The term calibration must be used cautiously here, as technically three calibration options exist that can be performed with the TMCS110x: a one-time layout sensitivity calibration, and then an "up to two step" process: a one-point offset device calibration, and a device level sensitivity calibration if additional accuracy is desired beyond the one point calibration routine. Note that the latter of these options is device specific, requiring unique values to be programmed in firmware, and thus adding complexity in applications requiring scalability.

The first option mentioned is a calibration based on layout and entry angle of the current into the TMCS110x device. A good example of the errors under consideration here are those discussed in Conduction Paths. The entry angle of the current, as well as other potential factors, may shift the sensitivity due to their proximity and contributions to the TMCS110x. To perform this calibration, two known currents are applied to the TMCS110x, and the corresponding outputs measured. From these points, the true sensitivity of the device is calculated via point-slope form, that is, sensitivity is calculated as ΔV_OUT/ΔI_IN. Once this slope is determined, it is held in logic and the output may then be data corrected in logic via this coefficient in place of the ideal device value from the data sheet. Once this calibration to the layout is performed, the expected variation is then limited from device to device by the maximum sensitivity error in the data sheet.

A second calibration step for additional accuracy is to then eliminate DC offsets potentially present in the system. This is typically performed once the device is in place (in terms of application, not layout), and will remain stationary, as the Earth's magnetic field or any other DC fields may change if the orientation changes. This calibration is performed via a zero current condition placed at the inputs, and measuring the corresponding output. In an ideal sense, under this condition, the output voltage is the reference voltage of the device. The deviation between this expected ideal and the actual output being measured is the DC offset under discussion. The expected output for each device is shown in TMCS1101 Data Sheet: Linear Operating Region and is corrected to this measured value in logic. Note, as mentioned before, this type of calibration is device and orientation specific and should be done for each device when the board is its expected use location.

The final calibration option may be implemented if the first calibration does not provide enough error correction, and additional accuracy is required. This calibration is simply repetition of the board level calibration, but is performed for each specific device, rather than a blanket coefficient to the layout itself. This eliminates the device variation in sensitivity, and removes the sensitivity at a given operation point. Note that unaccounted for external fields, or outside factors such as temperature drift may still introduce additional error here.