Batteries the Focus of AT&T Investigation

An anonymous reader writes "AT&T is focusing on the batteries supplied by Avestor as the cause of its 2006 equipment explosion in a suburban Houston neighborhood. The carrier says it has 17,000 of those same batteries still in its network. Some photos of the equipment that was shredded in the blast are also available."

Hydrogen + Spark = Bad

[effigiate]

I'm not sure what types of batteries they were using, but standard lead-acid batteries vent hydrogen during charging. If you don't make provisions for the removal of it and it builds up in the cabinet, one tiny spark and you've got yourself a little bomb sitting there.

Re:Hydrogen + Spark = Bad

[Anonymous McCartneyf]

And the phone lines are near the power lines, so lots of opportunities for little sparks...
But I hear that the batteries in question are li-poly. I don't think they vent hydrogen; they just appear to have unfortunate internal similarities to C4 explosive when they're made wrong.

Re:Hydrogen + Spark = Bad

[NoMaster]

Not so much of the "little sparks"; yes there are relays still inside that sort of gear, but they're reed relays with the contacts sealed in a nitrogen atmosphere.

What you do find, however, is cct breakers and contactors on the main power feed & internal distribution. But the usual explosive trigger is the sparks from the cells themselves as they self-destruct...

We used to get 1 or 2 incidents like this a year in Aus, mostly in the far north. The SLA batteries used here don't take kindly to temps above ~ 35C (easily found inside cabinets even in colder areas) - battery life is dramatically reduced, and the usual failure mode is thermal runaway causing internal shorts &/or case splitting. The end result was usually just acid vapour destroying everything inside the cabinet. Occasionally, as I said, we'd get one that went boom.

It certainly wasn't unusual for the batteries to to from "visually OK, passes all discharge / internal impedance test" to "case on the verge of rupture" in as little as 1 month. We usually checked them every 3 months, and that was considered by both the manufacturers and my superiors as "excessive".

Having said that, what freakin' idiot decided Li-Pol cells were a good idea in that sort of environment! The things are barely stable at room temperature, requiring very careful feeding and care of charge / discharge rates even then (which is why every single consumer-use Li-Ion/Pol battery sold has a charge controller & thermal monitor cct built-in to the battery itself). Essentially, from a non-technical POV, Li-Ion & Li-Pol cells are little containers of metallic fire that will self-destruct at the slightest provocation. I certainly can't imagine any really safe way of using them in a online / continuous float application like that...

(Ex-telco senior switching, CAN, and battery maintenance tech.)