# Induction lamps: fluorescent lighting's final form

## Метаданные

- **Канал:** Technology Connections
- **YouTube:** https://www.youtube.com/watch?v=SaKKzZRrPIg

## Содержание

### [0:00](https://www.youtube.com/watch?v=SaKKzZRrPIg) Segment 1 (00:00 - 05:00)

You’ve no doubt seen plenty of fluorescent lights, but have you ever seen one like this? It may not seem all that strange at first glance, after all there have been plenty of weird fluorescent lamp designs over the years. But this one’s missing a pretty crucial part: the electrical connections. Your typical fluorescent tube has little wires attached to pins which stick out of the glass at each end, and they connect to electrodes inside the tube. Those wires are then connected to a ballast and starter which together place a voltage on the electrodes to pump electricity through the lamp which makes it glow. But this… this is just a glass tube. How’s that supposed to work? Well, with these. The tube is designed to be held in place by a pair of… let’s call them electromagnets. That’s not exactly the right word which will become clear in a moment but each one of them is formed by wrapping a wire around a ferrite core several times. With the help of electronics inside this driver unit, a high-frequency alternating current is sent through the wires which then causes a magnetic field to repeatedly form, collapse, and form again inside the ferrite cores. Now, if you know a little bit about power transformers, then if we take the tube out of the picture for a moment and look at a core by itself, then what we’re looking at appears to be a toroidal power transformer that somebody forgot to finish. See, usually we’d have at least two coils of wire wrapped around one of these cores. When we run alternating current through the first coil of wire, called the primary winding, it will induce a rapidly changing magnetic field through the core which in turn induces a voltage on the other coil of wire known as the secondary winding. When that secondary winding is part of a circuit, current will flow and the upshot is that energy is transferred from the primary winding, through the core, and into the secondary winding. We often do this because the secondary winding will produce a different voltage when it has a different number of turns around the core compared to the primary. But here, there isn’t a secondary winding! It just doesn’t exist. That is, until I put the tube back. This lamp isn’t a literal wire, of course, but once I have it enclosed by the core it does indeed form a loop in the exact same way a single turn secondary winding would. It looks a little strange because it’s going way over there before looping back to the core, but that’s what it is. Now, because this lamp is so large, we need a second core positioned on the other end of the tube, but they’re wired together and so act as one. When the alternating current being sent through the wires by this driver unit induces magnetic fields in the cores, those fields then induce a voltage inside the fluorescent tube. And, because it’s formed in a loop, that results in current flow which causes the lamp to light. For that reason, this is known as an induction lamp. The actual light it produces is the result of the exact same principle as an ordinary fluorescent tube: a low-pressure arc discharge is produced by passing current through an ionized mixture of mercury vapor and a noble gas such as argon, and as electrons collide with the mercury atoms the mercury emits ultraviolet light which is then converted to visible light by the phosphors coating the glass. But rather than use electrodes reaching inside the tube in order to pass current through the gas, current flow is generated externally via electromagnetic induction. Why would we want that? Well first, it’s pretty cool! But also this particular lamp has a power rating of 200W and a light output in the neighborhood of 16,000 lumens. That is unusually powerful and bright for a fluorescent lamp, especially given its relatively small size. But brightness, power, and coolness weren’t really the main reasons these appeared - operating life was. Getting rid of the electrodes would mean weak point of the fluorescent lamp. In a conventional tube, the electrodes slowly wear out in a process known as sputtering. This happens the most at lamp start-up, but sputtering still happens throughout continued operation. Over time this causes the ends of the tube to darken - the dark spots are the material which used to be the electrodes, but as it sputters off it becomes deposited on the glass. That reduces light output slightly but the bigger problem is that as they wear away the electrodes lose their ability to emit electrons.

