Power Adapter Failure

[wilheldp]

I was looking at an RC helicopter today with two exploded lithium polymer cells.  It started a fire in somebody's house during the first charge of the battery pack out of the box.  They included the power adapter that came with the helicopter.  The battery pack was rated for 7.4 VDC at 1100 mAh.  The power adapter was rated for 8.4 VDC at 850 mA.  I couldn't tell if there was any charging protection circuitry on the battery pack due to the extent of the damage.

The power adapter was undamaged, so I plugged it in and measured the output voltage.  Open circuit voltage was 22.9x VDC.  When I loaded it with a 6.7k resistor, the output only dropped to 22.8x VDC.  I opened the adapter housing, and found some oxidation on the leads attached to an unidentifiable IC and a transistor...no heat damage or anything else amiss.  What could cause a power adapter to put out that much more voltage than its rating?  Did they just mis-rate a higher power adapter?  It's a Chinese POS, so the latter is certainly a possibility.

BTW, this is in America, so mains is 120 VAC, 60 Hz.

[NiHaoMike]

Sounds like it was mislabeled. As it is a safety related defect, report it to the CPSC.

[wilheldp]

I've been the spark behind a few CPSC recalls, but I have to wait for this case to settle first.

[Kohanbash]

Depends on the construction of the power supply.

There could be something that is shorted (transistor, etc..) letting it bypass the regulator or an intermediate conversion stage. Have you checked if there is any AC voltage on the output?

[wilheldp]

Depends on the construction of the power supply.

There could be something that is shorted (transistor, etc..) letting it bypass the regulator or an intermediate conversion stage. Have you checked if there is any AC voltage on the output?

No, I didn't think to check for AC voltage on the output.  I'd love to take it apart completely and troubleshoot it, but I'm not allowed to.

[wilheldp]

The power adapter is Jing Heng brand, model JH0840850, if anybody can find a schematic on it.  I've come up empty in my searches, but there is an Australian version of it on eBay right now.

[SeanB]

A SMPS will put out a very high voltage if the feedback is faulty and turns the driver on into it's maximum duty cycle. This is often caused by failing output capacitors ( or for a new one fakes or DOA capacitors) or by there being an open circuit feedback loop or a faulty optoisolator. I measured a 5V supply that was happy to deliver nearly 60v when open circuit, it needed new capacitors ( in put in some used reasonable ones to test it first at no cost) but the high voltage had destroyed the connected unit, so it was scrapped and replaced with a whole new kit, at a pretty steep cost.

[amyk]

I couldn't tell if there was any charging protection circuitry on the battery pack due to the extent of the damage.

Very likely not - if it had a protection circuit it would not have overcharged and exploded. Most of them for 2s packs can stand up to 30V at the inputs.

The power adapter was undamaged, so I plugged it in and measured the output voltage.  Open circuit voltage was 22.9x VDC.  When I loaded it with a 6.7k resistor, the output only dropped to 22.8x VDC.  I opened the adapter housing, and found some oxidation on the leads attached to an unidentifiable IC and a transistor...no heat damage or anything else amiss.  What could cause a power adapter to put out that much more voltage than its rating?  Did they just mis-rate a higher power adapter?  It's a Chinese POS, so the latter is certainly a possibility.

Try giving it a higher load than 3mA - if it drops significantly then it could be a stuck-open feedback loop, if it remains at ~23V then more likely that it's either been mislabeled or had the wrong components installed (need pictures/schematic to be sure - e.g. wrong divider resistor in a TL431 design could cause this)

[wilheldp]

I'm hoping the customer lets me do some additional testing on it, but I doubt they will because they can't sue the Chinese manufacturer anyway.  I may get to play with it after the case is over though.  I will post back with updates if I do get to do more testing.

[Zero999]

A SMPS will put out a very high voltage if the feedback is faulty and turns the driver on into it's maximum duty cycle. This is often caused by failing output capacitors ( or for a new one fakes or DOA capacitors) or by there being an open circuit feedback loop or a faulty optoisolator. I measured a 5V supply that was happy to deliver nearly 60v when open circuit, it needed new capacitors ( in put in some used reasonable ones to test it first at no cost) but the high voltage had destroyed the connected unit, so it was scrapped and replaced with a whole new kit, at a pretty steep cost.

In an application where fire, explosion and death could result from such a failure, there should be a protection circuit to stop that from happening, for example a simple crowbar circuit would stop that.

[SeanB]

Bean counters will say the crowbar adds unneeded cost, and will happily leave it out as it will still fail after the initial warranty period ( or after they have been paid for the devices, whichever comes first) so causing them to sell a new one at a profit.

[alm]

Until said bean counters figure out the costs of liability and bad publicity when their products damage their customers' property or injure/kill customers. Try selling Apple or Sony a charger $1 under the price of their other suppliers, with the notice that it will probably catch fire for 0.01% of the customers.

[all_repair]

More reliable does not mean it is more safe.  The fuse approach that feed straight into the capacitor is a less reliable design, IMHO, it is a safer design.  Surge, adnormality, just blow the fuse.  No need to add in so many parts which may make iy more reliable but less safe.  The added parts also can malfunction, and the adnormalities can be out of calculation.  Especially due yo mishandlings.

[Zero999]

Until said bean counters figure out the costs of liability and bad publicity when their products damage their customers' property or injure/kill customers. Try selling Apple or Sony a charger $1 under the price of their other suppliers, with the notice that it will probably catch fire for 0.01% of the customers.

If there's adequate current limiting, a crowbar would only brick it if the voltage feedback loop is broken.