A nasty question regarding current shunts.

just one thing abt resistive shunts and your gain , they give noisy signals.

on the 10k resistors put ome capacitance across them (say 1nF) to filter and HF noise

Dean Huster said:

, don't forget that you'll need to use a rail-to-rail I/O op amp since one reference is at the same voltage as your supply or ground rail

Click to expand...

You said you had LMC6081's lying around? According to the datasheet, they should be ok - they swing within 20mV of supply rails.

Pay close attention to the offset inherent to the opamp you select versus the accuracy you need. For example, if you get a TL082, it has an DC input offset voltage of 20 mV. That means that you can't reliably read the 200mV input any more accurately than 20mV. You could have 0mV in and the opamp reads it as 20mV. If you've got a 10x amplifier circuit, it would output 200mV with no current going through the shunt.

It is a mostly fixed error and there are ways to compensate the circuit for the offset, but the offset not only varies from part to part, but with temp as well. The temp variation shouldn't be as much though, looks like an additional 5mV when you include the temp extremes on the TL082 spec sheet.

Bottom line, make life easy and do some shopping on Digikey or wherever to look for low offset amps. There's something called a "chopper amp" which has extremely low offset, into the nanovolts for some. But no need to go nuts if your specs don't need it.

Phasor said:

Dean Huster said:

, don't forget that you'll need to use a rail-to-rail I/O op amp since one reference is at the same voltage as your supply or ground rail

Click to expand...

You said you had LMC6081's lying around? According to the datasheet, they should be ok - they swing within 20mV of supply rails.

Click to expand...

LMC6081 output swings within 20mv of supply rails. The input common mode range is not compatible with high side current sensing. You either need to pick an op amp whose high side input common mode limit is within a volt of the positive rail, or switch to low side current sensing.
As an alternative to a full diff amp, which requires 5 precision resistors (or a potentiometer) for good accuracy, consider **broken link removed**.

I had the same problem for a digital current sensor I'd like to do. High side sensing is tough! It's far easier to do low side but you may have reasons why you can't do that.

Low side problems:
1. If battery is discharging, the input is negative and the op amp can't amplify with without a neg supply (MAX1044 will solve this)

High side issues:
You can do a simple diff circuit with a single op amp. The low input impedance may be a problem, but if you've got a lot resistance shunt you're probably ok.
There's a 3 op amp circuit with high impedance inputs, but the offset errors are cumulative AND all the resistors are 1% at best, so accuracy is crap. The total offset is a killer for millivolt signal inputs.
You can do a charge pump on the shunt voltage. I.E. charge a capacitor across the shunt, disconnect both leads, and reconnect them against ground and op amp input. The amp must have low impedance to avoid discharging the cap. It's good for digital ADCs with a finite sample period, but its chopped output is a problem for analog readings.
You may not have an op amp capable of going to the positive rail on the input. A neg -5V voltage regulator (or zener) gets you 5V below the rail, and MAX1044 to invert the voltage to 5V above the rail, and you use that for the op amp supply. Problem solved!
Review your specs and see if there's any way you can float the ground. Use the battery + as ground for all your system and the neg voltage reg and MAX1044 to get +5 above the battery +. A simple analog meter always floats like this. There's a lot of possible reasons why you can't do this, and you really have to know what you're doing to spot all these reasons, so be cautious.