Request Development

Wearable Battery Life Estimator

See how many days your wearable runs on one battery, in about a minute.

  • Days on one cellA conservative, typical and optimistic range for your battery.
  • Your biggest drainThe part that uses the most current, and what to change first.
  • Values you can checkEvery current and margin traced to a datasheet or marked as our assumption.
Estimate battery lifeFree · No sign-up · Datasheet values as of October 22, 2026

The short answer

A Bluetooth LE patch with a PPG (optical heart rate) sensor, a motion sensor and a temperature sensor lasts 77 to 127 days on one CR2032 coin cell, about 108 in our typical case. The rest depends on how long the processor, radio and sensors stay awake, and on how much capacity the cell still gives when the current comes in short peaks.

How it works

  1. Pick your partsChoose the battery, radio, processor and sensors, or start from one of five typical devices.
  2. We add up the currentSleep, processor, radio, each sensor and the self-discharge of the battery become one average current, and the capacity of the cell is cut for its current peaks.
  3. You get days on one cellThree cases from conservative to optimistic, with the biggest drain and what to change first.

The estimate leaves out flash memory, displays, haptics, buzzers, GPS, NFC, the regulator’s own quiescent current and a LiPo protection circuit. Pairing, firmware updates and fast advertising after a lost connection are not counted either, so size the battery for your highest-drain mode; for LTE, each uplink is a proxy from Nordic’s lab energy, without your network’s settings, large uploads or eDRX. The margins assume a body-worn device, not cold use, and a mid-life cell in the typical LiPo case.

What OVA does here, and where our responsibility ends

220devices developed62engineers$125per hourISO 13485certified quality system

We design the electronics, firmware and mechanics of medical devices and take them from concept to design for manufacturing. We are not a regulatory consultancy and not a contract manufacturer: the regulatory strategy and the submission stay with you or your regulatory consultant.

The result is an engineering estimate from datasheet values, not a measurement of your device. Confirm it by measuring the current of your own board. The tool runs in your browser. Your answers are not sent to us; with your cookie consent, Google Analytics records only that the tool was used and the category of the result.

Assumptions and sources

Days = usable capacity / (average current x 24 hours). The currents are typical datasheet values at 25 °C, read on October 22, 2026. The three cases differ in current spread, sleep leakage at skin temperature, and temperature and aging margins; those margins are OVA engineering assumptions and are marked as such in the table. Other parts, such as the MAX30101, MAX86141, MAX30208, nRF54L15 and STM32, are not verified for this version.

