Ionisation Smoke Alarm Dismantled - What's Inside
HTML-код
- Опубликовано: 24 июл 2020
- What's inside an ionisation smoke alarm.
This example is a 230V powered item with 9V battery backup. Originally installed in 2008 and labelled for replacement in 2019.
► Support this channel:
Patreon: / jwflame
PayPal Donations: xo4.uk/?PPP
► Social Media:
Twitter: / jwflame
Facebook: / jwflame
Instagram: / jwflame
► Contact info, sending stuff in: etc.:
xo4.uk/?YTT
► Website - More on this video and many other subjects
Website: flameport.com 
Наука
I’m working on these exact alarms right now in Ei Electronics, Shannon Ireland and I must say it’s really interesting to watch someone else disassemble something I’ve put together 🤣
Love your videos and presentation style JW, you are so relaxing and informative - could watch for hours. Thank you!
John, you need a remedial class or a more in depth class in health physics. Radiation damage is a stochastic phenomenon. Am241 is primarily an alpha emitter but it ALSO emits gammas at 59.6 KeV and 26.3KeV peaks in addition to lesser amounts of gammas other energies. And what about the X-rays? When those alpha particles impact the electrodes or even the carrier substrate, the law of conservation of energy says the energy most go somewhere so the abundant kinetic energy of the alphas is transformed into Bremsstrahlung X-rays which are too low in energy to be seen with a Geiger counter but are EASILY visible to a scintillation counter. Safety councils have decided that the very real danger of the radioactive material is outweighed by the very real danger of a potential fire, it's not that the source is actually free of danger to be in the vicinity of. Again, radiation damage is a stochastic phenomenon and "safety" is a numbers game.
I used to design alpha and gamma scintillation spectrometers and I've watched the radiation from a 0.91uCi Am241 source turn white plastic brown with radiation damage. That typical detector source was 34,000 Becquerels at manufacture as your source is in this video, meaning that the material when new will experience 34,000 disintegrating nuclei per second with each one proving emission. That's not a small number. While the alphas are stopped in a few centimeters by interaction with air molecules or immediately by the plastic and metal components of the detector, the resulting low energy Bremsstrahlung X-rays are both penetrating and perfectly strong enough to damage DNA. John, I would HIGHLY encourage you to invest in a halfway decent used scintillation counter so you can with proper techniques (you'll need a "lead castle") be able to quantify the things you can't see before making too much commentary about them.
Sounds like you know your stuff would you say it's better to get rid of it? Erm I kind of smashed the last one when we were renovating and the dist kept triggering it I truly believe that wasn't the best idea.
I have always wondered is the new smart meter emitting any sort of radiation maybe you have had any experience in this field?
@@Naughty_Squad I can't tell exactly what you're asking but if you want to keep one of these radioactive sources removed from a detector for a test source, just keep it in a well marked, sealed container inside a metal can and don't store it near where people, pets, or animals will be. No worries. The problem comes when people store these things in their workshops where they spend time or bedrooms where exposure adds up. In a detached garage or a shed it's a non-issue. If you want to be totally safe with it, look up "lead pig" which is the standard name for a lead shielded storage container. They used sell small ones on Amazon. With a pig you could keep the source in your bedroom without any potential issue. It doesn't take much shielding to render it very safe.
Smart meters are something totally different. They use wireless radios that put out strong radio pulses every once in a while just like your cell phone, but using less energy than your cell phone. Think of them as cell phones that make a couple 1 second calls per day.
Hi Peter, you seem like you know a lot about the science of this subject. Could I ask you about a query I have?
I remember i used to take everything apart to harvest electronics, i had no ides radioactive things were in smoke alarms and the building I worked in was throwing them out so I opened one up to see what's inside. I don't remember any radiation warning signs in it, pretty sure I would've remembered that. Whats the chances that it could've been a wired by the mains ionisation smoke alarm in 2014 UK?
If so, what are the chances the chamber had no warning and I would've mistaken the amercium pellet for a screw and tried to unscrew it this releasing alpha particles on a screw driver I would've put back into my pants pocket?
Further more; in this unlikely event, what's the chances i'd still be alive with no cancer or sickness?
Classic JW quote alert, “As long as your not going to be shoving it in your body and, if you were to grind it up and inhale it, there’s no real risk here!!!”🤣🤣🤣👍
He's been to the internets enough to know people shove thing in their bodies all the time :)
The chip is actually really interesting. It saves energy by powering the system down most of the time: only checks for smoke for 10ms every 1 and 2 thirds of a second. Every 40 seconds i checks for a low battery. And it does the interconnect comms : allowing up to 125 devices to be interconnected.
Everything else is just support for the chip, or the powersupply, which looks like a capacitive dropper.
Thanks John 👍
Thanks for sharing 👍
0:25 Guinness Book of Records attempt for the most mounting holes possible in a Smoke Alarm
I ordered an ion chamber from an old smoke detector (older than yours, I'm sure). It arrived today, so I tested it with my Geiger counter (Terra-P +, by Ecotest). Nothing could be detected through the chamber housing, even with the cover removed from the Geiger tube, inside the Geiger counter.
