Troubleshooting TM6 Encoder Index Marker Calibration

Calibration will recover only when the controller captures the encoder index pulse during the homing search; normal motion and correct position counts prove only that the incremental count channels are working. On this TM6, the strongest fault boundary is between the Yaskawa output and the Vigilant PCI card's index-capture function, especially after the Windows 2000 installation was replaced. Restore and verify that signal path instead of bypassing the reference marker.

Symptom quantities and decision path

The number that matters is the count of valid index events accepted by the PCI card during its capture window. The axes reach both hardware limit switches, move back by two shaft turns, repeat the sequence twice, and stop on the third limit-switch contact with Axis calibration failed and X Axis encoder index mark not detected or the equivalent Z-axis message.

Changing sensing distance to 20 mm and moving the limit switch did not change the result. Because both axes fail together after the operating system was reinstalled, investigate shared PCI configuration, device drivers, input definitions, and restored machine files before adjusting two independent encoders.

Index-capture mechanism

An incremental encoder uses repeating position channels for motion feedback and a reference channel that occurs once per revolution. The position channels let the controller count displacement, but their count alone does not identify a unique mechanical revolution. Homing first obtains a coarse location from the hardware switch, then captures the reference marker to establish a repeatable machine coordinate.

A visible pulse at the Yaskawa output is not the same as an accepted event inside the PCI card. The pulse must reach the correct terminal, share the required electrical reference, cross the card's input thresholds with the expected polarity, and remain valid long enough for the input circuit and driver to latch it. The capture function must also be armed while the marker passes. Input threshold, required pulse width, polarity, and capture-window rules must be read from the installed drive and PCI-card documentation because no values are identified for this installation.

Swapping phase A with the reference channel caused the position counter to increase by two counts per revolution. That test shows that the reference waveform can reach a counter-capable path, but it does not prove that the dedicated index input, its mapping, or its latch logic is operating.

Signal-path diagnostic checks

1. Record the original wiring before changing conductors. Return phase A and index phase C to their normal terminals after each substitution test.
2. Use SigmaWin to confirm that the Yaskawa system can locate the marker. This separates motor-and-encoder generation from downstream capture.
3. Measure the index at the Yaskawa output, then at the Vigilant PCI input terminal, using the receiving circuit's reference conductor. Compare polarity, high and low levels, noise, and pulse width. Read the permitted levels and minimum accepted width in the manuals rather than inferring them from a visible oscilloscope trace.
4. Rotate slowly while observing the PCIQuad index indicator. Confirm that position counts change and note whether the red index indication occurs once per revolution.
5. Run the same observation during calibration. A pulse detected during manual rotation but missed during homing points to capture timing, search direction, speed, or arming logic. A pulse never detected points to wiring, electrical compatibility, channel mapping, driver, or card configuration.
6. Compare X and Z at the PCI connector. A common loss on both channels favors a shared reference, card setup, driver, or machine-configuration problem.

This is timing and electrical margin, not ordinary axis logic. A cold-dependent failure can indicate a marginal level, degraded connection, or temperature-sensitive electronics, but heating the drives does not repair that condition. Do not direct a fan heater at a drive; inspect ambient temperature, connections, supply condition, and signal margin through the equipment's approved diagnostic process.

Configuration restoration procedure

1. Identify the exact Vigilant PCI card and installed hardware revision from its label. Obtain the matching driver and configuration utility through the machine builder, MDSI, the card manufacturer, or another official support channel.
2. In Windows 2000, confirm that the card loads without a device conflict and that the intended driver is bound to it. A generic or mismatched installation may expose position counters while omitting index-latch behavior.
3. Restore the known machine configuration for OpenCNC and WinMax. Compare axis-channel assignment, index-input mapping, polarity, capture enable, homing direction, search speed, sensing distance, and any grid-shift setting against a backup or an equivalent working machine.
4. Use PCIQuad to test each encoder channel independently. Require one index indication per shaft revolution before attempting automatic calibration.
5. Run calibration at the configured search settings while monitoring the index input. If the marker appears electrically but the utility never latches it, stop changing mechanics and resolve the card, driver, or mapping fault.
6. After index capture works, restore any moved limit switch to its documented mechanical position and validate its safety and overtravel function.

A limit-switch-only bypass removes the fine reference used by this calibration sequence. It can shift the coordinate origin by switch repeatability, mechanical approach effects, and the unresolved encoder phase. Use such a mode only if the machine builder documents and supports it for this control.

Calibration verification

First verify the diagnostic layer: slowly rotate each motor through multiple full revolutions and observe exactly one reported marker event per revolution. Next verify the sequence layer: during calibration, record limit-switch transition, reversal, marker capture, and calibration completion in order.

Repeat a complete power-down and cold start, then run calibration without manually positioning the shafts. The axes should complete homing without the X or Z index alarms. Repeat calibration from several starting positions and compare the reported home coordinate or a fixed mechanical reference; repeatability matters more than one successful cycle.

If warming the axes by 10 or 20 traverses changes the result, capture index amplitude, pulse width, supply condition, and card detection at cold and warm states. Temperature-dependent success identifies insufficient margin or failing hardware, not a configuration value to conceal.

Recurring pitfalls

Normal movement is often mistaken for proof that the entire encoder interface works. Position counting and index latching are separate functions, so a controller can move, measure, and spin normally while calibration fails. An oscilloscope trace at the drive output also leaves the connector, receiving threshold, polarity, and software latch untested.

Increasing the sensing distance cannot repair an index input that never registers. With 20 mm travel and 10 mm pitch, the shaft rotates twice; repeated failure across that travel directs attention to capture rather than marker placement. Moving the hardware switch can also disturb the intended machine reference and overtravel margin.

Reinstalling the operating system can remove a machine-specific driver, configuration file, registry setting, or channel map while leaving basic PCI counting operational. Preserve the repaired system by recording card identification, driver details, configuration files, axis mappings, and a successful cold-start calibration test.

FAQ