WEBVTT Kind: captions Language: en 00:00:00.130 --> 00:00:04.203 This video contains a heckuva lot of flickering imagery throughout, 00:00:04.203 --> 00:00:08.310 and if that is a problem for you, you probably shouldn’t watch this one. 00:00:08.310 --> 00:00:12.556 Hello and welcome to No Effort November, a series of — 00:00:12.556 --> 00:00:13.587 what? 00:00:14.469 --> 00:00:16.314 It's December‽ 00:00:16.938 --> 00:00:18.896 Oh no, I'm not prepared for this! 00:00:18.896 --> 00:00:20.556 Uh, uh, uh OK 00:00:20.556 --> 00:00:26.538 well... mmmhmm I'm not doing that thing where I paint Christmas lights like some kinda madman. 00:00:26.538 --> 00:00:30.985 I know, it’s devastating to me, too, but in light of the season 00:00:30.985 --> 00:00:34.792 I will talk about a certain kind of light of the season. 00:00:34.792 --> 00:00:40.059 Some folks like to decorate their homes all subtle-like with a candle in the window sills. 00:00:40.059 --> 00:00:44.539 But real candles are annoying and also a fire hazard 00:00:44.539 --> 00:00:52.048 so some people will use these definitely convincing and not-at-all tacky light sockets on a stick! 00:00:52.048 --> 00:00:56.830 put a little night light bulb up there, and have electric candlelight. 00:00:57.131 --> 00:01:01.268 And honestly, on the right sort of house it does look pretty festive. 00:01:01.268 --> 00:01:03.684 At least from outside. 00:01:03.684 --> 00:01:07.984 But what if you want something that's just a little less subtle? 00:01:07.984 --> 00:01:12.441 Well, you get yourself one of these things. 00:01:13.560 --> 00:01:14.872 Ain’t it a beauty? 00:01:14.872 --> 00:01:20.834 It replicates the bright orange glow and incessant flickering of a real candle just perfectly! 00:01:20.834 --> 00:01:23.866 I mean, take a look at this side-by-side comparison. 00:01:23.866 --> 00:01:26.257 I can’t tell the difference! 00:01:26.946 --> 00:01:28.998 OK, I’m being a little hard on it. 00:01:28.998 --> 00:01:35.329 This thing is actually pretty cool, if not exactly faithful to what it’s attempting to mimic. 00:01:35.329 --> 00:01:41.110 This is a “flicker flame” lamp, though I’m not sure if it even has an official name. 00:01:41.110 --> 00:01:47.336 It’s a rather clever exploitation of a flaw that can occur in neon indicator lamps. 00:01:47.336 --> 00:01:49.310 This is one such lamp. 00:01:49.310 --> 00:01:55.507 It’s dead simple, just a glass envelope filled with a bit of neon (and often a hint of argon), 00:01:55.507 --> 00:01:58.363 with a pair of electrodes floating in the middle. 00:01:58.363 --> 00:02:03.361 Put a fairly high voltage across these electrodes, something a bit north of 100 volts, 00:02:03.361 --> 00:02:05.942 and you’ll get current to flow between them. 00:02:05.942 --> 00:02:12.216 And thanks to neon being neon that will cause a visible glow discharge on the cathode. 00:02:12.216 --> 00:02:15.862 Why exactly that happens we don’t need to get into but importantly, 00:02:15.862 --> 00:02:20.314 once you have exceeded the striking voltage and you get that glow discharge, 00:02:20.314 --> 00:02:24.331 you need a current-limiting resistor to keep the current in check. 00:02:24.331 --> 00:02:28.599 Without it, the lamp will basically short circuit and often explode. 00:02:28.599 --> 00:02:29.781 Fun! 00:02:29.781 --> 00:02:31.187 Wanna see that happen? 00:02:31.187 --> 00:02:32.336 Why not! 00:02:32.336 --> 00:02:37.160 Good thing I’ve got those switched outlets in this room so I can do this from a distance. 00:02:37.504 --> 00:02:39.411 Contact! 00:02:40.100 --> 00:02:43.166 That was significantly less violent than I was expecting. 00:02:43.166 --> 00:02:45.240 OK, we're doing this one more time. 00:02:45.455 --> 00:02:46.534 Contact! 00:02:46.534 --> 00:02:48.144 [rattling as glass bounces on table] 00:02:48.725 --> 00:02:50.236 Much better. 