Hacker News Reader

Comment by weaksauce | Hacker News Reader

might want to put it into context though... there are on average 160,000 passenger car fires per year and about 360 deaths due to that. i'd like to see that figure broken down per 100k miles driven or per 100k cars of the type but i can't seem to find that easily.

https://www.nfpa.org/vehiclefires

View on HN

Yes, but an important part of that context is that electric car fires also burn longer and hotter and take significantly more effort to put out. Too little water can actually make the fire worse.

Annecdotally, my recollection from other articles is that electric cars do in fact catch on fire less often. But it also seems like they may be more difficult to deal with. This may change as EV's become more mainstream & fire departments adapt: Graphite may be effective in putting out these fires, and maybe it will become a common tool in fire departments.

For the time being, rare or not, they post a much larger problem to deal with when they do occur.

> For the time being, rare or not, they post a much larger problem to deal with when they do occur.

It seems to me that being harder to put off but occurring less frequently, and most importantly not exploding (leaving significant amount of time for occupants to get out, if they are able) is a pretty good trade off.

It could be a good trade off, it depends on more details though. Do electric car fires flame up faster, giving passengers less time to exit the car? What about toxic fumes emanating from the burning batteries? Since they they burn hotter and for much longer, are they more likely to damage surrounding infrastructure, like a bridge, before being extinguished?

I don't know the answer to these questions. They could all be "no". But before judging the trade off those are important factors to understand.

Lithium batteries are far more explodey than modern gas tanks. Battery fires burn much hotter and faster than "regular" vehicle fires (which are generally just the contents of the vehicle burning, not the fuel).

Also a gasoline car actually exploding is exceedingly rare. It generally takes a bit of effort to get gasoline to explode, rather than simply catch on fire. The tiny explosions that occur in the engine are only possibly due to gasoline being aerosolized into a fine mist.

Take a gas can with a soaked rag & trail of gasoline leading to it & set it on fire and you'll have quite a fire on your hands, but no explosion. Now, add a tiny bit a black powder to the equation to disperse the flaming gasoline into the air and you've got a small fuel-air bomb on your hands. But that-- or similarly explosive circumstances-- just don't happen with cars much at all. Pop culture, where a few guns shots result in an explosion, gives the wrong impression here.

For the curious: The Mythbusters actually tested the "shooting a gun at a gas tank explodes the car" thing, and it didn't work. It only worked when they tried it with tracers, but those are a whole different thing than ordinary ammo.

I couldn't find a clip of it, but here's the Mythbusters Wiki page for the episode: https://mythbusters.fandom.com/wiki/Car_Capers

Don't try this at home though

I don’t know about giving occupants more time to get out. I have seen the carpark video of a Tesla going up in flames in China.

https://www.bloomberg.com/news/videos/2019-04-22/tesla-inves...

Thing went from 0 to 110 real quick.

From an exterior perspective of a parked car certainly that would look like the case. The interior software in an operating Tesla has very loud warnings as soon as the first firewall breaks between cells which according to multiple incidents happens at least 5-10 minutes before enough cells hit critical temperatures that the entire vehicle becomes fire. Far more warning than an ICE fire.

There's probably a case to make that there should be an effort to provide better, similar software warnings for parked/"off" vehicles to signal emergency services with that extra lead time.

> Tesla has very loud warnings as soon as the first firewall breaks between cells

I searched but wasn't able to find a video of this. Any reference?

> It seems to me that being harder to put off but occurring less frequently, and most importantly not exploding (leaving significant amount of time for occupants to get out, if they are able) is a pretty good trade off.

It's very rare to get a gas tank to explode (as others have pointed out). You need to have very little fuel (or better, none at all) in the tank before a spark will cause explosions.

Take a 2l plastic soda bottle, fill it 90% full with petrol and (with tongs from a distance) drop in a lit cigarette. I've tried to get the bottle to go bang multiple times and it never happens - the ciggarette is extinguished every single time.

Then empty that bottle until there is less than 5ml of petrol in it, close the bottle and shake it up to get a rich mixture, and then repeat the experiment (I've gotten fairly big bangs every single time, this way).

