A homing cycle is repeatable only when machine zero returns within the tolerance required by the next setup. Measure the same reference through multiple clean cycles and separate switch variation, approach behavior, mechanics, and measurement noise.
Define the Reference Point to Be Tested
Homing repeatability is the spread among repeated references, not whether the display reaches zero. Define a physical checkpoint or measurement method that can resolve the tolerance the job needs. The switch signal, axis deceleration, pull-off move, mechanical stop, and controller sequence all contribute to where the machine finally declares home.
Warm Up Only as Current Guidance Requires
Run a cold sequence and a sequence after normal operation if temperature or machine state may matter. Approach from varied starting positions within safe travel. Record the measured home location after each cycle without moving the indicator or reference target between runs.
Run Repeated Home-and-Return Cycles
Separate random spread from a consistent offset. Random results suggest switch actuation, connector, cable, contamination, mechanical looseness, or motion variation. A stable but wrong location may follow a changed fixture, reference target, pull-off setting, or coordinate relationship. The remedies are not interchangeable.
Compare Axes and Approach Behavior
Test axes individually where the controller allows a supported diagnostic method. A combined home cycle can hide which axis contributes to a shifted work origin. Never defeat an input or drive the carriage against a hard stop to finish the experiment. If return error grows with commanded distance, investigate scale through Calibrate CNC Steps per Millimeter Without Hiding Backlash before judging the homing switch.
Set a Fixture-Specific Acceptance Limit
Relate the measured spread to the downstream workflow. A loose-tolerance one-off may not rely on homing for part location, while a camera, jig, multi-session job, or automatic tool setting may require much tighter return behavior. Define the acceptance limit from that dependency before calling the result good or bad.
Store the Baseline for Drift Checks
Archive cycle count, starting positions, machine temperature or state, switch and pull-off configuration, measurement method, per-axis results, and final limitation. After switch replacement, frame movement, firmware changes, or transport, repeat the test before reusing fixture coordinates that depend on machine home.
