WEBVTT Kind: captions Language: en-US 00:00:00.080 --> 00:00:03.213 The compact fluorescent lamp is, essentially,   00:00:03.600 --> 00:00:04.704 no more. 00:00:04.704 --> 00:00:06.147 We’ve moved on. 00:00:06.147 --> 00:00:08.650 In our search for  more efficient sources of lighting, 00:00:08.650 --> 00:00:11.727 we’ve gone full-steam-ahead towards the glowing computer chips 00:00:11.727 --> 00:00:13.522 we call LEDs. 00:00:13.522 --> 00:00:16.213 I’m borrowing "glowing computer chips," by the way, 00:00:16.213 --> 00:00:18.751 from a vehemently anti-LED commenter 00:00:18.751 --> 00:00:21.473 who thought that was some sort of a derogatory quip, 00:00:21.473 --> 00:00:23.257 but honestly I think it’s awesome! 00:00:23.257 --> 00:00:24.650 Glowing computer chips? 00:00:24.650 --> 00:00:25.627 Yes, please! 00:00:26.000 --> 00:00:28.102 Now, I’m not about to go and make a video about 00:00:28.102 --> 00:00:31.689 “The Slightly Antique Light Bulb that’s Better than Yours” 00:00:31.689 --> 00:00:35.024 because today we are in an objectively  better place. 00:00:35.024 --> 00:00:38.889 I wouldn’t for a moment suggest that the CFL was a superior light source 00:00:38.889 --> 00:00:42.292 to honestly any LED bulb you can find today. 00:00:43.280 --> 00:00:46.507 Well, OK, perhaps better than the very dubiously-made 00:00:46.507 --> 00:00:49.555 LED bulbs that folks like Big Clive find on eBay, 00:00:49.555 --> 00:00:51.680 at least from shock hazards and 00:00:51.680 --> 00:00:54.171 prone-to-burning-your-house-down perspectives. 00:00:54.171 --> 00:00:59.618 But while these were flawed, problematic, and aesthetically… 00:00:59.618 --> 00:01:03.773 suboptimal, in hindsight I think they got an unfairly bad rap. 00:01:03.773 --> 00:01:06.240 And in this video, I’d like to explain why. 00:01:07.200 --> 00:01:10.880 Let’s begin with why these were  a big deal technologically.   00:01:10.880 --> 00:01:13.830 The reason fluorescent lighting was a thing  in the first place 00:01:13.830 --> 00:01:18.371 is that the technology produces much more light per unit of electrical energy 00:01:18.371 --> 00:01:20.907 than is possible with incandescent bulbs. 00:01:20.907 --> 00:01:25.424 Fluorescent tubes also had a much longer life than  the typical incandescent lamp. 00:01:25.424 --> 00:01:29.098 The efficiency and long life made fluorescent lights more economical, 00:01:29.098 --> 00:01:32.122 so in applications like office buildings, store locations, 00:01:32.122 --> 00:01:36.185 and other places where a lot of light  was needed for a long period of time, 00:01:36.185 --> 00:01:38.624 fluorescent lighting made perfect sense. 00:01:38.624 --> 00:01:42.295 But for many years, the form factor was… 00:01:42.295 --> 00:01:43.760 rather commercial. 00:01:44.720 --> 00:01:46.960 Long fluorescent tubes like these 00:01:46.960 --> 00:01:49.335 were the de facto standard for many decades. 00:01:49.335 --> 00:01:55.143 Sometimes we’d get a little clever and bend them into a  U-shape or maybe even a circle, 00:01:55.143 --> 00:01:56.417 but that was it. 00:01:56.417 --> 00:02:00.470 Now, these tubes are filled with a low pressure mixture  of various gases 00:02:00.470 --> 00:02:02.484 and a tiny bit of mercury. 00:02:02.484 --> 00:02:05.314 When an arc discharge is produced along their length, 00:02:05.314 --> 00:02:08.528 the mercury vapor emits a lot of ultraviolet light. 00:02:08.528 --> 00:02:11.942 A mixture of phosphors that coat the inside surface  of the glass 00:02:11.942 --> 00:02:16.648 then convert this largely invisible UV discharge into visible light that we can see, 00:02:16.648 --> 00:02:20.088 making the mercury discharge useful for our eyes. 00:02:20.088 --> 00:02:25.333 Additionally, the glass itself blocks nearly all of the harmful UV wavelengths. 00:02:25.333 --> 00:02:29.914 It takes special quartz glass to produce  UV-C germicidal lamps. 00:02:29.914 --> 00:02:35.233 Nevertheless many people wrongly assumed that ordinary  fluorescent lighting produced harmful UV. 00:02:35.760 --> 00:02:39.711 It didn’t, at least not in any remotely meaningful  amounts. 