### [5:00](https://www.youtube.com/watch?v=SaKKzZRrPIg&t=300s) Segment 2 (05:00 - 10:00)

That makes it more difficult for the ballast and starter to initiate the arc discharge across the tube, and eventually they will wear to the point the tube is no longer able to start and the lamp has failed. As the technology developed we got pretty good at making electrodes which could last a good while, and fluorescent lamps would typically have a rated life of 10,000 hours. But if we could do without the electrodes, the lamp could in theory last forever. And, well, that’s the point of this technology. There’s nothing to the lamp but a phosphor-coated tube filled with some argon gas and a pellet of mercury amalgam chilling in this little glass appendix. It has no electrodes to wear out and this means the tube itself has a theoretically unlimited lifespan. Nothing truly lasts forever, of course, and due to phosphor wear and mercury absorption these will lose light output over time, but these lamps have a rated lifespan of 100,000 hours - a tenfold increase over conventional tubes. Operating dusk-to-dawn, that would mean this should last nearly 23 years before requiring replacement. Now, obviously, that’s a pretty phenomenal lifespan and so it might be surprising how uncommon this technology is. I mean, there’s not a whole lot to this thing and it operates on the fundamentals of electromagnetism that have been understood since the 19th century. Sure enough, electrodeless lamps which operated via induction had been demonstrated many, many times in the past. Nikola Tesla was playing with it because of course he was. But a practical and effective application of the technology was surprisingly elusive. It wasn’t until 1967 that John Anderson filed a patent for a lamp like this one, and then it took until 1990 for it to actually be commercialized into a finished product. What took so long? Well, despite how simple the overall idea is on the surface, it turns out there were a LOT of different problems that needed solving before this could actually become a reality. For one thing, inducing current flow in a loop of gas, even if it is conductive, is a lot harder than a solid metal wire. To sustain the arc discharge in the tube, the magnetic field generated by the cores has to switch back and forth very a lot, so the frequency at which this thing operates needed to be very high. You’ll notice the patent shows the driver as “converter to radio frequency” and Anderson lists a preferred frequency range of 100 to 500 kilocycles. That’s almost reaching the AM radio broadcast band. By 1967 it was relatively trivial to produce high-powered oscillators which ran at that frequency, but that wasn’t the main issue. Because of the required high frequency, wrapping wire around an ordinary iron core as found in most power transformers wasn’t feasible. At those frequencies, the eddy currents which form inside the cores would cause it to get extremely hot. That’s the mechanism by which induction cooktops work. But the problem is, that would both limit the efficiency of the power coupling and how much power you could feasibly transfer into the tube before the cores started to melt. This meant ferrite compounds, which are extremely magnetically permeable but not electrically conductive and therefore don’t form eddy currents inside, were required for high-frequency power coupling. They weren’t invented until the 1930’s and even though they were fairly common by the time Anderson filed his patent, ferrite was still an expensive material. In fact, that’s the main reason this lamp gets narrower where the cores attach: making the tube diameter smaller requires less material in the cores. But here’s where the story gets a little weird. Although this patent is nearly a spot-on description of this lamp both in form and function, the first commercial products using the technology were very different from this. The QL line from Philips was the first commercial induction lamp and they looked… pretty much like an ordinary incandescent light bulb. They weren’t - they were in fact fluorescent lamps which worked by induction similar to this. But compared to this tube and its external cores, they were constructed inside-out. I don’t have one to show you, but thanks to a donation to the channel from Jeff, a long-time Patreon member

