 3000k option.

We also recognized and found luminaires that had more optical options to control the glare and to control the distribution of light on the work plane – both of those were also likely to reduce efficacy further.

The LED Lighting Facts Database, and similar databases, such as the DLC listings, can be an excellent source of information. But it is so important to understand that the specifier needs to investigate the performance of the specific product that they're ordering. Don't assume that the product characteristics listed in the listings are going to correspond to the specific product you're ordering. You must find out what the actual performance of the color of luminaire, the distribution of luminaire, whether the luminaire has any optical glass, or louvers, or baffles installed, and what the lumen range of your actual luminaire is. Because that will affect the actual performance.

You also need to consider other lighting quality issues. For industrial lighting, it still makes sense to consider visual comfort, flicker issues, light distribution – obviously – color quality. But be prepared for the inevitable tradeoffs in efficacy. Some of these tradeoffs are minor. Some of them are more significant. It's incumbent on you to do the homework.

So with that, I'm going to finish talking. And that gives Michael Myer a chance to ask me some pesky questions. Michael.

Thank you, Naomi. We had a handful of different questions. I'll try to correlate them to slide numbers when possible. And if not, I'll try to help you as well.

The first one – which was related to slide number 20, if that helps – was, "most of these fixtures can be ordered with or without a lens. It would be interesting to see how people would rate the importance of efficiency versus quality. Has there been any studies looking at this factor?"

So similar to what you did look at with your subjective study of the appearance. So they're asking about lens and efficiency.

It's an excellent question. Some of these luminaires had bare LEDs. And a couple of them had either an acrylic or a glass surface on the aperture. And that acrylic or the glass would help reduce dust. In none of these cases did it really affect the optics, but it would certainly reduce the transmittance of the LEDs.

So, yes. As soon as you put a piece of glass or plastic in front of those fixtures, it will certainly help make that fixture more dust resistant, and more resistant to bugs, for example. More resistant to any other issues that might be happening in the industrial space. But it will have some small reduction. Usually no more than 5% to 12%, but it will contribute to a reduction in efficacy.

Thank you. Another question somewhat related to glare – and I think this image is probably a good reference point – "should crossed baffles have been considered as well, and related, was the viewing position fixed, where they were walking with the linear baffles?"

Crossed baffles are an excellent option. It does reduce the glare from two directions, rather than just one direction. No question about it. But again, as soon as you add crossed baffles, you're also reducing the efficacy of the product somewhat.

Some of the products that we ordered here could have been ordered with crossed baffles. But because we were shooting for the high efficacy products, we didn't end up ordering those. The best we got was ones with linear baffles.

Thank you. Related to your comment about ordering. And I don't know if this was a base to selection, but a person wanted to know, "were all the fixtures chosen made in the US, or was that a criteria?"

That was not a criteria. Good question. And I can't – some of them were American manufacturers. In fact, I'm trying to think. Was it most of them? They are listed in the report if you're curious. So you can go to the link on the last page of this presentation and look at the report itself.

But whether they were made in the United States or not was not a consideration. But there were many American manufacturers.

Thank you. Always a next question. "Are there plans to continue this line of research?" And related, "was there any investigation into thermal performance and lumen depreciation at the higher ambient temperatures?"

That would be a wonderful study. We have not done it at this point in time. And we will consider that in further studies. Thanks for the suggestion.

OK. And hopefully there will be further studies as suggested in that question. Related to the appearance and glare of the smaller fixtures, was it that they were off when they were rated as better appearing. Do you think it was the actual style those fixtures were? Or do you think at some point, a large fixture is just not desirable in appearance?

Frankly, I think it was style. I think it may also have been size. The finish on those fixtures was a little less rough than some of the others.

OK. Finish always matters, so thank you. Related to procedure of protocol, were the observers given the performance results of the sphere testing to factor into their cost estimates?

No, they were not. They were given results of the sphere testing after they had made their observations. So we just discussed that later. And yes, there may indeed be inevitable tradeoffs that you make because you want the higher efficacy product. So that is data that you can get from the report. You can see both the ratings of the observers and the efficacy results, and make your own decision

Thank you. And yes, thank you for referencing to the larger more in-depth report. That's always a good reference for more information. Someone here was asking if you could elaborate related to the flicker index. You made a comment that you're finding flicker index combined with flicker frequency. Do you compare both columns, or do you multiply the values together? Could you elaborate on how you use those together? This person's asking.

