Resolving DG2S-16035 Overcurrent Fault on Mach3 CNC Retrofits

Problem Overview

A 1984 Tree 320 3-axis vertical milling machine retrofitted with new CNC electronics exhibits a critical startup fault: the Yaskawa DG2S-16035 servo drives refuse to bring up their green status LED, and the drive manual lists "no indicator lights" as indicating an overcurrent condition. This condition prevents any axis motion, blocks tuning, and leaves the operator unable to bring the machine into a controllable state. Because overcurrent detection in servo drives is intentionally conservative—designed to protect the output stage from catastrophic failure—a no-LED condition on first commissioning almost always points to wiring, motor, or power-supply issues rather than a faulty drive itself.

The motor, the encoder, and the drive are each suspect, and the diagnostic sequence must rule them out in the order that is both safest and most likely to isolate the fault. Working on a live bus without first checking the motor windings with a megohmmeter is a common path to a destroyed drive, so the procedure below starts with the de-energized bench tests and progresses to the powered tests only after the motor and cable have been cleared.

This article walks through the field-proven diagnostic sequence used to isolate the fault. The approach is derived from generic brushed-DC servo drive commissioning practices applicable to the DG2S-16035 family and similar amplifiers used in 1980s machine tool retrofits. Where the original Tree 320 documentation or specific Yaskawa manual page references are known, they are linked to the Yaskawa product portal or the Mach3 support site for verification.

System Architecture: Tree 320 Retrofit Components

The 1984 Tree 320 is a knee-type vertical milling machine originally equipped with analog servo drives and brushed DC servo motors with either resolver or encoder feedback on each axis. A modern hobby/professional retrofit typically replaces the legacy control with the following components:

• Mach3 motion control software running on a Windows PC, providing G-code interpretation, trajectory planning, and operator interface.
• SmoothStepper Ethernet motion controller from Warp9 Tech Design, replacing the parallel port to eliminate the 25 kHz step-rate ceiling and the timing jitter of a Windows non-real-time environment.
• C32 breakout board from CNC4PC, providing optoisolated inputs, relay outputs, charge-pump handling, spindle control, and analog I/O to interface the SmoothStepper's step/direction logic to the drive-level 24 VDC signals.
• DG2S-16035 brushed-DC servo amplifiers driving the original servo motors and feedback devices, retaining the motor-to-ball-screw mechanical interface.
• Original DC servo motors and feedback devices, preserved because the motors were tested and found mechanically sound and able to move freely under manual DC excitation.

The interface chain runs: PC (Mach3) → SmoothStepper (Ethernet) → C32 (BOB) → DG2S-16035 (servo drive) → DC servo motor. The C32 buffers the step/direction signals, the drive closes the velocity and current loops using the encoder/resolver feedback, and the motor turns the axis. Understanding which link is suspect is the first step in any "drive won't come up" complaint.

DG2S-16035 Overcurrent Indication: What "No LED" Means

The DG2S-16035 operator manual describes the front-panel indicator behavior as follows: a solid green LED indicates the drive is enabled and the bus is healthy; a flashing or solid red LED indicates a defined fault code (over-voltage, under-voltage, over-temperature, following error, encoder error); no LED illumination at all, immediately after applying logic power and main power, is the manual's specific signature for an unrecoverable output-stage overcurrent event that has latched the drive into a hardware-protect state.

In practice, this means the drive's internal current-sense circuitry has detected a fault condition severe enough that the controller has refused to energize the output bridge and is not even bringing up the normal fault display. Possible interpretations include:

• A short across the motor armature terminals (P to N, or phase-to-phase on brushless variants)
• A short from the motor output to chassis ground
• A short inside the drive's output transistor bridge
• A regeneration/absorber circuit failure causing the bus to spike and trip the hardware overcurrent comparator
• A logic-power brownout that left the protection latch in a stuck state from a previous event

The drive is generally salvageable in the first four cases; only in the fifth case (a failed output stage) is the drive itself the root cause. The diagnostic procedure is designed to differentiate these cases systematically.

Root Cause Matrix

The table below summarizes the most common causes of the no-LED/overcurrent condition during first commissioning of a brushed DC servo drive retrofit, ordered by likelihood based on field experience with machine tool retrofits of this era.