00:02:39.711 --> 00:02:42.995 This is a bit of a tangent, but the clearest evidence of this 00:02:42.995 --> 00:02:48.695 is that the plastic lenses of fluorescent ceiling fixtures generally didn’t yellow with time, 00:02:48.695 --> 00:02:54.274 which you’d expect them to and rather quickly if the tubes were producing ultraviolet light. 00:02:54.274 --> 00:02:57.042 UV is not kind to plastics. 00:02:57.600 --> 00:03:02.217 Anyway, you also need a ballast and starter circuit to drive these tubes.   00:03:02.640 --> 00:03:05.256 First, an arc discharge has to be established - 00:03:05.256 --> 00:03:09.593 after all, this is a long tube and getting electricity through plasma 00:03:09.593 --> 00:03:13.039 from this end to that end takes a bit of effort. 00:03:13.039 --> 00:03:16.001 There are many different ways to do this which aren’t  important right now, 00:03:16.001 --> 00:03:18.673 but once it’s struck you have a new problem. 00:03:18.673 --> 00:03:24.103 Being a discharge lamp, you need a  current-limiting device or else it will explode. 00:03:25.040 --> 00:03:29.840 OK, well maybe not explode, but discharge lamps  have negative electrical resistance 00:03:29.840 --> 00:03:34.833 and without a current limiting ballast they would rapidly enter a  sort of runaway condition 00:03:34.833 --> 00:03:36.667 and cause problems. 00:03:37.280 --> 00:03:41.743 So, with any given fluorescent fixture, you’ll  have a ballast and starter of some sort 00:03:41.743 --> 00:03:44.220 driving some number of large tubes. 00:03:44.220 --> 00:03:47.635 As a result, fluorescent fixtures tended to be… 00:03:47.635 --> 00:03:48.383 large. 00:03:48.640 --> 00:03:52.069 This is a basic shop light and as you can see it’s lar- 00:03:52.069 --> 00:03:52.856 long. 00:03:52.856 --> 00:03:54.781 Four feet to be precise. 00:03:54.781 --> 00:03:59.384 Shorter tubes were available, and eventually you also had  fun things like the circline tubes 00:03:59.384 --> 00:04:05.015 but in general fluorescent lighting was limited to overhead  applications where space wasn’t at a premium. 00:04:05.015 --> 00:04:06.694 So… shop lights. 00:04:06.694 --> 00:04:07.628 Troffer lights. 00:04:07.628 --> 00:04:09.137 Surface-mounted light strips. 00:04:09.137 --> 00:04:10.224 Et cetera. 00:04:10.224 --> 00:04:13.981 Applications that, for the most part, weren’t found in the home. 00:04:13.981 --> 00:04:17.337 And if they were, it wasn’t usually in a formal space. 00:04:17.840 --> 00:04:21.365 But wouldn’t it be great if we could have  the benefits of fluorescent light - 00:04:21.365 --> 00:04:23.936 that being longer life and lower energy use - 00:04:23.936 --> 00:04:25.510 in more places? 00:04:25.510 --> 00:04:26.930 Heck yes it would. 00:04:26.930 --> 00:04:29.162 But how to do that? 00:04:29.162 --> 00:04:33.607 Well, you could build this thing which I made a  video on quite a while ago, 00:04:33.607 --> 00:04:37.316 but that’s definitely not the most practical way to go about it. 00:04:37.316 --> 00:04:42.220 What we really needed was some sort of self-contained fluorescent lamp 00:04:42.220 --> 00:04:46.672 with an integrated ballast that is about the same size as an incandescent bulb, 00:04:46.672 --> 00:04:48.884 and with hopefully the same brightness. 00:04:48.884 --> 00:04:51.511 And so, these things were born. 00:04:51.920 --> 00:04:56.108 This now iconic curly-q design was  the final landing spot 00:04:56.108 --> 00:04:58.348 after a bunch of different experiments. 00:04:58.348 --> 00:05:02.774 See, the ultimate limitation to how much light you can get out of a fluorescent tube 00:05:02.774 --> 00:05:04.856 is its surface area. 00:05:04.856 --> 00:05:07.036 With these long fellas 00:05:07.520 --> 00:05:09.065 that isn’t much of an issue, 00:05:09.065 --> 00:05:12.374 but if you want to get the same amount of light as an incandescent bulb 00:05:12.374 --> 00:05:15.604 in the same amount of space using a fluorescent tube, 00:05:15.604 --> 00:05:17.591 it takes some twisted thinking. 00:05:17.920 --> 00:05:21.920 This 60 watt equivalent lamp has a tube which  begins here, 00:05:21.920 --> 00:05:23.768 travels in a helix to the top, 00:05:23.768 --> 00:05:26.800 then doubles back down the other side where it ends here. 00:05:27.520 --> 00:05:33.280 That allows it to occupy a volume smaller than  a cube 2 inches or 5 centimeters on each side   00:05:33.280 --> 00:05:37.840 while the tube itself is roughly 14  inches or 35 centimeters in length.   