### [10:00](https://www.youtube.com/watch?v=SaKKzZRrPIg&t=600s) Segment 3 (10:00 - 15:00)

I do have this variation which was manufactured by GE. This Genura lamp was released in 1994 and it is designed to replace a reflector flood bulb but it works on the same basic principle to the original Philips QL lamps. Here, the discharge tube isn’t so much a tube as it is... a vessel. And a single coil, once again in the form of a wire wrapped around a ferrite core, protrudes from the base of the lamp into a hollowed-out section of the glass vessel. This evacuation stem holds the pellet of mercury amalgam which is required to produce ultraviolet light and it rests inside the ferrite core when assembled. An oscillator circuit in the base of the bulb sends high-frequency pulses through the wire wrapped around the ferrite core which in turn creates a rapidly fluctuating magnetic field around it. That field reaches well inside the discharge vessel. As the field changes in intensity, current flows through the gas inside, and our desired mercury discharge occurs. In this lamp design, the single coil protruding into the vessel produces a magnetic field in a toroidal shape around its perimeter. That’s why the QL lamps from Philips look a lot like an ordinary light bulb. It’s not a linear discharge running through a tube but more of a fuzzy donut of ultraviolet light surrounding the induction coil, which a bulb-shape happens to encapsulate quite well. The phosphors on the glass convert the UV to visible light and the result is a compact yet very bright light source with an extremely long life. The original QL lamps ran at 85 watts and had a rated life of 100,000 hours. Now, this GE lamp isn’t quite so optimistic but that’s likely down to the fact that its drive electronics are built into the lamp itself. This was sold as a drop-in replacement light bulb, so everything had to get crammed into here. And since these are generally operated with the base facing up, heat from the discharge would rise right into the electronic components which shortens their operating life, especially capacitors. The QL line from Philips used an external driver muck like this thing does and the lamp itself was little more than a glass orb on a stick. Still, it wasn't quite as simple as it looks. Because the induction coil was shoved up the middle of the lamp, it gets pretty hot - particularly in the case of the QL lamps which operated at 85 watts. This was a problem because as the ferrite material increases in temperature, its magnetic permeability decreases which limited the strength of the magnetic fields it can produce and thus the amount of power it can send into the gas discharge. That reduces both brightness and energy efficiency. Philips solved this by incorporating heat-conducting material into the design of the ferrite core’s stem, which by the way they referred to as the antenna. From what I can tell looking at patents, GE didn’t solve this at all, but since the lamp only operates at 23 watts the induction coil probably never got hot enough to matter. But another issue with the single-coil designs is that it require an even higher frequency to function: around 2. 5 megahertz. And because these are operating at radio frequencies and pumping dozens of watts or more into the discharge vessel, well radio frequency interference was a significant problem. To try and help solve it, the glass is coated with a transparent yet conductive material, such as indium tin oxide, which absorbs radio frequency energy and keeps it from escaping. And in this lamp, the underside of the discharge vessel also has a metal shield which was bonded to this copper ribbon that allows the absorbed energy to return to the circuit ground. But, it’s clear that GE at least wasn’t making the strongest promises about its effectiveness. They warn specifically against using these on boats as they could disrupt maritime communications. And, uh, given that they provided a phone number for interference complaints... well I think it’s safe to say this was very much a beta test. Now, it might seem a little odd that the basic idea for this later design was patented in 1967 yet induction lighting didn’t get commercialized until after Philips and GE released their single-coil designs. But it might make more sense if you consider that these external-coil lamps... really aren't that different from conventional fluorescent tubes. In fact, many lower-power versions of this design incorporated circular tubes which look a whole awful lot like a standard circular fluorescent lamp.

### [15:00](https://www.youtube.com/watch?v=SaKKzZRrPIg&t=900s) Segment 4 (15:00 - 20:00)

They did offer higher light outputs, but the form was still very familiar. The fact that neither the QL lamp nor this Genura lamp look anything like a fluorescent tube is probably why the technology was first commercialized like this. Consider the GE bulb: the resulting discharge inside this vessel produces a compact but powerful ring of UV light just under a relatively flat piece of glass. This meant fluorescent technology could not only mimic the appearance of an incandescent reflector much more faithfully than sticking a coiled tube inside a fake bulb but it was also much more effective. This flavor of induction technology was able to create much more powerful yet also compact light sources than a linear fluorescent tube while still offering similar energy-efficiency. That was compelling enough on its own to pursue not only for aesthetic purposes but because it allowed the use of fixtures with optical systems that produced much more directional light. To a point at least - I’ll touch on that more in a bit. Induction technology also solved one of fluorescent lighting’s little annoyances: a slow warm-up. High-efficiency fluorescent tubes in particular only produce a fraction of their light output at first due to lower vapor pressure when cold and thus they can take several minutes to warm up. This lamp also starts with a lower vapor pressure and thus reduced light output, but the warmup time is just a few seconds. And the large external-coil lamp is similarly quick - it’s at full brightness in only 10 seconds. You can actually observe the mercury discharge migrating away from the amalgam pellet as the tube heats up and the pressure builds. But… well, that was at room temperature, and these lamps were sold to go outdoors. In fact street lighting is one of the most likely places you’re going to find these in the wild. And I live where the outside air sometimes gets colder than a freezer, so I put this whole thing inside a freezer overnight to see how quickly it would warm up from truly cold. The lamp had no issues starting but, like most fluorescent lights in this temperature, the mercury was barely contributing anything and we could see the purple glow of the argon starter gas. Interestingly it initially did the same thing where the light output seems to grow from the mercury pellet, but before long that bright spot disappeared and it all averaged out to a very dull grey. It took about 10 minutes to attain full brightness so it looks like cold-weather performance is only improved slightly. But at least it had no trouble starting! Speaking of starting, well now it’s time to come clean here - I don’t really understand the mechanisms by which these lamps start. See, the gas mixture in these tubes isn’t electrically conductive until it’s ionized. The electrodes in a traditional fluorescent tube emit electrons when heated and/or when a sufficiently high voltage is placed across them, and those electrons collide with the gas molecules inside and ionize them which causes dielectric breakdown. But with these, all we got is magnets. The 1967 patent mentions the use of electrodes in contact with the glass that, when a sufficiently high voltage is placed across them, will ionize enough gas inside the tube to initiate dielectric breakdown. But this device has no such electrodes. A later patent filed in 1995 describes this device nearly perfectly, and it even has a diagram which more clearly illustrates the rather confusing way the wires are wrapped around the two ferrite cores. But it also mentions a conductive strip in contact with the glass specifically for the purpose of creating ionization points for starting the lamp. But again, this lamp doesn’t have those. All the information I could find on how this lamp gets started was quite handwavey, but my best understanding is that the driver circuitry initially sends a much higher than usual voltage through the wires which produces a strong enough electric field that free electrons in the gas mixture are sort of thrown about, causing some ionization. Basically it’s the same way that fluorescent tubes can start glowing when you hold them near a Tesla coil. And once some of that gas is ionized the discharge across the tube can be completed after which the driver can switch to its normal operating voltage.

