Analog Angle Blog

When an unanticipated failure mode leaves you puzzled

It's not news that components eventually fail even in normal use; engineers are used to that. We also understand the likely failure modes: filaments burn out, film capacitors dry out, and batteries lose their ability to be recharged after some number of cycles.

But I recently had a modest failure experience which has me puzzled and a little worried. I smelled the classic odor of electronics getting a little too hot and starting to burn, but couldn't localize it. So, I went around closely sniffing my PCs, TVs, and other devices, looking like a slightly crazed escapee. A few hours later, a 13-watt compact fluorescent bulb (which had been in use for at least several thousand hours) burned out, which was not unusual by itself. I did not yet know there was a correlation between the smell (which had already disappeared) and the burned-out bulb.

But when I went to change the bulb in the fixture, I saw that the merging where the CFL's glass tube meets the base of the bulb had actually melted and charred, also allowing the tube to tip from its original position:


(click here to see enlarged image)


(click here to see enlarged image)

This is worrisome. Somehow, this low-wattage bulb had managed to go “critical”, getting hot enough to where, I assume, it could have started a fire if there was any easily combustible material nearby; it was also source of the smell I couldn't localize.

Yes, the bulb has that UL label, but so what? As we know from a recent EE Times story on counterfeit electronics, “IHS: Counterfeit parts represent $169B annual risk” and the many good comments it received, false UL certifications are fairly common. Or sometimes, vendors substitute non-approved components into a product after it has been certified (unintentionally, or maybe not so unintentionally—shocking but true).

I am still puzzled by what actually failed in my bulb and allowed a product of such low nominal dissipation to get so hot. On one hand, thirteen watts is not what I normally consider enough internal consumption and subsequent dissipation to start a fire; on the other hand, the bulb is connected to a 120Vac line and, hey, “stuff happens”.

Since I don’t have access to the necessary forensic tools or bulb-design documentation, I can’t determine exactly went wrong. Also, having the purported documentation probably wouldn't help, since the design, original BOM, and components (if they were ever approved) may not be the ones in the bulb as it actually shipped.

Still, I am puzzled by details of the failure mode which allowed a 13W, AC-line CFL to go into modest meltdown. It's got me worried about things just a little more.

Readers: Have you ever had a failure mode that not only surprised you, but that you were unable to adequately explain, whether due to lack of time, information, or insight on your part?

[Additional note : it's not just products with active components such as CFLs that can have basic problems; check out this interesting entry “Poor Lamp Socket Design Sparks Danger” by Jon Titus in the “Made by Monkey's” series of our sibling publication Design News —interesting and somewhat scary, too.]?

0 comments on “When an unanticipated failure mode leaves you puzzled

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.