It is my understanding that internal arcs are rather common when incandescent lamps fail and that the arcs can draw substantially more current than the filament. For that reason incandescent lamps (at least in the US) have an internal fuse wire that is supposed to keep the branch circuit fuse or breaker from opening.
I suspect the object that melted through the glass was part of the filament or support lead that was melted by the arc.
I have never seen an incandescent lamp blow a branch circuit breaker or fuse. Terry has more experience in this area and may be able to shed more light on the subject.
Its rare, but its possible. Most bulbs fail when the filament evaporates, develops a weak spot and thermal stress (usually the surge when turning on a cold lamp) finishes the job. This is a relatively low energy event, since the arc, in series with the remaining filament limits the current.
Its possible for the filament and/or supporting wires to fail structurally and result in a short circuit inside the bulb. That will dissipate quite a bit of energy by comparison and possibly trip the branch circuit protection. The folks on sci.engr.lighting might be able to shed some more light on this.
Ballotini (Italian: small balls;-) is one of the types of fuse used where the fusewire is in a tube filled with tiny glass balls to cool and quench the arc quickly. These are common, but so are bare fusewire fuses.
It's quite common in the UK where we commonly use dual element circuit breakers, and the magnetic (fault current) element trips faster than a lamp base fuse. It's rare when the branch circuit protection is provided by a real fuse.
| I have never seen an incandescent lamp blow a branch circuit | breaker or fuse. Terry has more experience in this area and | may be able to shed more light on the subject.
I had a "mini chandelier" pendant lamp with 3 bulbs blow all three once. It tripped the breaker when it did that. One of the bulbs suffered very catastrophic failure, with the glass cracked and the entire inside all blackened. The other two bulbs looked like just blown. Very little of the filament could be seen remaining of quite a lot they had (25 watts). Additionally, all 3 bulbs suffered melting in the screw part of the base, with the bulb most affected having the most. It looks like they did not have good enough contact in the base to support the fault current and an arc formed between base and socket.
There was only one loud pop sound. I don't know for sure if they all went at exactly the same time, but it sounded like it. But I wonder if the arc in the first bulb went out before zero crossing, leaving enough current to surge the other two and blow them.
Some years ago, I worked in a test lab where such phenomena were studied to help customers with the design of lamp circuits and, especially, electronic dimmers. As others have said, such dimmers are often the fuse when a lamp filament fails.
We caused the lamp to fail by using clear lamps and focusing a laser on various parts of the filament. There may even have been a paper on the results and I'll take a look through the old LS proceedings to see if I can find it.
But what I recall is that the fireworks inside the lamp were mainly a function of the point in the sine wave cycle when the filament opened and the "stiffness" of the power supply feeding the lamp. By "stiffness" I mean the ability of the circuit to quickly increase the lamp current. Break the filament just as the current wave was building in a circuit with plenty of current capacity and you had a chance for some spectacular lamp failures because of the arc that got started between the internal lamp parts. The most powerful arcs were between the filament mounts, but more usually between a piece of filament and one of the mounts. As Vic said, the fuse wire in one of the leads usually limited the duration of the arc, but fuses are not precise devices and there were often pieces of molten metal flying around inside the bulb and embedding themselves into the glass surface.
I'll guess that in the case Paul mentions above, a gas-filled lamp is involved and the filament arcs just at the right time and place to position a hot pocket of gas such that the filament melts and turns into a miniature rocket. I'm curious to know what the piece of material that hit the floor looked like. Did it still resemble a piece of filament or was it just a solidified blob of molten metal.
I also wonder if internal lamp arcs have the potential to be more powerful on 50 Hz systems since there is more time for the arc to start and heat up in a single cycle than on 60 Hz systems.
What I suspect happened, before I get much into this thread:
The lamp (lightbulb, that is) had a reasonably normal cold-start burnout with a "burnout arc" forming across the developing break in the filament and "blowing up" to quickly move and grow to have its terminations being the ends of the filament. With hardly any filament material in series with the arc, there is not much to limit its current.
One thing about most arcs: As you feed them more current, they become more conductive, often even disproportionately. Sometimes as a good explanation and sometimes to maybe usually as an oversimplified one, higher current makes an arc hotter, and higher current increases the percentage of the atoms in the arc being ionized.
So when a "burnout arc" forms in an incandescent lamp and "blows up", it is limited by any of the following:
1) A fusible link in one of the wires in the lamp (lightbulb) serves as a fuse and blows.
2) The branch circuit's fuse/breaker blows/trips.
3) The arc dies during the next zero crossing of the line voltage, preferably before it gets *Really Bad* if 1 or 2 does not occur first.
