Take apart a Compact Fluorescent Bulb

Step 6 What Breaks?

Having looked at the guts quite a few dead CFL bulbs, I feel somewhat qualified to point out a few of the reasons that they go bad.

First, of course, the tube itself can go bad, having leaked too much vacuum, or evaporated too much metal internally, they just stop working. When manufacturers quote you extreme lifetimes for CFL bulbs, this is the failure mode that they have in mind.

Unfortunately, a large number of CFLs seem to go bad in the ballast electronics. I've seen them smoke, emit bad odors, and even spark (scary, given the probably flammability of lamp shades.) I've taken them apart and seen obviously burnt components. I'd like to blame this on "cheap imports", but I've had a fair number of brand name CFLs with similar problems. Even some electronic ballasts in circleline fluorescent fixtures. Sigh. (It does seem to be getting better.)

Unfortunately, just because a component on the circuit board is burnt, doesn't mean that that's the component that went bad initially.

The major suspect seems to be the electrolytic capacitors that filter the HV DC. I've seen these with bulging and even burst casings. If you read capacitor spec sheets, you'll discover that such capacitors have a finite lifetime to start with, and that lifetime goes down relatively dramatically as operating temperature goes up. Inside a poorly ventilated casing with 20W of power being dissipated nearby makes for some pretty high temperatures. There ARE high-temp capacitors, but I've never seen one inside a CFL :-( Once the cap goes, the HV oscillator is getting pulsing current instead of DC, which I suspect it doesn't like, and it's not surprising that other things go wrong too.
Some, but not all, CFLs contain a fuse...

The inductors are pretty hardy things; they're probably good unless they show obvious signs of being burnt. The non-electrolytic caps are probably the same, and you can easily test them for shorts using a multimeter. I've never tested any of the transistors...

Comments

[ginbot86]

When it comes to harvesting components from CFL bases, be sure to test those transistors. I cracked one open that burnt out in my bedroom and found that the E13003 power transistors were short circuited (reading 0 ohms on a multimeter) between the base and collector, base and emitter, or all three pins.

Higher-wattage bulbs have a (somewhat) greater chance that their transistors would have survived the lamp failing; I have some nice E13007 transistors, and some really nice MOSFETS of which I can't remember the part number right now.

[westfw]

yes, you want to be careful harvesting components from bulbs that have actually burnt out in one way or another. The usual failure modes seem to take out several components. But these days the (subsidized) prices of CFLs are low enough that you can think seriously about harvesting brand new CFLs. (I think I've seen 4-packs for less than $5)

[dan_moran]

I had my first CFL die (early in my opinion), so as I do with incandescents, I wiggled and tapped on it to make sure it was getting a good connection in the socket. It was screwed in tight but it will light up for a couple seconds when I tap on it. What do you think is going on inside this bulb and is it salvageable?

[westfw]

One of the failure modes I've seen is that a positive-temperature-coefficient (PTC) thermistor (think of it as an inrush limiter, I think) will apparently overheat (perhaps due to other problems?) and have one of it's leads unsolder itself from the component body...

[gordzooks]

The PTC is what Chrysler used to use in the 60's as an ignition ballast resistor. It's not so muchan inrush item as limiting current once a circuit is in operation. When it's cold ( startup, or colder conditions ) the resistance is low allowing more current. Once the circuit starts operating, current warms the PTC, raising the resistance and lowering the current. This allows the PTC to cool, lowering the resistance. At some point it will reach an equilibrium state based on demand (varying in a car, stable in a CFL) and ambient conditions.

[Marshal Banana]

I never thought about the caps overheating and dying. I have several lights in my ceiling, and the CFL's only last a few months. Whereas the CFL's in normal lamps which get plenty of ventilation have never died on us. I was originally interested in taking these bad boys apart in order to make them into LED bulbs.