00:05:38.400 --> 00:05:42.578 This alone is a huge achievement in manufacturing  and design. 00:05:42.578 --> 00:05:44.824 This is glass, mind you. 00:05:44.824 --> 00:05:50.007 We needed a way to mass-produce hollow glass tubes formed  into this intricate shape, 00:05:50.007 --> 00:05:51.176 and we did it! 00:05:51.280 --> 00:05:53.668 But that wasn’t the only innovation, here. 00:05:53.668 --> 00:05:58.816 The tube is driven at a higher intensity than an ordinary tube like this one. 00:05:58.816 --> 00:06:02.564 That allows its phosphors to glow more much brightly. 00:06:02.564 --> 00:06:06.536 There seems to be an inverse relationship between the diameter of the glass 00:06:06.536 --> 00:06:08.913 and how intense the discharge is. 00:06:08.913 --> 00:06:12.230 You see this as we moved from the old T12 tubes down to 00:06:12.230 --> 00:06:14.913 T8 tubes and eventually T5 tubes. 00:06:15.280 --> 00:06:19.440 I’m not exactly sure why that is, hopefully  one of you will let us know down below.   00:06:19.440 --> 00:06:22.237 But there’s also one more trick up its sleeve. 00:06:22.237 --> 00:06:24.187 Its electronic ballast. 00:06:24.480 --> 00:06:28.800 Fluorescent lamps operate on the same  AC power that everything else does.   00:06:28.800 --> 00:06:33.680 Because the voltage crosses the zero  point 100 or 120 times per second,   00:06:33.680 --> 00:06:39.220 lamps run on traditional magnetic ballasts are  actually pulsing at that frequency. 00:06:39.220 --> 00:06:42.389 This periodic dark time limits their output, 00:06:42.389 --> 00:06:48.489 but electronic ballasts get around this by first converting the incoming AC power to DC 00:06:48.489 --> 00:06:52.160 and then driving the actual tube at very high frequency AC. 00:06:52.960 --> 00:06:58.400 The frequency is so high that the phosphors  don’t have time to decay between pulses,   00:06:58.400 --> 00:07:02.974 and the effect is the lamp is constantly lit and does not pulse. 00:07:02.974 --> 00:07:05.331 This not only boosts it output further, 00:07:05.331 --> 00:07:09.724 but also eliminated the potential for eyestrain issues that old-style fluorescent lights 00:07:09.724 --> 00:07:11.869 could cause in some individuals. 00:07:11.869 --> 00:07:14.320 I actually made a video about that quite a while ago now. 00:07:14.960 --> 00:07:17.503 So. These were pretty neat little things. 00:07:17.503 --> 00:07:20.587 I keep using the past tense but you can still buy them, 00:07:20.587 --> 00:07:23.560 they’re just very much out of fashion  at this point. 00:07:23.560 --> 00:07:24.960 But they were a big deal. 00:07:25.520 --> 00:07:28.882 Bringing the advantages of fluorescent  light to more places was, 00:07:28.882 --> 00:07:31.662 I would argue, a great thing. 00:07:31.662 --> 00:07:35.156 However, they weren’t received with unanimous praise. 00:07:35.156 --> 00:07:38.468 And deservedly so, we’ll discuss that shortly. 00:07:38.468 --> 00:07:42.000 At the same time, though, a lot of the hatred for these things just… 00:07:42.560 --> 00:07:44.126 didn’t make sense. 00:07:44.126 --> 00:07:46.298 At least, not holistically. 00:07:46.298 --> 00:07:48.370 The biggest problem with these things? 00:07:48.370 --> 00:07:49.830 They’re nuanced. 00:07:49.830 --> 00:07:53.398 And we know how easily the  general public engages with nuance. 00:07:54.080 --> 00:07:58.240 See, I have a pet theory that a large  percentage of the aversion to CFLs   00:07:58.240 --> 00:08:04.240 came from people who really just… didn’t  know that much about lighting in general.   00:08:04.880 --> 00:08:08.769 I would argue that the CFL was really the first  mainstream way 00:08:08.769 --> 00:08:12.934 to get non-incandescent color temperatures in the home. 00:08:12.934 --> 00:08:16.699 Sure, fluorescent tubes weren’t unheard of in the domestic environment. 00:08:16.699 --> 00:08:19.581 They found their place in garages and shop  spaces, 00:08:19.581 --> 00:08:25.189 sometimes kitchens had them and linear over-the-mirror sort of deals in bathrooms were kinda common. 00:08:25.189 --> 00:08:32.688 But prior to the CFL, those sorts of odd applications defined what “fluorescent” light was. 00:08:32.688 --> 00:08:38.720 Fluorescent light was that cooler, gray-blue light you saw at the  office or in the store or maybe the basement. 