Resolving NX100 Miller Axcess 300 Erratic Arc on 1mm Stainless

Problem Overview: Intermittent Arc Pullback on Thin-Gauge Stainless

A two-arm Motoman NX100 robotic cell driving four Miller Axcess 300 MIG welding power sources is producing intermittent arc instability on butted seams between 1 mm duplex stainless steel dome tops and cylindrical bodies. The defect signature is consistent: arc appears to "pull back" periodically, the bead cross-section narrows, and sidewall fusion becomes marginal. Between the unstable intervals the same seam welds cleanly with smooth arc, low spatter, and consistent bead profile. The inconsistency happens on the same seam and on both robots, which eliminates a single-arm servo or path-following issue and points instead to a process-side root cause shared by the cell.

This class of symptom on thin-gauge austenitic or duplex stainless is almost always process-related, not a controller fault. The two highest-probability causes on a Miller Axcess 300 / Motoman NX100 combination welding 0.8 mm wire onto 1 mm stainless are:

1. Variable friction in the wire delivery path (drive rolls, liner, spool drag, torch cable stress).
2. Inadequate or remote-only workpiece grounding, which makes arc current sensitive to torch-to-ground distance as the seam progresses around the boiler.

Both must be ruled out before changing weld schedules. The troubleshooting matrix in Section 3 lets you step through them in the correct order without throwing parameters at the wall.

System Architecture: NX100 Controller and Miller Axcess 300 Integration

The cell uses a Motoman NX100 controller paired with Miller Axcess 300 inverter MIG power sources and Miller automatic wire feeders. The NX100 typically handles arc welding through:

• INFORM III job programs with arc-weld macros (ARCON, ARCOF, WVON/WVOF, WSET).
• A weld interface board that maps NX100 I/O to the Miller power source: Arc On, Wire Feed, Gas On, contactor, and fault feedback.
• Optional DeviceNet or EtherNet/IP gateway for advanced Axcess 300 features such as waveform control and pulse parameter sets.

The Axcess 300 is part of Miller's robotic-automation MIG platform and supports both conventional CV and pulsed-MIG waveforms. The Motoman / Miller partnership is formalized through Yaskawa Motoman's status as a Miller Authorized Warranty Repair Center, which is the official channel for cross-vendor escalation when a robot-side problem shows up as a weld-side symptom.

Root Cause Matrix: Why the Arc Pulls Back Intermittently

The "pulls back / goes narrow" signature on thin stainless is mechanically driven. The table below ranks the candidates by how often they appear in this exact cell topology.

The first two rows explain essentially every intermittent pullback report on this cell topology. They also explain why "good weld / bad weld" alternates within a single 360° pass: feed friction changes with arm pose, and arc-circuit resistance changes with torch-to-clamp distance.

Mechanical Wire Feed Inspection Procedure

Wire delivery is the highest-yield inspection point. On a 0.8 mm stainless wire running through a Binzel liner and torch cable, even a small increase in drag is enough to modulate the burn-off rate and produce the observed arc pullback.

Step 1 — Spool and Reel Inspection

1. Decouple the spool reel and confirm the wire pays off with no perceptible drag. Sticky reel brakes or a misaligned adapter add load that the drive rolls cannot overcome consistently at low WFS.
2. Verify the spool sits centered on the hub; an off-axis spool drags wire side-to-side into the inlet guide.
3. Confirm the wire is the correct diameter for the drive roll and liner: 0.8 mm (0.030″) stainless wire typically uses a V-groove or U-groove drive roll labeled 0.030″ / 0.8 mm.

Step 2 — Drive Roll Setting

1. Loosen the drive roll pressure to zero, then increase until the wire just stops slipping under a moderate thumb load on the wire between drive rolls and torch.
2. Perform the standard "pencil test": with the wire feeding and the tip removed, the wire should protrude straight and resist being stopped with a gloved hand. If you can stall the wire, the drive is underset or friction is excessive.
3. Inspect drive rolls for wear or contamination. Steel drive rolls on stainless wire will groove and slip over time; replace when the wire contact face shows a polished groove.

Step 3 — Liner and Conduit

1. Pull the liner and inspect the inside diameter. A liner that is kinked, crushed at the table-mount fitting, or fouled with copper flashing from extended stainless wire use will cause surging feed.
2. Binzel liners are color-coded for wire diameter. Confirm the installed liner matches 0.8 mm wire (typically a blue or marked 0.030″ liner, depending on series).
3. Verify total conduit length and bend radius match the cable carrier routing. Sharp bends or excess length add friction that grows with robot pose.

Step 4 — Torch Cable and Liner Stress Test

1. Disable arc enable, jog the robot to the failing taught point, and hold the pose.
2. Manually feed wire through the torch with the drive rolls running. If the wire stalls or you feel a pinch, the cable is being stressed in that pose.
3. Lift the torch by 50 mm in World Z, repeat the feed test. If the pinch disappears, the cable carrier or torch mount is the culprit; re-route the cable carrier or add a torch-mount rotation clamp.

