27

Why are the Edison screws allowed to be used? They seem unsafe. Why are they designed that way?

They require a complete grounding system to ground devices for the rare situation where the hot wires somehow touches the metal case of the device

While they let you use a lamp fixture which has an exposed hot contact when you remove the lamp, and nothing is there to protect your finger from touching it.

I would expect this socket to have a different design which will cover the contact from accidental touch.

enter image description here

Sumurai8
  • 103
  • 3
Ronen Festinger
  • 435
  • 1
  • 4
  • 6
  • Use bayonet sockets instead, which isolate both terminals. https://en.wikipedia.org/wiki/Bayonet_mount "from the 1870s onwards". They are also a helluva lot faster to change. –  Jun 21 '16 at 22:22
  • 5
    Forget Edison bases, what about the standard Nema 1-15 or 5-15 outlets? Completely dangerous. – Passerby Jun 21 '16 at 22:31
  • 6
    @Passerby Sounds like the problem is that all electrical systems in US are silly dangerous. The split-delta windings is silly dangerous too... – Aron Jun 22 '16 at 01:59
  • 3
    In my experience they're a lot easier to change than bayonet, and the fitted bulb doesn't end up wonky / off-centre. – OrangeDog Jun 22 '16 at 09:04
  • Well, we have 110V for one, which isn't so lethal as what you've shown. Maybe that's why England switched to bayonet mounts? IAC we're taught to never stick your finger in a light socket at an early age, and to turn off the lamp before changing a bulb. – JDługosz Jun 22 '16 at 12:43
  • 1
    @JDługosz, turning off the lamp won't save you from getting poked if the cord is mis-wired so that the center button is neutral and the wall of the socket is hot. (Don't ask me how I know!) – Solomon Slow Jun 22 '16 at 13:21
  • 2
    Hmm, maybe I need to invent "Edison Gloves", to protect against glass breaking, shocks, and heat. – JDługosz Jun 22 '16 at 13:24
  • 2
    First of all, sticking a finger all they way in is hardly an accidental situation. Secondly, those are not your wall sockets, light fixtures are expected to be always populated with a bulb. – Agent_L Jun 22 '16 at 14:11
  • 29
    Why are they designed that way? They were designed earlier than 1910 near the birth of the electric light bulb. They had bigger things to worry about in those days than sticking your finger in the socket accidently - things like getting _electricity to the home_ so they didn't have to use _gas_ or _kerosene_ lighting with _open flames_ and the possibility of _accumulating gas_ which could _explode_ when you tried to light them.. – davidbak Jun 22 '16 at 17:46
  • 1
    @jameslarge if your illumination switch does not cut off both wires then you have a serious safety issue with your wiring. – dbanet Jun 22 '16 at 22:59
  • 5
    @dbanet -- LOLwut? the neutral in 99.9999% of US wiring is continuous all the way back to the service entrance -- and we somehow aren't shocking ourselves into oblivion over here! – ThreePhaseEel Jun 23 '16 at 00:26
  • 3
    The Edison base may have some deficiencies, but, having recently had to fiddle with both standard double-ended fluorescent fixtures and the flaky double-ended halogen sockets, I much prefer the Edison for ease of use and practical safety. If the Edison is not wired backwards then you'd have to really try to get shocked from it, and it's very easy to screw in a new lamp (and be confident it's in properly), compared to inserting lamps into other designs. – Hot Licks Jun 23 '16 at 03:23
  • 1
    @dbanet Only if your electrical provider is using a Split-Delta winding. – Aron Jun 23 '16 at 05:59
  • @ThreePhaseEel really? That's stupid. What if someone swaps neutral and phase wires, basically anywhere? – dbanet Jun 23 '16 at 12:46
  • 1
    @dbanet, I don't recall ever seeing a portable lamp with a dual-pole switch. (But then I'm a U.S.A'n, and the whole theme of this discussion seems to be the lax safety standards in the U.S.A.) – Solomon Slow Jun 23 '16 at 13:20
  • 2
    @JDługosz When did "England switch to bayonet mounts"? Putting it that way suggests that Edison screw was earlier. Yet Cassells "New Technical Educator" (UK, published in 1897) shows a bayonet socket, (it's not clear if it's to the current standard) calling it an improvement over bare platinum wires emerging from the glass - with no mention of the Edison screw. Wikipedia suggests the Edison screw "licensed in 1909" and "most common in the US" by 1908 but doest't place a date on it. –  Jun 23 '16 at 16:45
  • @Brian-Drummond Reading about bayonet mounts, it still doesn't prevent an accidental touching of the exposed contacts at the base of the mount. – Ronen Festinger Jun 23 '16 at 18:06
  • @RonenFestinger indeed you can push your finger into the socket. It would be difficult to do by accident, but not impossible. The sleeve however is not part of the circuit so does not become live even if mis-wired. –  Jun 23 '16 at 18:14
  • 3
    @dbanet -- 'tis why we have polarized outlets (and smack people with a copy of our electrical codes whenever they reverse neutral and hot) – ThreePhaseEel Jun 23 '16 at 22:15
  • @JDługosz it's not like England doesn't have screw mounts, they're just less common, especially for household ceiling fixtures. – OrangeDog Jun 24 '16 at 09:56
  • The pedantic answer to "why are they allowed to be used" is because they have not been prohibited. They're ubiquitous and cheap. – Adam Lawrence Jun 24 '16 at 15:28
  • 1
    Oh come on, a +25 score question with valid, research based answers, closed? – Passerby Jun 25 '16 at 03:34

