WEBVTT Kind: captions Language: en 00:00:00.070 --> 00:00:04.090 In my last video, we talked about the high pressure sodium lamp and its ubiquitous use 00:00:04.090 --> 00:00:05.620 in outdoor lighting. 00:00:05.620 --> 00:00:09.509 Already this type of lighting is starting to be phased out with various technologies, 00:00:09.509 --> 00:00:12.150 with new LED technology among the most common. 00:00:12.150 --> 00:00:16.289 Now, given the robust nature of the sodium lamp and its proven track record in providing 00:00:16.289 --> 00:00:20.939 an efficient light source reliably, does it make sense that we switch to newer light sources? 00:00:20.939 --> 00:00:24.170 Well, as usual in life, there are pros and cons. 00:00:24.170 --> 00:00:28.359 So let’s start with a basic question: why do we use outdoor lighting? 00:00:28.359 --> 00:00:32.520 For street lighting, specifically, the basic answer is to improve safety, particularly 00:00:32.520 --> 00:00:34.200 pedestrian safety. 00:00:34.200 --> 00:00:38.270 The odds of a crash of any kind are greater at night, because it’s harder to see. 00:00:38.270 --> 00:00:42.510 With the aid of artificial roadway lighting, a driver can see much farther than their car’s 00:00:42.510 --> 00:00:45.110 headlights shine, especially to the sides. 00:00:45.110 --> 00:00:48.170 Now of course there are disadvantages to large amounts of street lighting, which we’ll 00:00:48.170 --> 00:00:52.330 get to, but assuming safety is the goal, are sodium lamps good? 00:00:52.330 --> 00:00:53.630 Well, no. 00:00:53.630 --> 00:00:54.630 Not really. 00:00:54.630 --> 00:00:58.559 Remember how I said that the Sodium D-line is close to our eyes’ peak sensitivity, 00:00:58.559 --> 00:01:01.300 but only under photopic daylight conditions? 00:01:01.300 --> 00:01:05.810 Well it turns out that our eyes see quite differently at night than during the day. 00:01:05.810 --> 00:01:10.930 Under nighttime, scotopic lighting conditions, our eyes actually see bluish light better. 00:01:10.930 --> 00:01:14.780 And that makes sense--after all, moonlight and starlight are pretty bluish, so under 00:01:14.780 --> 00:01:19.159 these dimly lit conditions, having a greater sensitivity to blue light would mean we can 00:01:19.159 --> 00:01:20.640 see better. 00:01:20.640 --> 00:01:25.340 Under scotopic lighting conditions, only the rod cells in our eyes are activated. 00:01:25.340 --> 00:01:28.890 Rod cells cannot distinguish color, but they are much more sensitive to light than the 00:01:28.890 --> 00:01:30.130 cone cells. 00:01:30.130 --> 00:01:36.230 The peak sensitivity of the rod cells is around 498 nanometers, which is a green-blue color. 00:01:36.230 --> 00:01:40.259 Now of course under street lighting our cone cells are still active--we can after all see 00:01:40.259 --> 00:01:43.259 colors and are not exclusively using the rod cells. 00:01:43.259 --> 00:01:48.729 This dim-but-not-quite-dark lighting scenario is often called mesopic vision, a mix of the two. 00:01:48.729 --> 00:01:53.090 Still, stimulation to the rod cells will be far more visible and is more important. 00:01:53.090 --> 00:01:57.960 So then, how well does the light from sodium vapor lamps line up with our scotopic light sensitivity? 00:01:57.960 --> 00:01:58.960 Not well. 00:01:59.000 --> 00:02:03.340 This is the CIE 1951 scotopic luminosity function graph. 00:02:03.350 --> 00:02:08.190 The X axis is the wavelength of light in nanometers, and the Y axis is the eye’s relative sensitivity 00:02:08.190 --> 00:02:11.620 to these wavelengths under scotopic nighttime conditions. 00:02:11.620 --> 00:02:15.310 As you can see, peak sensitivity is around the 500 nanometer mark. 00:02:15.310 --> 00:02:18.020 And where’s the wavelength produced by a sodium vapor discharge? 00:02:18.020 --> 00:02:21.080 It’s about here, 589 nanometers. 00:02:21.080 --> 00:02:24.440 The rod cells are barely activated by a sodium discharge. 00:02:24.440 --> 00:02:28.500 While the discharge may be extremely efficient at producing visible light, at night time 00:02:28.500 --> 00:02:31.820 this light is fundamentally misaligned with our eye’s sensitivity. 00:02:31.820 --> 00:02:36.220 See, if you look at the response curves of scotopic and photopic conditions together, 00:02:36.220 --> 00:02:40.380 you can see that the sodium discharge lines up great with our photopic vision. 00:02:40.380 --> 00:02:44.910 But when our eyes are adjusted to nighttime lighting conditions, it’s actually pretty bad. 00:02:44.910 --> 00:02:48.370 This is why focusing on the sodium D-line emission’s seemingly perfect alignment with 00:02:48.370 --> 00:02:50.510 our vision is somewhat of a farce. 00:02:50.510 --> 00:02:55.270 While it’s true during the day, it’s literally quite far from the truth at night. 00:02:55.270 --> 00:02:59.640 Aside from the simple spectral misalignment of the sodium lamp, research shows that people 00:02:59.640 --> 00:03:04.060 can indeed see better under light sources with a bluer spectral content. 00:03:04.060 --> 00:03:07.990 In Peter Morante’s research for the Lighting Research Center at the Rensselaer Polytechnic 00:03:07.990 --> 00:03:13.410 Institute (link below), survey respondents strongly preferred the light from a 6500k 00:03:13.410 --> 00:03:17.650 correlated color temperature light source over that of high pressure sodium, with metrics 00:03:17.650 --> 00:03:22.540 of visibility, brightness, safety and security, color rendering, and overall preference all 00:03:22.540 --> 00:03:28.080 favoring the newer light source, which also used only 55% of the energy of the high pressure 00:03:28.080 --> 00:03:30.260 sodium lamps it replaced. 00:03:30.260 --> 00:03:35.360 The study, which by the way is really comprehensive and worth taking a look at, was done in 2008, 00:03:35.360 --> 00:03:39.180 which was a tad before LED technology became economically viable. 00:03:39.180 --> 00:03:43.110 The study compared induction lighting, which is sort-of like fluorescent lighting (and 00:03:43.110 --> 00:03:47.310 worth a video on its own one day because it’s pretty neat if not so practical any longer) 00:03:47.310 --> 00:03:52.080 as well as metal halide lighting to the sodium lamps for the purposes of creating a recommendation 00:03:52.080 --> 00:03:54.300 to a local utility company. 00:03:54.300 --> 00:03:58.490 The conclusion was that metal halide didn’t make sense due to higher maintenance costs, 00:03:58.490 --> 00:04:00.760 but induction would make tons of sense. 00:04:00.760 --> 00:04:03.000 But I also liked this final recommendation: 00:04:21.140 --> 00:04:25.480 Well, it’s been about 10 years time since that study, and LEDs are economically viable. 00:04:25.490 --> 00:04:29.230 In fact, they’ve become incredibly economically viable. 00:04:29.230 --> 00:04:32.930 But before we move on to them, let’s go over the main issue once more. 00:04:32.930 --> 00:04:37.140 Because of the spectral misalignment of the sodium vapor lamp with our scotopic (and mesopic) 00:04:37.140 --> 00:04:42.330 light sensitivity, it takes more light output from a sodium lamp to produce the same visible 00:04:42.330 --> 00:04:44.690 light level from a bluer light source. 00:04:44.690 --> 00:04:49.160 In fact, Dr. Alan Lewis of the New England College of Optometry measured individual’s 00:04:49.160 --> 00:04:52.980 response time to a hazard approaching from the sides, and found that in well-lit areas 00:04:52.980 --> 00:04:57.910 such as a major motorway, a high pressure sodium light system needs to produce 3.9 times 00:04:57.910 --> 00:05:02.420 as much light as a metal halide source to achieve the same response time. 00:05:02.420 --> 00:05:07.510 The effect is even greater in dimly lit areas, where he found that 7.8 times as much light 00:05:07.510 --> 00:05:12.270 from high pressure sodium was required to match response time under cooler, metal halide 00:05:12.270 --> 00:05:13.270 light sources. 00:05:13.270 --> 00:05:16.890 So it seems as though the high pressure sodium light is less efficient that it appears on 00:05:16.890 --> 00:05:21.050 paper, and that roadway safety is greatly increased when light sources are used that 00:05:21.050 --> 00:05:24.400 are tuned to our scotopic and mesopic light sensitivity. 00:05:24.400 --> 00:05:26.940 It’s looking pretty bad for high pressure sodium. 00:05:26.940 --> 00:05:32.060 But there’s one thing that HPS technology doesn’t do that bluer light sources might. 00:05:32.060 --> 00:05:35.520 That, my friends, is circadian rhythm disruption. 00:05:35.520 --> 00:05:39.340 Research has suggested that the color of light we are exposed to has a great impact on what 00:05:39.340 --> 00:05:42.360 time our biological clocks think it is. 00:05:42.360 --> 00:05:46.880 The shorter wavelength of daylight sun and blue sky may help keep us awake by suppressing 00:05:46.880 --> 00:05:51.610 melatonin production, and the long wavelength light of the sunset may signal our bodies 00:05:51.610 --> 00:05:55.550 to sleep by allowing melatonin to seep into our blood streams. 00:05:55.550 --> 00:05:59.401 To help us understand the impact different light sources can have on circadian rhythm 00:05:59.401 --> 00:06:03.380 disruption, light sources are characterized by their melanopic content. 00:06:03.380 --> 00:06:07.230 There’s a great source from the US Department of Energy linked down below which goes over 00:06:07.230 --> 00:06:11.950 this in much greater detail, but as a general overview, with high pressure sodium technology 00:06:11.950 --> 00:06:17.160 normalized at 1 for both scotopic light content and melanopic light content, a metal halide 00:06:17.160 --> 00:06:23.200 lamp with a color temperature of about 4,000K will have about 2.5 to 2.8 times as much scotopic 00:06:23.200 --> 00:06:30.050 light content, but also produces 3.16 to 3.75 times as much melanopic light content. 00:06:30.050 --> 00:06:34.120 If you look at the various color temperatures of LED light sources, you can see that as 00:06:34.120 --> 00:06:39.960 the color temperature increases, both scotopic light content AND melanopic light content increase. 00:06:39.960 --> 00:06:44.690 However, melanopic light content increases at a greater rate than scotopic light. 00:06:44.690 --> 00:06:46.760 What this means is, there’s a trade-off. 00:06:46.760 --> 00:06:48.380 And it’s complicated. 00:06:48.380 --> 00:06:52.020 The higher the color temperature, the more scotopic light it produces, which means you 00:06:52.020 --> 00:06:57.540 could use less light (and thus less energy) to get the same perceived brightness and safety levels. 00:06:57.540 --> 00:07:02.440 But, because melanopic light content increases at a greater rate, although you may need less 00:07:02.450 --> 00:07:07.800 light of a higher color temperature, it will disrupt circadian rhythm to a greater extent. 00:07:07.800 --> 00:07:11.520 You can see that high pressure sodium has among the lowest melanopic light content of 00:07:11.520 --> 00:07:17.240 any light source, with only amber LEDs and low pressure sodium producing less melanopic light. 00:07:17.240 --> 00:07:19.480 So then, we’re presented with a choice. 00:07:19.480 --> 00:07:23.730 We can clearly use less energy and expect a safer nighttime driving experience with 00:07:23.730 --> 00:07:28.500 light sources of higher color temperatures, but this may cause unwanted side-effects that 00:07:28.500 --> 00:07:30.950 sodium lighting largely doesn’t. 00:07:30.950 --> 00:07:33.740 And this is completely ignoring aesthetic preferences. 00:07:33.740 --> 00:07:38.230 I myself generally detest cooler lighting, as I find it harsh and unpleasant. 00:07:38.230 --> 00:07:40.810 But I can acknowledge its safety advantages. 00:07:40.810 --> 00:07:45.370 In my area on Interstate 88, many of the roadway lights have been changed from high pressure 00:07:45.370 --> 00:07:47.050 sodium to LED. 00:07:47.050 --> 00:07:52.180 The change is dramatic, with visibility greatly enhanced once under the cooler light. 00:07:52.180 --> 00:07:56.550 As much as I don’t like the color, I can tell it’s a lot safer. 00:07:56.550 --> 00:08:00.169 Visibility in my periphery is tremendously better, and I’m certain these new lights 00:08:00.169 --> 00:08:03.020 are using less energy than those they replaced. 00:08:03.020 --> 00:08:06.980 Before I move into the conclusion of this video, let’s discuss the issue of light pollution. 00:08:06.980 --> 00:08:10.700 Light pollution is exactly what it sounds like--excess light in our environment that 00:08:10.710 --> 00:08:15.500 is irritating, unnecessary, poorly distributed, or in general unwanted. 00:08:15.500 --> 00:08:19.730 Outdoor lighting is by far the most prolific source of light pollution, and it has gotten 00:08:19.730 --> 00:08:24.240 so bad that people living anywhere near a city can barely see the night sky. 00:08:24.240 --> 00:08:29.440 There’s even an anecdote about a widespread blackout in Los Angeles causing many panicked 00:08:29.440 --> 00:08:34.039 911 calls from lifelong residents who had never seen the Milky Way before and were a 00:08:34.039 --> 00:08:35.339 little scared of it. 00:08:35.339 --> 00:08:39.750 Light pollution is a complicated problem, but the LED may actually help to solve it. 00:08:39.750 --> 00:08:43.270 Now to be clear, the solutions I’m about to offer aren’t exclusive to LEDs, but the 00:08:43.270 --> 00:08:48.020 way light is emitted from an LED chip makes controlling it relatively easy. 00:08:48.020 --> 00:08:53.300 So first, let’s discuss one of the biggest causes of light pollution; lights that point up. 00:08:53.300 --> 00:08:56.580 This is less obvious than it seems, but it’s incredibly important. 00:08:56.590 --> 00:09:01.830 Right outside my apartment are drop-lens cobra luminaries containing high pressure sodium lights. 00:09:01.830 --> 00:09:06.450 Because the lens protrudes downward from the fixture, a lot of light escapes to the sides 00:09:06.450 --> 00:09:07.640 and indeed upwards. 00:09:07.640 --> 00:09:12.270 I’m on the 4th floor of my building, and my eye-level is above these street lights, 00:09:12.270 --> 00:09:15.150 but I can still see the source of the light. 00:09:15.150 --> 00:09:17.450 This is not ideal for a number of reasons. 00:09:17.450 --> 00:09:22.890 First, a lot of light is being wasted by lighting up things that are not the road. 00:09:22.890 --> 00:09:24.270 That’s kinda dumb. 00:09:24.270 --> 00:09:27.350 But secondly, a lot of this light is going up into the sky. 00:09:27.350 --> 00:09:31.380 Granted, this style of fixture isn’t the worst offender, but a better design would 00:09:31.380 --> 00:09:36.460 be a flat lens that does not allow light to escape above the horizontal. 00:09:36.460 --> 00:09:40.460 This may still throw light farther to the sides than necessary, but none of it will 00:09:40.460 --> 00:09:42.560 end up lighting the sky. 00:09:42.560 --> 00:09:46.730 The worst offenders for this kind of light pollution are lights that illuminate buildings 00:09:46.730 --> 00:09:51.760 by shining upwards, wall-pack lights without shielding, and these decorative fixtures. 00:09:51.760 --> 00:09:54.940 I’ll admit they’re pretty, but they’re really wasteful. 00:09:54.940 --> 00:09:57.560 Recently I was on an airplane flying into Chicago at night. 00:09:57.560 --> 00:10:01.730 I took some video as we landed, and you can see the difference between a well-managed 00:10:01.730 --> 00:10:03.870 light and a poorly managed light. 00:10:03.870 --> 00:10:05.300 This roadway is lit well. 00:10:05.300 --> 00:10:10.570 I cannot see the actual light sources, I can only see the reflected light from the road. 00:10:10.570 --> 00:10:12.009 That’s what we want. 00:10:12.009 --> 00:10:16.649 As we got closer to the airport, these neighborhoods had tons of lights that were visible from 00:10:16.649 --> 00:10:17.870 above. 00:10:17.870 --> 00:10:21.970 Much of the light produced by these lamps is shining into the sky and being wasted. 00:10:21.970 --> 00:10:27.490 I should not be able to see the actual light source from an airplane, yet I can. 00:10:27.490 --> 00:10:29.890 This is contributing to skyglow. 00:10:29.890 --> 00:10:33.779 Skyglow is what makes the night sky hard to see when you’re close to a city. 00:10:33.779 --> 00:10:37.640 This is probably the most widespread light pollution problem, and while it’s not caused 00:10:37.640 --> 00:10:42.260 exclusively by poorly designed fixtures, they are a major component. 00:10:42.260 --> 00:10:46.180 But once again, the solutions to skyglow are complicated. 00:10:46.180 --> 00:10:51.089 Largely because light sources that cause the least skyglow are high and low pressure sodium. 00:10:51.089 --> 00:10:56.029 In fact, low pressure sodium is used widely around large astronomical observatories because 00:10:56.029 --> 00:11:01.040 their nearly monochromatic light output can easily be filtered out, eliminating any skyglow 00:11:01.040 --> 00:11:02.040 they create. 