fuse alternative for a solenoid?

[meeotch]

I'm going to be controlling several (15 or so) solenoids using an arduino, and I'd like to protect them from getting stuck in the energized position - e.g., if the arduino crashes or something shorts out. I can't do it with a single fuse on the power supply, since there's no telling how many may be on at any given time & thus the total current is unpredictable. I could fuse them individually, but I'm trying to avoid that to save space/time/cost. Also, it's difficult to size the fuses, because the inrush current is much higher than the steady-state current when the coils activate. (The coils may pull a couple of amps each when first activated, but only one amp when held in the on position.)

Is there some sort of alternative device or circuit that will act like a slow-blow fuse? That is, one that will react to the *time* that coil has been energized, rather than only the amount of current going through it?

I thought some sort of R-C circuit might do the trick, but these are AC coils, and my head exploded when attempting to puzzle it out.

Power source is a 120V to 24VAC transformer, and the coils are the only things drawing from it. So the ideal solution would be one that sits next to the transformer, and cuts the circuit if it draws 1A for more than, say, 30 seconds - but will not react to a draw of 10A for less than a few seconds.

 

[jiripolivka]

There are some overcurrent protection devices made by Rayovac. They pass e.g. 1.0 A current but "open" when the current exceed 1.5 A.

---------- Post added at 09:53 PM ---------- Previous post was at 09:52 PM ----------

They operate by heating a resistive compound, and they reset once the current drops.

 

[meeotch]

Thanks for the reply. Are you referring to PTC's (positive temperature coefficient devices)? I had wondered about these... The thing is, the condition I'm trying to avoid isn't strictly an overcurrent condition. The current will be at the normal operating current of 1A * the number of energized coils. It's the amount of time spent in this state that I need to limit.

If the PTC (or other device) has a super slow, cumulative response then it might work. That is, if 5-10A for a few seconds would not trip the device, but 1A for 10x that amount of time would trip it. On the other hand, if it simply trips at 1.5A, then I'd have the same difficulty as with a fuse: fast-blow would not work at all, due to the inrush current; slow-blow would work, but I'd have to fuse each coil separately, and the tolerances would be tricky. (For comparison - I tried a 2A slow-blow fuse already, and it correctly held during the inrush, but the 1A steady state current wasn't enough to blow it, no matter how long I kept it energized.)

btw - usage on this application: the coils will be firing in groups of 1 (min) to 6 (max), repeatedly for probably 10sec or less. Followed by a cool-down period that's likely to be at least 30sec or a minute.

 

[FvM]

A simple hand calculation clarifies, that your specification can't be met by a thermal fuse. They are basically acting on averaged I² (RMS current) respectively I²t for short times. This means 10 A has a hundredfold effect of 1 A, so the trip time for 10 A would be shorter (at least, cause there's also a threshold effect) by a factor of 100.

Normally, people manage circuit protection with thermal fuses (or circuit breakers) though, but if you keep your requirement strictly, it can't work. The somewhat vague idea of using a RC circuit isn't bad at all, but has to be translated to a real circuit. For the said reasons, it doesn't work with a pure thermal fuse design, circuit breakers with mechanical (e.g. pneumatic) delay are in use, but an electronic fuse is probably the most universal way. Simply think of a current detector with a threshold and a relay, acting with a delay.

 

[jiripolivka]

I am not sure the exact mechanism of those Raychem current protectors, but they do heat up to trip "open". I think you would have to get some samples and experiment with some.
One possible approach to your time problem may be to use either a heat sink to extend reaction time, or a thermal insulator on the device to shorten it.

You can certainly develop a circuit with all flags and whistles but it would be much more complex. Adjusting time intervals of minutes can consume a lot of time. You can use a current sensor and temperature sensor, or a combination with a timer. I prefer simpler things if possible.