I took the cover off the ion chamber and tested it again. With the cover on the Geiger tube (that's inside the Geiger counter), still nothing showed up, but when I tested it again, without the cover over the Geiger tube, I got a reading of 2.4 micro Sieverts (it varied wildly, but that was the maximum reading I got).
For reference, the normal background radiation here is about 0.17 micro Sieverts. The alarm goes off when it reaches about 0.28 micro Sieverts (or close to that).
Nice video
You said it suffered no apparent damage from 1000V, and this is true. But...
Yeah, a conscientious designer might come up with something that can withstand 1000V just in case somebody sticks a megger on it. It's a selling point, even. Until the production engineer takes a look and sees that the circuit could be simplified, or cheaper components used, and it still handles 240V fine. But maybe there are no cost savings to be made that way, so no change to the design, so it can handle 1000V. Until management learns that it can withstand accidental megger application and realizes that more units will be sold if it *cannot* cope with 1000V.
But maybe this design wasn't subjected to management and it really can handle 1000V without degradation. That doesn't mean every such unit on the market can. Maybe it's only the ones from this manufacturer. Maybe it's only that particular model from this manufacturer.
Most likely of all, though, is that voltage ratings of some components have been exceeded and their lifetime has been reduced. Putting a megger on it means it will fail sooner than it otherwise would have. Maybe you even got lucky and another unit of the same design from the same batch would have failed instantly.
The moral: *never* exceed the ratings. Not even if you saw somebody on RUclips do it and get away with it.
Smoke alarms do seem to have good long-term reliability, I recently replaced one on the upstairs landing, (mainly because the originally white plastic case had become very yellow) which had been in place for about 30 years, just a standard battery powered one, still worked when I tested it too!
I bet the optical battery powered ones usually work for donkey's years because there isn't really anything there to wear out unlike these pieces of rubbish with a radiation source and a failing mains PSU. I don't remember ever seeing a faulty one in my house no matter how old they are.
Oddly enough I have these exact alarms. I also have a Triton Cara shower, the model which Clive dissected a few years ago.
JW, I'm therefore guessing you and Clive once worked together on a block of studio flats in Anfield...
Well covered, still I was hoping you would go a little into the networking bit and how it's actually using mains to power a 9v chip and the battery changeover on power loss.
You can find smoke detector chip datasheets online.
It's a rather spesific part made to barely use power until it needs to be loud. How the big change in power consumption is handled on the supply side could be interesting.
Great vid John. 🐵
I was hoping you would look for the source of the "cracking" sound when you applied 1000V earlier. Obviously this device uses a capacitor-dropper on the AC side, so chances are the capacitor got damaged or maybe some arcing?
Probably arcing made the noise. Probably also capacitor degradation with holes punched through the plastic film. Self-healing, but the capacitance drops with each punch-through.
@@bdf2718 That could well be the biggest failure mode of these things. That cap dropper is going to age and drop below the C needed to power the device and perhaps the megging will decrease the lifetime of the power supply below the expected operating life even if it still works initially.
The chip is from Allegro Micro Systems: www.allegromicro.com/-/media/files/datasheets/a5348-datasheet.ashx
Interestingly, code here requires dual-sensor smoke detectors. Doesn't stop local retailers from selling single-sensor ones though. And some of the fancier ones (like spoken alarms) are only available in single sensor versions, so I compensate by making sure all the others are dual and are interconnected. (One of the two kinds is preferred for kitchens as less prone to false triggering, don't recall which)
Pet hate: Posts that refer to "here" without saying where "here" is.
The sensor is wired directly to the bent-up lead of the chip to avoid leakage across the board introducing noise. The two adjacent pins are probably zero volts to reduce leakage across the package of the chip.
You could avoid the bent lead by putting a guard rail around the sensor input, but then you'd need a double-sided board because you need the guard rail on both sides. It's cheaper to have a single-sided board and bend the pin. Slightly greater assembly time, but presumably still cheaper than going to a double-sided board just for that.
Bending the leg of the chip is still a bodge though and not the correct way to do it. Its the lazy way. They just saved 0.00000001 pence by not needing a short link wire.
@@simontay4851 No, they *had* to bend the leg to keep it away from the board. Leakage.
Short vertical link wire. Soldered at both ends. Yeah, that's gonna be easy. And unnecessary.
There is someone in China who's only day job is bending millions of pins for these at the factory :)
@@simontay4851 The datasheet most likely tells you to do exactly that to that pin. It would not surprise me if the chip is delivered with the pin bent that way from factory.
@@simontay4851, you seem to be affected by Dunning-Kruger effect in this instance. That design has to do with performance and is very intentional to achieve that end.
Is the Americium a solid or an encapsulated powder? Any risk of a powerful compressed airblast accidentally dispersing it?
MY pet hate is the battery types ! test button just proves the button works, battery has some life in it ! and that the sounder is still operable !, but carn't prove the actual detector is still working !! , general public need to be educated on how to correctly test the equipment installed in their home don't you think ? , ie Smoke or CO2 spray ,
Brilliant informative videos though !! , keep'em up !! best wishes Tartan Sparky.
So no signs of the 1kv test ,crackling sound withing this?