5 Answers5

59

The Edison lamp base design dates predates all twentieth-century safety regulations (because it predates the twentieth century altogether). Having light bulbs screw in and out is not great in environments where vibration is a problem, and replacing a bulb which has just burned out while in use may require use of a cloth to hold it, but it is generally advantageous to have a lamp which can be inserted or removed by handling the bulb rather than having to mess with the holder. Making a metal thread assembly that can mate reliably with a metal socket is easier than trying to mold threads into glass, and having one contact in the center of the socket is easier than trying to have two concentric contacts.

While I doubt the Edison base would be approved by any safety agency if it were being introduced as an entirely new product, it has been effectively "grandfathered in" because it has been used for a long time, people are familiar with it. A "safer" design that people aren't familiar with might lead to more accidents than the century-old design which, for all its imperfections, is well understood.

supercat
  • 45,939
  • 2
  • 84
  • 143
  • 23
    gotta love "(because it predates the twentieth century altogether)" !! – placeholder Jun 21 '16 at 22:58
  • 6
    Not only is is "well understood," but everyone has lots of practice with it. Plus, if the fact that this construct isn't known for being dangerous in daily use is any indication, it would seem that it's been refined into a reliable design over the years (something like how lithium batteries have been made fairly safe despite being [extraordinary dangerous](http://electronics.stackexchange.com/a/230164/114682) by nature). – jpmc26 Jun 22 '16 at 08:49
47

Here's your opportunity. The market's looking for that right now so build a better mousetrap.

The USDOE and California CEC want to murder the Edison base to finally stop people from using incandescent bulbs, and enable fixture designs that don't have to worry so much about dissipating heat. They mandated GU24 in 2008, which solves some of your concerns. Take a look at how that's going 8 years later. LOL.

There are several flaws in the GU24 that you should address in your new design.

  • Ease of installing "blind" when you just can't see the socket or it's deep in a recess.
  • Equipment Grounding Conductor.
  • 3-way lamp support.
  • Or since dinosaurs called and want their dual-filament bulbs back... how about a standard for a signal pin and protocol to command the bulb to "dim". In track lighting, the signal line could be bussed to each outlet and controlled by a single dimmer.
  • Multi-voltage, either standardize that all bulbs must be multi-voltage, or have different keying for 120V, 220-240V and 277V.

Good luck!

