Ebike battery fire

[guerney]

Again there is scant detail in the article. They don't say which battery was used, which bike, which charger, how old the battery was, where it was bought, what cells and BMS it contained, what motor and controller the bike had and how they were set up etc. etc. Do they omit such details for the same reason they obscure brands of condiments and other ingredients in cooking shows?


Horrifying video shows moment e-bike battery erupts into huge fire as flames engulf house

8 March 2023

https://www.lbc.co.uk/news/lithium-battery-ebike-explosion-house-fire/

from video:

 

[guerney]

its one of the rules like 250lb 60v batt and 250mph max weapon speed most use custom rc batts and cost a fortune you can spend 50k on just building one of those things.

Project Thor is radio controlled, can I use:

Loading…

en.wikipedia.org

 

[soundwave]

no you have to start in the red or blue square in the battle box and cant spin up the weapon b4 it starts.

you can have a drone but total waight for both cant exceed 250lb /113.398kg.

 

[guerney]

no you have to start in the red or blue square in the battle box and cant spin up the weapon b4 it starts.

you can have a drone but total waight for both cant exceed 250lb /113.398kg.

Finally a battery explosion at 2m 04s! Maybe the other battery packs burst into flames later on, while being transported home after being brutalised?

 

[soundwave]

some have flame throwers when a batt gets hit the ones they use vent white smoke and burn to the ground.

wepon and drive use there own batts and controllers so one hit in the wrong place is all it takes.

 

[guerney]

some have flame throwers when a batt gets hit the ones they use vent white smoke and burn to the ground.

wepon and drive use there own batts and controllers so one hit in the wrong place is all it takes.

Would a very sticky microwave ray emitting battery exploder limpet bot be allowed?

 

[soundwave]

 

[guerney]

 

[soundwave]

robots have tits now even have rubber ones to change them

 

[Bikes4two]

Not necessarily.

Under a fault condition, there could be enough current flow to cause other parts to get hot and catch fire, but not enough current to melt the wire.

OK but I'm trying to garner from other's experiences what those fault conditions might be so that on my current (no pun intended) and next battery build, I can mitigate the occurrence of those fault conditions if possible.

And of course my original questioning was prompted by a couple of posts saying something (in relation to battery packs in bags rather than solid containerts) along the lines of "only one wire needs to short ..." etc etc to which I am at yet unconvinced.

 

[StuartsProjects]

OK but I'm trying to garner from other's experiences what those fault conditions might be so that on my current (no pun intended) and next battery build, I can mitigate the occurrence of those fault conditions if possible.

And of course my original questioning was prompted by a couple of posts saying something (in relation to battery packs in bags rather than solid containerts) along the lines of "only one wire needs to short ..." etc etc to which I am at yet unconvinced.

The originalpoint was made by @saneagle applies here, a battery bouncing around in a bag is bad, accident waiting to happen really, it has enough mass to damage just about all the wires in a typical battery\controller bag.

Wires only bouncing around is not so much of a problem if the wires are properly insulated and they would normally be secured also.

 

[Bogmonster666]

I don't have to rubbish self drive cars. Not just Tesla, they all destroy their own future by failing to perform remotely adequately.

Just come back from holiday. Was chatting to a German software engineer I met there who works on vehicle automation. I rather mischievously asked him when I could get a car to drive me back from the pub whilst I slept in the back seat. His answer was somewhere between a long time off and never. Reasonably well controlled environments like motorways are one thing, driving around a typical European city is quite another.

There is a lot of talk about AI today and LLM's can do seemingly impressive things, but 'general intelligent' AI eludes everybody at the moment. AI today isn't sentiment and can't 'think' and isn't well positioned for novel problem solving - and this is likely necessary in complex road environments.

Augmented / assisted driving is one thing, full automation is a very much bigger problem. One way to solve the problem is remove the variables - separate the pedestrians and the bikes from the autonimous vehicles. Separate the autonomous vehicles from the manual cars, allow autonomous cars to communicate directly with each other to share intent etc. However, how do you do this in a brownfield environment?

 

[flecc]

allow autonomous cars to communicate directly with each other to share intent etc. However, how do you do this in a brownfield environment?

Fully agree with your entire post.

