How to sense a small current while handling larger ones

Question

For one of the projects I'm working on, I need to monitor a DC power line for reverse current; the solution has to output a 24V signal when around 10A or more of current is flowing in reverse through the line (forwards current can be ignored), but the line may carry up to 400A in either direction.

The main problem seems to be that for any solution, say a shunt resistor with a sense line connected to the gate of a MOSFET, anything that will give a workable output with 10A going through the line would fry at 400A, while anything that works at 400A would give a near negligible output at 10. In general, how does one usually measure something that could range much higher than the target measurement value?

For this specific application, we would love to just use something off the shelf, but the general question is something that seems handy to know.

Answer 1

10A is 2.5% of 400A, which is more than 'near negligible'. However a shunt with reasonable loss at 400A will have very low output voltage. The answer is to use a low-drift op amp to amplify the voltage enough to drive a switching device.

Let's say you can accept a shunt loss of 4 watts. At 400 A that corresponds to 10 mV, with a shunt resistance of 0.01 / 400 = 0.025 mΩ. 10 A would then produce 10 * 0.025 = 0.25 mV or 250 μV. The AD8551 is an example of a so-called 'zero-drift' op amp, with an offset voltage of 1 μV and drift of 0.005 μV/°C. This could theoretically measure that 10 A to an accuracy of 0.4%.

Another alternative that doesn't require a shunt is a Hall-effect current transformer. The HT500M is an example which uses closed loop sensing for greater accuracy. It has an output current of 200 mA at 400 A, and offset current of +-200 μA. At 10 A the output would be 5 mA +-0.2 mA = +-4% - 10 times worse than the current shunt, but with the advantage of non-contact sensing.