This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

  • Resolved

[FAQ] INA181: Understanding the Output Swing of Current Shunt Monitors and Amplifers

Part Number: INA181

How to interpret and utilize the output swing specifications of current sense amplifiers so I can get the most out of my current measurements?

  • Knowing how close an amplifier’s output voltage (VOUT) can swing to its positive and negative power rails is extremely important for design because the larger the swing, the more meaningful measurements you can acquire from the output of the amplifier. In general, the output stage of an amplifier is physically limited in how close it can swing to the rails because of the output stage transistors’ saturation voltages.


    The VOUT swing of an amplifier or current shunt monitor (CSM) can usually be found in the datasheet’s Electrical Characteristics table, specifically the “Output” section. Here’s an example from one of our CSM’s, the INA181.

    Figure 1: INA181 Electrical Characteristics Voltage Output Specifications


    There are two significant details to note about the specifications in Figure 1. One is that the test was done with one load (RL = 10kΩ), which means that the output current (IOUT) is small (~5V/10kΩ = 500uA). Two is that the inputs of the amplifier were being overdriven to create the most maximal swing to the rail possible, which means that the amplifier is not operating in its linear range. Overdriven means the differential input voltage (VIN = VIN+ - VIN-) was greater than the maximum allowable input (e.g., 250mV) or less than the minimum input (e.g., -250mV). This is not immediately apparent, but is standard practice and is further elucidated with the “(3)” note at bottom of page.

    As an aside, if designers want to know the output stage’s linear operating range, they can inspect the testing conditions for various performance specs. For a standard operational amplifier they would look to the VOUT range used in generating the open-loop gain spec. Since CSM’s are closed-loop, you can look for the linear VOUT range used to generate gain error (EG) or Nonlinearity error specifications (Figure 2). Note: 500mV from the rails is a very conservative linear range of Vout. Realistically, most amplifiers are linear up to 100mV from each rail especially when considering typical operating temperatures. You can read more about the linear range of amplifier output stages here.

    Figure 2: Output Specifications for INA181


    If the designer knows his or her design will work with heavier loads (lower RL) and thus higher IOUT current levels, then he or she will have to defer to the Output Voltage Swing vs. Output Current plot or referred to visually as the “claw curve”. This plot will give the designer an idea on how the VOUT range will be affected by the load. Figure 3 shows this curve from the INA181 datasheet. You can see that as the IOUT increases, the VOUT swing decreases and this is always this typical case for amplifiers.

    Figure 3: Vout Swing vs. Output Current


    One other consideration for VOUT swing is how close it can swing to ground under zero-current conditions versus the input common-mode voltage (VCM). Note zero-current conditions means VIN = 0V, while overdriven mean VIN<0V or VIN>full-scale. This is important for when the CSM is in a unidirectional operation (sensing shunt current flowing in one direction) and thus the device reference pin can be driven to ground to achieve full output range. Figure 4 below shows how low you can expect the part swing when the part is referenced to GND.

    Figure 4: Zero-Current Output Voltage vs. Common-Mode Voltage


    If the CSM’s input offset (VOS) is greater than zero, then the VOUT swing to ground will be limited by VOS *gain and this will always become the case as VCM increases even if VOS is initally negative . Since VOS varies with VCM, we can determine the swing to ground under zero-current conditions using Figure 4 from the datasheet.


    Here's an example using the INA181A3, which has a gain of 100V/V, and we power it with 5V.  Let's say we want to measure between 1A and 50A, and we choose a 1mΩ shunt resistor.  This means we will see 1A * 1mΩ = 1mV, which we then multiply by 100V/V to get the expected output of 100mV.  The swing-to-GND of the INA181 is GND+5mV, so an expected minimum output of 100mV (that is greater than 5mV) is acceptable.  However, at 50A, the sense voltage on the input will be 50A * 1mΩ = 50mV, which we multiply by 100V/V to get the expected output of 5V.  The swing to rail spec is Vs-0.03V, so the maximum output of the INA181A3 in these conditions is only 4.97V, and 5V exceeds that.

    So, what can be done to make this circuit work?  There are a few options.

    1. Increase supply by at least 30mV over 5V (but less than the max of the part).  This isn't usually an option in designs as power supplies are generally fixed at common values like 1.8V, 2.5V, 3.3V, 5V, etc.

    2. Choose a device with lower gain.  The INA181A2 has a gain of only 50V/V, so the maximum output voltage needed would be only 2.5V, well under the 4.97V limit.

    3. Choose a smaller shunt resistor.  Reducing the value of the shunt will decrease the size of the input signal and subsequently the output signal.  Switching from 1mΩ to 0.5mΩ accomplishes the same thing as option 2.

    4. Choose a different amplifier with more suitable swing-to-rail specs.  For instance, the INA199 has a swing-to-Vcc spec of Vcc-0.2V.  Switching from the INA181 to the INA199 would reduce the available output range for the circuit as the output approached Vcc.  However, the INA199 can support power supply voltages up to 26V, where the INA181 maxes out at 6V, making option 1 worth investigating again.

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.