Hello:
I know that what explodes in the american CFLs at their EOL, is an electrolytic capacitor and that such CFLs have a voltage doubler to fuse the filament at EOL, which causes this explosive EOL. But is this the main electrolytic capacitor of the ballast that explodes, or that the voltage doubler have its own electrolytic capacitor, that sets the lamp with fire at EOL?

I ask this, since my Osram Duluxstar 8W/827 took more than 20 mins to fuse the deplated electrode, yet there was no smoke or fire, and even the glass in that area didn't crack.

Exploding components are not a safety problem. They blow to pieces inside the ballast enclosure, nothing except smoke gets out. The safety problems start, when something cannot "explode" quickly but instead burns up slowly, with enough time to get very hot, burn something else with this heat, catch fire and so on



In a CFL without voltage doubler, the rectified line voltage and the charged capacitor are connected in parallel to the load

In a voltage doubler, there are 2 identical capacitors : In each half wave one of them is charging, and the other is in series with the line, supplying the load. Therefore the voltage the load gets = line peak voltage (from the power line) + line peak voltage (from the capacitor) = 2x line peak voltage. In the next half wave, the other capacitor will be supplying power and this one charging up

Electrolitic capacitors are not a very strong energy source. The energy in them is by far less than what can be supplied from the power line



Now, imagine that there is a short circuit in the circuit inside the CFL

In the CFL without doubler, the short circuit is connected in parallel to the capacitor and to the rectified line. So the power that can be supplied to the short circuit (from the line) is unlimited, and is sufficient to blow up everything very quickly

In the CFL with doubler, the short circuit is connected in series to the capacitor and to the rectified line. So the power that can be supplied is effectively limited by what the capacitor can let through.. Which is not allways sufficient to blow up things. Especially in an aging CFL, where the capacitor is dried up and limits the power even further down vs what was designed in the 1st place

When nothing blows, there is time for various components to heat up...

 - If there is no enough power to blow the cathode but only to make it overheat, the result will be significant melting of the CFL housing near the cathode and possibly popped tube

 - If there is no enough power to blow the fuse resistor, it will only heat up and possibly melt a hole in the CFL base

 - If there is no enough power to blow up electronics, there might still be enough to set them on fire instead.....

So are 240v CFLs safer at EOL (less chance of fire) than 120v CFLs with the voltage doubler?

I think so. Imagine what happens when something shorts in a 230V CFL : Even without the Elco at all, the 230V can go simply through 2 diodes of the rectifier bridge to the short circuit. Nothing limits the current much. So it will easily blow something (the shorting component, one of the rectifier diodes, or the fuse)

But not in all failure modes there is something pulling enough current to be blown. 10 years ago I met a CFL that had some failure that caused everything to heat up a lot, but apparently not pull too much excessive current..



At the time i lived for a while in dorms of a technology institute. There was a corridor with no windows, one door entrance from the stairs, several doors entrances to rooms. The lighting was 2 enclosed Plastic meltlights with CFLs

When i only got there, both CFLs were working

Eventually one EOL'd. The other still worked, so nobody was too bothered to relamp the EOL one

Few days later the other EOL'd. There was full darkness, in the darkness i could see that none of the CFLs emit any glow at all (in both positions of the switch, as it was a 3 way i could not tell which position is on or off, but i tried in both)

So i climb up to replace the CFL (with some GLS which i unscrewed from an unused light in the room). The one that EOLd first and was supposedly dead for a few days allready

I touch the luminaire cover, it is hot

I opem it.. All the corridor is filled with strong burning CFL smell

I touch the CFL... The tube is not hot, but as i touch it, the ballast casing disintegrates and it remains the CFL tube, hanging on wires from the round circuit board, hanging on wires from the screw cap (still in the E27 socket)

I disconnected the wires from the back of the socket (not touching the ends, assuming they are live as i cant know in which position the switch is off) and this way unscrewed the CFL base from the socket

In the CFL ballast the choke, 2 transistors and electrolitic capacitor were very hot, the transistors were scorching. But apparently it stayed like that few days wihout blowing anything..



In 120V, the small CFLs (with low voltage tube, i think around ~11W and below) might suffice with 170V DC, so perhaps they dont have a doubler either

CFLs should never have been designed to do anything like that where something has to blow up at the end. Very careless designing and very unsafe for anyone. LED, Halogen and Incandescent are all better types to use.

The exception is those older magnetic preheat CFL's

It is designed as disposable, so there is no effort to preserve the ballast. The simplest means of EOL shutdown is to make something inside act as a fuse

The intended component for that is the lamp cathodes, and if something else failed shorted in the ballast itself, then a resistor placed in series with the line input (where the fuse would normally go). The choice of resistor over a fuse (dont think of Glass fuse, but of making a fuse out of a circuit board track - which is far cheaper) is probably to limit the current, so the CFL will burn out without tripping the circuit breaker or damaging the light switch

What i think happens is : The available energy for the fireworks was calculated (by the designers) counting on good capacitors in the voltage doubler. In reality by the time the tube is EOL the capacitors are dried up, and all they can do is heat up and hiss in series with the shorting circuit (on the voltage doubler that is), allowing exactly the limited power at which things go up in smoke and catch fire instead of going out quickly

Thanks for the detailed info Ash. I never knew they were designed to fail in such a way. Would have been much safer to put a real fuse inside like a magnetic ballast has a thermal cutoff inside.