Trouble with the quoted part though is hacking, the intent could be malicious.

Or the alternative, a car on each side of a single lane hump back bridge unable to decide or concede who goes first, so stuck there all day.

I don't think the AI protagonists have even begun to understand the complexity of the problem.
.

 

[Bikes4two]

The originalpoint was made by @saneagle applies here, a battery bouncing around in a bag is bad, accident waiting to happen really, it has enough mass to damage just about all the wires in a typical battery\controller bag.

Wires only bouncing around is not so much of a problem if the wires are properly insulated and they would normally be secured also.

I'm thinking we're not going to agree on this one buy hey, that's what forums are all about.

I have absolute confidence in my own battery build that (a) the battery isn't subject to any more 'bouncing' than the OEM cased battery I have, and (b) even if it were, the construction methods I've used (gleaned from experienced builders on YT like the fellah from EbikeSchool.com ) will mitigated the risks mentioned in these posts.

However, every day is a school day as they say, and I'm still open to other viewpoints. In the meantime I'm off to Fogstar for another batch of 21700's.

 

[StuartsProjects]

Do remember that most all wires\cables for stuff like brake sensors, displays, throttle, PAS, Lights etc will have an individual ground connection in them.

Now, after a few years spent in electronics design and production, I can envisage circumstances where another wire shorted to ground could cause something to overheat.

Now when electronic components overheat, they do normally fail, and when they fail they can sometimes fail in a way that causes large amounts of current to flow which in turn can cause other components to fail etc, etc, etc.

Trying to guess what might happen when indivdual wires touch or fail is not so simple.

I could recant the tale of the time a tiny wire only about 0.2mm diameter was fracturing (physical shocks) in power supplies and on two ocaisions came very close to burning down the factory where I worked.

 

[Woosh]

Fully agree with your entire post.

Trouble with the quoted part though is hacking, the intent could be malicious.

Or the alternative, a car on each side of a single lane hump back bridge unable to decide or concede who goes first, so stuck there all day.

I don't think the AI protagonists have even begun to understand the complexity of the problem.
.

The problem of communication for self driving car shouldn't be complicated. Look up CSMA/CD and /CA. The wifi is a case in point.

 

[flecc]

The problem of communication for self driving car shouldn't be complicated. Look up CSMA/CD and /CA. The wifi is a case in point.

Just wait 'til the cars get hold of it, will an Aston Martin give way to an Audi?

Seriously, it's hacking in that could be the big problem, organising ever bigger motorway pileups or head on crashes. It's just such a tempting target for the malicious.
.

 

[Bogmonster666]

https://www.theguardian.com/us-news/2023/jun/01/us-military-drone-ai-killed-operator-simulated-test

I agree malicious use is an enormous risk and AI is notoriously easy to manipulate. Take LLMs as an example and look at how hopelessly they are vulnerable to attacks like prompt injection:

Prompt Injection Cheat Sheet: How To Manipulate AI Language Models

A prompt injection cheat sheet for AI bot integrations

blog.seclify.com

And promt injection is just the start. Feeding LLM AI with biased data intentionally is going to be a massive problem. Imagine a state sponsored AI churning out misinformation into regular media outlets and online content that is then hoovered up by the LLMs that everybody else is using and you can massively skew and bias the models. AI is even more susceptible to propogandavthan humans.

I'm not an AI expert and the above examples relate to generative AI, but all AI is struggling with similar problems.

 

[Woosh]

Just wait 'til the cars get hold of it, will an Aston Martin give way to an Audi?

Seriously, it's hacking in that could be the big problem, organising ever bigger motorway pileups or head on crashes. It's just such a tempting target for the malicious.
.

computer viruses pose a similar problem.
I remember well looking everywhere on every machines for traces of the Stoned boot sector virus in the late 80s on the early DOS machines. Back in those days, the technique was scanning the floppy discs for a particular string. Now viruses are much more pernicious, the antivirus measures are more AI led, spotting bizzare out of place pieces of code. Viruses evolve giving rise to multi billion pound computer security industry.

 

[Woosh]

Imagine a state sponsored AI churning out misinformation into regular media outlets

we have already twitter and FB bots. They still can't replace lying politicians though.