00:02:50.731 --> 00:02:55.635 Now because AC line voltage exceeds the striking voltage of these lamps, 00:02:55.635 --> 00:03:01.109 all you need is that resistor to keep ‘em from, y'know, splodin' and they’ll work. 00:03:01.109 --> 00:03:05.318 They also have very long lives and consume very little power. 00:03:05.318 --> 00:03:11.032 So, they are often used as indicator lights in simple devices hooked up to mains power. 00:03:11.032 --> 00:03:16.768 Since LEDs are so stinking cheap now we don’t see these quite so often as we used to, 00:03:16.768 --> 00:03:18.726 but they’re definitely not gone. 00:03:18.726 --> 00:03:24.236 Simple kitchen appliances without electronics, for example, will often still use them because 00:03:24.236 --> 00:03:31.597 if you don’t need a low-voltage power supply for logic circuitry, you might as well not bother with driving an LED. 00:03:31.597 --> 00:03:36.717 Oh, and there are also versions of these filled with other gasses which produce different colors of light, 00:03:36.717 --> 00:03:44.345 including a bit of UV which when combined with a phosphor coating on the glass means you can make pretty much any color you like. 00:03:44.345 --> 00:03:46.654 Blue and green are quite common. 00:03:46.654 --> 00:03:50.710 As a matter of fact here’s a blue one in the switch of my cheap Walmart kettle, 00:03:50.710 --> 00:03:53.290 and there’s a green one in this hot plate. 00:03:53.290 --> 00:03:57.530 These are not LEDs - they’re tiny little discharge lamps. 00:03:57.530 --> 00:04:02.047 Another fun use of this principle is in novelty lamps like this. 00:04:02.047 --> 00:04:07.033 Here we find electrodes which are shaped in novel ways to entice tourists to buy it. 00:04:07.765 --> 00:04:08.901 I mean... 00:04:08.901 --> 00:04:10.008 it worked! 00:04:10.008 --> 00:04:15.018 And in here we don’t find neon but instead something else. 00:04:15.018 --> 00:04:17.823 I’m leaning towards mostly argon but I’m not sure. 00:04:17.823 --> 00:04:21.636 Anyway, whatever it is produces a violet discharge on the turtles 00:04:21.636 --> 00:04:28.591 which conveniently also causes a phosphor on the Hawaii script to fluoresce green so we get two colors! 00:04:28.591 --> 00:04:30.656 Could have even more if we wanted. 00:04:30.656 --> 00:04:32.784 Anyway, back to the neon ones. 00:04:32.784 --> 00:04:37.769 Earlier you might have caught me saying that the glow discharge occurs on the cathode. 00:04:37.769 --> 00:04:41.198 But here, both electrodes appear lit. 00:04:41.198 --> 00:04:50.355 That’s because this is powered by AC, so which electrode is the cathode keeps flipping back and forth really fast and both appear lit. 00:04:50.355 --> 00:04:54.778 If I put a diode in series with the lamp we’ll keep one of them from glowing, 00:04:54.778 --> 00:04:58.780 and by reversing the polarity we can make the other half glow. 00:04:58.780 --> 00:05:06.425 Making cathodes in a bunch of different shapes, then stacking them together with a space in between and running high voltage DC to them 00:05:06.425 --> 00:05:09.790 is how Nixie tubes happen, by the way. 00:05:09.790 --> 00:05:13.414 How have I not done a video on Nixie tubes yet? 00:05:13.414 --> 00:05:14.754 That’s weird. 00:05:14.754 --> 00:05:20.865 Uh, anyway, as these indicator lamps age they can start to become a little unstable. 00:05:20.865 --> 00:05:24.233 If driven lightly they can last practically forever, 00:05:24.233 --> 00:05:31.560 but if they’re overdriven by a resistor that’s too low in value the electrodes can become damaged with time. 00:05:31.560 --> 00:05:33.924 Sometimes doing this is OK; 00:05:33.924 --> 00:05:42.632 for appliances which only get used occasionally, it’s not the worst thing to squeeze some extra brightness out of the lamp in exchange for longevity. 