I feel like these companies traded consumer safety for an extra buck. I'm glad I do not use CFLs at all anynore

They were not designed to fail with fire and lots of smoke if that is what you are asking. It is meant to be a silent EOL in case of normal EOL, and a "bang" and a little smell in case of something going wrong

The more scary effects are results of (what i suspect is) overlooked design flaws, and of cost cutting



The fuse in CFL is there to go off in case of a ballast failure, not normal tube EOL. So in most CFLs it will never be called upon to blow (up to and after the lamp EOL), but sometimes it will. What choices we have :

 - Glass Ballotini fuse. Those were actually used in CFLs of the 90s

 - Metal film resistor

 - Ordinary resistor (like whats used for electronics)

 - Thin track on the circuit board making a "fuse"



What we get with each option (sorta..) :

	Glass fuse	Metal film resistor	Ordinary resistor	PCB track fuse
Protection from "high current" short circuit	Good	Good	Moderate	Moderate
Protection from "limited current" short circuit	Good	Moderate	Absolutely No !!! (it catches fire)	Bad (it may leave a carbonized PCB behind, risk of arcing)
Fuse itself limits the short circuit current	No	Yes	Yes	No
Cost	High	Moderate	Low	Free

We want the fuse to be effective against all sorts of faults, work safely, and we do want it to limit the fault current - so the CFL handles its faults on its own and does not trip breakers

(with 90s CFLs a ballast short would be more likely to cause a welded light switch or tripped breaker, but the ballasts were by far better built - which really minimized the chance of such failure to happen in the first place)

Quite obviously, the Metal film resistor is fair choice for the later, cheaper (higher chance of ballast failure) CFLs. But then there are some things to consider..

 - If there is a low current fault, the resistor will heat up, but not enough to fuse. The hot resistor itself must not present a danger : Melting the CFL Plastic enclosure, setting it on fire, and so on. The resistor must be positioned so it isnt outright pushed against the Plastic from inside, and the Plastic must be non flammable (at temperatures to which it might get heated by the resistor)

 - When the resistor fuses, it can shoot some (small) sparks. The CFL ballast enclosure gotta be made of sturdy enough Plastic to not melt through, and not have an opening right in front of the resistor so the sparks dont come out (and land on the user's bed below the CFL...)



One of the most common things i seen is use of ordinary resistor instead of the Metal film one. That is pure cost cutting, potentially even done by beancounters instead of by engineers : The engineers said X ohm resistor, so we found an equivalent one thats cheaper than that expensive one they specified...

But the fault in which the difference between the 2 resistor types will show up (limited current) is quite rare, so most CFLs (atleast here) have the ordinary resistor yet scary stories that come down to the wrong resistor type being installed are very rare



My personal thoughts on that are :

 - Incandescents were ok all those years, but usually i want higher K temperature of the light and less energy use

 - If not my parents ("Ashy, a lamp for $20 is WAY too expensive !!") back in the very early 00's (Osram Dulux EL's from Germany), we'd probably be using CFLs allready then - because i seen them somewhere and liked them. In reality i switched my house to CFLs around the mid 00's, when their prices started going down and i finally managed to convince my parents. Yep, the tradeoffs made to get the CFLs cost less are the ones that made them fail in nasty ways sometimes...

 - Decent CFLs do contain their faults, even if they are not the best. They are not going to be dangerous

 - I dont feel a need to be "safe" from a potential scary (but harmless) sudden "bang" surprise.. (which havent yet happened to me in ~12 years). If it happens, it happens..

 - Most other lamps may fail with a bang too - Incandescents (arcing), LEDs (electronic driver short circuits), so CFLs are not particularly worse

The only resistor types, which may be used as protection elements are those specifally rated as "fusible resistors" and so bearing the corresponding safety certification (UL,...; have to meet the same requirements and certification tests as any other fuses).
These are then safe even for mild overloads, as they are certified to go safely into an open circuit, once they overheat.
These were in the old days in the form of wirewound resistors in a tall ceramic "brick", with attached melting solder thermal cutout stripe.
Modern types are usually in the form of a thin film metal resistor with part of the deposited stripe being heat fusible.

Any other resistor is just a resistor, so no one could count on them to act as fuse or so. Regardless what technology that uses.

So for the suitability as a fuse the resistor technology is not relevant, relevant is only the resistors rating (and that could be limited to given mounting style - e.g. the wirewound "bricks" needed sufficient clearance in the direction of the spring release, the metal film ones require either clearance or heat resistance of the surrounding,...)