Unsupported arm raise fatigues quickly
Aliases: gorilla-arm fatigue · unsupported mid-air hold · deltoid burnout
What it is
Both arms held out toward an invisible pane of glass. Within minutes the deltoid burns and the shoulders drop. That unbraced upper-limb pose has a circulating name: gorilla arm. Unsupported arm raise fatigues quickly: the task is not hard; the arm’s weight is being carried by the shoulder girdle alone, and the anti-gravity maintenance power spends local endurance in short order.
Quickly is relative to a desk. The same hand on a tabletop can work an hour; held in the air, it often fails to outlast a demo. Remove the support and the time scale falls from hours to minutes.
Why it happens
When the upper arm abducts or flexes, deltoid and supraspinatus must continuously produce a moment against gravity; the moment arm is the horizontal distance from shoulder to the arm’s centre of mass. Larger abduction, longer moment arm, more muscle force for the same arm weight. Near-isometric contraction piles metabolites faster than motor units can rotate; the ache and the dropped shoulder arrive together.
A desk, a chair arm, a belt, the other hand — any of these can hand that moment arm to an external structure and drop the shoulder’s maintenance power toward zero. XR’s default pose often refuses those structures on purpose: the hands must point, grab, and gesture in air. So “natural 3D operation” puts people in a posture already abandoned on touch screens — a vertical surface, both hands hanging. Fatigue is not a side effect. It is the posture’s main result. People drop the elbow first, then drag the interface in, then simply put the hands down; the interaction starts changing shape before the muscle is done.
Studying it
Use methods such as consumed endurance, which maps joint angles of a pose onto a rate of endurance spend, or deltoid surface EMG and a Borg rating, and compare time-to-failure with and without support, at different abduction angles.
Independent variables: shoulder abduction / flexion angle, elbow rest present or not, one hand versus two, mass (empty hand / controller). Dependent variables: time until the pose cannot be held, drop in EMG median frequency, rated ache, centimetres of shoulder drop.
Time-to-failure falling steeply with abduction is the most stable result. Do not take a five-minute demo as “not tiring” — the gorilla-arm curve steepens after those minutes. Controller mass cuts endurance of the same pose further; empty-hand numbers do not transfer to a held device.
Where it stops holding
Park the elbow on a desk edge, the belly, or the other hand, and the premise drops; hold time returns to desktop scale. In water, lying down, or with an exoskeleton taking the arm, the gravity term is cancelled or moved. Operations with tiny abduction, hands hanging near the hips, barely trigger gorilla arm; the degrees that take the upper arm off the torso do. Individual differences are large: trained shoulders, younger people, and anyone who works overhead can last longer — that is not a product session length. Lab curves under “hold as long as you can” last longer than the field, where nobody is competing with themselves on endurance.
Applying it
- Default to upper arms by the torso and elbows down. Do not put the main surface at a height that needs a horizontal raise.
- When a brief raise is required, budget its duration in minutes and give a legal drop: hands down means idle, not a system shouting to bring the hands back.
- Allow real edges (desk, chair arm, one’s own knee) as elbow rests. Do not design them out in the name of “pure mid-air.”
- How to check: have people work in the product’s default pose, in silence, and watch when the elbows start to fall. Elbows dropping before the task ends means the raise is already fatiguing fast — lower the surface until the drop appears after a session length you are willing to claim.