Detents per revolution cap the granularity a single turn can achieve

What it is

A knob with 12 steps per turn, twisted through a comfortable half turn of the wrist, yields at most six countable landings; 30 steps per turn yields more in the same motion, with each step narrower. Detents per revolution cap the granularity of a single turn: software can step finer, but if two events cannot be felt apart, the hand’s ceiling is still that circle’s step count. Damping, pawl versus motor, and the resistance curve are separate. This layer only asks how dense the steps are.

Why it happens

A comfortable finger twist is limited by forearm angle, roughly 60°–120°; few people turn a full 360° and stop precisely in one motion. The granularity ceiling is therefore “comfortable angle × steps per turn,” not the encoder’s electrical pulse count. Too sparse, and one twist spans too much parameter, so trim becomes many tiny motions. Too dense, and neighboring drops crowd two-point threshold, counting degrades from “step, step” into “a stretch of sand,” and quantized feedback exists in name only. Whether software maps a step to 0.1 dB or 5% is parameter alignment; it does not change how many steps the hand can count. Electrical resolution above the mechanical step count produces pulses with no tactile event, and people will not use them.

Studying it

Have people complete a named number of units in one twist (“up 3,” “up 15”) across different steps-per-turn.

Independent variables: steps per revolution, comfortable twist angle (sit/stand, knob diameter), parameter increment per step. Dependent measures: steps actually traversed in one twist, counting error, switching to many small twists, abandoning counting for the display.

Report electrical pulse count and feelable step count separately. Time-to-target makes overly dense steps look fast—people are watching the display, not counting.

Where it stops holding

An endless encoder with an acceleration curve (faster turning makes each step a larger jump) voids “steps per turn” during a flick, substituting another granularity. End-stopped pots that do not span a full turn have an effective count between the stops, not 360° times density. Gloves smear dense steps first, so effective steps per turn drop. Ten-turn pots on instruments spread granularity over many revolutions; the ceiling of a single twist is still set by steps per turn, only the total range grew.

Applying it

• Back-solve steps per turn from “how many units a comfortable twist must cover,” not from how fine the encoder can report.
• For eyes-free counting, keep steps per turn in the band the pad can separate one by one; finer steps go to a second stage or a hold-to-accelerate, not a packed circle.
• Do not make software steps finer than the hand’s steps unless a numeral that must be watched sits beside them.
• Verify: hide the readout, ask for N units in one twist, sweep N and steps per turn. When counting smears or people look at the display, density has passed the ceiling. In gloves, record the largest still-separable steps-per-turn as the field spec.

Related

• Same group: C10.10.2 Mechanical detents come from a physical pawl; electronic detents are motor-simulated, and they feel different · C10.10.3 Too little damping lets a knob turn by itself under vibration · C10.10.4 The resistance curve between detents decides whether people can count steps by feel alone
• Adjacent: C10.02 Knobs and Continuous Adjustment · C1.03 Control-Display Gain
• Search: detents per revolution · encoder resolution · adjustment granularity