### [20:00](https://www.youtube.com/watch?v=SaKKzZRrPIg&t=1200s) Segment 5 (20:00 - 25:00)

That may also further explain why the internal core lamps were commercialized first. Based on readings of patents, it was initially thought that a secondary winding would be needed on the ferrite core to produce a high voltage capable of initiating a glow discharge, but that feature seems to have gone away once it was decided to use the very high frequency found here. This patent filed in 1982 seems to suggest that simply by inducing the magnetic field at a frequency of 3 megahertz, a sufficiently strong electric field is produced inside the vessel which can ionize the gas. But I still haven’t gotten a truly satisfactory answer, here, and would welcome one in the comments! It might have something to do with the metal mesh found here, and a closer look up the middle of this Genura’s discharge vessel shows a piece of wire at the top of the evacuation stem. But truthfully, I gave up trying to find an answer. So, with their extremely long operating life and newfound applications for energy-efficient fluorescent lighting, it might seem strange that this technology is so obscure. I mean, today we have LEDs and this technology is obsolete but knowing that back in 1990 we had a lighting technology with a lifespan that essentially matches the best LEDs we have today, you would think it would have taken the world by storm. Yet it didn’t. Why? Was there another one of those light bulb conspiracies afoot? Yeah, no. See, here’s the thing - the tech is really cool and opens up some new possibilities, but not that many and it also came with some significant downsides. For one, cost. The QL line of lighting from Philips was very much a specialized commercial product and I’ve been having a heck of a time finding original pricing, but you can bet it was quite high. Not only had they created a new kind of fluorescent lamp, but they had to manufacture specialized electronics and the power couplers. The lighting system really only made sense in applications where relamping was an operational headache or its operating method offered improved safety - apparently they saw success in the oil and gas industry due to regulations concerning explosion-proof lighting. GE’s take on the technology was sort-of consumer focused, and the Genura bulb retailed for about $30 in 1994, equivalent to $65 today. But that’s a pricey light bulb especially considering its expected life isn’t all that great. It’s certainly much better than the 1 or 2,000 hours you could expect out of an incandescent flood bulb, but it was barely an improvement over existing fluorescent technology. You needed to really, really want fluorescent lighting in this specific form-factor, and I find it interesting that GE really didn’t bother to explain what’s so cool about this technology on the box. They're just calling this an electronic compact fluorescent lamp. Which is true but certainly underselling it. But the biggest issue, and what I think is likely the main reason this technology never took the world by storm, is that it’s still fundamentally fluorescent lighting. Yes it was longer-lived, but it still had kinda meh light quality. And, because fluorescent lights are very much not point-sources of light, it was hard to direct the light these emitted. This 200W lamp was originally housed in this flood fixture, and while you could point this in a direction, it still casts a very wide beam of light. I don’t even think you could describe it as a beam. The QL lamps were more compact, but they still emitted diffused light which is difficult to control. Take street lighting as an example. I've encountered a few induction street lamps in the wild, but they all have about as much beam control as a circline fluorescent light stuck in the middle of the kitchen ceiling. Because that’s pretty much what they are. And the inverse square law means that you can’t place the lamps too high above the street or they just won’t offer effective light output. This ultimately means you need many more of them placed closer together. High-intensity discharge lamps like high pressure sodium and metal halide produce all their light in a small arc tube which allows fixtures with optical systems to focus the light they produce into a narrow beam, permitting the use of fewer fixtures by placing them higher in the air. And actually, the Genura lamp demonstrates this downside quite well. While it mimics the basic appearance of an incandescent BR flood very faithfully...

