For accuracy measurement there are three components considered: electrical offset error, magnetic offset error and sensitivity. The electrical offset error (also referred as voltage offset error), $V_{OE}$ , is the deviation from the ideal $V_{out}$ when no input current is flowing. This test was run by inputting zero current and measuring the zero current output voltage, $V_{out,0A}$ . $V_{OE}$ was then calculated by using Equation 1. For the magnetic offset error, the devices were saturated to 110% of their maximum current, returned to the quiescent point, then $V_{out}$ was captured at no current. The offset voltage was calculated to determine if there were any changes due to any residual effects caused due to the saturation of magnetic fields. Lastly, sensitivity is the change from the sensor $V_{out}$ due to the change in input current, $I_{IN}$, which can also be seen as the slope as shown on Figure 1. Sensitivity testing was done by measuring the sensors $V_{out}$ at no input current flowing and then loading a different currents and recording the corresponding $V_{out}$ to then calculate the slope. Then the error of the sensitivity, $e_{SENS}$, was calculated using Equation 2.

From the results we can see that TI TMCS1143 has the lowest offset voltage and similar to all other competitors apart from the closed-loop current transducer (Comp5), no hysteresis is observed after the device was saturated. It's interesting to note that Comp5 shows a 3mV difference since this is a closed-loop current transducer and it is known that hysteresis effect is minimized for this type of method. For the sensitivity results we can see that the sensors with the lower sensitivity like Comp4, TMCS1143, and Comp5 could be a better solution for high current applications where this is important to allow the full current range without saturating.

Figure 2-1 Sensitivity, Offset and Nonlinearity Error