00:11:02.040 --> 00:11:06.330 I saw a large number of people saying that high pressure sodium can also be filtered 00:11:06.330 --> 00:11:10.570 out, but I don’t think that’s true due to the pressure broadening and their spikier 00:11:10.570 --> 00:11:11.620 output. 00:11:11.620 --> 00:11:15.110 Someone correct me if I’m wrong but I could only find references to low pressure sodium 00:11:15.110 --> 00:11:16.810 being used around observatories. 00:11:16.810 --> 00:11:21.520 Anyway, where it gets tricky is that an LED light source has about three times as much 00:11:21.520 --> 00:11:25.020 sky glow impact than a high pressure sodium light. 00:11:25.020 --> 00:11:29.899 But also, you need less of it, so perhaps the sky glow impact is similar. 00:11:29.899 --> 00:11:35.110 In addition, the sky glow impact of incandescent lighting is barely higher than low pressure sodium. 00:11:35.110 --> 00:11:39.870 So it could be that LED street lighting correlated to a 2700K color temperature causes less sky 00:11:39.870 --> 00:11:42.120 glow than high pressure sodium. 00:11:42.120 --> 00:11:44.330 But I think further research needs to be done there. 00:11:44.330 --> 00:11:48.690 In any case, what makes LEDs potentially much better at reducing skyglow is the optical 00:11:48.690 --> 00:11:50.770 systems that can be combined with them. 00:11:50.770 --> 00:11:56.529 Early LED fixtures may have used a large number of small 1 watt LEDs and tiny lenses to direct 00:11:56.529 --> 00:11:57.839 their light. 00:11:57.839 --> 00:12:02.440 Some really bad designs may have simply had an exposed chip, prodiving little directional 00:12:02.440 --> 00:12:03.440 control. 00:12:03.440 --> 00:12:05.839 You still see this a lot in cheap flood light fixtures. 00:12:05.839 --> 00:12:10.000 But newer fixtures like these from Cree will use large chip-on-board emitters, like these 00:12:10.000 --> 00:12:14.580 10W chips but larger, and because they only emit light in one direction, it’s very easy 00:12:14.580 --> 00:12:16.860 to control their output with a lens. 00:12:16.860 --> 00:12:20.880 You can see on the spec sheet that there are 5 lenses in total, though larger fixtures 00:12:20.880 --> 00:12:22.210 have more. 00:12:22.210 --> 00:12:25.130 Most importantly, the optic system is customizable. 00:12:25.130 --> 00:12:30.290 Depending on fixture height and spacing, you may need a wider spread of light or a shorter one. 00:12:30.290 --> 00:12:33.610 Due to the customizable optics, you can get wonderfully consistent lighting on a road 00:12:33.610 --> 00:12:35.910 surface such as this area here. 00:12:35.910 --> 00:12:40.160 This may also help to prevent light pollution because less overall light is needed. 00:12:40.160 --> 00:12:44.660 The hotspots of light you see from above here mean some areas get too much light, and others 00:12:44.660 --> 00:12:45.850 get too little. 00:12:45.850 --> 00:12:50.000 A more scotopic light source with better and more consistent control may not only reduce 00:12:50.000 --> 00:12:54.020 light pollution, but may use less energy and provide safer driving. 00:12:54.020 --> 00:12:58.410 Now many of these developments are rather new, especially the newer optic designs. 00:12:58.410 --> 00:13:03.650 But the advantages of LED lighting become confused when drop-in replacement bulbs are used. 00:13:03.650 --> 00:13:07.149 I don’t have any major issues about going this route--after all replacing the entire 00:13:07.149 --> 00:13:11.170 fixture can be costly, and some drop-in designs are fairly good. 00:13:11.170 --> 00:13:15.480 But the optic system of a sodium, mercury vapor, or metal halide fixture is designed 00:13:15.480 --> 00:13:21.920 to reflect and project light emanating from a tiny arc tube, which an LED drop-in can’t recreate. 00:13:21.920 --> 00:13:25.280 Many of the complaints regarding poor light distribution in LED replacement lamps may 00:13:25.290 --> 00:13:27.600 simply be from the use of these replacements. 00:13:27.600 --> 00:13:30.490 Then there’s the issue with existing ballasts. 