Adaptive Sampling Charge Budget

Ada re-derives this chapter’s own numbers step by step, at full precision

ADA · CALCULATION AUDIT

Adaptive Sampling Charge Budget

The adaptive-sampling widget defaults to a stable interval of 60 seconds and an active interval of 10 seconds, split 70% stable, drawing on a sensor that pulls 30 mA for 3 seconds and a radio that pulls 170 mA for 2 seconds against a 2000 mAh battery. The chapter frames this as a way to cut energy by only sampling fast when a rolling standard deviation of the last 10 readings crosses a threshold, transmitting only when the value changes by more than 5%. This audit asks the question that design invites: how many days does the fixed 10-second strategy actually buy against the adaptive 70/30 split, once every mAh is added up?

Companion to the chapter Power Accuracy and Field Hardening — every number here comes from that chapter.

Ada: Treat the power model as charge accounting. The default widget values already give the sample intervals, stable/active split, threshold-filter rule, load currents, and battery size; the check below only turns those values into mAh per day.

The fixed strategy samples every 10 s, so it takes 24 x 3600 / 10 = 8640 readings per day.

• Sensor charge: 8640 x 30 mA x 3 s / 3600 = 216 mAh/day.
• Radio charge: 8640 x 170 mA x 2 s / 3600 = 816 mAh/day.
• Sleep charge: 0.010 mA x 24 h = 0.24 mAh/day.
• Fixed total: 216 + 816 + 0.24 = 1032.24 mAh/day; a 2000 mAh battery lasts 2000 / 1032.24 = 1.94 days.

The adaptive default spends 70% of the day at the 60 s stable interval and 30% at the 10 s active interval:

• Stable readings: 24 x 3600 x 0.70 / 60 = 1008.
• Active readings: 24 x 3600 x 0.30 / 10 = 2592.
• Total readings: 1008 + 2592 = 3600, which is (8640 - 3600) / 8640 = 58.3% fewer readings than the fixed case.
• Threshold-filtered transmissions: 3600 x 0.30 = 1080.
• Adaptive sensor charge: 3600 x 30 mA x 3 s / 3600 = 90 mAh/day.
• Adaptive radio charge: 1080 x 170 mA x 2 s / 3600 = 102 mAh/day.
• Adaptive total: 90 + 102 + 0.24 = 192.24 mAh/day; a 2000 mAh battery lasts 2000 / 192.24 = 10.4 days.

The physics conclusion is narrow: adaptive sampling helps because it reduces both sensing time and radio bursts, but the 1 deg C change threshold is still a data-quality constraint. Lower energy is only a win if the threshold does not hide the event the sensor is supposed to catch.

Every number above is taken from the chapter’s own material and re-derived step by step.