I stumpled over the following article comparing the following difference amplifier circuits in terms of errors due to input offset voltage and resistor tolerances.

enter image description here

To my understanding the right-most circuits gets away with fewer and possibly also less accurate, hence cheaper resistors. This made me think wether I could adapt the circuit to measure much larger differential voltages such as the the cell voltage in a stack of batteries.

I played around in LtSpice and it seems that it might work, however, I really don't understand the rightmost circuit. Where is the feedback path? What influence does the transistor have? Spice indicates that a PMOS also works, but how and why? Are there device that alreay include the transistor and what are they called?

I found some shunt current devices which seem to implement the rightmost circuit, but those are only specified for small differential voltages. Why is that?


I really don't understand the rightmost circuit. Where is the feedback path?

You have to consider the voltages on the op-amp's inputs and realize (as in all linear op-amp circuits) that they have to be equal (for an ideal op-amp). For this to happen (when current flows through the 0.1 ohm shunt resistor), negative feedback ensures the same volt drop across the 1 kohm resistor i.e. the op-amp output drops a little to turn on the PNP BJT and dump current through the 20 kohm resistor.

So if 1A is flowing through 0.1 ohms causing a volt drop of 0.1 volts then there also has to be a 0.1 volt drop across the 1 kohm. This means a current through the 1 kohm of 100 uA.

You can probably see immediately that the ratio of the two resistor values is inversely the ratio of the two currents.

The 100 uA (minus a tiny amount that flows through the base emitter junction) also flows through the 20 kohms. This means that with 1 A flowing to the power load, 2 volts appears across the 20 kohm resistor.

A PMOS would work more effectively in that there is no base current. Also, don't forget that the op-amp positive power rail has to be higher than 12V unless you are specifically using an op-amp designed for high-side sensing.

| improve this answer | |
  • \$\begingroup\$ How does the choice of PMOS or PNP affect the accuracy especially the influence of ambient temperature? \$\endgroup\$ – Arne Jun 8 '16 at 10:02
  • \$\begingroup\$ The accuracy is mainly down to choice of tight tolerance resistors. Temperature changes affect the resistances so choose ones with low temperature coefficient. PNP steals current via the emitter is this represents about a 1% inaccuracy compared to PMOS. \$\endgroup\$ – Andy aka Jun 8 '16 at 10:15
  • \$\begingroup\$ Are there devices that already incorporate the PMOS or PNP and are suitable for larger >4V differential voltages? How are they called generally? \$\endgroup\$ – Arne Jun 8 '16 at 11:19
  • \$\begingroup\$ Linear Technology make these type of things - try searching for "high side current monitors": google.co.uk/… \$\endgroup\$ – Andy aka Jun 8 '16 at 11:34
  • \$\begingroup\$ I checked out the Linear Technology parts and it seems that these are taylored towards small differential voltages ~100mV. When measuring cell voltage the differential voltage is much larger 3V...5V. Any ideas? \$\endgroup\$ – Arne Jun 8 '16 at 15:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.