00:08:39.360 --> 00:08:41.376 And see, here’s where my theory comes in, 00:08:41.376 --> 00:08:44.080 if that was how you imagined fluorescent light, 00:08:44.640 --> 00:08:47.178 maybe you weren't even aware of the CFLs 00:08:47.178 --> 00:08:52.430 like this that mimicked the light from an incandescent bulb fairly well. 00:08:52.430 --> 00:08:56.496 Yes, I know there is and can even see the difference, 00:08:56.496 --> 00:08:59.735 but I would put money on there being lots of  folks out there 00:08:59.735 --> 00:09:02.321 who wouldn’t know this was a fluorescent light source 00:09:02.321 --> 00:09:04.868 unless they could see the bulb itself. 00:09:04.868 --> 00:09:06.729 Their imagination is that 00:09:06.729 --> 00:09:13.360 *this* is fluorescent, and when the very idea of color  temperature is a new concept to a lot of people, 00:09:13.920 --> 00:09:17.218 I’d bet there was a lot of confusion that led to  undeserved 00:09:17.218 --> 00:09:20.694 and sometimes misplaced hatred of these little fellas. 00:09:20.694 --> 00:09:25.486 No matter how hard store displays  worked to show how warm the light actually could be. 00:09:26.080 --> 00:09:29.884 Again, this isn’t meant to be a love letter to the CFL. 00:09:29.884 --> 00:09:32.891 They were bad in a fair number of ways. 00:09:32.891 --> 00:09:38.341 But even I, a really picky person when it  comes to quality of light and color temperature, 00:09:38.720 --> 00:09:41.448 don’t have a problem with most CFLs. 00:09:41.448 --> 00:09:44.053 Some were better than others, yeah, but at their peak 00:09:44.053 --> 00:09:47.644 a nice 2700K CFL in a table lamp like this? 00:09:47.644 --> 00:09:48.987 No problem. 00:09:48.987 --> 00:09:54.391 About the worst thing I ever notice is a slightly pink, occasionally green cast to the light, 00:09:54.391 --> 00:09:58.433 but it’s really only noticeable when you have another light to compare it to - 00:09:58.433 --> 00:10:00.800 if an entire room is lit with the same CFLs, 00:10:01.440 --> 00:10:05.360 my eyes would acclimate to it and I just  couldn’t see that color cast anymore. 00:10:06.080 --> 00:10:08.920 Of course, color rendering index isn’t fantastic. 00:10:08.920 --> 00:10:12.878 It’s really expensive to develop near-perfect phosphor formulations 00:10:12.878 --> 00:10:16.499 so nearly all lamps were good but not great. 00:10:16.499 --> 00:10:19.888 But I don’t think I’ve ever run across a warm white CFL 00:10:19.888 --> 00:10:24.292 that messed with colors to the point that I could notice without being hypercritical. 00:10:24.292 --> 00:10:30.120 I still have some CFLs kicking around without any intention to replace them until they fail. 00:10:30.120 --> 00:10:33.030 In many applications their limitations just don’t matter 00:10:33.030 --> 00:10:35.920 so I’m not about to bother upgrading them needlessly. 00:10:36.640 --> 00:10:39.627 But, and this was in my opinion their  Achilles heel, 00:10:39.627 --> 00:10:43.947 the suitableness of CFLs was very situational. 00:10:43.947 --> 00:10:48.155 They wanted to be and were sold as general purpose light bulbs 00:10:48.155 --> 00:10:51.268 but couldn’t actually be used as such. 00:10:51.268 --> 00:10:54.902 Some fixtures were absolutely unfit for CFLs, 00:10:54.902 --> 00:10:58.235 despite the fact that CFLs could fit inside them. 00:10:58.235 --> 00:11:01.398 Similarly, some uses for them were impractical, 00:11:01.398 --> 00:11:03.603 and others would shorten their life considerably 00:11:03.603 --> 00:11:06.682 leading many folks to believe that they didn’t last as long 00:11:06.682 --> 00:11:10.311 as the incandescent lights they were supposed to outlive tenfold. 00:11:10.311 --> 00:11:12.660 And yeah, often they didn’t! 00:11:12.660 --> 00:11:16.482 But in most cases that wasn’t really their fault. 00:11:16.880 --> 00:11:19.331 Let’s start with those untimely deaths. 00:11:19.331 --> 00:11:22.930 Here’s a fixture that seems perfectly appropriate for CFLs. 00:11:22.930 --> 00:11:25.833 Plenty of clearance for the lamps, not a problem. 00:11:25.833 --> 00:11:28.142 Except yes a problem! 00:11:28.142 --> 00:11:30.519 This fixture is enclosed. 