### [25:00](https://www.youtube.com/watch?v=SaKKzZRrPIg&t=1500s) Segment 6 (25:00 - 30:00)

at least from the face of it, there’s no directionality to the light it produces. Even a frosted lamp like this produces a somewhat narrow beam of light thanks to fact that the filament is placed far down the reflector and that is very important to both the character and effectiveness of the light it produces, especially when recessed in ceiling can fixtures. This lamp simply cannot recreate that directional light, and I would imagine that presented problems from both an aesthetic perspective and a functional one, too. It may be the same number of lumens, but when fired in all directions and not mostly downward, the room might be underlit compared to before. And let’s not forget that conventional fluorescent technology saw lots of innovation, too. If any of the lighting nerds out there have been wondering why I never touted the flicker-free light output created by the high-frequency drivers in these things, that’s because electronic ballasts for traditional tubes did the same thing. They didn’t operate anywhere near these frequencies, but they were fast enough for the persistence of the phosphors to produce truly continuous light output. That also increased the energy-efficiency of fluorescent lighting. As a matter of fact, a bog-standard T8 tube is more energy-efficient than either of these induction lamps. And then of course there’s the CFL. Those, too, got electronic ballasts and all the benefits thereof and they were available in much more modest power outputs than induction lamps. Sure, the Genura lamp is a more elegant solution than sticking a coil in a fake bulb, but… is its elegance worth the extra cost? The market clearly decided no it wasn’t. I mean, this lamp was sold long before the old curly-q CFL went mainstream. Once we figured out how to make those cheaply, small induction lamps like this just hardly made any sense at all. And when ordinary fluorescent lights already had a 10,000 hour lifespan, well frankly you just had to be extremely committed to the idea of a lamp which could last 20 years without being touched. That is, if it makes it that long. The weak point in the induction lamp isn't the discharge tube, it's the electronics inside this box. And while it certainly is possible to produce electronic circuitry which lasts 100,000 hours, well let’s just say there are never any guarantees. Honestly props to GE for being so realistic with these lamps. The QL line from Philips probably had the best quality drivers of any commercial system and since they were separated from the lamp they didn’t have to deal with heat. But had this tech gone mainstream, we’d no doubt have the same race to the bottom that plagues so many things. But of course now this is all moot. The LED has trounced every lighting technology that we’ve ever made on energy efficiency, quality of light, and flexibility. We still have the same race-to-the-bottom nonsense going on, unfortunately, and it’s not like there haven’t been some high-profile issues with LEDs (such as purple streetlights) but in general lighting is now just solved. We’re there. Aside from some cheap builder-grade light bulbs I still have kicking around, I personally haven’t needed to replace a light bulb in several years. That’s no doubt helped by the fact that I generally only buy high-quality bulbs and I use dimmers nearly everywhere so rarely are any of my lights at full-brightness, but yeah. I’m certainly not pining for the days of old. But luckily our path to get here was filled with all sorts of wacky nonsense like this. ♫ electromagnetically smooth jazz ♫ And because it’s formed in a loop, that induced voltage results in current flow which causes the lamp to light. Except it didn’t. Probably because… you need to be secured in place. Hopefully I didn’t just kill ya, that would be very problematic. For that reason, this is known as an induction lamp. And I cannot be sitting here with this in front of my eyes. …chilling in this little glass appendage. Is that the right one? Nope. I pointed at the wrong one. I mean, there’s not a not… not a… dada da! And in this lamp, [glass scraping sounds] the underside… that I’m sure sounded horrible. …have been demonstrated many times over the past. In the… in… ugh. One would have to call this technology... totally tubular, amirite? That one hurt to the core. Both of them, actually. Hey did you know that Michael used to be a taxi driver?

### [30:00](https://www.youtube.com/watch?v=SaKKzZRrPIg&t=1800s) Segment 7 (30:00 - 30:00)

He switched to a career in science since he only got one fare a day.

---
*Источник: https://ekstraktznaniy.ru/video/23399*