Answer 8—The actual peak error value depends on the specific ratio of the resistor values in the divider, but it will never be worse than ±10%.

The likely error is somewhat less than this. If the distribution of the errors in the individual resistors is assumed to be uniform over the ±5% range (not really true in practice, but a good worst-case assumption), then the combined error of the voltage divider has a triangular probability density function (PDF). Seventy-five percent of the time, the error will be less than half the peak value calculated above, and half the time, the error will be less than approximately 30% of the peak value.

As a general rule of thumb, it’s fairly safe to assume that the overall error of the voltage divider is roughly equal to the error of the individual resistors.

Contributor: David Tweed

Published October 2003