ComponentValueConditionsSource
CR2032 coin cell235 mAhTypical capacity to 2.0 V, 15 kOhm continuous (0.19 mA), 21 °CEnergizer CR2032 datasheet [S27]
CR2450 coin cell620 mAhTo 2.0 V, 7.5 kOhm continuous (0.39 mA), 21 °CEnergizer CR2450 datasheet [S28]
CR1632 coin cell130 mAhTo 2.0 V, 15 kOhm continuous (0.19 mA), 21 °CEnergizer CR1632 datasheet [S29]
Coin cell self-dischargeAbout 1% per year (CR2032: 0.27 µA)Room temperature storageEnergizer datasheets [S27] [S28] [S29]; current derived
Coin cell pulse derating, peak up to 10 mA0.83 / 0.80 / 0.77 of capacityCR2032, 1 ms pulses, period 25 ms or longer, end point 2.0 V. Peak = radio, processor, LED and sensors at onceNordic and Energizer white paper [S31] p.8 to p.9; 0.80 interpolated (OVA assumption)
Coin cell pulse derating, peak 10 to 30 mA0.77 / 0.75 / 0.73Same conditions[S31] p.9: 175 to 185 of 240 mAh
Coin cell pulse derating, peak 30 to 80 mA0.73 / 0.50 / 0.29Same conditions[S31] p.8: about 70 mAh at 80 mA peaks; 0.50 interpolated (OVA assumption)
Coin cell, pulses faster than every 25 ms0.5230 mA pulses every 10 ms, end point 2.0 V[S31] p.6
Coin cell, end point 2.4 VUp to 10 mA: 0.73 / 0.73 / 0.70; 10 to 30 mA: 0.73 / 0.58 / 0.42; above 30 mA: 0.29; under 25 ms: 0.42Used when a part needs more than 2.0 V: the ADS1292R analog supply needs 2.7 V[S31] p.9, [S20] p.1; 0.70, 0.58 and the rows above 30 mA and under 25 ms are OVA assumptions
CR2450 and CR1632 pulse behaviourPeak current x 0.67 (CR2450) or x 1.56 (CR1632), then the CR2032 rowsRate capability scales with electrode area, about the radius squaredOVA assumption based on [S31] p.3
LiPo capacityYour nominal capacity; conservative case x 0.974Rated at 0.2C to 3.0 V, 20 °C; 0.974 = 190 mAh minimum / 195 mAh nominal of the example cellRenata ICP402035 datasheet [S32]
LiPo cycle aging1.00 / 0.90 / 0.80More than 80% of minimum capacity after 500 cycles at 0.5C[S32]; split into cases by OVA
LiPo self-discharge1% / 2% / 3% per month2% per month of 195 mAh is 5.3 µAOVA assumption
Temperature margin on capacity1.00 / 1.00 / 0.90Device worn on the body. For scale: a CR2032 gave about 66% of its capacity at 0 °C and 1 mAOVA assumption; Energizer handbook [S30] p.4
Coin cell shelf aging1.00 / 0.99 / 0.981 or 2 years on the shelf at about 1% per year[S27], [S30] p.7; shelf time is an OVA assumption
Active current spreadx 1.0 / 1.1 / 1.3Typical to maximum in datasheets: MAX30003 109 to 120 µA, TMP117 3.5 to 5 µAOVA assumption from [S19], [S20], [S23]
Sleep current spreadx 1.0 / 1.3 / 2.0Leakage rises with temperature: STM32L476 Stop 2 with RTC 1.67 µA at 25 °C, 4.29 µA at 55 °COVA assumption from [S25] Table 35
nRF52832Sleep 2.2 µA; CPU 3.7 mA; radio TX 7.1 mA, RX 6.5 mA; peak 9.2 mASleep: System ON, RTC, full RAM (1.9 µA without RAM plus 0.3 µA, derived). CPU 64 MHz from flash, cache on. BLE 1 Mbps, 0 dBm. 3 V, DC/DCnRF52832 PS [S01] [S02]
nRF52840Sleep 3.16 µA; CPU 3.3 mA; radio TX 6.40 mA, RX 6.26 mA; peak 8.6 mASleep: System ON, RTC, full 256 kB RAM. CPU CoreMark 64 MHz from flash. BLE 1 Mbps, 0 dBm. Peak = CPU + RX. 3 V, DC/DCnRF52840 PS [S03] to [S07]
nRF5340Sleep 1.7 µA; application CPU 3.3 mA; radio TX 4.1 mA, RX 3.7 mA; peak 7.4 mASleep: System ON, 64 kB network RAM, network RTC. Peak = application CPU + TX (derived). 3 V, DC/DCnRF5340 PS [S08]
Bluetooth LE airtime8 µs per byte; advertising packet 16 + data bytes on 3 channels; 230 µs listening per channel when connectableConnection event: window widening 70 ppm + 16 µs, 150 µs gap, 80 µs empty packet. Data adds TX current x 8 µs per byteBluetooth Core Specification as commonly applied; OVA assumption
Bluetooth LE ramp-up and overheadRamp-up 40 / 90 / 140 µs; 1.0 / 2.0 / 3.0 µC per advertising event; 0.5 / 1.0 / 1.5 µC per connection eventCrystal start-up, processor wake-up and stack workOVA assumption, to compare with the Nordic Online Power Profiler [S10]
nRF9151 sleep floor2.7 µAPower saving mode, includes the application core idle; 3.7 V, 25 °CnRF9151 PS [S11] [S12]
nRF9151 application CPU2.7 mACoreMark at 64 MHz from flash, HFINT clock, cache onnRF9151 PS [S34]
LTE-M uplink91 / 94 / 97 mJ per uplink, then x 1 / 2 / 5 for a live networkNordic’s lab energy for a power saving mode tracking area update, 23 dBm. Charge = energy / battery voltage[S12]; the uplink proxy and the multiplier are OVA assumptions
NB-IoT uplink343 / 359.5 / 376 mJ per uplink, then x 1 / 2 / 5Same, NB-IoT[S12]; same OVA assumptions
LTE supply and peaksMinimum 3.0 V, nominal 3.7 V; peaks about 300 mAThe tool blocks LTE on coin cellsnRF9151 PS [S13]; hardware guidelines [S14]
PPG, ams OSRAM AS705825 Hz, tINT 58 µs: 17 µA + LED 15 µA; tINT 117 µs: 28 µA + LED 29 µA; power-down 1.98 µA1 LED at 10 mA, 1.8 V rails, LED supply 5 V, 25 °CAS7058 datasheet [S15] Table 5
PPG LED supplyLED current x 5 V / (battery voltage x 0.85); 10 mA peak at the LEDLED rail boosted to 5 V at 85% efficiencyOVA assumption
ECG, MAX30003100 µA running; 0.51 µA shutdown1.8 V, 25 °CMAX30003 datasheet [S19] p.6 to p.7
ECG, ADS1292R237 µA running (205 analog + 32 digital); about 0.33 µA power-down (1 µW at 3 V)AVDD 3 V, DVDD 1.8 V, 500 SPS; analog supply 2.7 V minimumADS1292R datasheet [S20] p.1, p.10; battery-side figures derived
Motion, Bosch BMI270Accelerometer low power 25 Hz 10 µA, normal 210 µA; with gyroscope low power 25 Hz 420 µA, normal 685 µA; suspend 3.5 µA1.8 V, 25 °CBMI270 datasheet [S21] Table 1
Motion, ST LSM6DSOAccelerometer ultra-low power 52 Hz 9.5 µA, low power 52 Hz 26 µA; with gyroscope, high performance 0.55 mA; power-down 3 µA1.8 V, 25 °CLSM6DSO datasheet [S22] Table 3
Temperature, TI TMP117Every 1 s with 8x averaging (default) 16 µA; every 1 s, single 3.5 µA; one-shot 135 µA for 15.5 ms, then 0.15 µA3.3 V, 25 °CTMP117 datasheet [S23] p.6, p.14; one-shot averages derived
Regulators for 1.8 V and 3 V railsBattery current = load currentLDO assumed; its own quiescent current not includedOVA assumption
A value written “a / b / c” is the optimistic / typical / conservative case. Currents are typical datasheet values at 25 °C unless stated. [S..] points to the Sources below. Data as of October 22, 2026.

