ABZ Signal Troubleshooting for Incremental Encoders: A Systematic Field Guide

Start static: power, levels and termination

Before touching a scope knob, verify the supply at the encoder connector with the cable installed and the encoder running: voltage drop across a long cable or a tired connector is the most common cause of marginal signal levels. Measure each signal's DC state where the encoder idles—differential outputs should show a valid logic level across each pair, and open-collector outputs need the controller's pull-up to show anything at all. Confirm the termination: RS-422 pairs want their characteristic-impedance termination at the receiver end; missing termination shows up later as rounded, ringing edges at speed, while termination where the datasheet forbids it overloads the driver.

These static checks eliminate half the fault families in minutes and, just as importantly, establish which half remains. If supply, continuity and levels are all correct, the fault is in timing, noise or mechanics—and the oscilloscope becomes the tool that decides.

The three-channel view: A, B and Z together

Trigger on A and display all three channels with the axis jogging slowly and steadily. Read four things in order. First, quadrature: B should lag or lead A by a quarter of the signal period, consistently in both directions; a phase error near 90 degrees that varies with position points to mechanical alignment, while a fixed wrong phase points to wiring or a damaged channel. Second, symmetry: high and low times of each channel should be equal; asymmetry indicates signal integrity problems or, on analog-derived interfaces, amplitude imbalance. Third, the index: Z should appear once per scale period, with the datasheet width, and always at the same phase relationship to A and B. Fourth, cleanliness: any glitch on Z between its legitimate pulses is a homing failure waiting to happen.

Direction confusion deserves its own note: if the axis counts backwards, either A and B are swapped at the controller, or the phase convention differs from what the controller expects. Swapping one pair—or configuring the controller's count direction—is the fix; reversing both wires of both channels changes nothing, since that only inverts each logical level.

• Quadrature phase checked in both rotation directions
• Channel symmetry and edge quality inspected at operating speed too
• Z width and Z-to-A phase compared against the datasheet
• Glitches on Z hunted with persistence display and slow sweep

Index pulses: the classic failure

The Z channel fails in characteristic ways. Its pulse can be narrow—a few electrical degrees—so marginal cabling or a slow receiver edge easily erases it; check the width specification and measure it, do not assume it. It can be mistriggered: noise on Z at the moment the controller samples produces spurious home positions, which is why well-designed homing routines qualify the Z edge with the state of A and B rather than trusting Z alone. And it can be consumed by configuration: some controllers expect one index per revolution, others per encoder cycle, and a mismatch between the encoder's reference mark spacing and the controller's expectation looks exactly like a missing index.

When homing is inconsistent, the procedure is to capture Z with A and B displayed, count Z pulses per travel unit against the expected datum spacing, and then check how the controller qualifies the edge. In most cases the encoder is innocent—the fault is in width, noise or configuration—and the scope proves it in one capture.

Faults that only exist in motion

If the encoder is perfect on the bench and faulty in the machine, the remaining suspects are the cable and the environment. Flex the cable by hand at various points while watching the signals on the scope—a dropping channel is an internal conductor fracture, often within centimetres of a connector boot. Watch for glitches synchronized with drive switching or with contactor operation: that is EMC coupling, and the remedy is shielding, grounding and cable separation, not a different encoder. Check that the shield is bonded at the correct end per the interface specification and that encoder cabling does not share a tray or drag-chain compartment with motor power lines.

Document the reproduction conditions before changing any hardware: speed, direction, cable position, nearby equipment. A fault that reproduces predictably is a fault that can be fixed once; a fault chased by parts-swapping without conditions recorded tends to return with the next vibration or the next summer heat.