Checkpoint: Power Math

Checkpoint: Power Math

animation
contact-lenses
connected-health

Smart Lens Power Budget

Scenario: An IoT startup is designing a glucose-monitoring smart contact lens for diabetic patients. The lens must measure tear glucose every 5 minutes and transmit accumulated readings to the patient’s smartphone every 30 minutes via NFC.

Given:

  • RF harvesting antenna: delivers 40 uW at 13.56 MHz when phone is within 5 cm
  • Glucose biosensor: 2 uW during measurement (150 ms per reading)
  • MCU (Cortex-M0+): 12 uW active, 0.5 uW sleep
  • NFC transmitter: 30 uW for 50 ms burst (6 readings per burst)
  • Temperature sensor: 1 uW for 10 ms per reading
  • Target: 18-hour wear time, readings every 5 minutes

Steps:

  1. Calculate number of readings per day:
    • Readings per wear period: 18 hours x 60 min / 5 min = 216 readings
    • NFC transmissions: 216 / 6 = 36 bursts
  2. Calculate energy per glucose reading:
    • Sensor: 2 uW x 0.15 s = 0.30 uJ
    • MCU active (processing): 12 uW x 0.05 s = 0.60 uJ
    • Temperature (compensation): 1 uW x 0.01 s = 0.01 uJ
    • Total per reading: 0.91 uJ
  3. Calculate energy per NFC burst:
    • NFC transmitter: 30 uW x 0.05 s = 1.50 uJ
    • MCU active (packaging data): 12 uW x 0.02 s = 0.24 uJ
    • Total per burst: 1.74 uJ
  4. Calculate daily energy budget:
    • Sensing: 216 readings x 0.91 uJ = 196.6 uJ
    • Transmission: 36 bursts x 1.74 uJ = 62.6 uJ
    • MCU sleep: 0.5 uW x 18 hr x 3600 s = 32,400 uJ
    • Total daily: 32,659 uJ = 32.7 mJ
  5. Calculate required average power:
    • Average power: 32,659 uJ / (18 x 3600 s) = 0.50 uW
    • RF harvesting delivers 40 uW when active
    • Need RF reader proximity for only: 32,659 uJ / 40 uW = 816 seconds = 13.6 minutes/day

Result: The lens requires only 0.50 uW average power, well within the 40 uW RF harvesting capability. The patient only needs their phone near their face for ~14 minutes total per day to fully power the lens. In practice, this happens naturally during phone calls, texting, or deliberate 30-second NFC taps every 30 minutes.

Key Insight: With the given budget, reducing 0.5 µW sleep power to 0.1 µW changes average draw from about 0.504 to 0.104 µW, roughly a 4.8× reduction; biofuel-only operation would still need a demonstrated source and storage budget.

TipAda, Applications and calculation-audit guideCheckpoint: Power Math

You know:

  • An 18-hour wear period with readings every 5 minutes produces 216 readings and 36 six-reading NFC bursts.
  • The chapter’s example totals 32,659 uJ, or 32.7 mJ, which averages to 0.50 uW across the wear period.
  • Because MCU sleep contributes 32,400 uJ, or 99.2% of the total, sleep current is the main lever even though the RF harvest path can deliver 40 uW during reader proximity.

Engineering qualifications

The startup example applies the same five-minute sensing and six-reading upload pattern to Amir’s imagined use. It is a worked energy estimate, not evidence that the required storage or reader contact has been demonstrated.

Result: The estimate requires about 0.50 µW on average, compared with 40 µW harvested during reader proximity. Dividing total energy by harvesting power gives 816 seconds, or 13.6 minutes (about 14 minutes), of ideal reader contact over the wear period. That energy total alone does not show that the lens can run between contacts. It needs a demonstrated storage capacity, peak-power budget, and harvesting efficiency.

The source example proposes 30-second NFC taps every 30 minutes. Its 36 uploads would then provide 1,080 seconds, or 18 minutes, of nominal contact. Phone calls or texting do not establish that a 13.56 MHz reader remains within the stated 5 cm harvesting range. The practical design must show how energy arrives, survives storage, and reaches each subsystem when needed.

Key Insight: With the given budget, reducing 0.5 µW sleep power to 0.1 µW changes average draw from about 0.504 to 0.104 µW, roughly a 4.8× reduction; biofuel-only operation would still need a demonstrated source and storage budget.

MCU sleep contributes 32,400 µJ, or 99.2% of the estimated total. This is why the apparently small continuous draw matters more than the short measurement bursts. The 0.5 µW MCU sleep assumption here differs from the earlier 0.2 µW sensor-only illustration; they describe different budgets and should not be substituted for each other.

Back to the chapter