What voltage would you expect to get in alarm on the interconnect with respect to neutral?
As it still needs to work with the mains power removed my guess would be the interlink wire is 9vdc wrt neutral. Its certainly not mains voltage.
Most interesting part is the use of a 14? Pin DIL micro controller rather than an 8pin or black blob.
Dedicated smoke alarm IC not uC and they've been around for decades. Others have posted a link to the datasheet for the device used here.
Smoke detector chips are very specialized low power chips that pre-date microcontrollers. A microcontroller is useless baggage in this application. Google "allegro A5348 datasheet" for more.
@@Peter_S_ I've just dived into the library and dug out the Motorola European CMOS Selection databook from 1979. It has the five smoke detector ICs in it: MC14461 -2, -4, -5 & -6.
Datasheet for the chip. It's an Allegro A5348CA pdf1.alldatasheet.com/datasheet-pdf/view/55126/ALLEGRO/A5348CA.html
I remember i used to take everything apart to harvest electronics, i had no ides radioactive things were in smoke alarms and the building I worked in was throwing them out so I opened one up to see what's inside. I don't remember any radiation warning signs in it, pretty sure I would've remembered that. Whats the chances that it could've been a wired by the mains ionisation smoke alarm in 2014 UK? If so, what are the chances the chamber had no warning and I would've mistaken the amercium pellet for a screw and tried to unscrew it this releasing alpha particles on a screw driver I would've put back into my pants pocket? Further more; in this unlikely event, what's the chances i'd still be alive with no cancer or sickness?
Most smoke alarms in 2014 were ionisation types.
The amount of radioactive material is tiny, and alpha particles are blocked by skin, paper and other thin materials. Unless you made a special effort to eat the radioactive substance or ground it into a powder and breathed it in, the risk is essentially zero.
@@jwflame thanks for the reassurance John, this has been a concern of mine for years. So Americium is not brittle enough to break apart when scratched by one of those steel hexi screw drivers? It would have to be grinded to break into a dust?
Data sheet for the chip is here www.allegromicro.com/~/media/Files/Datasheets/A5348-Datasheet.ashx
Buy optical smoke alarms. They detect fires better
I think these smoke detectors are not as good as to the photoelectric ones i possibly due to not having proper air circulation into the chamber.
2:28 more like Amɪˈrɪsiəm?
More like Am-er-rick-ium. It's named after the Americas. Unless you call them the Amerissas.
Smoke 💨 Alarm 🚨, This time we’re going to actually take this apart and see what’s inside, mainly because we’ve got it here so we might as well!!! Woohoo 🙌🏻🥳 JW!!! Smoke 💨 Alarms 🚨 Are Fun To Mess With & Also Really Handy For Those Days When Projects Just Want To Be Sparking ⚡️ & Arcing ⚡️ And Starts A Small Fire 🔥!!! Oops 😬 ⚡️🔥, My Bad 🤗!!!!
So cheaply made infact that they didn't bother to populate the MOVs and transistors that are supposed to be there. Whats the point of designing the PCB to have those components if they're never included. They would cost 0.00000000001 pence. The battery is probably always connected and the 9V from the battery and the mains is probably ORed with diodes.. I see no way that it could be only connected when the alarm is connected to mains - there are not enough pins.
Bying a MOV and installing it on a board will inevitably cost more than 0.00000000001 pence. Maybe closer to 0,1-1 and that very certainly adds up when you make them times 10 million or more. I'm pretty sure that the designer and manufacturer of this product had a better cost estimate for his BOM than you do and had more to go on with their decision than a random internet stranger commenting on a random internet video. Pretty much every product you will find has a circuit board inside has unpopulated component locations just because they do cost-effectiveness calculations and do several different models of the product with the same board, as stated in the video if you cared to listen to it.
Different model or market. It wouldn't be the first time a board has been designed with build options.
PCBs can also have unpopulated component positions in cases where the designers weren't sure if a given component would be required for some application or other, and it was easy enough for them to add a component footprint that can be populated if it ends up being needed, rather than forcing them to update the design and have a new batch made.
After all, adding an extra footprint during the design stage is effectively free (assuming it doesn't increase the overall size of the board, cause complications elsewhere, etc.); having an updated batch of boards made very much isn't :)
Things like that may be eliminated in future design revisions, but not always: why risk breaking the design and necessitating more testing if removing the unneeded footprint isn't required?
@@AndrewGillard Way back before '86 when I left the company concerned, Sales were given the chance to completely spec a product after arguing it was them that had to sell it and they knew what the market wanted. However we in Development realised they'd got it wrong, put in the required components as a build option and then kept quiet.
At the prototype stage Sales somehow found out what we'd done and insisted the PCB was redesigned without our build option. That's how it went into production and as expected it didn't sell - customers wanted what Sales had claimed wasn't needed.
We had to design an add-on board for existing stock and future products were specced the way they'd always been, by Development and the Tech Director.
[rolls eyes] That PCB is able to be built as either an ionisation detector or an optical detector. In the factory it's called a build option. It allows a single PCB and housing to be certified as a safety device and allows the maker to build whichever version they want under that single certification.