  • 4
    Sorry for stupid comment, but where is 277 V in common use? – AndrejaKo Jun 22 '16 at 13:00
  • 21
    @AndrejaKo 277V is single phase of 480V 3-phase. In USA factory and workshop equipment often runs off 3-phase 480V, so 277V is the lowest voltage you can easily get in such system. Edison-screw 277V bulbs are quite easy to get over the internet. – Agent_L Jun 22 '16 at 14:06
  • 3
    Given that LEDs in well design fittings, last a long as most fittings, I expect we will move away from replaceable bulbs. – Ian Ringrose Jun 22 '16 at 18:02
  • Traditionally incandescent light bulbs are already multi-voltage. They just aren't as bright when operated with lower voltage. – John Dvorak Jun 22 '16 at 23:21
  • A 1-extra-pin *de facto* standard dimming protocol already exists -- it's called switched/dimmed hot and is already widely used with commercial/institutional dimmable fluorescent ballasts. As the name says -- one wire carries a constant 120/...V hot when the light is ON while the other carries a triac-dimmed output. – ThreePhaseEel Jun 23 '16 at 00:28
  • 1
    @ThreePhaseEel Well, that's half of it then... now we just need a plug. Unrelated, there's now a requirement to put a connector between fluorescent ballasts and their line-voltage supply source. There is no standard as to the shape. Some are 3-pin, and now I know why. – Harper - Reinstate Monica Jun 23 '16 at 00:39
  • 1
    @AndrejaKo Yes, it's one "wye" leg of 480V 3-phase. It's mainly used on sodium/mercury/halide lighting and fluorescent ballasts. With MWBC, it lets you put a **lot** of light in a ceiling cheaply: 3 lighting circuits on 4 wires (conduit is ground), 9 circuits per conduit without de-rating (neutrals don't count in MWBC). 40KW on twelve 12AWG wires. Not bad. – Harper - Reinstate Monica Jun 23 '16 at 00:50
  • 1
    @AndrejaKo All the lighting in the office I'm in right now (fluorescent tubes) is 277VAC, so the switches and ballasts are rated for that. The boxes in the wall are no different than usual for 120V wiring. – Spehro Pefhany Jun 24 '16 at 15:54
  • 1
    For the 120V/220V/240V/277V compatibility thing a mandated 450V active PFC stage eliminates all that. – Maxthon Chan Jun 24 '16 at 16:18
23

Nobody answered why were they designed that way?

I recall it was for practicality and cost of manufacturing.

I heard about it indirectly, on a TV show about Tesla. He was contracted to provide lights as Edison's competitor, but could not use Edison's patented sealing/base method and had to use his own which lacked the advantages.

Trying to find more about it, I think the point works like this: you seal the glass capsule and shape it like a bottle top easily enough. The fitting needs to slip over that, and be simple and cheap to make. The contacts along the outside surface of the simplest cap shape is the simplest. Naturally the contacts will be co-axial, and you need at minimum 2 metal areas separated by insulation. Making one of those parts be the thread as well saves components and joints. It also becomes easy to connect the emerging wires to the cap without any alignment.

In short, it's the simplest possible fitting to manufacture.

From the link above, the author quotes Alan Makkos,

The familiar screw-in base was a whole other can of worms. Edison’s first bulbs slipped into their sockets without a way to secure them, until he was struck by a utilitarian design while having lunch in his workshop. “Edison saw a can of kerosene on a shelf, and said, ‘Oh, the lid for that kerosene would make a dandy screw base for a light bulb.’ So they got the can down, cut the lid off with a band saw, and made a light-bulb socket out of it,” says Jenkins. “By the time they were in production in 1885, they had reduced the size of the base quite a bit, and it looked a lot more like modern bulbs.”

JDługosz
  • 675
  • 1
  • 4
  • 12
  • Has the metal base ever sealed the enclosure? Making air air-tight seal between metal and glass that won't leak under changing temperature conditions is hard. By my understanding, a sealed glass bulb is constructed with two wires passing through the glass; those wires are then attached to the base. – supercat Jun 24 '16 at 22:35
10

Today's sockets for incandescent or CFL lamps (and now some LED lamps) come with a black and a white wire. The white wire is neutral and is connected to the outer "threaded" part. The black (hot) wire goes to the center contact. I have checked this myself with an ohmmeter to prove it.