I have heard of complaints of some incandescent lamps lacking fusible links in their internal wires. The usual complaint of those is that burnout arcs draw current surges that blow fuses/pop breakers, ruin dimmers, or cause the wiring in the lamps (lightbulbs) to vaporize internal glue (or whatever gets vaporized) and bulbs get popped from their bases by the pressure of this vapor. Thankfully, such problems are rare! I would hope that manufacturers that "cheap out" by not including those fusible links lose sales as a result!
Now for what else could happen:
I did once overvoltage a 500 watt photographic lamp (ECT), as in giving about or slightly over 140 volts to a 110-120V 3200K 500 watt unit. The filament broke from melting, and without a big bright flash, and apparently one end of one half of the freshly broken hot filament touched and melted through the glass bulb material.
The glass bulb got a small hole melted into it. The filament did not get stuck to the bulb, although I do have memory traces of other incidents where filament fragments stuck to the inner surface of a bulb.
One more datum: In the above incident with an ECT photoflood lamp, the lamp was powered up without overvoltage and the voltage was increased gradually over several seconds.
Back to your 40W incandescent:
I suspect that, if your symptoms did indeed occur as you described, you had some extremely rare bad burnout where a piece of hot filament (or molten portion thereof) both broke away and melted through the bulb.
I consider this improbable and at most very rare, since I have yet to hear of this or experience it, even after spending a goodly chunk of my youth as "The Mad Lamp Abuser"!
Having a burned out lamp (lightbulb) with an actual melted hole showing signs of being made by an exiting small hot object sounds like strong evidence to me. If I ever had such a thing, I would find a way to make an arrangement to loan it to the manufacturer so that they can use it for feedback but then I get back such a rarity. If the lamp was of a brand or model easily identifiable as subpar, especially if purchased from a dollar store, then my next move would be to find out if my municipality, county and state has fire marshalls and report this to every fire marshall possible. I would also report it to the Consumer Product Safety Commission.
This was a plain old-fashioned rewireable fuse. No trace of the original fuse wire, except under the washers at ech end.
Looking at the lamp:
Everything seems present including most of the filament. The end of one third of which seems to be "welded" to an intermediate support. Ditto for another third. And what appears to be the remaining third is lying loose inside.
There are several other marks that look like metal droplets embedded in the glass, on the inside. So, it looks like molten metal was sent in many directions, one droplet only getting through the glass.
No trace in the carpet of what came out. Hardly surprising, it would have been pretty small.
No fusible link is apparent.
I must admit to still being very puzzled over exactly what happened. The glass is intact, appart from the pin hole.
I wish I could take decent photos that I could show you - I have never seen anything remotely like this before.
Did you induce failure while the filament was hot? That will severely limit the arc current, until the arc expands/crawls/blows-up around the filament.
I suspect the arc could expand faster if the filament breaks from an "end-of-life cold start failure" where most of the filament has not yet reached its normal steady-state temperature.
Then again, as I understand it, most cold start failures have most of the filament achieving a peak temperature most of the way to their steady-state operating temperature.
But if you had a way to apply power and then sever the filament late in the first half of a half cycle, then you would have an "acid test" (test reasonably able to match or exceed the severity of *The Really Especially Bad* filed failures).
On the other hand, I have seen some stable burnout arcs!
I have heard of this occurring with 120V T3 halogen lamps used in torchiere fixtures. I have even seen one failed T3 lamp in such a fixture that showed severe bulb blackening around the break.
2) I have seen a 60W 120V A19 with a stable burnout arc in progress. The filament was a C-9 (the multi-supported C-shaped one with singly coiled wire). The lamp operated 24/7 and I give slight chance that it could have developed an unstable burnout arc, and greater chance that it would have developed a stable burnout arc that gets broken by cold-start-burnout-related filament movement, if it failed during a cold start.
It's the (only) web page of the company that produces/markets them. You wrote, "If the lamp was of a brand or model easily identifiable, etc" and the page both identifies the company and shows that it is, "members of both ASTA and BSI and hold accreditation to ISO9002"
I realised that you were not from these parts. However, the offer is there for you to acquire this light bulb to act as a conversation piece for years to come. Like you, I wouldn't have believed it if, I hadn't actually seen it happen and/or seen the lamp in question.
In alt.engineering.electrical TimR wrote: | Probably this is a dumb question, but: | | I know NFPA now requires arc fault breakers for some circuits in the | US, but I went overseas before I actually saw one used. | | Will a light bulb arc trip an arc fault breaker?
Never say never, but generally very unlikely. I've seen one case of a bulb burnout trip a regular breaker. Presumably had that been an arc fault breaker, it would still be tripped at least for the overload part of the fault. An arc fault breaker generally includes GFCI protection as well as traditional magnetic and thermal overload protection.
I think you're right. From my web reading, it looks like an AFCI should be protective against parallel arc faults but not likely to trip on a series one, nor anything on an applicance cord, extension cord, etc.