Hello,
Our customer use the ONA226, estimate accuracy.
If the customer assumption that the error of the BUS voltage measurement is 1% and
the error of the shunt voltage measurement is 1%,
how about the error of the electrical power calculation results?
Is it root mean square?
Best Regards,
Naoki Aoyama
Naoki,
If customer is expecting 1% error for the current and bus voltage measurements for whatever reasons may be (e.g., 1% shunt resistors or using a resister divider for the VBUS pin with 1% resistors), then using the RSS error is probably fine. Although customer may want to go with a more conservative total error calculation, especially if the error sources for the shunt and bus measurements are correlated. Since both measurements come from the same internal ADC it is possibe they share the same gain error, meaning if the ambient temperature rises to 100C the gain errors for shunt and bus voltage measurements will increase by the same delta. Although according to Figure 7 shunt voltage gain error does vary with the common mode voltage while the bus voltage gain error does not, so they are not correlated in this way.
RSS = sqrt( 0.01^2 + 0.01^2) = 1.41%
Maximum total error = 2%
Overall I would go with the maxium total error since this is the upper bound. The maximum total error can be realized through working through a scenario. Assume ideal current value = 1A and ideal bus voltage = 12V. If the both measurements experience +1% error, then they will be reported as 1.01A and 12.12V.
Power = V*I = 12.12V*1.01A = 12.2412 Watts
Maximum total error = 100* (Ideal - Measured)/Ideal = 100*|12-12.2412|/12 = 2.01%