On the old 'Edison' sockets you have 2 brown wires. The light brown wire with ridges is the neutral, and connects to the threaded part of the socket, allowing a bulb to be replaced with no shock hazard. The other darker brown wire is the 'hot' wire to the center connector of the socket, coming from the light switch.

The socket is wired this way so that whilst screwing in a bulb the neutral makes contact first, and the neutral line should have just a few volts on it at most, depending on if it is a 'shared' circuit with other loads. All neutrals are grounded at the main breaker panel to keep the voltage low (<10vac) on the neutral wires even with active loads throughout a house, apartment, etc.

Touching the threaded part should not cause a shock - ever! If it does then the wires to the socket may have been reversed, either at the socket or the light switch for that socket. Even if it is a socket for a 3-way bulb, the threaded metal part is connected to neutral.

NOTE 1: A high wattage device such as a space heater far from the breaker panel could have 10vac on the neutral connection at the AC outlet, but 10vac is not a shock hazard even if you found a way to test if it is a shock hazard.

NOTE 2: Without a bulb screwed in and the power turned on, there is a 220vac shock hazard present. For European light sockets, it is 4 times the energy of American sockets running at 120vac if someone touches the center contact. Most light fixtures are above the reach of children, but not desk-top lamps. Best solution is to always keep a bulb screwed in-don't leave light sockets open.

For information on what constitutes a shock hazard in terms of neutral voltage, I found this article here.

