To find the cause of the Galaxy Note7 incidents, Samsung examined every aspect of the Galaxy Note7, including hardware, software and related processes over the past several months. Samsung’s investigation, as well as the investigations completed by three independent industry organizations, concluded that the batteries were the cause of the Galaxy Note7 incidents. The causative factors are further explained in the infographic below.
The presentation last night was quite informative, and both Samsung and the three independent organisation got time to explain their findings in quite some detail. Sadly, they removed the VOD of the livestream from YouTube, so there’s no way to rewatch it (edit: someone uploaded the VOD), but some of the slides can be found at the bottom of the linked article.
I haven’t watched it, but I think this is the presentation:
https://www.youtube.com/watch?v=Iu18CykEH9o
I can’t watch the presentation right now but do they explain exactly *what* in software they’d be looking for as a source of an exploding battery?!
Edited 2017-01-23 23:37 UTC
A bug might be shorting out some cells.
PJBonoVox,
According to the presentation, the investigation revealed there were no flaws with the samsung software or electronics that contributed to this. It was a physical battery flaw.
As I understand it, unlike NiCad or NiMh which can be charged safety with basic circuits, lithium ion batteries are less stable and require more dedicated circuits and even firmware to charge & monitor them safety.
Here’s a very detailed paper from blackhat where they did a teardown of a mac battery, electronics and all.
https://media.blackhat.com/bh-us-11/Miller/BH_US_11_Miller_Battery_F…
I’m having trouble finding it now, but I had read an article where someone was deliberately hacking the battery firmware in an attempt to get it to explode. (he failed btw).
Edited 2017-01-24 10:35 UTC
Hi,
Does the presentation explain how Samsung’s QA testing was so inadequate (for a potentially explosive device) that they failed to realise they were shipping faulty batteries (regardless of who made them) long before any consumer received one?
– Brendan
The problem with inadequate testing is that you only know it was inadequate once something goes wrong.
The faults were INSIDE the battery on a microscopic scale. An assembler of parts cannot look inside a SINGLE part, much less ALL OF THEM. And at less than 1 failure per 10,000, even Q&A at the battery factory wouldn’t have caught this problem.
What you say makes sense until one considers that it’s might odd that all the faults ended up in the same Samsung device model. To me it doesn’t sound like a random fault in every one out of ten thousand, but perhaps a bad batch out of many batches.
darknexus,
If only certain batches of batteries had a defect, then I suspect the recall would only apply to specific ranges of Note 7 serial numbers rather than all of them.
Apparently they used two different battery manufacturers, one had a defect were the cells got bent physically and the others were missing insulation and got damaged during welding. I doubt either fault would be apparent to Note 7 product line Q/A – just as if you built a product and sourced a lithium battery from another manufacturer. The videos show that they x-rayed the batteries to see the problems, so maybe they’ll have to x-ray lithium batteries in the future before trusting new production lines.
It’s not clear to me whether the battery manufacturers were Samsung owned manufacturers or if they were independent contractors. Anyone know?
It sucks that batteries are no longer replacable, since rather than just a battery, users had to replace/reinstall the whole phone.
Edited 2017-01-24 22:55 UTC
It was an independent battery manufacturer that got bought by Samsung in order to get a better price on orders. Really worked out well for them, didn’t it?
The reason it was specific to one phone is an indictment on nearly all electronics today – parts are made for ONE MODEL ONLY and then discontinued. This makes it much more difficult to get cheap replacement parts, but by the same token, if there’s a problem with a part, you’re stuck as you can’t get parts from someone else. That’s why Samsung had to eat a few BILLION dollars worth of phones because of a scant handful of bad batteries. Live by the unique part, die by the unique part.
JLF65,
I happen to dislike the AA/AAA/C/D battery standards, they’re awkward and space inefficient for anything other than cylindrical flashlights, but it is so nice that they are at least standard, like being able to pick up batteries for a camera or kids toys at the corner store. I would hate it if all those toys and gadgets required custom batteries.
With lithium, and every product having a custom battery, I feel it’s a lost opportunity for the industry to benefit from having standardization there too. We had a camera that was <2 years old, I went to best-buy (who stock all kinds of specialty batteries) and they only stocked batteries for cameras that were in stock…doh! I had to order manufacturer batteries for $60 (on an obsolete $150 camera) or $20 3rd party clone batteries from a chinese company.
Edited 2017-01-24 23:23 UTC
Considering the physical size of the manufacturing defect, it’s easy to see how the problem is random rather than specific to a batch. When you have extremely small tolerances, a single piece of dust getting in the way can be the difference between no problem and a serious problem.
You have quite the point, beer lover [%]D
Tolerances were quite easy to meet at brick phone times. New ‘fashion’ demands, specially from the few, big clients -who are difficult to turn down- has brought this State of disarray.
These over stressed battery designs are a liability. Note7 wasn’t that much a phone fiasco, but a battery industry in complaisance with her market, over the engineering & governmental normativity].
As Alfman comments, as long as [battery]formats don’t standardize and detach from oligopolistic or monopolistic liaisons, regulatory will REMAIN A MESS.
Hi,
You can’t look inside a single part; but you can take every 1000th phone (or one from each batch, or…) and do stress testing (e.g. see what happens at full load and check it still meets the phone’s specs, etc), and figure that something might not be right if it explodes during that testing.
– Brendan
When you have a failure rate of less than 1 in 10,000, testing every 1000th phone will very likely get you nowhere fast. This didn’t happen because of negligence on Samsung’s part. Unless you were looking for a very specific problem on a microscopic level, nobody would’ve caught the problem.