00:11:30.519 --> 00:11:35.440 CFLs didn’t get anywhere near as hot as  the incandescent bulbs they replaced, 00:11:35.440 --> 00:11:37.344 but they still got hot. 00:11:37.344 --> 00:11:39.598 Trap one of ‘em, or worse multiple, 00:11:39.598 --> 00:11:42.880 inside a glass bowl with no airflow at all 00:11:42.880 --> 00:11:45.741 and you’ve made a torture chamber for them. 00:11:45.741 --> 00:11:50.532 The electronic components in their integrated ballasts are only rated to a certain temperature, 00:11:50.532 --> 00:11:53.360 which can easily be exceeded in fixtures like these. 00:11:54.000 --> 00:11:58.544 And so, premature lamp failure was almost a  given in enclosed fixtures, 00:11:58.544 --> 00:12:04.800 and to be fair the packaging for these usually warned you that  they weren’t suitable in enclosed fixtures   00:12:04.800 --> 00:12:06.881 (but who reads the packaging?). 00:12:06.881 --> 00:12:10.279 This is actually still something of a problem with LED bulbs, 00:12:10.279 --> 00:12:16.175 but many of them on the market today have  electronics and temp sensors designed to limit their output and prevent overheating 00:12:16.175 --> 00:12:17.760 which allows them to be used without worry. 00:12:18.480 --> 00:12:22.160 But even simple orientation  could be a problem with CFLs.   00:12:22.160 --> 00:12:26.983 If a fixture had its socket at the top and thus the bulb was hanging, 00:12:26.983 --> 00:12:31.317 a CFL would find itself with its ballast above the tube. 00:12:31.317 --> 00:12:35.120 The tube is the part that gets hottest, and since warm air rises, 00:12:35.680 --> 00:12:38.384 well the ballast gets pretty hot. 00:12:38.384 --> 00:12:43.341 This wasn’t always a problem - there could be enough airflow depending on the fixture's design. 00:12:43.341 --> 00:12:45.571 But often enough it was. 00:12:45.571 --> 00:12:47.436 More premature failures. 00:12:47.840 --> 00:12:49.736 Another thing that would kill them quickly? 00:12:49.736 --> 00:12:51.805 Frequent short use. 00:12:51.805 --> 00:12:56.385 Now, this problem didn’t necessarily need to be a problem. 00:12:56.385 --> 00:12:59.398 But, because we humans are impatient fickle things, 00:12:59.398 --> 00:13:03.017 we didn’t like bulbs that behaved like this one. 00:13:03.017 --> 00:13:06.750 A delay between switch-on and light is certainly annoying, 00:13:06.750 --> 00:13:11.209 but this is much easier on the electrodes at  the ends of the tube. 00:13:11.209 --> 00:13:15.252 This programming allows thermionic emissions from the filaments at the ends 00:13:15.252 --> 00:13:16.749 to help strike the arc. 00:13:16.749 --> 00:13:18.463 It’s the easiest way to do that, 00:13:18.463 --> 00:13:21.122 and the most gentle on the electrodes. 00:13:21.122 --> 00:13:24.746 This particular bulb spent many years in a fixture on a timer 00:13:24.746 --> 00:13:28.740 and has undoubtedly racked up many  thousands of hours, 00:13:28.740 --> 00:13:30.193 yet it still functions. 00:13:30.640 --> 00:13:33.285 But to the average person this was just annoying 00:13:33.285 --> 00:13:35.760 so instant-start ballasts became the norm. 00:13:36.320 --> 00:13:40.877 These essentially just blast a high-voltage  charge right through the tube. 00:13:40.877 --> 00:13:43.597 And this does make for a delay-free start, 00:13:43.597 --> 00:13:46.988 but it comes at the expense of harming the electrodes. 00:13:46.988 --> 00:13:49.811 Each start wears them down a tiny bit, 00:13:49.811 --> 00:13:53.333 and while a long operating session may help  restore them, 00:13:53.333 --> 00:13:58.490 when used in places like a closet or bathroom where the light is only on for a few minutes at a time, 00:13:58.490 --> 00:14:01.384 that opportunity simply doesn't exist. 00:14:01.384 --> 00:14:04.544 The great irony here was that the instant start ballasts 00:14:04.544 --> 00:14:08.398 made them seem more practical for these short-use applications, 00:14:08.398 --> 00:14:11.280 but in actuality it made them much worse. 