Skin contact at the wrist makes haptics reliable

What it is

A watch presses its actuator against skin with a strap. Contact is the default, not a maybe of grip. A phone can sit in a pocket, a bag, or on a desk, and the buzz often never arrives. If the watch is still on, a short tap usually reaches the same patch of skin. This entry is about whether coupling is stable. It is not about whether haptics can stand in for a visual confirm, and not about how many rhythms you can encode.

Why it happens

To be felt, energy must cross the case and strap, reach skin, and clear that site’s detection threshold. Dorsal wrist skin is thin and close to bone, so a short pulse crosses threshold more readily than on the palm. Preload from the strap keeps the case on one contact patch; a handheld device changes the transmission path with every grip. A watch can therefore run a smaller amplitude than a phone and still read as a tap. Reliability is conditional: a loosened strap, a face slid toward the ulna, a sleeve between case and skin, or a winter wear outside the cuff lengthens or breaks the path. The actuator is not broken. The mechanical contact premise is gone. People do not interpret a miss as “loosen the strap.” They interpret it as the system not responding.

Studying it

Measure detection rate under real wear, not acceleration in a jig. Have people wear the watch at everyday tightness and play the same pulse while walking, after sweat, and under a sleeve; they report whether they felt it. Contrast the tightest hole with the loosest hole that still does not fall off.

Independent variables: strap preload, contact site (dorsum / toward ulna), sleeve, skin wetness, pulse amplitude. Dependent variables: detection rate, misses reported as “no system response,” walking shocks mistaken for haptics.

Gel pads and clamps in the lab overestimate detection. Walking’s low-frequency shocks raise the noise floor; thresholds must be calibrated in motion, not copied from sitting.

Where it stops holding

Off-wrist, on a charger, or fashion-loose enough to spin, the premise fails and haptics cannot be the primary channel. Residual limbs after amputation, severe peripheral neuropathy, or a strap on a wound: contact is not the same as sensation. Diving and thick gloves put material in the path. A child’s strap that can only be set to “won’t fall off” rather than “flush” will detect like a phone in a pocket. This is not about how sharp a waveform a motor can make—sharp still has to land.

Applying it

• Make must-feel events (payment done, timer end, safety alert) short pulses, assuming the watch is being worn; keep a visible or audible backup for when contact is poor.
• Do not copy a phone-strong amplitude to the wrist: too strong hurts and reads as a fault. Calibrate the just-above-threshold step at the target preload.
• Offer a repeatable “try haptic” in settings so people can confirm feel at their current tightness, rather than passing only a factory jig.
• Verify by having people walk on their own hole setting with pulses inserted at random. If misses cluster on loose holes, sleeves, or post-sweat, fix wear guidance or the backup channel before raising amplitude.

Related

• Within the group: K4.04.2 Haptics can confirm without looking at the screen · K4.04.3 The number of haptic patterns is capped by discriminability
• Adjacent: D3.02 Vibration Intensity, Duration, and Rhythm · D3.14 Cross-device Haptic Capability Differences · A4.02 Two-point Discrimination Threshold
• Search terms: haptic coupling · wrist-worn · detection threshold