A wearable is a heater strapped to a person. The battery is one source of that heat; the SoC, display driver, and charging circuit are others. The design goal isn’t to minimise temperature — it’s to keep the skin-facing surface below a threshold the user’s nerves won’t complain about.

The number you have to respect

IEC 60950-1 and IEC 62368-1 both define the skin-contact limit for continuous contact with a metal surface at around 41 °C in a room-temperature ambient. Plastic or glass enclosures get a few degrees of slack, but most product teams anchor to 41 °C because it’s the strictest number they might be tested against, and because users begin to describe devices as “warm” around 40 °C and “hot” around 42 °C.

Cosmetics aside, going above 43 °C for extended periods starts to cause low-grade thermal damage to skin — a real regulatory concern for medical wearables.

Where the heat actually comes from

SourceTypical loadDominant during
Battery internal resistance10–150 mWHigh discharge (TX burst, motor)
SoC / compute0.5–3 WActive use, wake events
Display driver0.3–1.5 WAlways-on display, high brightness
Charging (losses)0.2–1 WCharging to 80–100% SOC
Radios (BT/Wi-Fi)50–300 mW avgStreaming, sync

On most AR glasses and smartwatches, charging is the thermal worst case, not use. The cell is hot from I²R losses in the CC phase, the charger IC is bleeding the CV phase as heat, and the device is sitting still on a cradle with no natural airflow.

Three heat paths you control

  1. Spread. Copper or graphite foil across the inside of the enclosure turns a point source into a large radiator. A 30 × 20 mm graphite pad on top of the cell reduces peak local temperature by 4–7 °C compared to bare plastic.
  2. Insulate on the skin side. A 0.15–0.3 mm aerogel or silicone pad between the cell and the skin-facing surface creates a small thermal gradient, pushing peak skin temperature down by 2–4 °C. Costs runtime indirectly (slightly worse cell cooling), so tune carefully.
  3. Throttle at the BMS. The cheapest way to cap temperature is to stop charging fast. Skin-contact temperature limiting is now a standard feature of modern fuel gauges — trigger a charge-rate reduction when the battery-surface NTC crosses 38 °C; full pause at 41 °C.

The charging profile that actually works

A three-stage profile survives most real-world environments:

  • 0–80% SOC: charge at up to 0.7C, constant-current.
  • 80–95% SOC: taper down to 0.3C, entering CV.
  • 95–100% SOC: trickle only. Most users never need the last 5% in a hurry, and stopping at 95% more than doubles cycle life.

Overlay a thermal envelope on this: if the NTC ever reads above 40 °C, divide the current-at-that-stage by 2. If it hits 43 °C, pause entirely for 60 seconds. User-perceived charge time barely changes, but the fail-case skin temperature is controlled.

On-wrist vs off-wrist

A smartwatch charging on a cradle behaves differently from a watch charging while being worn (over-night on the wrist, for example). The wrist acts as a heatsink but also an insulator — net effect varies by about ±2 °C. If your device supports both, you must sense which mode you’re in (typically via the PPG sensor or a capacitive skin-contact electrode) and adjust the thermal envelope.

Testing: the 30-minute session

The most useful thermal test we run is a 30-minute continuous-worst-case session: maximum brightness, continuous streaming audio, GPS on, BT on, ambient at 30 °C. Measure at five surface points every 10 seconds. If any point crosses 41 °C during the session, the design is not done. Repeat at 40 °C ambient for the outdoor-use worst case.