I'm trying to use multiple current sense circuits to measure currents over a wide dynamic range. The circuit has to be able to measure current ranging from 0.75A to 30uA. This can be achieved by using two high side sense resistors in series, 17 ohm and 3 ohm, each connected to their own op-amp current sense circuit with a gain of 20V/V. The issue i'm facing is that at 0.75A, the 17 ohm circuit will be railed which is fine, but the power dissipation will be almost 10W. I'm trying to figure out a way to bypass the 17 ohm current sense circuit after it has railed. I have done a fair amount of researching and I keep coming up empty. I keep on having "genious" breakthrough, but every time i realize it wont work. Does anyone have any experience with multi-stage current sense? Any input would be greatly appreciated.

 

I'm trying to use multiple current sense circuits to measure currents over a wide dynamic range. The circuit has to be able to measure current ranging from 0.75A to 30uA. This can be achieved by using two high side sense resistors in series, 17 ohm and 3 ohm, each connected to their own op-amp current sense circuit with a gain of 20V/V. The issue i'm facing is that at 0.75A, the 17 ohm circuit will be railed which is fine, but the power dissipation will be almost 10W. I'm trying to figure out a way to bypass the 17 ohm current sense circuit after it has railed. I have done a fair amount of researching and I keep coming up empty. I keep on having "genious" breakthrough, but every time i realize it wont work. Does anyone have any experience with multi-stage current sense? Any input would be greatly appreciated.

Why don't you use lower resistance current sense resistors and higher gain? Also, instead of two separate current sense resistors, use two different levels of op-amp gain to get your whole dynamic range?

Note, current sense resistors of 0.01 ohm are possible. Your voltage drop and heat generated will be minimal if you pick a reasonably sized resistor.

 

Follow up questions: What are your bandwidth and common mode requirements?

I agree with gopher, especially if you do not need a high frequency response. For example, if the purpose of the current sensing is to capture fast current transients, then more work is needed. But if all you are doing is recording or reporting the average value, then a single slow, quiet, high gain amp with intentionally limited bandwidth and stepped gains will handle the range.

Right now it looks like you get an output of 0.0102 V at 30 uA through 17 ohms with a gain of 20. A 1 ohm resistor and a gain of 340 (51 dB) gives the same result. At 0.75 A, the max power dissipation in the resistor is 0.56 W. Many IC instrumentation amps are characterized at gains of 1000, and some have programmable gain-switching built in. Also, consider integrated current sensor ICs from Maxim and Linear Tech.

ak