  • 7
    This doesn't answer the question. – user57709 Jun 21 '16 at 21:57
  • 1
    @user57709. Added paragraph 2 to define wiring of the old Edison light sockets. I am telling the OP that the sockets are safe to use if wired correctly. –  Jun 21 '16 at 22:23
  • @user57709 it directly answers the question in that the question comes from a false premise. – Passerby Jun 21 '16 at 22:35
  • 2
    @passerby The question is about the exposed live contact when no bulb is fitted. I didn't downvote this, btw. – user57709 Jun 21 '16 at 23:35
  • 4
    Yes, I'm not talking about the netural connection, I'm talking about the hot contact at the bottom , people can accidently touch it while replacing a bulb or leaving a socket without a bulb. – Ronen Festinger Jun 21 '16 at 23:42
  • 5
    @Sparky256 backstabbing? What? Lets calm down here. Your score on this question is currently +1 and -1, so 0. If someone partially up voted you, I'm guessing it means someone undid a down vote as far as I can see on the timeline. +10 -2 is the rep score, as down votes cost you 2 points but up votes give you 10. You're still 8 in the green. The **reason someone down voted you is wholly up to them, and they don't have to comment or explain why.** That's how SE is designed, and it's just a matter of fact here, you either grow a thicker skin or let it get to you. Make that +2 / -1 votes. – Passerby Jun 21 '16 at 23:42
  • 1
    @RonenFestinger you should edit that into the question. And in that case, it's even simpler. The sockets are not meant to be changed or left empty, when the switch the socket is connected to is left on **ON**. But yea, legacy/grandfathered standards from a more carefree era. – Passerby Jun 21 '16 at 23:44
  • @Passerby. My apologies for rocking the boat. The answer you put in bold type explains it all. There are times when I get +12 on a simple answer, and do not understand why. Now I do. It all evens out in the long run. –  Jun 22 '16 at 02:08
  • @Sparky256 exactly. It's a unique system here, for all its pros and cons. :) – Passerby Jun 22 '16 at 02:33
  • 3
    While I agree that the answer as-is doesn't cover the center contact risk, the reason for it is pretty obvious in terms of history. I'm upvoting this because of the useful and rare information about stray voltage levels on neutral -- this is seldom discussed or understood even among electricians. – stevegt Jun 22 '16 at 03:37
  • 1
    @Passerby. I wrote my answer with the logical assumption that no one would put a finger into a open socket with the power on. In such a case any socket on the planet would be a shock hazard. If wired correctly and just screwing in a new bulb, there should be no shock hazard. That was the logic I used to guide my answer. Based on the OP's question about an exposed center contact, all light sockets would be dangerous, Edison or those built now. Sigh... –  Jun 22 '16 at 06:03
  • Note that our system of using only 120VAC makes it difficult to kill ourselves if we get a shock (if not actually inpossible). This is based on the amount of current you can get across the chest. The EU's use of higher voltages makes their systems dramatically more dangerous. – Robert Endl Jun 22 '16 at 08:20
  • 1
    @Sparky256 If you were to bring out a new 220V (for example) connector system in the EU, you would have to satisfy the Low Voltage Directive. This would entail making sure that a human finger cannot touch live parts, which you would check with an IEC test finger. It wouldn't be enough to assume no-one would do it. Connectors that don't protect stray fingers are inherently more dangerous. – user57709 Jun 22 '16 at 08:39
  • @Passerby I did mention "an exposed hot contact" so I wasn't talking about the neutral contact. Which I know is grounded so no risk of touching it. – Ronen Festinger Jun 22 '16 at 12:50
  • 3
    "For 220vac european sockets (and some other countries) there is no neutral" - this sounds strange - could you elaborate? (I'm in a 220vAC European country - GB - and we certainly have a neutral.) – peterG Jun 22 '16 at 20:38
  • 1
    @peterG. I was in Germany for 2 years so my statement that their is no neutral only applies to loads that only use the 220vac hot wires. We had adaptors to get 120vac from the 220vac outlets. They used just one power pin and the neutral/ground tabs. We called them 'cheaters', because it was not a proper use of the 220vac socket. That was 40 years ago. Sorry for the confusion. –  Jun 22 '16 at 20:50
  • Split-phase 220 does not require a neutral to *function*. But it requires an EGC to be *safe*. – Mazura Jun 22 '16 at 22:29
  • Electrical installations in Germany normally have three wires: a black or brown one for the hot wire, a blue one for neutral, and a green and yellow one for ground. If you use a two-wire socket, the ground one is not installed. Could someone elaborate on where the difference to the US-system is? I am lost here... – Dux Jun 23 '16 at 08:01
  • @Dux *Most* US household wiring is the same as you describe, except we use black for hot, white for neutral, and hot is only 120VAC. However, the feed to the house is ["split phase"](https://en.wikipedia.org/wiki/Split-phase_electric_power), which means certain appliances (e.g. those that have a heavy-duty electrical heater inside, like a clothes dryer or an oven) can be designed to run on 240V; they get it by bridging the phases. Older installations wouldn't bother with a neutral wire for those circuits, but, as Mazura points out, that's unsafe, and no longer building-code-acceptable. – zwol Jun 23 '16 at 13:52
  • @Dux However, nobody runs lighting off 240v circuits in the USA. – zwol Jun 23 '16 at 13:54
  • 2
    220V in EU means that one must be neutral. 2 lives in Europe yield 380V. The adapter you describe could only work in **American** 220V socket (eg dryer socket). If it was 30 years ago in Germany then possibly it was USArmy building, wired for US appliances, with European sockets retrofitted on typical American 2-phase installation. Such imports are not representative of European wiring, certainly. Bulb screw sockets across the world are almost same except voltage. – Agent_L Jun 23 '16 at 18:13
  • @zwol ECG is **not** a neutral wire. Those are 2 separate wires, even in US installations. US 220V equipment mostly doesn't need a neutral, but it does need ground. – Agent_L Jun 23 '16 at 18:23
  • @Agent_L. That was a military installation I was based at. I only knew that the outlets supplied 220vac, and a 120vac stereo would blow a fuse if plugged in. Had to buy local stereos/TV's that had a 120/240 switch. –  Jun 23 '16 at 22:16
  • @Sparky256 I understood you used 110V stereos with those cheater adapters. – Agent_L Jun 23 '16 at 22:55
  • @Agent_L. Yes, but they were not trusted. They were sealed in plastic so i did not take one apart or use them myself. I did use one and my stereo burned up, so i never used them again. I bought new stuff with the 120/240 switch built in. –  Jun 23 '16 at 23:08
  • 1
    @Sparky256 That sounds suspiciously like regular transformer converter, those were bulky, heavy and had a reputation of doing just what you've described. I suspect you might have seen the ground pin 'cheater' back at home in a dryer socket, because that's pretty much how US installations are designed to work and there is no way to get 110V from European installation by simple rewiring. – Agent_L Jun 23 '16 at 23:24
  • 1
    Note 1 is wrong. All European 230 V AC light sockets have a neutral connected. Many of them also have a ground connected to any extraneous metal parts. What the US military install in its US bases in Germany might bear no relation to normal German standards and practice. – RedGrittyBrick Jun 24 '16 at 10:03
  • 1
    @peterG European systems are mostly TN, i.e. they do have a hot and a neutral wire, but most plugs in use are not polarized (Schuko and the Italian ones) or there are conflicting standards (France and the Czech Republic use the same polarized plugs but have contradictory standards over which pin should be neutral). Notable exceptions are the British Isles and Switzerland (which have polarized plugs and an unambiguous standard of which wire goes where). For the rest, the net result is: looking at the two wires of a desk lamp, you have no way of knowing who's hot and who's not. – user149408 Jun 24 '16 at 10:42
  • @user149408. I have deleted my original Note 1 about European sockets not having a neutral. –  Jun 24 '16 at 16:45
3

