Hi all,

I am trying to find a current sense amplifier for an audio frequency circuit.
The load is inductive and has a resistance of 10 ohms.
The current is unidirectional and changes from 100mA to 1.4A
Rsense is 100m Ohm, so Vdrop across it will vary from 10mV to 140mV
Rsense is on the low side of the load
The voltage applied to the load is 14V
Gain needs to be 100.
The output will be applied to an oscilloscope to view the current waveform, with the peak being around 14V.

I've never used a CSA before. I've narrowed it down to either an INA193 or an INA199.
I also considered just using an Op amp like the TLV2372.

Thoughts/Opinions?

 

The challenge will be both the common mode voltage that the amplifier can work with, and the output voltage capability of the amplifier. Consider that 100 x 140 millivolts is 14000 millivolts=14 volts. An opamp powered by 5 volts will not be able to do that.

 

The challenge will be both the common mode voltage that the amplifier can work with, and the output voltage capability of the amplifier. Consider that 100 x 140 millivolts is 14000 millivolts=14 volts. An opamp powered by 5 volts will not be able to do that.

Yes. And considering my power supply is 14V I need something that's rail to rail from what I've read.

 

So Rsense is in the ground connection at zero volts with no current?
If so, then a RRIO op amp configured as a non-inverting amp with a gain of 100 should work.

Below is the LTspice sim of such a circuit.
The TLV9152 is similar to the TLV2372 but the TLV9152 has significantly lower input offset voltage of 0.8mV vs 6mV max.).
Either one should work, but lower offset is better for this application since, with a gain of 100, the output offset voltage will be 80mV vs. 600mV max.

 

So Rsense is in the ground connection at zero volts with no current?
If so, then a RRIO op amp configured as a non-inverting amp with a gain of 100 should work.

Below is the LTspice sim of such a circuit.
The TLV9152 is similar to the TLV2372 but the TLV9152 has significantly lower input offset voltage of 0.8mV vs 6mV max.).
Either one should work, but lower offset is better for this application since, with a gain of 100, the output offset voltage will be 80mV vs. 600mV max.

View attachment 320190

Thank you!!!

Yes, one side of Rsense is connected to ground while the other side is connected to the load.

I didn't know the input offset was multiplied by the gain? Thanks for explaining that.

One thing I do wonder, can the offset be lessened by changing the resistor divider slightly (is that what R2 is for), or will that just push the offset further in one direction (low measurement end vs high measurement end)?

Edit: added (is that what R2 is for)?

 

didn't know the input offset was multiplied by the gain

Yes, the input offset looks like a signal, so is multiplied by the circuit gain (technically the non-inverting gain).

can the offset be lessened by changing the resistor divider slightly (is that what R2 is for)

No, the offset is random and can be zero to plus or minus the maximum, so the only way to cancel it is to add a pot that can be adjusted to cancel this offset after the circuit is built.
Since that is not desirable, using a low offset op amp is usually preferable.

R2 is just to give a nominal gain of exactly 100, since standard 1% resistors give an ideal gain that is about 1% off, not including the inherent additional 1% accuracy of the resistance values.
So adding the 10Ω resistor means the gain is then 100±2%.

 

There are so many low-offset op-amps these days, that it is much easier simply to choose a low-offset op-amp than to go to the trouble of adding and adjusting a preset.
But don't forget that input bias current also causes an offset equivalent to the input bias current (from the datasheet) multiplied by the input resistor (R4+R2), multiplied by the gain.

 

But don't forget that input bias current also causes an offset equivalent to the input bias current (from the datasheet) multiplied by the input resistor (R4+R2), multiplied by the gain.

The TLV9152 has an input bias current of 10pA, so is not a factor in this (or most) circuits.

 

The TLV9152 has an input bias current of 10pA, so is not a factor in this (or most) circuits.

Just worth mentioning, in case he were to choose an opamp with lower offset voltage such as an OP07, which has 4nA bias current, so he would be no better off.