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.

INA193: Vout formula

Part Number: INA193
Other Parts Discussed in Thread: INA281

Hi Guys,

This is our customer's inquiry. Kindly help to provide the formula to calculate the Vout for INA193, if Vin+ and Vin- is known, along with Rshunt.

Attached circuitry for reference:


(I) load varies from 300mA to 1A . Where as Vinput will be from 5 to 6.6V

Thank you in advance for your support. 

Best regards,

Jonathan

  • Hi Jonathan,

    Vout = I x 82mR x 20

    Kai

  • Jonathan,

    Kai's expression above is correct. The "20" in his equation is the gain associated with the INA193, and the I*Rshunt is the sense voltage generated between the pins of the device. This means for your given min and max, the expected outputs on the amplifier would be

    Vout@300mA = .492V

    Vout@1A = 1.64V

    Is the customer using the chosen shunt here for thermal benefit? Based on a 1A maximum measurement and a 3.3V supply, they could realistically use up to a 165mΩ shunt here (although I would choose something like a 150mΩ to help prevent device saturation). As the size of the shunt increases, the generated sense voltage on the inputs increases in kind, which will help reduce the error against the offset of the device at the lower end of the measurement. This comes with the tradeoff of additional heat being burned on the resistor, but they will need to determine an amount of error that is acceptable here.

  • Hi Jonathan,

    If we use higher value of rshunt , the power dissipation and the drop due to higher value of resistor will lead to other issue. Also the size of the resistor chosen to be more , due to pcb real estate issue we are not able to chose a higher value of resistor

    Out typical Current will be around 0.3A to 0.5A for which Vsense will be approximately 24mV to 41mV . Vsense above 20mV is not having much error right .

    Vsense > 20mV is fine right for measurement.

  • Hi Asha,

    see figure 5 of datasheet Thinking

    Kai

  • acc to datasheet , for vsense of 25mV we are seeing a error of 1.25% . If error of 1.25% is there for vsense = 25mV . In my application it is fine to have a error until 2%. Please confirm for Vsense = 25mV , error will be around 1.25%

  • Asha,

    What figure/spec are you referring to when you say you'll see 1.25% error at 25mV? The error is not going to be static here, but will shift downward as you increase the signal against the offset voltage. 

    Is this a design that is planning to be scaled? Operating this low in the curve, the offset voltage is going to potentially dominate the signal here. While the typical offset of the INA193 is only 500uV RTI, the maximum is potentially 2mV. Integrated into your 24mV signal, this is nearly 8% error from the input signal alone dependent on the distribution of the offset from part to part. You may be able to improve these specs by moving to the newer member of this family, the INA281.

  • Hi Carolus,

    I think it's figure 5 of datasheet of INA193 Asha is referring to:

    Kai

  • Thanks for pointing this out Kai.

    Asha,

    While Figure 5 does show 1.25% at a Vsense of 25mV, keep in mind that this curve demonstrates typical behavior of the device (evidenced by the 0.75% at a Vsense of 100mV). The error curve I show above is for worst case behavior, and does not take into account test limits of the device. As per the EC table, the total output error is tested at a Vsense of 100mV, and any device demonstrating >2.2% error would be rejected from the lot, but this does not necessarily extrapolate 1:1 at the 25mV mark for the purposes of worst case analysis. This is why I ask you above if this is a design intending to scale.

    Here's the same curve I generated extended to a Vsense = 100mV condition. With an 80mΩ shunt, you would generate 100mV sense voltage at around 1.25A. Notice in the worst case analysis, we are right around that 2.2% error mark. However, the worst case offset has much more effect on the error at a Vsense of 25mV: