Hi Guys,

I'm hoping to sense the current through a P-channel Mosfets Source (12V+) giving a range of 0 - 3.3V on the output of a diff opamp setup. One consideration is the voltage of the 12V is automotive and will change with the alternator running or not, possbily up to say 14V also if a car battery discharges to 12V or less (god forbid). I do realise considering this is differential config, it should be fine, but I just wanted some opinions on the differential opamp resistor values. I was hoping to use it for short circuit protection (over 35 to 40A). The reason I'd like the output to be less than 3.3V is to trigger a uC to switch off the mosfet until it's power cycled (to prevent an intermittant short from continuously arcing). I was also considering using a TL431 on the output as an "instant" mosfet gate switch.

i found this calculator online and reaslise it's for when ground is connected to the sense resistor, but i wasn't sure that would be a problem considering we're just dealing with differential and gain.

Any help would be appreciated.

 

You are talking about a "high-side" current sensing application, where the shunt resistor is at, or near the supply voltage.
The differential amp idea requires a high "Common Mode Rejection Ratio" to function accurately, which requires the resistors to be extremely well matched to perform well.

Why not use a part designed for this application?

MAX4375TEUB-TG075?

 

You are talking about a "high-side" current sensing application, where the shunt resistor is at, or near the supply voltage.
The differential amp idea requires a high "Common Mode Rejection Ratio" to function accurately, which requires the resistors to be extremely well matched to perform well.

Why not use a part designed for this application?

MAX4375TEUB-TG075?

Thats an expensive suggestion. I'm going for a solution for short circuit protection. Not an ammeter. Like I said between 25-30 amps cut off.

 

Thats an expensive suggestion. I'm going for a solution for short circuit protection. Not an ammeter. Like I said between 25-30 amps cut off.

Is this for mass production? is that why the cost is a factor?

If you just need rough overcurrent sensing, a resistor with a single transistor can do the job (roughly) at very low cost.
At 25 A, a shunt of 24 milliohms will have a drop of 0.6V - the Vbe of a transistor.
This would respond faster than an opamp too.

 

Is this for mass production? is that why the cost is a factor?

If you just need rough overcurrent sensing, a resistor with a single transistor can do the job (roughly) at very low cost.
At 25 A, a shunt of 24 milliohms will have a drop of 0.6V - the Vbe of a transistor.
This would respond faster than an opamp too.

But would waste 15W.

 

Why don’t you use one of the several Allegro ACS-series hall cell based current sensors?

 

But would waste 15W.

Did I miss the part where the nominal current is specified?
The idea assumes the nominal current is much less than 25A?

 

Well, I would not expect someone to set a current limit of 35 or 40A so a circuit that normally draws 1A. But yes, we should ask that.

@gkeep: What is the range of current in normal operation.