00:13:30.490 --> 00:13:34.500 Some drop-in lamps claim to work with existing ballasts, but I have a few misgivings regarding 00:13:34.500 --> 00:13:38.470 how well their power supplies deal with the voltage the ballast provides--particularly 00:13:38.470 --> 00:13:43.490 if the high voltage ignitor sends some crazy voltage spikes to the LED drop-in. 00:13:43.490 --> 00:13:47.110 I’m sure they can be designed to cope, but it still worries me somewhat. 00:13:47.110 --> 00:13:50.660 So then, we have a series of complicated choices to make. 00:13:50.660 --> 00:13:54.910 Sodium vapor lights are pretty efficient, have only a moderate contribution to sky glow, 00:13:54.910 --> 00:13:59.170 cause only minor circadian rhythm disruption if any, and have a proven track record of 00:13:59.170 --> 00:14:00.560 reliability. 00:14:00.560 --> 00:14:06.260 But their color also makes them far less effective at improving safety, and due to the misalignment 00:14:06.260 --> 00:14:10.890 of their output with our scotopic light sensitivity, they require more light (and thus use a lot 00:14:10.890 --> 00:14:13.950 more energy) than a whiter light source. 00:14:13.950 --> 00:14:18.700 Perhaps less an issue but still important is that they contain both mercury and elemental 00:14:18.700 --> 00:14:22.390 sodium, meaning their disposal is far more dangerous and complicated. 00:14:22.390 --> 00:14:26.899 If we were to switch to a white LED source with a color temperature of about 5,000 K, 00:14:26.899 --> 00:14:29.730 nighttime visibility would be greatly increased. 00:14:29.730 --> 00:14:33.399 Studies have shown that people see hazards far sooner under this light, and as the bluer 00:14:33.399 --> 00:14:37.870 wavelengths match our scotopic and mesopic color sensitivity more closely, we can use 00:14:37.870 --> 00:14:41.220 less of it while also achieving a greater safety benefit. 00:14:41.220 --> 00:14:43.090 This reduces the need for energy. 00:14:43.090 --> 00:14:47.660 However, this bluer light contributes more to circadian rhythm disruption and skyglow, 00:14:47.660 --> 00:14:52.020 but some of that is mitigated by the lower output required by this light source. 00:14:52.020 --> 00:14:56.670 Still, many people may not find the aesthetics of this light source pleasing. 00:14:56.670 --> 00:15:00.920 One possible compromise would be to use a warmer color temperature LED light source. 00:15:00.920 --> 00:15:06.330 Data from the US Department of Energy tells us that a 3000K LED light would produce 1.89 00:15:06.330 --> 00:15:11.060 to 2.39 times as much scotopic light as a high pressure sodium lamp, while increasing 00:15:11.060 --> 00:15:15.630 melanopic content between 2.1 and 2.99 times. 00:15:15.630 --> 00:15:19.720 Because of its greater scotopic output, a 3000K LED replacement should only need to 00:15:19.720 --> 00:15:23.550 produce about half as much light as a high pressure sodium lamp. 00:15:23.550 --> 00:15:27.580 This effectively cuts the circadian rhythm disruption potential in half, too, placing 00:15:27.580 --> 00:15:30.290 it near about the same as high pressure sodium. 00:15:30.290 --> 00:15:34.920 The greatest downside to using a warm color temperature LED is in their efficiency. 00:15:34.920 --> 00:15:39.360 The efficiency of these lights is very similar to that of an average high pressure sodium 00:15:39.360 --> 00:15:40.360 lamp. 00:15:40.360 --> 00:15:45.250 At 67 lumens per watt, these 3,000K LEDs are just slightly less efficient than the sodium 00:15:45.250 --> 00:15:47.180 lamp featured in my last video. 00:15:47.180 --> 00:15:51.350 Although you would need only about half as much light output, there are HPS lamps which 00:15:51.350 --> 00:15:53.970 approach 150 lumens per watt. 00:15:53.970 --> 00:15:58.529 This would mean that a 3000K LED will use just about as much energy as a very efficient 00:15:58.529 --> 00:16:00.060 HPS lamp. 00:16:00.060 --> 00:16:03.780 I’d still call that good, but it makes replacement less compelling. 00:16:03.790 --> 00:16:07.140 To normalize the effects of scotopic light content, I’ve multiplied the lumens per 00:16:07.140 --> 00:16:11.180 watt number by the scotopic light content for the following light sources. 00:16:11.180 --> 00:16:15.500 As you can see, the normalized efficiency of the LED goes up considerably as the color 00:16:15.500 --> 00:16:20.240 temperature does, due to both luminous efficiency and greater scotopic content. 