00:05:42.632 --> 00:05:45.148 You just don’t need it to last that long. 00:05:45.148 --> 00:05:50.331 But for something like the switch of a power strip which may be turned on forever, 00:05:50.331 --> 00:05:53.121 you’ll start to get problems eventually. 00:05:53.121 --> 00:05:59.968 The glow discharge will often become unstable with damaged electrodes, causing the indicator to flicker. 00:06:01.066 --> 00:06:02.703 Wait a minute… 00:06:02.703 --> 00:06:04.361 This thing flickers! 00:06:04.361 --> 00:06:06.402 It flickers a heckuva lot! 00:06:06.402 --> 00:06:12.844 And if we look at it a little more closely, we’ll see that this is effectively the same thing as a neon indicator. 00:06:12.844 --> 00:06:18.043 We’ve got two electrodes separated by a gap connected to wires leading out of the envelope. 00:06:18.043 --> 00:06:23.700 Really the only thing that makes this different from an ordinary indicator is the shape of the electrodes. 00:06:23.700 --> 00:06:26.002 They’re meant to mimic a flame, 00:06:26.002 --> 00:06:28.800 so they’re shaped like a flame. 00:06:28.800 --> 00:06:31.272 The shape of the glass itself is also evocative 00:06:31.272 --> 00:06:32.529 of a flame. 00:06:32.529 --> 00:06:39.377 And inside the base of the lamp is that current limiting resistor so we can operate this from ordinary AC power without — 00:06:40.087 --> 00:06:41.429 that. 00:06:41.429 --> 00:06:43.410 So why does it flicker? 00:06:43.410 --> 00:06:45.880 Because it was built wrong! 00:06:45.880 --> 00:06:47.517 On purpose! 00:06:47.517 --> 00:06:55.709 The electrodes in here are not trying to create a uniform glow, as a matter of fact that’s exactly what we don’t want. 00:06:55.709 --> 00:07:01.305 Now, details into how exactly this is made flickery are sketchy. 00:07:01.305 --> 00:07:04.067 I’ve found conflicting information in a few places, 00:07:04.067 --> 00:07:11.528 for instance a patent for this lamp design claims to use a barium azide coating in order to enhance the flicker effect 00:07:11.528 --> 00:07:19.655 but an earlier patent for an unrelated device claims the same substance increases discharge uniformity. 00:07:19.655 --> 00:07:21.702 That doesn’t make sense! 00:07:21.702 --> 00:07:28.610 Plus the turtles in the novelty lamp appear to have a very similar coating, yet they don’t flicker at all. 00:07:28.610 --> 00:07:34.279 A different source I found claims the sides of the electrodes that face each other inside the lamp 00:07:34.279 --> 00:07:40.621 are coated with an insulating varnish in order to force the glow discharge onto the outside surfaces. 00:07:41.030 --> 00:07:47.624 And, y'know, I suppose makes a little sense but that patent doesn’t mention the need to do that. 00:07:47.624 --> 00:07:56.810 Plus, it’s not like the electrodes on the indicators only glow on the halves that face each other so I don’t see why that's necessary. 00:07:57.241 --> 00:07:58.536 Wait a minute, 00:07:59.203 --> 00:08:02.019 Through the Magic of Buying Ten of Them, 00:08:02.019 --> 00:08:04.607 I can afford to sacrifice one of these and find out! 00:08:04.607 --> 00:08:05.660 [pop] 00:08:05.660 --> 00:08:09.675 Nope, there doesn't appear to be a difference at all between the two sides. 00:08:09.675 --> 00:08:15.566 Though interestingly, these electrodes quickly turned white once exposed to air. 00:08:15.566 --> 00:08:20.414 I don’t know what exactly I was just touching so you can bet I washed my hands pretty thoroughly after this. 00:08:20.414 --> 00:08:26.405 Oh, uh, if there was an insulating varnish on one side (which again - doubtful) 00:08:26.405 --> 00:08:30.791 the leads on my multimeter scratched through it so, you know, 00:08:30.791 --> 00:08:32.692 I don't think it's there. 