Hi I'm building a project that uses a kanthal heat sealing wire. The wire is on for about 500ms. It uses an external 12v power source that's controlled by an Arduino using a relay. The problem I'm facing is more of a safety concern. I want to make very sure that the power is not on for longer than 1000ms, regardless of if the relay is on or not. I'm thinking that a bug or malfunction in the Arduino could potentially leave the wire heated long enough to cause damage or maybe even fires, and I'd like to prevent that with some sort of simple hardware solution. So my question is if there is any such component as a circuit breaker that can break the power after a certain amount of time? It would of course not have to be connected to the Arduino, since this would be a safety device incase the Arduino fails. Any suggestions much appreciated Thanks
Slow blow fuse ?
Connect Arduino output to edge detector, edge detector to 555 monostable which in turn is set to produce a 500 ms pulse.
You will have a problem if the Arduino retriggers frequently, or if you want to allow for an output that lasts a bit longer or shorter than 500 ms. To allow for all that, you need a microcontroller, such as an Arduino, to read the signal and produce a 500 ms output with a certain minimum time in between two of those 500 ms pulses.
UKHeliBob: Slow blow fuse ?
Is a show blow fuse slow enough? Maybe a polyfuse with optional heatsink (*) can be tuned more precisely?
(*) I mean for added thermal mass, not for cooling.
Karma for giving this issue serious thought. That's wise.
I would consider swapping a MOSFET for the relay. Relays have nasty pathologies, such as contact welding. Uncommon, but it does happen and should be considered if the consequences are dire.
I'd consider crowbarring the heater power to ground, through a fast fuse which will blow. The crowbar would need to be controlled by a completely independent time delay circuit, triggered by application of heater power and activating the crowbar after 1 sec. I'd use a 555 (as suggested by wvmarle).
Finally, a slo-blo fuse should also be present, more as a backup than the primary protection. The time constant might be difficult to control. S.
How much effort are you willing to put in to accomplish this ? Everything you need to know is in this book. With almost no electronics experience, I used it to design and build this controller for a VIP lighting project: FYI, note the small white wire labels on every wire. This controller consisted of about 30 timers , both astable and multivibrator, in addition to analog switches. At the time I had only been employed in the capacity of an entry level electronics test technician testing 5 1/4" floppy disk drives (for the IBM PC) at Tandon Magnetics. The lighting controller technical consultant was a side gig. The controller was for the Los Angeles Museum of Fine Arts Contemporay Art Council Awards Banquet (Tribute-82) with Leonard Nimoy and William Shatner as guest speakers. The lights it controlled were a full size flying saucer landing special effect designed by the late Lighting Designer Sam Salde, famous in Hollywood for VIP lighting at the time. (did Frank Sinatra's Birthday Party , Ronald Reagan's Inauguration Banquet, Aron Spelling's wife's birthday party, the Acadamy Awards Banquet (two years in a row), the Mafia New Year's Eve Party at the Colosseum Room at Ceasar's Palace (I worked that party) and many more. Everyone who was anyone in Hollywood used Sam for their parties. (William Shatner autographed my copy of the IC TIMER COOKBOOK)
Is 555 more reliable than a MCU running a simple code?
Is less likely for MOSFET to fail shorted than for a relay?
Another concern: the more complex the system, the higher the chance that something fails, for the simple reason that there are more things that can fail.
Smajdalf: Is 555 more reliable than a MCU running a simple code?
Is less likely for MOSFET to fail shorted than for a relay?
Good questions. I chose a 555 precisely because it is not an MCU and therefore should not be susceptible to the same failure modes. I chose a MOSFET to eliminate a failure mode known (and which I have experienced personally) for relays. But your point is well taken -- as wvmarle also observed, the trick is to avoid simply adding complexity without actually enhancing reliability. S.
srturner: I chose a MOSFET to eliminate a failure mode known (and which I have experienced personally) for relays.
I've never experienced a relay fail, but rarely use the things, MOSFETs (for DC - not tried yet for AC switching) and TRIACs (for AC) are my preferred switches. I have experienced MOSFET fails. Usually they failed open, but also have had them fail closed. I must admit that those fails were pretty much all due to my own mistakes (such as overvoltage, improperly placed flyback diode, reverse voltage).
What I would do: Arduino and safety device both control load. Both must turn it on for power to be delivered. 1) Arduino tries to power the load and checks it is not powered: this confirms control device is able to turn off the power. 2) Arduino asks control device to turn on the load. Check the load is now powered (meaning Arduino can detect the device is powered). 3) After required pause Arduino turns off the load. Check if it truly lose power (ensuring Arduino's switching device is also working). 4) Control device turns off its switch some time later as it wishes.
If Arduino goes crazy it can turn on the load permanently and keep triggering the control device quickly. There should be a lockout period in the control device to prevent this!
wvmarle: I've never experienced a relay fail, but rarely use the things, MOSFETs (for DC - not tried yet for AC switching) and TRIACs (for AC) are my preferred switches. I have experienced MOSFET fails. Usually they failed open, but also have had them fail closed. I must admit that those fails were pretty much all due to my own mistakes (such as overvoltage, improperly placed flyback diode, reverse voltage).
Relay contact fusion and coil magnetization was specifically what I was thinking of. I've had it happen, but not recently, as I've also tried to avoid using them in new designs. Relays also feature an inductor which can produce voltage spikes. Those are supposed to be mitigated by the flyback diode, but why not just avoid the issue altogether (unless there is some other concern, like a need for isolation). I've also had triacs fail closed on several occasions. It was disturbing and the cause was not always evident. This was a long time ago, so maybe modern devices are less prone to this. I had thought that MOSFETs would usually fail open, but I could be wrong about that... S.
Your actual concern appears to be over the temperature, not time, so why not have a thermal cutoff instead of a time-based one? Doing this indirectly with a time metric is only going to cause problems. Suppose you leave it on for 900 ms, stop it for 1 ms, then turn it on again. How long can you allow it to remain on? What if it was one for 400 ms, how long does it have to stay off to be safe? That’s not so easy to answer.
If you’re worried about safety, directly monitor the thing you’re worried about instead of a nonsense indirect metric.
Sealing plastic bags? Perhaps you should use a current regulator to adjust or limit the maximum temperature of the sealing wire. Any more information on your project?