00:14:12.080 --> 00:14:15.760 The CFL was really best suited for general  living space illumination 00:14:15.760 --> 00:14:18.961 that would be left on for long periods of time. 00:14:18.961 --> 00:14:22.994 But of course, much of that came from ceiling light fixtures like these 00:14:22.994 --> 00:14:26.176 which would harm the bulbs but in a different way. 00:14:26.176 --> 00:14:29.189 Where they could really shine were in table and floor lamps, 00:14:29.189 --> 00:14:32.835 a lot of outdoor fixtures like  these carriage things would do OK also… 00:14:33.124 --> 00:14:38.240 Really, just any fixture where it wasn’t upside  down and had decent airflow would do alright. 00:14:38.960 --> 00:14:41.186 But what about recessed cans? 00:14:41.186 --> 00:14:43.985 Those are always hanging bulbs upside down. 00:14:43.985 --> 00:14:47.042 Well, this wasn’t the only style of CFL. 00:14:47.042 --> 00:14:52.774 In their heyday, we were brute-forcing  CFL tech into all sorts of weird and interesting places, 00:14:52.774 --> 00:14:55.435 like these imitation flood lamps 00:14:55.435 --> 00:14:58.542 (some were even made to look like actual PAR lamps) 00:14:58.542 --> 00:15:00.903 or, my particular favorite, 00:15:00.903 --> 00:15:04.417 these teeny tiny things that IKEA sold at one point 00:15:04.417 --> 00:15:08.590 (with an intermediate base to be just  that extra bit of annoying) 00:15:08.590 --> 00:15:12.491 Still, even in suitable applications where they wouldn’t  overheat and die, 00:15:12.491 --> 00:15:14.720 CFLs still had drawbacks. 00:15:15.360 --> 00:15:19.166 A big, unsolvable issue was that when you  first turned them on, 00:15:19.166 --> 00:15:22.036 they hadn’t yet gotten their Fulbright scholarship. 00:15:22.036 --> 00:15:24.753 Fluorescent lamps take time to warm up 00:15:24.753 --> 00:15:28.095 as not all the mercury is a vapor when they’re cold. 00:15:28.095 --> 00:15:31.781 For general purpose lamps like these at room temperatures, 00:15:32.320 --> 00:15:34.156 this wasn’t a huge deal. 00:15:34.156 --> 00:15:37.433 They’d be about half their normal brightness when cold, 00:15:37.433 --> 00:15:41.173 certainly not at their best but not unusable by any means either. 00:15:41.173 --> 00:15:45.334 And usually within just a minute or so, they’d be as bright as they’d get. 00:15:45.334 --> 00:15:50.274 But, for whatever reason, most of the specialty bulbs that came about over the years 00:15:50.274 --> 00:15:53.213 would start out with next-to-useless brightness. 00:15:53.760 --> 00:15:56.676 Just take a look at this little “flood” thing. 00:15:56.676 --> 00:15:59.925 This is at room temperature but hasn’t been started in a while. 00:15:59.925 --> 00:16:01.561 Turn it on 00:16:01.561 --> 00:16:03.534 and that’s barely anything. 00:16:03.534 --> 00:16:08.437 In fact you can see the pink tinge brought about by the bit of neon in there used as a starter gas. 00:16:08.437 --> 00:16:12.440 Sometimes this would be purple suggesting argon was the starter gas. 00:16:12.440 --> 00:16:17.302 Anyway, after a while it is nice and bright, but this is annoying. 00:16:17.302 --> 00:16:22.749 In my experience, nearly every non-standard CFL  behaved just like this 00:16:22.749 --> 00:16:25.911 be it flood. Globe. A19 mimics. 00:16:25.911 --> 00:16:27.697 These weird little things. 00:16:27.697 --> 00:16:31.377 Really anything that wasn’t the standard exposed curly q. 00:16:31.377 --> 00:16:36.557 I’m really curious about why cold starts seemed  to be much worse in these specialty bulbs, 00:16:36.557 --> 00:16:41.749 especially since these floods, for example, are  really just a standard lamp with a special shade. 00:16:42.160 --> 00:16:48.391 I have no idea why these should have darker starts  than these ordinary lamps yet they do. 00:16:48.391 --> 00:16:55.200 Only thing I can think of is that having the tube enclosed led to  higher operating temps thanks to a lack of airflow 00:16:55.200 --> 00:17:00.642 which required different vapor pressure characteristics  which ultimately meant weak, slow starts. 00:17:01.200 --> 00:17:06.077 And then, regardless of type, using CFLs  outdoors was… 00:17:06.077 --> 00:17:07.440 interesting. 00:17:07.440 --> 00:17:14.487 In cold temperatures, they’d be *very* dim at startup and could take several minutes to reach full brightness. 00:17:14.487 --> 00:17:17.020 For something like a porch light that’ll be  on all night, 00:17:17.020 --> 00:17:20.851 well that’s not too much of a problem but when it gets *really* cold, 00:17:20.851 --> 00:17:24.234 well they start to lose brightness. 