00:16:20.240 --> 00:16:25.540 This is likely why most LED street lamp installations are done with the blueish 5700K and higher 00:16:25.540 --> 00:16:27.250 color temperatures. 00:16:27.250 --> 00:16:30.730 You can use the least amount of energy to produce the same amount of perceived brightness 00:16:30.730 --> 00:16:31.730 and safety. 00:16:31.730 --> 00:16:36.690 However, the normalized efficiency of even the 3000K LED is very similar to that of high 00:16:36.690 --> 00:16:41.880 pressure sodium, and few HPS lamps actually output 150 lumens per watt. 00:16:41.880 --> 00:16:47.230 Also, the calculated lumens per watt of the LED is based on the input power of the fixture, 00:16:47.230 --> 00:16:51.320 so the losses in the ballast (which are fairly high for high pressure sodium) aren’t accounted 00:16:51.320 --> 00:16:52.320 for here. 00:16:52.320 --> 00:16:56.270 To conclude, although the high pressure sodium light is very efficient, its primary output 00:16:56.270 --> 00:16:59.400 color is misaligned with our nighttime visibility. 00:16:59.400 --> 00:17:03.980 Only about a quarter of its light is actually effective at stimulating the cells in our eyes. 00:17:03.980 --> 00:17:08.140 Although the cool color temperature of many LED replacements is harsh and aesthetically 00:17:08.140 --> 00:17:12.320 displeasing, studies have shown that it is not only more efficient but also makes driving 00:17:12.329 --> 00:17:13.929 at night safer. 00:17:13.929 --> 00:17:17.379 There is however the potential for greater circadian rhythm disruption and larger amounts 00:17:17.379 --> 00:17:19.809 of skyglow using these bluer light sources. 00:17:19.809 --> 00:17:24.640 Still, it seems clear that the high pressure sodium lamp is on its way out. 00:17:24.640 --> 00:17:28.179 Advancements in LED technology are happening at a breakneck pace. 00:17:28.179 --> 00:17:30.840 Just 10 years ago they weren't seen as viable. 00:17:30.840 --> 00:17:35.269 But today, even the least efficient of LED replacements ends up meeting the efficiency 00:17:35.269 --> 00:17:39.159 of high pressure sodium when scotopic light output is considered. 00:17:39.159 --> 00:17:43.359 As it stands in 2018, we are faced with a choice of efficiency over aesthetics. 00:17:43.359 --> 00:17:48.510 I’m pretty sure I’d enjoy roadways lit with the relatively warm 3000K LEDs, and these 00:17:48.510 --> 00:17:50.679 also wouldn’t disrupt sleep much. 00:17:50.679 --> 00:17:55.690 But you can save a lot more energy (and potentially have safer roadways) with 5700K lighting. 00:17:55.690 --> 00:18:00.850 Either way, it seems clear that LED technology will very soon overtake the tried-and-true 00:18:00.850 --> 00:18:06.269 high pressure sodium lamp, just as the HPS lamp itself replaced the mercury vapor lamp. 00:18:06.269 --> 00:18:11.020 And in 40 or 50 years, who knows what technology might light our roadways. 00:18:12.240 --> 00:18:15.490 So I have a couple of things to close out, first you may have noticed in my chart that 00:18:15.490 --> 00:18:20.341 the mercury vapor lamp had a normalized efficiency of over 100 lumens per watt, and the 50 watt 00:18:20.341 --> 00:18:24.300 sodium lamp in the last video was only 78 lumens per watt. 00:18:24.300 --> 00:18:29.580 Mercury vapor bulbs do have considerable operating disadvantages compared to HPS, most notably 00:18:29.580 --> 00:18:33.539 their steady decrease in light output as they age, but I think it is somewhat humorous that 00:18:33.539 --> 00:18:37.760 our current understanding of the visual system suggests that sodium light may have been a 00:18:37.760 --> 00:18:39.710 step backwards in some situations. 00:18:39.710 --> 00:18:43.269 You may have noticed that I didn’t talk about the blue light from LEDs and how this 00:18:43.269 --> 00:18:48.799 is supposedly ruining our eyes--that’s because the “science” behind this is questionable 00:18:48.799 --> 00:18:50.259 at best. 00:18:50.259 --> 00:18:54.530 You can clearly see in this chart that there is blue-light content in nearly all light 00:18:54.530 --> 00:18:58.619 sources, and lower color temperature LEDs have less blue-light content than their higher 00:18:58.619 --> 00:19:00.590 color temperature varieties. 