# The easy math that explains why accurate IoT current measurement is hard

Blog Post created by BradJolly on Feb 24, 2017

The never-ending drive to increase IoT battery life is great for customers, but it poses extraordinary challenges for design engineers. As sleep mode currents edge ever-closer to zero, the challenge of making measurements across a wide dynamic range becomes increasingly difficult.

Consider wireless transceiver chips, illustrated in the chart below. Each line represents one transceiver, and the lines go from a given device’s lowest sleep current to its highest operating current. The dynamic range, of course, is the ratio of the two current levels, and the base two logarithm of that ratio indicates how many bits are required to represent the dynamic range.

Merely representing the dynamic range, however, is not sufficient for accurate measurements. For example, if your measurement instrument uses 18 bits to represent a 250,000:1 dynamic range that spans 25 mA (middle of the chart), then your measurement resolution is approximately 100 nA. When you measure relatively large currents in the mA range, this is fine, but when you measure the 100 nA sleep current, your accuracy is ±100 percent – a rather coarse value.

For 25% accuracy, you need two additional bits, because the four possible values of two bits divide your resolution accordingly. Similarly, for 10, 5, and 1 percent accuracy, you need 4, 5, and 7 additional bits, as summarized in the following table, which uses non-integer base two logarithms to reduce the number of bits in some cases.

It is, of course, difficult to find instruments that provide accurate current measurements with 20 or more bits of resolution in a given measurement range. The best solution is to use an instrument with seamless ranging that avoids glitching during range changes, or to use a dual range current probe with an instrument that has the intelligence to use the appropriate range as the current level changes.