00:17:24.234 --> 00:17:27.763 They can’t actually reach their operating temperature when it's that cold, 00:17:27.763 --> 00:17:32.692 and on cold enough days they may not even be able to start at all. 00:17:32.692 --> 00:17:34.239 Not ideal. 00:17:34.239 --> 00:17:38.455 And, as you can imagine, their proneness to death by overheating 00:17:38.455 --> 00:17:42.720 meant that in hot climates they’d be even  more at risk depending on the situation. 00:17:43.360 --> 00:17:46.130 Oh, and of course there’s the elephant in  the room. 00:17:46.130 --> 00:17:47.419 Mercury. 00:17:47.419 --> 00:17:49.517 It’s not great stuff! 00:17:49.517 --> 00:17:53.027 Now, in some senses this problem was overblown. 00:17:53.027 --> 00:17:57.090 The amount of mercury that’s actually in these is miniscule, 00:17:57.090 --> 00:18:00.380 and it’s not like if you broke one  you’d need to exit the room in a panic 00:18:00.380 --> 00:18:04.492 and call a HAZMAT team else you’ll wake up the next day  with an extra toe. 00:18:04.492 --> 00:18:08.142 Accidental breakage was not a four-alarm fire kind of event. 00:18:08.142 --> 00:18:12.840 But at scale, disposing of these in landfills is not good. 00:18:12.840 --> 00:18:16.810 They should be recycled, and unfortunately efforts to do this 00:18:16.810 --> 00:18:20.160 (particularly in the US it seems) were pretty sloppy. 00:18:20.880 --> 00:18:24.565 Some stores kept receptacles for you to return  dead bulbs, 00:18:24.565 --> 00:18:26.889 IKEA was particularly good at this, 00:18:26.889 --> 00:18:31.200 but plenty of places left you in the dark on how  to deal with a bulb that left you in the dark. 00:18:31.840 --> 00:18:35.244 You can never get 100% capture with anything, 00:18:35.244 --> 00:18:39.328 and while estimates for how many ended up being recycled are all over the place, 00:18:39.328 --> 00:18:42.140 in the best case it’s quite bleak. 00:18:42.140 --> 00:18:45.805 The majority of these were disposed of improperly. 00:18:45.805 --> 00:18:47.270 Yay. 00:18:47.270 --> 00:18:50.106 There are only two ways to fix that problem: 00:18:50.106 --> 00:18:53.536 make it easier or perhaps incentivize recycling them 00:18:53.536 --> 00:18:56.137 (shame on policymakers for not getting that together) 00:18:56.800 --> 00:18:59.380 or… eliminate them entirely. 00:18:59.380 --> 00:19:01.440 And luckily that’s pretty much where we’re at. 00:19:02.160 --> 00:19:07.101 LED bulbs are still e-waste, to be clear, but they don’t  have mercury in them 00:19:07.101 --> 00:19:08.926 and that’s a huge plus. 00:19:09.120 --> 00:19:14.225 In general, the LED bulb fixed all the problems that CFLs had 00:19:14.225 --> 00:19:17.465 (except maybe for RF interference problems) 00:19:17.861 --> 00:19:20.911 One particular thing that LEDs brought us that 00:19:20.911 --> 00:19:25.382 really just never got figured out with CFLs was dimming. 00:19:25.382 --> 00:19:31.846 Sure, dimmable CFLs did exist but in my experience they always sucked. 00:19:31.846 --> 00:19:33.939 I don't know what was so hard about that - 00:19:33.939 --> 00:19:38.241 at my old job we had TONS of ceiling fixtures in the lobby that had 00:19:38.241 --> 00:19:41.040 fixed ballasts and replaceable lamps like these, 00:19:41.600 --> 00:19:44.002 and they handled dimming just fine. 00:19:44.480 --> 00:19:49.200 I guess it was just too expensive to figure out  how to do that in lamp-integrated ballasts. 00:19:49.840 --> 00:19:52.528 Now, dimmable LEDs are everywhere, 00:19:52.528 --> 00:19:56.001 and they come in all sorts of great styles with various features. 00:19:56.001 --> 00:19:59.085 Some of them are even… smart. 00:19:59.085 --> 00:20:01.411 I think they know algebra. 00:20:01.411 --> 00:20:05.694 We’ve gotten good at designing LED lamps to dissipate the heat they generate well, 00:20:05.694 --> 00:20:08.997 and to keep themselves in check when that’s not possible. 00:20:08.997 --> 00:20:11.860 Color rendering these days  is generally much better, 00:20:11.860 --> 00:20:13.137 even in basic lamps, 00:20:13.137 --> 00:20:16.243 and can be truly excellent in premium products. 00:20:16.243 --> 00:20:19.200 We are definitely in a better place today. 00:20:19.200 --> 00:20:20.642 But, these things? 