Pre-Power Safety Checklist

Before applying any power to the drive, perform the following bench tests. All measurements are made with the drive de-energized and the motor leads disconnected from the drive.

1. Visual inspection. Look for pinched wires under the drive mounting screws, solder bridges on the C32 breakout board, missing or broken ferrule crimps, and any signs of arcing on the contactor. Pay particular attention to the motor junction box; coolant migration into the junction box is a common cause of insulation breakdown on knee mills.
2. AC input verification. Confirm the drive is wired to the correct AC tap. A 1.6 kW class DC servo drive of this vintage is typically configured for 230 VAC single-phase on the main bus, but the specific tap and voltage class must be verified against the manual's nameplate. A drive set to a 460 V tap fed from 230 V will produce an under-voltage condition; the reverse will produce a destructive over-voltage.
3. DC bus discharge. Confirm the bus capacitor is discharged through a 10 kΩ resistor before touching any power terminals. The DG2S-16035 can hold lethal bus voltage for several minutes after power-off, and the internal bleeder resistor may have failed.
4. Motor resistance. Measure armature resistance lead-to-lead with a quality DMM. For a healthy DC servo motor in this class, expect values typically between 0.5 Ω and 5 Ω depending on the motor frame and winding. Open circuit (OL) indicates a broken brush lead or commutator open; a value near 0 Ω suggests a turn-to-turn short.
5. Motor insulation. Use a 500 V megohmmeter (megger) between each armature lead and motor frame (chassis). Healthy insulation reads >100 MΩ. A reading below 1 MΩ indicates moisture, contamination, or insulation breakdown and will cause the drive to detect a ground fault on first enable.
6. Cable insulation. With the motor cable disconnected at both ends, megger each conductor to ground and conductor-to-conductor. Pay special attention to the shield, which must be bonded at one end only to avoid ground loops.
7. Encoder cable continuity. Inspect the encoder cable for continuity on each of the 5 V supply, A, B, and index lines. A shorted encoder cable can pull the drive's 5 V rail down and prevent logic initialization.

Step-by-Step Diagnostic Procedure

Step 1: Drive Standalone Power-Up

With the motor leads and encoder cable disconnected from the drive, apply logic power only (typically +24 VDC control power). A healthy drive should bring up the green LED (or a defined fault LED code) within 2-3 seconds of logic power application. The display will often sequence through a self-test, briefly illuminating all LED segments.

If the LED does not come up with the motor and encoder disconnected, the fault is internal to the drive or in the logic supply. Proceed to the power supply verification section.

If the LED comes up cleanly with the motor and encoder disconnected, the motor, motor cable, or feedback device is suspect. Reconnect the motor and proceed to Step 2.

Step 2: Motor Reconnect Test

Reconnect the motor armature leads only (leave the encoder disconnected). Apply logic power. If the LED now fails to come up, the motor armature or the motor cable is shorted. Verify with the megger tests from the pre-power checklist.

If the LED remains healthy with motor reconnected but encoder still disconnected, the encoder or its wiring is suspect. Proceed to Step 3.

Step 3: Encoder Reconnection

Reconnect the encoder cable. Apply logic power. If the LED fails at this point:

• Measure the encoder supply voltage (typically +5 VDC) at the drive's encoder connector. The supply should be present and clean (ripple < 50 mV).
• If the supply is missing or low, the encoder cable has a short or the drive's 5 V regulator is loaded.
• If the supply is present, check the A, B, and index channel signals with an oscilloscope. A shorted encoder output transistor can pull the line low and confuse the drive's position counter, which in some servo drives can manifest as an overcurrent trip during the initialization handshake.

Step 4: Main Power Application

With the drive passing the previous tests, apply main AC power. The bus capacitor charges through the soft-start resistor; a healthy charge takes 1-3 seconds. After charge, the drive should be in a "ready" state and the green LED should be lit steadily. The drive is now ready to accept an enable command.

If the LED drops at the moment main power is applied, the soft-start resistor, the bus capacitor, or the rectifier is suspect. Measure the DC bus voltage at the test points specified in the manual. A 230 VAC input typically produces a 310 VDC bus (±10%).