The TLDR of it all is that there is danger of explosion when the tank is emptied, not when it is full, because the fumes cause explosions, not the liquid state of the fuel (which just catches fire).

How long would you suggest those tongs should be, to ensure a safe distance? For the ones that go bang.

> How long would you suggest those tongs should be, to ensure a safe distance? For the ones that go bang.

I'm hesitant to answer in case some tries this and gets hurt.

I used a flexible claw (one like this: https://www.amazon.com/Flexible-Claw-Automotive-Mechanic-Gen...) inserted into a standard copper pipe (used for taps and other plumbing) to stiffen the spring and held it at arms length.

So, about 2m was safe for me. I also suggest wearing protection for your ears as these are really loud bangs, and protection for your eyes as well (it is an explosion, after all).

I think, given the HN community, the appropriate answer is "build a robot to do it for you"

Yes, as I understand it, the fuel vapors give a relatively minor boom that disperses the rest of the gasoline enough to become essentially the second stage. Not that it takes long enough to really discern two stages: aerosolized gasoline will go up fast.

The “explosions” you hear from car fires are almost always first the tires popping, and later the airbags going off. I watched a car fire near me last year and the firefighters didn’t even flinch when the loud bangs happened.

Yep. Moreover, the issue isn't new cars today.

You know when you saw that old Chrysler Grand Caravan burning on the side of the road, and you hear the story that the person has been holding it together solely with thoughts and prayers and maybe $500 between tires and oil for the last 7 years of its 19 year life?

That's the current inevitability. I'm not worried about someone's 6-month old Bolt or 2-year old Model 3. The issue is when we're eventually talking about a 22-year old Model 3 that's had 13 years of deferred maintenance, is totally clapped out, and is only on the road because someone paid some mechanic off to look the other way during a state vehicle inspection for the past 5 years.

Germany is starting to standardize on towing an electric car away to dunk it in a tank to ultimately handle fires.

> Germany is starting to standardize on towing an electric car away to dunk it in a tank to ultimately handle fires.

I suppose if the problem is widespread enough, every city will have a mobile dunk truck at the ready for electric vehicle fires.

And what do you do if those fires cause other things to burn in the vicinity?

Not if, when.

Same thing you do when an ICE car sets a building on fire. Apply water and put it out.

Except for an ICE you only need to dump about 300 gallons on it. For an EV, it can take 24 hours to get the thing fully dealt with.

That's only relevant for the EV itself, not for the stuff it lit on fire.

You put out the fire, then dunk the EV in a tank when it's out to keep it from reigniting, while not having to monitor it.

Bring the regular truck filled with water to hose it down.

This is yet another good reason to decouple the battery from the rest of the vehicle. Hit the emergency release, battery drops out of the car, and push the rest of the vehicle away.

> and is only on the road because someone paid some mechanic off to look the other way during a state vehicle inspection for the past 5 years.

Making mechanics liable - maybe jointly as a business and an individual - for their failure to inspect vehicles properly would be a pretty simple mechanic to reduce this risk.

These inspections are nothing but pointless revenue generation for the state, checking for the check engine light and inspecting 1 wheel at most usually. Lots of states don't even have them.

In Germany they are very serious about vehicle inspections, and insurance requirements.

Didn't the big German auto manufacturers literally develop the hacks to bypass diesel emissions inspections? I'm not sure I'd hold anyone up in a pedestal on this score...

They got caught.

Not by any German regulator or agency.

Those have nothing to do with each other: Germany indeed has a company level corruption problem (especially in combination with the conservative party), but pretty low corruption at the citizen level, i.e. where vehicle inspections happen.

Those cars still passed the inspection. Which means they were physically sound. A battery isn't going to shift into an alternate mode where it becomes a spontaneous fire hazard.

That sounds more like an argument for effective inspections being universally required more than anything else? Apologies if I've misinterpreted.

We get by fine without rigorous inspections, and they amount to a regressive tax like a lot of other fees related to keeping a car.

Death by burning from an exploding battery from an uninspected, poorly-mainteained car in the garage of your shared low-rent apartment building is also a regressive tax.

I'll take the inspections.