00:20:20.915 --> 00:20:22.543 They didn’t suck. 00:20:22.543 --> 00:20:23.857 They were important. 00:20:23.857 --> 00:20:28.064 Miniaturizing fluorescent lights into these tiny things was a feat of its own, 00:20:28.064 --> 00:20:29.760 and I appreciate it greatly. 00:20:30.400 --> 00:20:33.046 CFLs also finally spurred development 00:20:33.046 --> 00:20:36.181 of warm white phosphor formulations that weren’t terrible. 00:20:36.400 --> 00:20:39.582 I mean, this circline tube is just horrid! 00:20:39.582 --> 00:20:43.256 That is not a warm and pleasant glow, GE. 00:20:43.256 --> 00:20:44.686 It’s sickly. 00:20:45.040 --> 00:20:48.423 These were just… misunderstood. 00:20:48.423 --> 00:20:52.450 I bet many of you can think back to a few of these that just kept on truckin' 00:20:52.450 --> 00:20:55.572 in fact maybe you’re like me and still have some going strong. 00:20:55.840 --> 00:20:58.985 When they were used in an appropriate application, 00:20:58.985 --> 00:21:00.834 they really weren’t half bad. 00:21:00.834 --> 00:21:06.163 If anything is to blame for their reputation, I think it’s that  their abilities were oversold. 00:21:06.163 --> 00:21:09.869 A lot of them died too young simply because of how they were used, 00:21:09.869 --> 00:21:12.399 and it’s not their user’s fault! 00:21:12.399 --> 00:21:14.781 Manufacturers just want to sell you light bulbs, 00:21:14.781 --> 00:21:17.751 so they’re not likely to help you figure out those details. 00:21:18.080 --> 00:21:20.772 And again, and I know I’ve said this like a  million times, 00:21:20.772 --> 00:21:22.720 it’s not like I miss these. 00:21:23.360 --> 00:21:27.111 Even in the best of cases, they would still lose output as  they age 00:21:27.111 --> 00:21:29.068 which often lead to color shifting. 00:21:29.068 --> 00:21:32.659 They still take time to warm up, particularly when they're cold. 00:21:32.659 --> 00:21:36.532 Dimmability is still rare and awful when it is there. 00:21:36.532 --> 00:21:40.171 But I don’t look back on these with pure hatred. 00:21:40.171 --> 00:21:44.400 In fact, I’m quite fond of what they managed to accomplish. 00:21:44.995 --> 00:21:47.549 The end. 00:21:48.391 --> 00:21:50.971 ♫ underappreciatedly smooth jazz ♫ 00:21:51.937 --> 00:21:56.507 I wouldn’t for a moment suggest that  the CFL was a superior ligh… 00:21:56.507 --> 00:21:57.494 darnit! 00:21:57.494 --> 00:21:59.703 I missed a pause. 00:21:59.703 --> 00:22:01.506 Goin’ REAL great! 00:22:01.506 --> 00:22:04.966 ...to produce UV-C germicidal lambs. Neverth - 00:22:05.840 --> 00:22:07.924 Lamps. Not lambs. 00:22:07.924 --> 00:22:10.362 Germicidal lambs? 00:22:10.362 --> 00:22:11.476 What are those? 00:22:11.476 --> 00:22:14.311 First, and arc discharge ha- [tube hits table and clanks] 00:22:18.127 --> 00:22:20.901 Don’t mind me, just thinking about how I’m gonna  put this down. 00:22:20.901 --> 00:22:21.538 Anywa… 00:22:21.538 --> 00:22:25.159 Anyway, you al…. [mumbling] nah I don’t wanna be too dismissive of myself. 00:22:25.159 --> 00:22:29.851 First, an arc discharge has to be established [tube clanks on table again] after… dang  [laughter] 00:22:29.851 --> 00:22:31.760 how many times will I record this line? 00:22:32.640 --> 00:22:39.200 From this end to that end takes a bit of  effort. There are many… [grumbling under breath] 00:22:39.200 --> 00:22:42.400 There are many different ways to do  this which aren’t important right now,   00:22:42.400 --> 00:22:45.042 but I missed a word so I have to restart. 00:22:45.042 --> 00:22:48.178 Anyway, you also need a ballast and starter [lamp clanks]... 00:22:48.178 --> 00:22:49.557 Anyway, you also need - 00:22:49.557 --> 00:22:54.798 Anyway, you also need a ballast and starter circuit to drive [lamp clanks]... 00:22:54.798 --> 00:22:57.054 this is REALLY annoying. 00:22:57.054 --> 00:23:01.012 With these long fellas, that wasn’t much of an issue [tube clanks repeatedly] but if you want to get... 00:23:01.012 --> 00:23:03.846 Stop making so much noise! 00:23:05.630 --> 00:23:08.792 Did you enjoy this light content? 00:23:08.792 --> 00:23:11.197 I know I did. 00:23:11.197 --> 00:23:14.988 It got my mind swirling with all sorts of twisted ideas. 00:23:14.988 --> 00:23:17.678 OK that sounds bad. 00:23:17.678 --> 00:23:20.866 Yikes.