00:08:32.692 --> 00:08:38.514 Given that information around these is contradictory and apparently sometimes flat-out wrong, 00:08:38.514 --> 00:08:46.429 my explanation here on how it works is more than a little bit speculative and synthesized from stuff I’ve observed and read online. 00:08:46.429 --> 00:08:52.903 If we have any subject matter experts in the audience who can confirm or challenge this please do chime in. 00:08:52.903 --> 00:08:57.010 Whatever the coating is on the electrodes, it’s not uniform. 00:08:57.333 --> 00:08:59.618 That’s clear just by looking at it. 00:08:59.618 --> 00:09:08.050 And this is probably going to change how easily a gas discharge can happen on any individual point across its surface. 00:09:08.050 --> 00:09:13.305 In fact, we can see that the pattern in the flickering often follows those imperfections 00:09:13.305 --> 00:09:16.720 so clearly those affect what regions will glow. 00:09:17.430 --> 00:09:22.890 When the lamp first strikes, then, only some parts of the electrode can light. 00:09:22.890 --> 00:09:28.648 The electrodes are not, I guess, pure enough for a neat, uniform glow. 00:09:28.648 --> 00:09:31.563 But I don’t think that’s the only thing going on. 00:09:31.563 --> 00:09:37.770 Looking back at the novelty lamp, the sea turtle-shaped electrodes are not perfectly coated either, 00:09:37.770 --> 00:09:42.766 however, they do glow uniformly and (importantly) stably. 00:09:42.766 --> 00:09:48.984 Perhaps this is a better coating than in the flame lamps but there are a few key differences here. 00:09:48.984 --> 00:09:51.510 Firstly, this ain’t neon. 00:09:51.510 --> 00:09:53.970 So the gas composition is different. 00:09:53.970 --> 00:09:58.092 But I think more importantly, this is a very large envelope 00:09:58.092 --> 00:10:03.089 and the electrodes are nowhere near as close to each other as in the flame lamps, 00:10:03.089 --> 00:10:05.385 and in fact they don’t even overlap. 00:10:05.385 --> 00:10:12.053 In the flame lamp not only do they completely overlap but there’s only a tiny gap between the electrodes, 00:10:12.053 --> 00:10:16.863 and that is probably causing instability due to Paschen’s law. 00:10:16.863 --> 00:10:18.007 Or "pass-kins?" 00:10:18.007 --> 00:10:19.429 However you pronounce that. 00:10:19.429 --> 00:10:24.145 See, once lit the area near the glowing discharge will heat up. 00:10:24.145 --> 00:10:29.316 And the electrodes are sandwiching a small quantity of gas between them. 00:10:29.316 --> 00:10:37.548 That gas is going to get warm very quickly once there’s a discharge, so it will expand and sort of jet out the sides. 00:10:37.548 --> 00:10:45.242 And thanks to science reasons, the voltage required to maintain the discharge changes based on the gas pressure. 00:10:45.242 --> 00:10:51.216 That’s going to cause the area that’s most likely to glow to move once the gas expands, 00:10:51.216 --> 00:11:01.510 but every time it moves that next area gets warm, too, so the gas in that region expands, and the spot most likely to glow gets pushed somewhere else. 00:11:01.510 --> 00:11:04.385 This just keeps happening over and over again, 00:11:04.385 --> 00:11:09.470 and while it is somewhat random it is also very clearly cyclical. 00:11:09.470 --> 00:11:14.771 And again, these patterns seem to follow the imperfections on the electrode surfaces, 00:11:14.771 --> 00:11:20.670 so while I don’t think that’s the only factor here, it probably matters a lot. 00:11:20.670 --> 00:11:26.430 I think the strongest argument for that comes from the fact that very old, worn-out indicators 00:11:26.430 --> 00:11:28.667 exhibit this same phenomenon. 00:11:28.667 --> 00:11:32.935 And luckily, I have one right here! 00:11:32.935 --> 00:11:37.350 After looking over all my power strips I finally found one with a flickery lamp. 00:11:37.802 --> 00:11:41.189 I had to destroy the switch to get a good look at it but hey, 00:11:41.189 --> 00:11:42.814 For Science! 