00:19:00.590 --> 00:19:04.899 I don’t doubt that blue light can disrupt circadian rhythm--that much seems certain. 00:19:04.899 --> 00:19:09.409 But considering that our eyes can withstand the intensity of sunlight, which is far far 00:19:09.409 --> 00:19:14.690 greater than any normal artificial light source and also has a lot of blue light (and ultraviolet 00:19:14.690 --> 00:19:19.309 which definitely IS harmful), I think the blue light thing is just fear mongering. 00:19:19.309 --> 00:19:23.609 If someone can point to some verified, peer-reviewed research supporting this, and not a dodgy 00:19:23.609 --> 00:19:25.970 website, I’ll consider changing my stance. 00:19:25.970 --> 00:19:30.879 In any case, the high flexibility of LED technology means that it can be tuned in pretty much 00:19:30.879 --> 00:19:32.070 any way you like. 00:19:32.070 --> 00:19:35.919 I also want to give a shoutout to VWestlife for the suggestion of LED fixtures with both 00:19:35.919 --> 00:19:40.289 high and low color temperatures that will switch to the warm light later in the night. 00:19:40.289 --> 00:19:44.519 I think that’s a great idea, though obviously it would add expense to any fixture. 00:19:44.519 --> 00:19:49.119 However, I was surprised to learn that the Cree LED fixtures I’ve been using as a reference 00:19:49.119 --> 00:19:54.639 are all capable of dimming, and they have a 0 to 10V control input to enable this. 00:19:54.639 --> 00:19:59.120 Reducing light levels to perhaps 50% of normal after midnight might become a common practice, 00:19:59.120 --> 00:20:01.190 and I think that would be pretty wise. 00:20:01.190 --> 00:20:04.749 Maybe this will get combined with technology similar to Philip’s warm-glow and we’ll 00:20:04.749 --> 00:20:07.370 get incandescent-like lighting at night. 00:20:07.370 --> 00:20:11.210 For those worried about light pollution for astronomical observatories, there are amber 00:20:11.210 --> 00:20:15.470 LED street lights available designed to replace low pressure sodium lights. 00:20:15.470 --> 00:20:20.139 This is also great news for wildlife--many animals cannot see the wavelength of light 00:20:20.139 --> 00:20:24.809 produced by low pressure sodium, so this light source is used where lights may be disruptive. 00:20:24.809 --> 00:20:28.720 One particular example is near beaches where sea turtles lay their eggs. 00:20:28.720 --> 00:20:32.340 After they hatch, baby sea turtles follow moonlight to the ocean, and street lighting 00:20:32.340 --> 00:20:35.100 was confusing the poor things and they were travelling inland. 00:20:35.100 --> 00:20:39.549 Since they cannot see the wavelength of a low pressure sodium light or its amber LED 00:20:39.549 --> 00:20:44.419 equivalent, they aren’t confused and successfully make it to the ocean where they belong. 00:20:44.419 --> 00:20:48.979 As a last little tid-bit, the spec sheet from Cree says that their LED cobra head replacements 00:20:48.979 --> 00:20:55.440 should produce at least 95% of their original light output after 100,000 hours. 00:20:55.440 --> 00:20:59.399 Assuming the driver and heat sink are robust enough, these fixtures should last well beyond 00:20:59.399 --> 00:21:01.340 20 years. 00:21:01.340 --> 00:21:02.519 That is impressive. 00:21:02.519 --> 00:21:04.499 Thanks for watching, I hope you enjoyed the video! 00:21:04.499 --> 00:21:08.289 If you haven’t subscribed to Technology Connections yet, and you like my nerdy deep-dives 00:21:08.289 --> 00:21:11.149 into whatever floats through my head, what are you waiting for? 00:21:11.149 --> 00:21:12.250 Hit the button! 00:21:12.250 --> 00:21:13.250 Please? 00:21:13.250 --> 00:21:16.909 As always, thanks to everyone who supports this channel on Patreon! 00:21:16.909 --> 00:21:20.139 You are all making a big difference and it’s super awesome of you. 00:21:20.139 --> 00:21:23.701 If you’re thinking of joining these awesome people and becoming a patron yourself, why 00:21:23.701 --> 00:21:25.749 not check out my Patreon page? 00:21:25.749 --> 00:21:26.749 Thanks for consideration. 00:21:26.749 --> 00:21:27.880 And I’ll see you next time!