There are several ways to disconnect the hall sensor (why is that necessary?) and interrupt current to the relay coil when the load current exceeds a preset level. When the "trip point" is reached, the relay de-actuates, causing it to interrupt the load circuit. Problem is, that function needs to be latching, so once it "trips" it stays tripped, else the lack of current in the load would be a signal to turn everything back on again. Depending on other parameters (like relay contact-transfer speed), the circuit would rapidly oscillate on and off as long as the load fault existed and the Hall sensor was telling the relay to energize.
So, is a manual reset acceptable, or do you want a function that waits awhile and then re-enables the trip function again the next time the hall sensor activates the relay coil? And just how much current are the relay contacts switching? AC or DC? What voltage is being switched by the relay contacts, and what voltage at how much current does the relay coil require to actuate?
What you want to do is easily accomplished with an
Allegro current sensor chip whose output is monitored by a
PIC10F206 microprocessor. The PIC microprocessor can monitor the Hall-effect sensor output as well as the Allegro output (which represents current as sensed and isolated by the chip). The PIC would then control the power relay that services the load, turning the relay on when the Hall sensor detects the magnet if, in the process of doing so, the load current remains below a preset value. If a fault condition exists (load current exceeds trip point), the relay would be immediately de-energized within a few milliseconds and stay de-energized until reset. If everything starts up okay, but later the load current exceeds the trip value, the PIC firmware would de-energize the relay coil, regardless of the Hall sensor output, and await a reset condition.
After the PIC firmware has detected a fault condition, you have a choice of "automagical reset" after a programmed time delay, or manual reset with a push-button switch whose state the PIC firmware monitors. You could even do both, the push-button switch overriding the programmed delay. All this two-chip, ten dollar, circuit requires is a 5 V DC power supply capable of providing about a hundred milli-amperes of current to power the Allegro chip, the microprocessor, and the relay coil. Maybe less, depending on the relay coil. You could also use a
solid-state-relay (SSR) in place of an electro-mechanical relay if your load requires high current. Most SSRs require only a few milli-amperes to actuate.
The Allegro chip functions in a way similar to your current meter. A short length of heavy copper is connected internally between terminals on the Allegro package. Passing an electrical current through this internal, low-resistance, connection creates a local magnetic field that an internal Hall-effect sensor detects. It is a linear sensor, so the Hall-effect output is linearly proportional to both the magnitude and the direction of the current passing through its current-sensing terminals. The output is a voltage the varies from zero to approximately the power supply voltage (0 to +5V) as the current varies from a maximum negative value to a maximum positive value. At zero current the Allegro chip output is one half the power supply voltage. That means you should AC-couple and then rectify and filter the Allegro output if the load is driven from an AC power source. Or you can easily do that function in software, thereby eliminating the need for a coupling capacitor, a diode, and a filter capacitor, looking for a peak value of the Allegro output corresponding to either a positive or a negative (or both) peak in the current being sensed. Programmer's choice, but I would opt for the software solution to eliminate the extra components.
The PIC I suggested has a built-in analog comparator function that allows you to compare a zero to +5V input with an external or internal reference voltage and respond accordingly. Or you could use a somewhat more sophisticated PIC, such as the PIC10(L)F320, which has a built-in analog-to-digital converter, obviating the need for a comparator. Either way, the cost of the two chips is small, even in onsie-twosie quantities. The Allegro chip is available in a ±200 A range, which may be way more current than your load would ever draw, but it has the advantage of through-hole leads which are easier to wire up. There is a surface-mount version, the
ACS712, that has a range of ±50 A, which is probably what you would want to use.
If the above sounds doable for you, but you lack the circuit design or programming skills, we can help you with both. There is a rather lengthy thread here that describes how one former newbie successfully completed a similar Allergo-based design.