Measuring sleep current without fooling yourself

Sleep current is easy to quote and hard to trust. A fresh lab supply, a long USB cable and a meter in the wrong range will happily show microamps while the modem is still negotiating attach.

We treat sleep measurement as a ritual: same firmware image, same antenna, same temperature, and a log of what the radio was doing in the minute before you cut power to the debug path.

Use a source that can switch between mA and µA ranges without dumping the rail. Lift the UART adapter off the board — leakage through the programmer is the most common “mystery” microamp.

Give the device five minutes after boot before you trust the number. Modems and GNSS engines have warm-up states that look like sleep if you measure too early.

We record three numbers: average sleep over ten minutes, worst spike during a scheduled wake, and the first minute after a forced reboot. That triplet tells you whether the budget is honest for a field deployment.

If the average and the spike diverge, the bug is usually scheduling — not the regulator.