Well, that's a good question. And I have to give you a tentative answer. A number of us have looked at different kinds of flicker metrics and plotted different kinds of luminaires, their flicker performance, on different scales. The flicker index on an x-axis and the flicker frequency – the fundamental frequency – on the y-axis gives you a distribution of flicker performance.

And you can draw something close to a straight line through the points, and kind of differentiate between those that are likely to be noticeably flickering and those that aren't. Now this is something that we're exploring with the CIE right now. So the jury is still out. Stay tuned.

Well, this person also asked, related to more studies as well. They're interested in PoE, which is Power over Ethernet, or digital low voltage lighting, specifically a similar study related to flicker and especially power distribution related to those characteristics of beyond this study. Are there other future studies related to newer lighting technologies in the works?

Good question. I know that there are studies here looking at PoE at PNNL. I don't think that it combined the issue of flicker, or some of the other issues that we studied here. But it would make a great study

Continuing on some of the earlier slides you had, this person wanted to know – you made a comment that lumens were no more or less than 10% above or below the value. They were hoping, how many were below? How many were on the lower end of the threshold? Was that a very common thing, or less common?

Actually, less common. I believe only two out of the seven luminaires under represented their lumen output. Or excuse me, overrepresented their lumen output. it was more typical to actually see higher lumen output than the manufacturer stated.

Getting into the nitty gritty, this person wants to know if actual fixtures were taken apart, or cracked open the hood, as they say. And were the driver manufacturers compared? Were there any trends in terms of flicker performance getting down to the actual nuts and bolts of the components?

That's an excellent question. That was not within the scope of this work. But it would have made a great addition to the study. Maybe we'll do that next time.

And the questions are coming in, and thank you for everyone who is submitting. This person wanted to know if, going back to that singular round fixture that you showed briefly, if shape affected efficacy in any way.

No, I don't think so. It was also high in efficacy. These were basically LEDs on a white piece of metal. And the LEDs were spaced half an inch apart, three quarters of an inch apart, a quarter of an inch apart. It really didn't have to do with whether it was in a round circle or in a linear array.

OK. And we really just have one or two more questions, it appears. Going back to another person about the power. A 7% lower power draw – they were just trying to clarify – that's a good thing, right? There's nothing really bad if they're under-reporting their power. Or their power is under-reported in relation to what – OK.

Yeah. What may have happened is that we were basing our analysis on the current photometric reports of these luminaires that were available, and from the manufacturer. And often, the manufacturer will change the chipset that they're using, and they'll be better chips.

So they may be putting out more light. Or they may be using less power. That's generally a good thing, as long as they're controlling glare at the same time.

And the final question. This person was wondering if there was a relationship in the perceived overall dollar value related to the efficacy performance.

No. No, because the highest efficacy was – I'm trying to remember. It was either the S4A or S5A. And they did not receive a higher rated value estimate. So no, there was no relationship between efficacy and perceived value.

Here are the observer rankings. So obviously, the top one was the luminaire that had the diffusion over the LEDs. All the others had exposed LEDs. The least preferred was the circular fixture that had very high luminance.

Thank you. And sorry, I just got one more. People are very interested in how you would design future studies. This one is related to, if you were conducting the same study today, would you choose luminaire based on the DesignLights Consortium dataset, or the Lighting Facts dataset?

Well, that's a good question. I think ideally I'd probably want to research the two sets, and see how they differed. And perhaps do a combination of both. The nice thing about the DesignLights Consortium dataset is that it may include a little more information that could be useful in evaluating luminaires that specifiers could use in evaluating luminaires.

Perfect. And at this time, there are no other remaining questions.

Well, we're pretty close to an endpoint then.

Yes, I would say we have three minutes to go. So I will defer to you on that.

OK. I'm going to zoom to my last slide again so that we can make sure that you'll have the link to the final report. And we'll send you – or if you access these online slides when they're available, you'll have the link to the final report.

Although the easy way to get to the final report is just key in "DoE top efficacy performers," and you'll get there. Thank you all for attending.