Questions

How long does a CR2032 last in a Bluetooth wearable?

It comes down to the average current. Our worked example, a patch with PPG, motion and temperature sensors, averages 67 µA and lasts 77 to 127 days. An ECG patch that streams its signal all the time averages about 190 µA and lasts 30 to 48 days.

Why does my coin cell die sooner than its mAh rating suggests?

The rating is measured at a light, steady load, and the peaks from the radio, processor and LEDs eat into what the cell can give. In tests by Nordic and Energizer, a CR2032 that gave 240 mAh at 0.5 mA continuous delivered 175 to 185 mAh with peaks below 30 mA and about 70 mAh with 80 mA peaks. Parts that need more than 2.0 V also leave capacity in the cell.

Can an LTE-M or NB-IoT wearable run on a coin cell?

No. The nRF9151 modem needs at least 3.0 V, runs at 3.7 V nominal and peaks at about 300 mA, while a CR2032 is 3.0 V nominal and rated at a 0.19 mA drain. Use a LiPo rated for those peaks.

How accurate is this estimate?

Read it as a range: it is a model built from typical datasheet values at 25 °C. Nordic says its Online Power Profiler for Bluetooth LE is typically within 5% for its reference parts. Your board adds regulators, pull-ups, memory and firmware behavior, so measure your first prototype.

Does temperature change battery life?

Yes. In Energizer’s tests at a 1 mA drain, a CR2032 delivered about 228 mAh at 21 °C and about 151 mAh at 0 °C. Sleep current also rises with heat, which is why the tool assumes a device worn on the body.

Related pages

Sources

Data as of October 22, 2026. By Lisa Voronkova, CEO of OVA Solutions. Updated on October 22, 2026.

Request Development