Probe faults divide into unintended triggers, no contact, intermittent contact, and wrong reference values. Test the input state and physical contact separately with the tool clear, then verify repeatability before trusting the probe for Z or tool length.
Distinguish No Contact From False Contact
Classify the failure before probing stock. A noisy probe changes state without contact, a stuck input never changes, and a contact problem triggers inconsistently at the surface. Keep the tool clear and use the controller's supported status view so the first test cannot drive an uncertain circuit into the work.
Verify the Physical Contact Chain
Inspect the complete electrical path: probe body, plate, clip, tool conductivity where relevant, connectors, strain relief, cable routing, contamination, and controller input. Coated tools, anodized surfaces, dust, and residue can prevent electrical contact even when the mechanism touches correctly.
Check Input State Before Probe Motion
Move the cable through its normal range only while the machine is in a safe diagnostic state. If the input flickers with cable position, repair the connection or cable rather than filtering the symptom in software. A probe signal that is stable on the bench but fails near the spindle may also indicate routing or interference.
Read the Input Before Testing Motion
| Observed state | Likely branch | Safe diagnostic action |
|---|---|---|
| Triggered before contact | Short, wrong polarity, contamination, or stuck input | Inspect the circuit and documented configuration with motion disabled |
| Never triggers | Open circuit, insulating contact, disconnected lead, or wrong input | Test each segment of the exact contact chain |
| Flickers when the cable moves | Broken conductor, connector, strain, or interference | Repair the position-dependent fault before probing |
| Stable state but inconsistent height | Surface, mechanics, approach, or repeatability problem | Repeat on a clean fixed reference and measure spread |
A probe circuit has two expected states: open before contact and triggered at contact, or the documented inverse for the exact controller. Confirm those states in the sender or controller view while the cutter is safely clear. If the state is already active, motion testing is the wrong response; find the unintended contact, wiring fault, contamination, or configuration mismatch first.
Next test the complete conductive chain actually used by the job. Touching a loose clip directly to the plate can bypass a coated cutter, dirty shank, insulating surface, damaged lead, or poor connection through the holder. A partial circuit test should not be mistaken for a successful probe setup.
Look for Interference and Moving-Cable Faults
After correction, perform repeated touches on a clean, stable reference using the documented approach speed and pull-off. Record the spread rather than accepting the first successful trigger. Compare results at the center and relevant work positions if cable motion or surface geometry changes across the bed. A cable-dependent probe transition can be isolated with the same state-based reasoning used in Diagnose False CNC Limit Switch Triggers.
Protect the Probe Move With Conservative Travel
Include a pre-contact check in the operating procedure. Many controllers and senders can verify that the input is open before motion and detect a triggered state before starting. The exact behavior depends on the system, so confirm it in current documentation instead of assuming the probe command is fail-safe.
Approve the Complete Zeroing Result
Release the probe with connection method, reference surface, approach conditions, repeat count, result spread, controller or sender context, and stop criteria. A probe that passes once but remains position-sensitive or intermittent should stay out of production until the cause is resolved.
After the signal is stable, the sender's macro still needs its own dry validation: direction, maximum travel, feed, retract, offset math, and response to a missing contact. This article stops at proving the input and physical contact chain. Treat macro logic as a separate layer so a wiring repair is not credited with validating motion commands it never tested.
Do not bypass protective inputs or rely on an online macro whose behavior has not been matched to the exact controller and sender. If the pre-contact or no-contact response is uncertain, keep the tool away from the work until current documentation resolves it.