It's not like states without inspections have an endemic of exploding cars as it is. I don't think the justification for inspections is really there.

Fuel-based cars tend to burn after an extreme adverse event, usually a collision.

Battery fires are more frequently spontaneous, or occur during charging.

How many barbequed children do you find an acceptable number?

Battery fires usually happen after collisions too. Fires while charging are rare but they grab headlines because they're novel.

I think trying to use inspections to solve that is a case of barking up a really stupid tree. In the case of EV charging fires there's not really much a mechanic can inspect that some electronics cobbled onto the pack couldn't. In the case of battery pack damage an annual inspection isn't going to do much good since most of those cases will cause problems before that or not at all.

"Parked Teslas Keep Catching on Fire Randomly, And There's No Recall In Sight"

https://www.thedrive.com/news/28420/parked-teslas-keep-catch...

Article list 9 cases. None involved accidents. Several were during charging, one during transport (on a flatbed truck).

All are of

In Ireland, roughly half of the vehicles going for the NCT fail on the first try, and ultimately about 5% don’t pass at all after retries.

Of that, I’d say half of the initial failures are simple things that don’t require a mechanic, (dead bulbs, bad/low tires) and the rest require a me hanic to do something.

My state has inspections, but does not charge for them.

Towing the car in a tank seems a sensible solution, but I'm wondering if it wouldn't be simpler to tow the tank near the vehicle, lift it and dunk it there?

It could be hard to get a car up into the tank... connecting a winch might not be possible due to the heat, but I wonder how they get around that to tow it?

Also, needing to plan for larger vehicles like a pickup truck would mean > 4,000 lbs of water to submerge one, base an approximate dimensions of such trucks & the corresponding volume of water.

Than again with batteries generally in the bottom of the car, maybe you'd only need to submerge a few feet?

I can definitely see the need for new emergency response equipment in the EV future.

> It could be hard to get a car up into the tank... connecting a winch might not be possible due to the heat, but I wonder how they get around that to tow it?

Seems like something similar in design to this would do the trick, maybe with sufficiently thick steel plating on the side:

https://www.youtube.com/watch?v=GaYN6yC4bnI

> The issue is when we're eventually talking about a 22-year old Model 3 that's had 13 years of deferred maintenance, is totally clapped out, and is only on the road because someone paid some mechanic off to look the other way during a state vehicle inspection for the past 5 years.

Except that the 22 year Model 3 also has a computer inside the battery that should be keeping things in check. Assuming the battery can even hold any charge at that age.

And said computer was likely receiving software updates and sending telemetry for most, if not all, of that time.

Or it have been hacked either to send fake data or no data at all... Knowing the general level of security in industry and even with some high end smartphones, I expect it to happen.

If it’s a Chrysler product it doesn’t need thoughts prayers or old age to spontaneously combust. A brand new one will sometimes so it all on its own due to bad design

I kind of hope that gets to be an (infrequent) problem.

As it is, I've lately been wondering about what happens to 14 year-old electric cars with dead batteries that cost more than the car is worth.

People like me buy them for $500 and put the motor in a grass kart

I had no idea what a 'grass kart' was, so I looked it up.

Cool stuff! And thanks for repurposing the motors.

Some countries have mandatory regular inspections for critical vehicle functionality like brakes. We will probably require EV battery checks like this as well, if the BMS / firmware is not enough to keep the battery pack safe.

GP covered that: "(...) and is only on the road because someone paid some mechanic off to look the other way during a state vehicle inspection for the past 5 years".

Safety inspections are close to worthless for safety on a macro level (they are great if you own a BHPH dealer or are a government regulator trying to give shops a reason not to fudge emissions inspections though).

You implement mild inspection laws and they do nothing because people can just pay the fine or whatever.

You implement invasive inspections where people are rich and they do nothing because rich people already don't drive on bald tires and whatnot.

You implement invasive inspections where people are poor and you create incentives to circumvent the law and hardship among those who don't ignore the law.

Mechanical failure is a negligible cause of accidents and injury so applying lots of effort co chase something that's only a source of the minority of problems is kind of a fools errand to throw a ton of societal resources at it.

