new invention?

[T3sl4co1l]

FYI, you don't have to write short replies; this isn't chat, and it isn't Discord.

The basic thing missing from your scenario is: how does your fuse know what voltage it's at?  Voltage is a difference between points, and the voltage drop across a practical conductive fuse must be normally small so its power loss is small.  All the bits of wire know is how much current is flowing through them; they could be at 0V to their surroundings, or a million!  There's no way to tell what voltage is across the actual load, so it can't possibly control that voltage.

In fact, this concept can be used to "game the system", say to test equipment used at high power.  Consider a monitoring system used on medium-voltage distribution lines: it might have 14kV AC on it normally, and sense currents over 100A.  That's over 1.4MW of continuous power!  (2.4MW even, since it would be three phase.)  How do you test that without your own power station?  We can apply the 14kV AC to the voltage-sensing terminals, and loop a different wire through the current sensor -- well insulated so it doesn't care about the 14kV nearby, and can be sourced from an ordinary transformer or whatever.  Typically such a system would use a magnetic current sensor (current transformer (CT) or Hall-effect), so that a wired connection isn't required to sense current, just a loop around the wire being sensed.  The voltage source then only needs to deliver ~mA, and the current loop ~mV, so the total power dissipated by the test equipment is small (~W?), but the EUT (equipment under test) reads up to full power.  You've successfully fooled the system!  The catch is, the current being measured is not, in fact, part of the full loop across the 14kV, it doesn't go out and return through the two points being measured across -- and the EUT can't know that, it just assumes you've wired it the right way.  But in fact you've made a sneaky loop around it, so you don't need to dissipate anywhere near a megawatt to test it.  Handy, eh?

Which also means, to make such a fuse, it needs to be wired correctly.  If it's sensing voltage, but switching current, it must be a three-terminal device.  There's no guarantee that the user wires it across the appropriate three connections, but you would document your device accordingly of course.  But anyway, we need to somehow construct it to do that.

You could make such a thing by simply putting a heating coil around the thermal fuse bit: the more voltage is applied, the hotter it gets, until it melts and opens.  This is a very slow-acting and crude method, but just to hand-wave a possibility.

There is also another interpretation to your question.

In the same way that a "current-sensitive fuse" operates (when current goes too high, current is forced low and voltage high), presumably a "voltage-sensitive fuse" operates exactly inversely: when voltage goes too high, voltage is forced low and current high.  That is, instead of blowing open, it "blows" shut; an anti-fuse!

And, note that an anti-fuse would be a two-terminal device, applied in parallel to the load, exactly opposite of a fuse applied in series with it.

And indeed, such a component is useful for protecting current sources, where the open-circuit voltage could be dangerously high, and shunting it to ~0V significantly reduces the power output of the source, just as open-circuiting a voltage source (what are commonly in use) significantly reduces its power output.

The deeper truth here is: circuit theory, at least, is symmetrical around voltage and current, that is, we can swap them around, and exchange series and parallel connections, and get an equivalent system.  We can use fuses in series with voltage sources, or antifuses in parallel with current sources.  This is also where the Thevenin/Norton ideal source equivalence comes from.

(As it happens, there are physical reasons that prefer voltage sources, and we typically have to go out of our way quite a bit to make a good and proper current source as such, but circuit theory at least makes no such distinction, and we can draw circuits that work equally well with CCS or CVS.)

Tim

[Zero999]

oh, I get what you mean. you're saying that the circuit will break before the fuse "blows"

No, just that all voltages are relative and it's the difference between two voltage which is important, rather than the total voltage.

Have you every noticed that birds can safely sit on high voltage electric wires, without getting zapped? It's because the voltage at all pares of the bird's body are equal, so no current flows through the bird.

The situation is similar for a fuse which hasn't blown. It only sees a tiny voltage.

The total voltage across the circuit is only important, when it blows, which is why the fuse has to have the correct voltage rating.

[jonpaul]

May be not new.... search patents from Littlefuse, Buss, etc.

Gallium: Costly, even in small qty. Will not work over comm or induystrial tem range0-70  or -25..+85C

Patent: Expect $20..100K cost, about 5 yrs from stat to grant in USA.

j

[ejeffrey]

I believe the normal technique is to use a Zener diode to turn an over-voltage into over-current and blow the fuse. Alternatively, use the Zener to trigger an SCR for larger current ability.

That's a bit old fashioned. It still works in many cases of course but causes a lot of stress on all the components that need to supply the short circuit current.  It doesn't work well with many modern  power supplies that are designed to limit current internally.  In this case the power supply may fail to clear the fuse in an acceptable amount of time and just end up driving it's short circuit current into an SCR indefinitely.

These days its generally better to use a MOSFET to just disconnect the load when the input voltage is out of spec.  A lot of equipmemt will already have an input MOSFET for inrush limiting, hot plug capability, under voltage lockout, and reverse polarity protection.

TI for instance makes a line of eFuse ICs which when paired with suitable MOSFETs provides all these functions.

This is specifically when you are trying to prevent moderate over voltage from destroying electronics.  If you have a surge which also takes out the pass FET it can still fail. If you need to protect safety there is more to it.  You also still probably need a real fuse somewhere to handle genuine over current events and avoid starting fires.  But it can then be sized for preventing fire risk without worrying if it will play nice with your overvoltage protection circuit.

[Infraviolet]

Gallium is expensive, comparably, and it can damage metals with a corrosion like effect (particularly steel and aluminium, Mercury, the other room temperature liquid metal, also attacks them similarly). This seems functionally equivalent to a fuse, but with added complications and cost.

The one advantage which this gallium fuse has, over a normal fuse, is the ability for a human operator to collect the resolidified gallium and put it back in place for re-use without needing to swap in a spare unit.

But where you want to be able to replace a fuse, with manual intervention but no need for spare parts, there are already electromagnetic coil and relay based breakers which can be reset. Or even polyfuse "fuses" which self-reset after a cooling period.

[coppercone2]

just keep in mind that thing with MOSFET open is for a regulated supply protection. not the mains, not a crazy generator , those sources have no mercy. Because a power surge like close call lightning might fry a MOSFET and then the normal power restored goes right through. its really not nearly as robust as any kind of fuse.

but that limits your area of use by alot, because there are many regulated things that need protection still.

[alligatorblues]

Complexity adds higher failure. My cousin is a materials engneer. Materials seldom retain the same properties when run at different ambient temperatures, different types of enclosures, different geometries, etc. But there are desireable applications for fuses in which current and heat fail. Very low currents would bemvery desireable to fuse,, especially from a cost standpoint.

That might warrant materials research and applying eletrical properties in different ways that heating by current. Actually, I just ran into an ingenius use of NTC thermistors to damp in rush current. At normal temperatures, NTC thermistor resistance is high, so it drops in rush current. The running current proceeds to heat the thermistor due to resistance, which drops with temp, and lowers the temp. It's balances out to an inexpensive, efficient  means to handle the in rush current, and maintain high efficiency thereafter.