Workpiece Grounding Strategy for Stainless Steel Assemblies

Stainless steel has roughly four to six times the electrical resistivity of mild steel, so the arc loop is far more sensitive to clamp placement and clamp count. A single ground clamp at one end of a 2 m × 1.5 m boiler shell will produce measurably different arc voltage as the torch progresses around the circumference, and that voltage shift shows up as a narrow, wandering bead.

Recommended Grounding Topology

1. Use at least three to four ground clamps distributed around the boiler shell at roughly 90° intervals, sized for the Axcess 300's maximum output (300 A class — use 300 A-rated Miller grounding clamps).
2. Star the ground leads back to a single, clean copper ground stud on the welding power source. Do not daisy-chain clamps to each other.
3. Remove paint, anti-spatter film, and oxide at every clamp bite. Stainless oxide is insulating; a clamp on mill scale is electrically invisible.
4. Confirm the work table is in the ground path. If the boilers are fixtured on an insulating mat or a painted table, the only return path is the cables you bolted on — and that path degrades with distance.

Voltage Drop Verification

1. With the arc established on a test coupon, measure between the workpiece near the arc and the power source ground stud using a true-RMS voltmeter.
2. Voltage drop should be under ~0.5 V at the working current. A drop above ~1.0 V indicates clamp loss, undersized cable, or insufficient clamp count.
3. Repeat the measurement as the robot moves around the shell. A varying drop confirms distance-dependent grounding as the root cause; install additional clamps where the drop peaks.

Process Parameter Tuning for 1 mm Duplex Stainless

Duplex stainless (typical 2205 grade) has a narrow heat-input window. Too little heat produces lack of fusion and the "narrow bead" symptom already observed; too much heat destroys the duplex ferrite-austenite balance and drops corrosion resistance. The Miller Axcess 300 in pulsed-MIG mode is the correct platform for this material, but only if the schedule is inside the duplex heat-input envelope.

Heat input (kJ/mm) = (Voltage × Current) / (Travel Speed × 1000). On a 1 mm butt, target the low end of the envelope and let pulsed-MIG control the arc energy. If the schedule is in CV mode and the arc looks unstable, switch to the Axcess 300's pulse table; the pulse waveform dramatically improves arc stiffness on thin stainless and is one of the most common fixes for wandering-arc complaints.

Robot Path and Torch Angle Optimization

On thin-gauge butt joints, torch angle is not optional. The two angles that matter:

• Work angle (in the plane of the joint): 0–5° from vertical for a square-butt joint. Excess push angle lifts the arc off the leading edge and produces undercut or wandering.
• Travel angle (in the direction of travel): 5–10° push or drag. Push angle deepens penetration but raises heat input; drag angle flattens the bead. On 1 mm duplex, prefer a slight drag to keep heat input down.

In the NX100 INFORM job, the welding macros WVON / ARCON start the arc at the taught start point. Confirm:

1. The start point sits 2–3 mm off the joint line and ramps onto the joint over the first 5 mm of travel — not directly on the seam, where a small position error becomes a wire-alignment error.
2. The seam-tracking or through-arc-seam-tracking (TAST) offset is enabled and calibrated. The Motoman arc-sensing option uses the Axcess 300's arc voltage feedback to track the joint left/right and up/down; without it, a 0.5 mm positional error on a 1 mm joint will produce exactly the intermittent narrow bead signature observed.
3. There is no abrupt pose change at the failing taught point. A wrist flip or extended reach that torques the torch cable is a hidden feed-friction source.

Vendor Support Resources: Yaskawa Motoman and Miller

When in-cell troubleshooting does not converge, escalate through the official channels rather than guessing at parameter edits:

• Yaskawa Motoman runs the cell-side robotics and the arc-sensing / seam-tracking options. Their field engineers can pull arc voltage traces, INFORM job logs, and I/O state from the NX100 to confirm whether the seam tracker is following the joint or losing it.
• Miller's robotic welding group owns the Axcess 300 power source, the wire feeder, and the weld schedule. They can bench-test the power source against a known load to rule out an internal regulator or pulse-board fault.
• Yaskawa Motoman is a Miller Authorized Warranty Repair Center. A single escalation covers both the robot and the welder and avoids the finger-pointing that happens when each vendor only owns half of the stack.

Field Commissioning Verification Checklist

After applying the mechanical, electrical, and parameter fixes above, run the following verification sequence before signing the cell back to production.

1. Hand-pull test on every robot pose that previously showed the defect. Wire must pull freely at every pose with the drive rolls running.
2. Voltage-drop sweep around the boiler shell at the working current. Every measurement must be below ~0.5 V from workpiece to ground stud.
3. Three consecutive coupons welded back-to-back with no operator intervention. Bead width, penetration, and arc sound must be visually and audibly identical across all three.
4. Cross-section one coupon. Confirm full penetration with no LOF (lack of fusion) on the sidewalls and a ferrite count in the duplex weld metal within the 30–70 FN range.
5. Arc-voltage trace review (NX100 arc-sensing log). Voltage standard deviation across a 360° pass must be < ~5% of the setpoint. Higher variance means the seam tracker is chasing the joint or the grounding is still moving with torch position.
6. One full production shift on the original part. Monitor tip-change interval; if tips are burning back faster than the documented 4–8 h arc-time baseline, the grounding or schedule still needs work.

Preventive Maintenance Schedule for This Cell