If inspections really make the roads safer it would be an insurance thing and you'd see differences in premiums that reflect differences in law, insurers setting up their own programs where none exist and lobbying states for inspection programs. You basically never even hear a peep out of them about safety inspections unless it's in the context of getting people to upgrade to a newer car with better safety tech. Contrast with intoxication, distraction and driver training which are issues that insurers and their lobbyists constantly weigh in on.

For everyone interested, there is an interesting study to be found researching EV vs gas driven car fires [1].

"Nevertheless, existing test results have revealed that the heat release rate of EV fire is comparable to that of the fossil-fuelled vehicle fire, while EV fire may release more toxic gasses like HF from burning Li-ion batteries."

[1] https://www.researchgate.net/publication/338542510_A_Review_...

How many of those are with the car off and at rest ?

Is that actually relevant though, or just an implementation detail? Perhaps you could argue a moving car seems more potentially hazardous? But then that's only learned from extant behaviour anyway?

A car being driven is guaranteed to have a person behind the wheel. (For now...) A electric car in an attached garage spontaneously bursting into flame at 2 am is a recipe for burning a house to the ground.

In the former case there's definitely a person inside the thing that's on fire, and in the latter case there almost definitely isn't a person inside the thing (I guess once in a while somebody sleeps in their car but this isn't recommended) and that seems like a much safer situation.

Houses are replaceable, that's why people have insurance.

Humans are replaceable, that's why people have insurance.

Just kidding.

How is it safer for a car to catch fire in your garage while you're asleep? This is clearly the more dangerous situation.

Safer than being in a car which is on fire. That's what we're comparing here.

I still dont understand how that is safer. If the car is on fire, you can get out. I have been to many car fires when I volunteered with the FD and unless it's a wreck with entrapment, there's never any injury. In comparison, there are plenty of deaths from fires starting while people are sleeping.

Safer if you don't expect the person to get out when they see black smoke and flames coming from under their hood. Cars don't usually catch fire explosively.

The major issue in these fires is the lack of someone attentive. We can leverage the existing horn in the same way as anti-theft systems to fix that. We need a way to detect the issue, but I would hope there's a thermometer somewhere we can use to sense dangerous thermals.

Even more interestingly, the car could communicate with your smoke detectors to set off the fire alarm. It would be a great value proposition for something like Nest. It can let you know that there's going to be a fire before it even starts.

I would hope that honking the horn in case of fire is something we could achieve with a software update, but maybe not.

The pipe dream would be for the car to be able to get itself somewhere safe-ish in bad conditions. We're probably a good ways off from it, and safer batteries may take over first, but worst case scenario we may be able to program some kind of "escape path" into the cars. My car has the ability to save garage door opener codes. It's not too far fetched to be able to tell it to roll up the garage door and pull out into the driveway in case of potential fire situations. You might need an automatic ejector for the charger cable, but that doesn't seem insurmountable either.

I do agree that electric cars are more dangerous for now, and will continue to be even after the updates. We can narrow that gap with software, and even more with some integrated hardware.

It's not all doom and gloom for EVs, though.

So automated ejection of the warp core in case of a pending breach?

Well at this rate the refrigerator will text you if your house is burning down.

My dad's gas car (20 years ago) caught on fire in the garage while he wasn't there. The house was fine b/c of a a code required fire break between the garage and house. It's probably more rare nowadays, but not impossible.

Catching fire when no-one's around seems like it could go a lot worse than catching fire with a driver present.

I would much rather my car catch on fire when I am not in it.

Well I’d really prefer it not catch on fire at all but you get the idea…

There have been a couple of big recalls in the not-to-distant past for ICE vehicles spontaneously combusting. They have electric components too, which can short out and start a fire.

Saturn had a recall over generator fires when the car was off. When running the generator would over heat, when the car was off the residual heat would cause a fire since there was no cooling airflow.

My sister had her Chrysler LeBarron go up in flames in her parking garage.

Happens more often than you think

https://www.nfpa.org/-/media/Files/News-and-Research/Fire-st...

56 vehicle fires / billion miles driven.

15 vehicle fire deaths / hundred billion miles driven.

Reply on news.ycombinator.com