Current Sensing with Single-Supply LM741?

[tineras]

I'm trying to setup a simple current sensing circuit with an LM741. The circuit will be powered by 12V car cigarette lighter. Because of this, I'm trying to avoid having to create a negative voltage source. I had no trouble making a current sensing circuit using a dual supply, but with a single positive supply I cannot get the output to change much when a load is attached in series with the sense resistor.

Is the LM741 an okay choice for this? (it's all I have except for a couple of LM386 chips)
How (in general) is the output calculation different from that of a dual-supply circuit? (Gain * Difference)

When the load is connected, it is using approximately 500mA @ 5V. The positive source is 5V and the negative is connected to ground. The sense resistor I have is 0.1 Ohm (or about 1/100 of the load resistance).

I have purposely not included an image of my exact circuit because I have tried so many, but the variations I have tried are similar to this: https://lh5.googleusercontent.com/-LBe4cije9IA/T95mwidwFuI/AAAAAAAAApk/Ulw5NJ8I-6o/s800/Single%2520Supply%2520Current%2520Sensor.jpg without the 2N3904. I have also tried low and high side variations.

I know I'm probably doing something silly. If anyone could give me some pointers, I would really appreciate it.

[amspire]

The LM741 seems to be about the least appropriate choice for a 5V supply current sense circuit. You will need to buy something better for the purpose.

You need an opamp that can run properly from a low voltage supply (like 5V) and you want the opamp inputs working at either the positive rail (for a current sense resistor on the positive rail), or negative rail (for a current sense resistor on the negative rail).

Look for rail-to-rail opamps, and check out the data sheets before buying.

Look at the common mode input specification and check it can handle the inputs up at the rail that the current sense resistor is on.

The NPN transistor helps the output to go all the way down to 0V - it means that the opamp only has to be rated for an output that goes to less then 0.6V.

Richard.

[tineras]

The LM741 seems to be about the least appropriate choice for a 5V supply current sense circuit. You will need to buy something better for the purpose.

You need an opamp that can run properly from a low voltage supply (like 5V) and you want the opamp inputs working at either the positive rail (for a current sense resistor on the positive rail), or negative rail (for a current sense resistor on the negative rail).

Look for rail-to-rail opamps, and check out the data sheets before buying.

Look at the common mode input specification and check it can handle the inputs up at the rail that the current sense resistor is on.

The NPN transistor helps the output to go all the way down to 0V - it means that the opamp only has to be rated for an output that goes to less then 0.6V.

Richard.

I had my suspicions about using the LM741 at a low voltage like this. I'm slowly but surely building up a stockpile of parts, but don't have any other op amps around. I'll see what I can find online and read up on rail-to-rail opamps. I'll post an update here when I get some more appropriate parts to work with.

Thanks for the reply!

[codeboy2k]

what amspire says is correct, but if you can power the 741 off the 12v supply, not the 5 v supply, and sense high side it should be ok to use.  Note that the opamp is a difference amplifier, the VCC voltage is not important, and thus the VCC to the opamp doesn't need to be well regulated.  So you can power your op amp off the unregulated 12v - 14v automotive and you can power your circuit off the regulated 5V.

You can't sense low side because then the - input would be at 0 ( at the rail ) and that won't work.  You need to sense high side, and then your + input is at 5v, your  (-) input will be at max 50mv below that, and both inputs are well away from the ground and the +12 supply.

The 741 also has minimum of 1mv offset up to max 5mv, so you will get 2% - 10% current sensing error. 
You can improve that to 0.2% up to 1% error by using a  1 ohm sense resistor, instead of a 0.1 ohm resister. This gives you 0-500mV at the opamp inputs for a 0-500mA input.  If you do this, then your voltage drop across the sense resistor might be too high for the load (5V - 0.5V = 4.5V) and you need to compensate for this.

You can compensate for the voltage drop across the sense resistor if you change the the 5v regulator to give a little more than 5v to say 5.5 v, if your load can take the extra voltages when drawing less current (at lower currents the voltage drop will be less and the load will see more voltage).  If the load is not voltage sensitive than it could work for you.

[tineras]

what amspire says is correct, but if you can power the 741 off the 12v supply, not the 5 v supply, and sense high side it should be ok to use.  Note that the opamp is a difference amplifier, the VCC voltage is not important, and thus the VCC to the opamp doesn't need to be well regulated.  So you can power your op amp off the unregulated 12v - 14v automotive and you can power your circuit off the regulated 5V.

You can't sense low side because then the - input would be at 0 ( at the rail ) and that won't work.  You need to sense high side, and then your + input is at 5v, your  (-) input will be at max 50mv below that, and both inputs are well away from the ground and the +12 supply.

The 741 also has minimum of 1mv offset up to max 5mv, so you will get 2% - 10% current sensing error. 
You can improve that to 0.2% up to 1% error by using a  1 ohm sense resistor, instead of a 0.1 ohm resister. This gives you 0-500mV at the opamp inputs for a 0-500mA input.  If you do this, then your voltage drop across the sense resistor might be too high for the load (5V - 0.5V = 4.5V) and you need to compensate for this.

You can compensate for the voltage drop across the sense resistor if you change the the 5v regulator to give a little more than 5v to say 5.5 v, if your load can take the extra voltages when drawing less current (at lower currents the voltage drop will be less and the load will see more voltage).  If the load is not voltage sensitive than it could work for you.

I wanted to let you know and it worked exactly as you stated. I will still look at some different, more appropriate, current sensing devices. Thank you both for your help!

[codeboy2k]

I wanted to let you know and it worked exactly as you stated. I will still look at some different, more appropriate, current sensing devices. Thank you both for your help!

Good to hear.

There are dedicated current sense IC's, and these work really well but might be too expensive sometimes.  You care more about offset voltage than anything else. You want to choose an opamp such that its maximum input offset does not add to your error. Remember, you'll amplify that offset voltage, it will appear at the output, so choose something that is 10x less than your error budget if you are gaining it up x10. 

You don't need high speed or large bandwidths, just low-offsets and low bias currents. Inputs for offset nulling can be helpful to calibrate your sensing solution. A low voltage opamp is also needed if you want to power the opamp off the low voltage regulated side, but it's not necessary to do that. You can use the unregulated side, or a simple zener regulator to operate your op amps at a higher voltage than your common mode voltage at the inputs.  And then you no longer need to restrict yourself to rail-to-rail devices any more.  You would then consider the op-amp's PSRR (power supply rejection ratio) as an important data-point; you want PSRR to be pretty high when using an unregulated or lightly regulated power supply to the opamp. So then you might want to choose an opamp with the best combination of low offset voltage, low bias currents, and high PSRR.