Step 5: Enable Command

With the green LED on and main power applied, apply the enable (S-ON) input. On most DG2S-style drives, the enable is a 24 VDC signal referenced to the drive's common, or a dry contact closure depending on configuration. Apply the enable and observe:

• LED stays green, no motion: Normal. Drive is ready for step/direction commands.
• LED drops out or goes red: The drive detected a position error, overspeed, or another latched fault during enable. Read the fault code from the 7-segment display (if equipped) or the PC-based diagnostic tool.

Power Supply and Bus Voltage Verification

The DG2S-16035 requires stable logic power and a clean DC bus. Common power-related failure modes include:

• Undersized AC input. Long cable runs from a 30 A branch circuit can cause the input voltage to sag below the drive's UV threshold at startup. Use a clamp meter to measure the AC voltage at the drive terminals, not at the panel, during a charge cycle.
• Missing neutral. Some DC servo drive families derive the control supply from one leg of the main AC. A floating neutral will produce half-voltage control power and a "no LED" condition that mimics an overcurrent fault.
• Bus capacitor failure. Aged bus capacitors in drives of this vintage often exhibit high ESR. Measure the bus voltage with the drive disabled but main power applied; the voltage should hold within 1% of nominal for 30 seconds after disconnect. A rapid droop indicates a capacitor that has lost capacitance.
• Crowbar/regen fault. If the regeneration resistor is shorted, the bus will not reach nominal voltage. Inspect the regen resistor for signs of overheating.

Motor and Cable Integrity Testing

The Tree 320 used brushed DC servo motors with permanent-magnet fields. These motors are robust but have several failure modes that produce exactly the symptom of a "no LED" condition on the drive.

Commutator Contamination

Oil or coolant contamination on the commutator creates a low-resistance path between adjacent bars. This is the most common cause of a "motor reads fine cold but shorts when hot" condition. Clean the commutator with a non-residue contact cleaner and a non-abrasive brush, then re-measure. On a knee mill that has seen decades of cutting fluid exposure, the commutator should be inspected and cleaned as a matter of course during any retrofit.

Brush Wear

Worn brushes reduce contact pressure, increasing commutator resistance, but they do not typically cause shorts. However, a broken brush lead can momentarily open the circuit at certain rotor positions and confuse the drive's current loop. Inspect the brush holders for cracked springs and proper brush length. Most DC servo motors of this era have a minimum brush length specification in the manual; below that length, replace as a set.

Turn-to-Turn Shorts

These develop from thermal cycling and age. The only reliable test is the megger plus a low-voltage high-current test (a "growler" or a regulated current supply that ramps the armature current while monitoring for a sudden drop). The original poster's observation that the motors "move nicely with applying DC" indicates the field and armature are at least partially functional, but does not rule out intermittent shorts. Repeat the megger test after running the motor under load for 15 minutes; a marginal insulation will fail when hot.

Cable Chafing

The motor cable on a knee mill sees constant flexing at the cable carrier. Inspect the cable along its entire length, paying special attention to the cable carrier and the entrance to the motor junction box. A single nick exposing copper to coolant is enough to produce a ground fault that the drive will detect as overcurrent. Replace the cable with a continuous run of shielded flexible servo cable rated for cable-carrier use; do not splice in the cable carrier.

Control Signal and Enable Wiring

The C32 breakout board from CNC4PC is typically configured to source an enable signal to each drive from Mach3's outputs. Common wiring errors include:

• Enable polarity reversed. Most DG2S-style drives enable on a high signal, but some enable on a low. Confirm the manual's spec for the S-ON input.
• Charge pump not enabled. Mach3's charge pump output must be active for the C32 to pass through enable signals. A misconfigured charge pump frequency (typically 12.5 kHz, configurable in the C32 plugin) will hold the drives disabled without raising an explicit error.
• Step/direction outputs floating. The drive's step input is typically a 24 V differential or single-ended input. A floating input can couple noise and look like spurious step commands, which can cause the drive to enter a follow-error state and latch off.
• Common ground loop. The drive's control common, the C32's ground, and the SmoothStepper's ground must all reference the same 0 V. A ground loop in the control wiring can introduce common-mode voltage that confuses the drive's differential inputs.