I suppose the real answer is because we don't live under a completely tyrannical government. People have been living with this type of socket for over a hundred years. This design has a lot of hours on it, and people understand how to use it, and are generally aware of the risks. Any competing design would have to convince, not only the government officials responsible for building codes, but enough of the population so that those government officials don't get replaced come next election. There is a market, and when people realize a clear benefit to any given solution, they will switch. Although you have pointed out some flaws in the legacy design, you must realize that each user does not attach the same weight to those flaws as you do. There is also a potential well that a new design must overcome, where it's benefit must outweigh the additional cost or inconvenience of using the new design.

I think you are looking at if from the perspective that it is the place of the government to only allow those behaviors this citizens which it deems fit. The United States System of Government, is based upon limiting the power of the government relative to the citizens, so that there must be a compelling reason for the government to act and restrict the freedom of it's citizens. Looking at if from that perspective, lets evaluate the apparent flaws in the prevailing design:

  1. Flaw: They require a complete grounding system to ground devices for the rare situation where the hot wires somehow touches the metal case of the device Mitigation: This is a one time cost, upon fixture installation. Magnitude of the cost is not that great, as it involves one extra wire in the cable.
  2. Flaw: While they let you use a lamp fixture which has an exposed hot contact when you remove the lamp, and nothing is there to protect your finger from touching it. Mitigation: Sparky's answer and user training do a great deal to eliminate this apparent flaw. It may appear that this is a dangerous design, but the facts don't really back it up. The Consumer Products Safety Commission evaluated electrocution deaths over a 7 year period from the last decade. They found only 25, 3-4 per year, were attributable to lighting equipment (https://www.cpsc.gov//PageFiles/108404/2008electrocutions.pdf). This does not mean that these 3-4 deaths would be prevented by a different fixture design, only it puts an upper limit on the risk from this design. My guess is that no one is killed by this design, because even it is wired incorrectly, it just gives you a small shock. Clearly though the risk is quite small.
  3. Flaw: I would expect this socket to have a different design which will cover the contact from accidental touch. Mitigation: It is only really an issue while the circuit is energize without a bulb present, so users have a sufficient work around. Again, the realized risk in this situation is diminishingly small.
Mike Vonn
  • 131
  • 3
  • "They're allowed to be used, because that's what we've always used." In essence. – MattyZ Jun 23 '16 at 18:07
  • I think a better way of putting it is that they are allowed to be used because there is not a compelling reason to disallow them. – Mike Vonn Jun 23 '16 at 18:09
  • 1
    I think the risk could be easily lowered from the beginning if it was designed so that the base of the bulb was a narrow rod and the the base of the socket was insulated with a small hole in the center which the rod will fit into. – Ronen Festinger Jun 23 '16 at 18:17
  • Perhaps, take a look at my updated answer. We are talking about a very small risk here, so there does not seem to be much incentive to fix it. – Mike Vonn Jun 23 '16 at 18:33