I think this is a product with simply not enough testing.
Makes sense - at the time they were a hardware startup needing to get a hockey stick sales graph to please investors, so I'm sure the engineers were pushed to make a minimum viable product and then fix bugs in software. Then with the increased sales and Google acquisition there was no time to fix the bugs because they had new products to work on.
End-of-life and/or battery warning for smoke detectors is a solved problem at least amongst established manufacturers serving the European market. Once this mode is triggered the detectors emit a mildly irritating chirp every 30 seconds or so.
Having a safety critical system simply stop working is very bad design and should be punished by the relevant regulator. The design of a system has to take in to account the fact that the components are going to time out and do something reasonable for maintenance. In the case of something like a fire sensor there has to be a "replacement required" interval at the end of life where the system still works and the sensors can be swapped out one at a time.
What surprises me is how people think it seems like a reasonable idea to electrically connect catastrophe detection devices together to ensure a common point of failure to begin with, anyone care to tell me how that's supposed to be more secure than each device being disconnected, offline and entirely self-reliant ?
seems like a good way to introduce additional failure modes, hopefully the law also mandates that the alarms be made in such a fashion that the individual alarm will function regardless of whether the "smart" part of it is getting hit with 200 kV over the wire, or if the smart part is wireless but the radio crashes due to.. I don't know.. overheating issues?
To be clear, this has already been a requirement in the UK as a whole for any new or modified construction for quite a while. This new regulation mandates installing a similar system in existing homes.
This is common and standard practice. It's quite possible that an alarm positioned at the other end of a house, with several closed doors in between, could easily fail to wake someone up. The goal is then to ensure that if one alarm triggers, they will all do so. Since they'll already likely be connected to the same electrical circuit anyway, this usually takes the form (in the UK at least) of a third core that connects the alarms together – and is a very simple system.
I won't speak to the "online" nature of any detectors, but I would expect standards mandate that they function fully and independently by design regardless of connectivity.
If it's mains powered, there's your opportunity to fry everything during the same episode that's likely to start a house fire (if not from the actual fire alarms bursting into flames in a fit of cosmic irony).
If they're linked by radio, hopefully the circuitry is independent of the operation of the basic "detect smoke -> make noise" circuit, so for instance, when your radio shorts out, it does not blow out the power supply to the detector and noise making mechanism.
Yeah, it's too hard to design this to trust that commercial options are safer than dumb alarms.
20 comments
[ 3.2 ms ] story [ 65.9 ms ] threadAnyway, it is exact same thing as the submission but actually caught on camera and posted on youtube.
Makes sense - at the time they were a hardware startup needing to get a hockey stick sales graph to please investors, so I'm sure the engineers were pushed to make a minimum viable product and then fix bugs in software. Then with the increased sales and Google acquisition there was no time to fix the bugs because they had new products to work on.
Those are not safe to use and so they selfdestruct for replacement.
Feels like an over-reaction to me, especially in response to a disaster that killed "only" seventy people, or a third of a day's UK COVID deaths.
I won't speak to the "online" nature of any detectors, but I would expect standards mandate that they function fully and independently by design regardless of connectivity.
"Interlinked" doesn't necessarily mean there's a wired connection, the basic ones are triggered by the sound of another alarm.
If they're linked by radio, hopefully the circuitry is independent of the operation of the basic "detect smoke -> make noise" circuit, so for instance, when your radio shorts out, it does not blow out the power supply to the detector and noise making mechanism.
Yeah, it's too hard to design this to trust that commercial options are safer than dumb alarms.