00:11:42.814 --> 00:11:49.720 Speaking of science, this lamp’s flickering is so borderline that it only flickers in the dark. 00:11:49.720 --> 00:11:55.471 Believe it or not, photons entering the glass envelope can help the neon gas to ionize, 00:11:55.471 --> 00:12:00.610 and even just a tiny bit of light from my phone flashlight gets this to stop flickering. 00:12:00.610 --> 00:12:02.070 Ain’t that neat? 00:12:02.070 --> 00:12:06.800 Anyway, if we look at this thing up close we can see that it’s darkened. 00:12:06.800 --> 00:12:10.938 This lamp was almost certainly being over-driven at least a tad, 00:12:10.938 --> 00:12:18.000 though to be fair it’s probably been on for well over a decade at this point so it still did quite well. 00:12:18.000 --> 00:12:23.016 Now, though, one of the electrodes can’t quite maintain an even discharge 00:12:23.016 --> 00:12:31.339 and sure enough, at least with the lights out, the spots that glow move around in a similar fashion to the flicker flame lamp. 00:12:31.339 --> 00:12:36.194 With the darkened glass a sure sign of material loss from the electrodes, 00:12:36.194 --> 00:12:42.149 it’s likely that the instability is being caused by damaged and thus imperfect electrodes - 00:12:42.149 --> 00:12:47.300 which seems to give more credence to the uneven coating theory for the flame lamp. 00:12:47.515 --> 00:12:51.567 However, it might not simply be the electrodes causing the flicker - 00:12:51.567 --> 00:12:57.448 the original patent for the flame lamps claims that the gas composition affects how the lamp flickers, 00:12:57.448 --> 00:13:01.920 and that could also explain what’s going on with the flickery indicator. 00:13:01.920 --> 00:13:08.079 Perhaps after years of service the gas mixture has become impure or damaged somehow. 00:13:08.079 --> 00:13:13.978 But regardless we know for sure that neon lamps are not *meant* to do this, 00:13:13.978 --> 00:13:16.760 but over time they often start. 00:13:16.760 --> 00:13:24.910 So in a sense, the flicker flame lamp is an exercise in perfecting the imperfect lamp. 00:13:24.910 --> 00:13:28.650 Through some combination of a non-uniform electrode coating, 00:13:28.650 --> 00:13:30.877 the shape and spacing of the electrodes, 00:13:30.877 --> 00:13:33.202 the size of the envelope relative to them, 00:13:33.202 --> 00:13:36.106 and even the specific composition of the gas mixture, 00:13:36.106 --> 00:13:41.027 this thing is deliberately bad at being a neon lamp. 00:13:41.027 --> 00:13:43.860 That just happens to look kinda like a flame, 00:13:43.860 --> 00:13:47.398 so make the electrodes flame-shaped and you’ve done it! 00:13:47.398 --> 00:13:50.762 But I had one more theory which I thought to test. 00:13:50.762 --> 00:13:56.555 What if the current-limiting resistor of this lamp is deliberately too restrictive? 00:13:56.555 --> 00:14:02.992 Maybe it simply doesn’t have enough current flowing through it to cover the entire electrodes with that glowy goodness 00:14:02.992 --> 00:14:07.196 and so the spots moving around is just a side-effect of that. 00:14:07.196 --> 00:14:13.977 Well, I rigged up these indicators in this definitely OSHA-approved testing device so we can find out. 00:14:13.977 --> 00:14:18.144 You can see that they decrease in brightness from left to right. 00:14:18.144 --> 00:14:25.709 The first lamp is powered through a 33 kiloohm resistor, which is too low of a value so it’s being over-driven. 00:14:25.709 --> 00:14:30.502 It glows quite brightly, which is nice, but it would likely become damaged over time 00:14:30.502 --> 00:14:33.579 and darken like we saw with the power strip. 00:14:33.579 --> 00:14:40.986 Generally, you want to drive this size of lamp with about a 100 kiloohm resistor for 120V AC power. 