Verify the C32's charge pump signal with an oscilloscope at the C32's output pin. The signal should be a clean 12.5 kHz square wave with 50% duty cycle. If the signal is missing, the most common causes are (in order): Mach3 charge pump output not enabled in Ports & Pins; SmoothStepper plugin not loaded; C32 jumper set to ignore charge pump (some C32 variants have a bypass jumper that should be removed for safety).

Encoder Feedback Verification

The encoder (or resolver) feedback is what closes the position loop in a servo system. If the feedback is wrong, the drive cannot regulate current properly and may detect an apparent overcurrent as it tries to correct an unstable position.

Verify each of the following with the drive powered and the motor stationary:

• Encoder supply voltage at the drive end: 5.0 VDC ±5% (or as specified). A reading below 4.75 V indicates a loaded 5 V regulator or a cable with too much voltage drop.
• Encoder supply voltage at the motor end: Should be within 0.25 V of the drive-end voltage under load. A long encoder cable with too-small conductors will drop voltage under load and cause intermittent position errors.
• A and B channel waveforms: Use an oscilloscope to view the A and B channels while manually turning the motor shaft. You should see clean 5 V square waves with a 90° phase relationship (quadrature) at a frequency proportional to motor speed. A duty cycle far from 50% indicates a misaligned encoder disk or a failing LED/photodiode pair.
• Index pulse: Once per mechanical revolution, you should see a brief pulse on the index channel. Some drives require the index pulse to be present for proper initialization; a missing index can cause a position-tracking fault that the drive interprets as overcurrent during the first enable.

Drive Configuration, Tuning, and Verification

After the hardware checks pass and the drive brings up the green LED, the next step is to configure the drive for the specific motor and load. The DG2S-16035 typically has a bank of DIP switches or a serial configuration interface to set:

• Motor armature current limit
• Motor armature voltage rating
• Encoder resolution (counts per revolution)
• Velocity loop gain
• Current loop gain
• Enable polarity
• Fault output polarity

If the drive was pulled from a working machine, the original configuration may be appropriate. If the drive is new-old-stock or from a different machine, it must be configured for the specific motor before tuning. The original poster's report that the drives came from the Tree 320 retrofit suggests the original configuration should match, but verify each parameter against the motor's nameplate before applying power.

Tuning Procedure

1. Set all gains to minimum (proportional gain = 0, integral gain = 0).
2. With the drive enabled and the motor unloaded, command a small step input (1-2 mm) through Mach3's MDI panel.
3. Slowly increase the velocity-loop proportional gain until the motor begins to oscillate audibly, then back off 30%.
4. Add integral gain to remove steady-state error. Back off 30% from the oscillation point as before.
5. Add current-loop gain if the drive allows independent adjustment. Most modern servo drives auto-tune the current loop.
6. Repeat for each axis. Run a circular interpolation test (G-code circle) and observe the contour. Tighten the following-error bandwidth until the contour is clean, then back off 20% for safety margin.

Verification

The retrofit is complete and verified when:

• All three drives show steady green LEDs with the machine idle for at least 10 minutes (catches thermal drift issues)
• Each axis responds to a step command with a clean, damped response and a small following error (typically < 0.05 mm)
• A full 30-minute machining test produces no overcurrent, over-temperature, or following-error faults
• The full-axes circular interpolation test produces a circle within 0.05 mm of nominal at the commanded feed rate

If the LED drops out during any of these tests, return to the diagnostic procedure with attention to thermal failure modes: motor overheating, drive overheating, cable insulation breakdown at temperature, and brush wear under load. The C32 breakout board from CNC4PC also benefits from a final visual inspection after the system has been running for an hour, looking for any signs of warmth on the optocouplers that might indicate marginal isolation.

Safety Notes

Servo drives of this vintage store lethal energy in their bus capacitors. Always:

• Wait at least 5 minutes after power-off before opening the cabinet
• Verify the bus is discharged with a high-voltage probe before touching any power terminal
• Never work alone when working on the main power section
• Use one hand when probing live circuits to avoid providing a path through the heart
• Wear ANSI Z87.1-rated safety glasses when working near rotating machinery

The Tree 320's mechanical brake and counterbalance system should be verified functional before any electrical commissioning, since a failure of the Z-axis counterbalance during testing can cause a rapid uncontrolled descent. Reference the original Tree 320 documentation for the brake release and counterbalance adjustment procedures before applying any motion commands.