Measuring current & 'active' over current protection

Evening all,

I've just had a thought - I'm currently protecting a design I'm working on from overcurrent by using polyfuses, so any component failures that lead to shorts to ground, or any external sensor and/or wiring faults that cause a short to ground will cause the polyfuse to trip and the power to be cut.

However, if I were to introduce a FET into the power supply (infact, I've already got a p-channel FET providing reverse polarity protection), and directly measure the current drawn then I could have the controller switch off the FET if a current threshold is exceeded for a certain amount of time.

How does one go about measuring current drawn by a device, and in such a way I can convert this to an analog voltage and feed it in to an ADC pin? I'd imagine I'd filter the signal first as there are plenty of current spikes in my application?

It seems a shunt resistor is my only real option?

Poly fuses are very slow to react and there charastics change every time they blow. They are not good news and have been removed from the later versions of the Raspberry Pi.

To turn a current into a voltage you use a resistor.

Yes I have heard they're not the best news - hence the thoughts to replace them! :)

So, simply a shunt resistor in series with the load, measure the voltage before and after the shunt, feed these into an opamp, then feed the output into an ADC to measure the voltage drop and have the controller applying the appropriate action, i.e. switch off power.

Here's a thought... Maybe you could use the transistor's RDS(on) as the shunt resistor. Would that work?

Only if the transistor is actually a FET and then it might be too small to measure. You need a rail to rail op amp powered from the unregulated side. Also do the shut down in discreat components rather than a controller for speed and reliability. Alternatively look for what is known as "high sided current monerting" chips.

Using the FETs on resistance is a possibility, if I use very high gain in the subsequent amplifying circuitry. However, given the FETs on resistance changes with voltage, current & temperature - it seems I'd probably have to use the controller to do the shutdown, not passives, if I want to cut power at an accurate current level.

I'll have to compare the variations between a FETs resistance, and a shunts resistance given the above changes. Thanks for the heads up thus far.