00:14:40.986 --> 00:14:43.831 The next lamp is driven at that correct value 00:14:43.831 --> 00:14:48.348 (or close, anyway, I don’t have a resistor of that exact value on-hand). 00:14:48.348 --> 00:14:57.160 But the next two lamps have very high-value resistors, with the one on the right having about 500 kiloohms of resistance in total. 00:14:57.160 --> 00:15:04.579 Although it is glowing very dimly, it’s not flickering and the glow around the electrodes is mostly complete. 00:15:04.579 --> 00:15:09.759 So it appears increasing the resistance value mainly affects the brightness of the discharge, 00:15:09.759 --> 00:15:12.209 and not how complete it is. 00:15:12.209 --> 00:15:19.160 In fairness, these are really tiny indicators with a small fraction the electrode surface compared to the flame lamp, 00:15:19.160 --> 00:15:22.314 so this may not be conclusive. 00:15:22.314 --> 00:15:25.497 Honestly, this entire video isn’t conclusive. 00:15:25.497 --> 00:15:32.200 There’s things about these lamps that seem to make perfect sense yet other things don’t make any sense at all. 00:15:32.200 --> 00:15:39.410 But given how they’re built to behave essentially exactly like the failure mode of a neon indicator, 00:15:39.410 --> 00:15:46.070 I think the most likely explanation is simply that they’re deliberately made kinda wrong. 00:15:46.543 --> 00:15:51.920 It’s probably not the best idea to put two flat electrodes right next to each other. 00:15:51.920 --> 00:15:57.420 The gas mixture might be just a bit off to ensure the glow discharge is uneven. 00:15:57.420 --> 00:16:02.075 And the coating on the electrodes is certainly less-than-perfect. 00:16:02.075 --> 00:16:06.111 But I do know there’s one thing we can learn from this: 00:16:06.111 --> 00:16:12.190 Sometimes things that are a little bit broken have a beauty all their own. 00:16:12.190 --> 00:16:13.849 Thanks for watching. 00:16:14.753 --> 00:16:17.265 ♫ inconclusively smooth jazz ♫ 00:16:18.901 --> 00:16:26.006 Hey, so I had forgotten that I have this other novelty bulb where the glowing electrodes are in front of one another, 00:16:26.006 --> 00:16:31.959 but here there's a substantially larger gap than we find in the flicker bulbs. 00:16:31.959 --> 00:16:36.630 That could just be because whatever factory made these didn’t have great tolerances, 00:16:36.630 --> 00:16:41.860 but the electrodes going through the stem are deliberately quite well spaced. 00:16:41.860 --> 00:16:50.981 Perhaps if these were too close to each other we’d get that flickering - and maybe that happens because the glowing areas can meet? 00:16:51.885 --> 00:16:53.776 I dunno, but I thought I’d bring this up. 00:16:55.046 --> 00:16:56.819 Aloha! 00:16:56.819 --> 00:16:58.856 Sometimes doing this is OK. 00:16:58.856 --> 00:17:01.373 For appliances which only get used occazzzzzzioonally 00:17:01.373 --> 00:17:03.251 what the heck was that? 00:17:03.681 --> 00:17:05.061 [coughs] 00:17:05.061 --> 00:17:07.605 So, this might no be concplusive. 00:17:07.605 --> 00:17:09.009 Conclusive! 00:17:09.009 --> 00:17:15.120 These are really tiny indicators with a small fraction the electrode surface compared to the flame ramp... 00:17:17.466 --> 00:17:19.170 Did I say "frame ramp?" 00:17:21.602 --> 00:17:23.355 I did, didn't I? 00:17:24.302 --> 00:17:27.315 So, of course *after* I made the video I find out Big Clive covered these a while back. 00:17:27.315 --> 00:17:29.691 I mean, that seems so obvious in hindsight and why didn't I look for that? 00:17:29.691 --> 00:17:33.747 Although we're mostly on the same page - he also thought there was an insulating coating on one side but destroying the lamp put that into question. 00:17:33.747 --> 00:17:38.781 However, a higher resistance did make the glowing spots smaller so that probably has something to do with it. 00:17:38.781 --> 00:17:40.426 Anyway, Happy Decemberween!