How to Fix FSW Tunnel Defects with Parameter Control
Table of Contents
- Why FSW Tunnel Defects Appear in Aluminum Welds
- Root Causes Behind FSW Tunnel and Void Formation
- Parameter Combinations That Eliminate FSW Tunnel Defects
- Tool and Fixture Conditions That Reinforce the Fix
- Correcting FSW Tunnel Defects Without Repeating Scrap
- Common Questions About FSW Tunnel Defects
- Can a tunnel defect be fixed by increasing spindle speed only?
- Will a visual inspection catch a tunnel defect?
- How do I know whether the tunnel is caused by parameters or by tool wear?
- Which FSW process record signals should I review first?
FSW tunnel defects are a containment failure, not just a weld quality issue. They appear when plasticized material stops consolidating behind the pin, and the void location tells you which side of the process window moved. Most of the time, the fix is a controlled adjustment to spindle speed, travel speed, or downward force, not a blind increase in heat input. The sections below cover the root causes we work through on 6xxx aluminum projects and the parameter corrections that close tunnels without creating flash or excessive thinning.
Why FSW Tunnel Defects Appear in Aluminum Welds
In a sound friction stir weld, the tool shoulder constrains softened material while the pin shears and transports it around the probe. A tunnel defect is a continuous internal cavity left along the trailing edge when that flow path does not close before the material cools and regains strength. The cavity is not the same as a spherical gas pore or a surface groove. It is elongated, frequently opens on the advancing side, and can run for tens or hundreds of millimeters inside a seam that still looks acceptable from the cap side.
The first diagnostic step is not to change a parameter. It is to record where the defect begins and ends. Tunnels that start at the weld start often point to insufficient initial consolidation or a cold tool. Tunnels that appear after a steady weld section point to a process variable that changed mid weld, such as plate gap opening, force drift, or tool wear. Tunnels at the exit usually point to inadequate runout and are often the least complicated to correct.
This defect is primarily a material containment problem. Heat input participates, but it is not the only variable. A parameter change that adds heat without restoring shoulder contact or workpiece clamping can simply turn a tunnel into flash and undercut. That distinction matters because it changes the correction order.
Root Causes Behind FSW Tunnel and Void Formation
Before changing any parameter, we rule out causes in the order below because different root causes require different corrections. The most common root cause in our 6xxx work is travel speed set too high relative to spindle speed. Downward force problems come next, then tool condition.
| Field observation | Root cause to rule out first | Correction direction |
|---|---|---|
| Tunnel on advancing side, smooth external surface | Travel speed too high for spindle speed | Lower travel speed or raise spindle speed |
| Intermittent tunnel with slight edge mismatch | Insufficient downward force or loose clamping | Increase Z-axis force and tighten fixture |
| Tunnel after a mid-production tool change | Worn pin features or reduced pin length | Inspect and replace the tool, then recheck plunge depth |
| Root void under a sound cap | Pin length too short or excessive tilt | Increase plunge depth or correct tilt angle |
Travel speed is the first variable we check because it controls the time available for material consolidation. Downward force is the second check because, in position control, the shoulder can lift slightly as the plate deflects, and a 0.1 mm lift changes containment pressure enough to open a tunnel. Tool condition is third, not because it is uncommon, but because it should not be changed until the process record has been evaluated.
Parameter Combinations That Eliminate FSW Tunnel Defects
Parameter corrections work only when they match the root cause. Use the table above as a starting point, then work through this sequence for a 3 to 6 mm 6xxx aluminum butt joint.
- Lower travel speed first. If the tunnel is continuous, reduce travel speed by 10% to 20% while holding spindle speed constant. The extra dwell time under the shoulder gives softened material time to consolidate behind the pin.
- If the travel speed reduction does not close the void, raise spindle speed in 50 to 100 rpm steps. More rotation adds heat and material transport, but too much rotation can overheat local material and create another flow defect.
- Increase downward force or plunge depth. On a force control machine, add force in 1 kN increments and watch spindle torque and the weld surface. On a position control machine, increase plunge depth in 0.05 mm increments.
- Check tilt angle. For most butt joints in 2 to 6 mm aluminum, a tilt of 1 to 3 degrees toward the trailing edge keeps the shoulder in control of the softened zone without dragging the surface.
- Record spindle speed, travel speed, downward force, and spindle torque for every step. Without the record, a successful fix cannot be repeated on the next fixture or the next shift.

Spindle characteristics also set the practical correction window. A high torque spindle holds a commanded speed under load, which makes a travel speed change repeatable. A lower torque spindle may sag in speed as the tool enters, and that sag can open the same tunnel even when the commanded speed looks correct. If the actual spindle speed cannot be held within a small range, the fix belongs in the spindle or tool holder selection, not in the weld recipe.
If your tunnel defect appears with 7xxx or dissimilar stacks, it is worth confirming the tool shoulder and pin combination before locking the parameter window, because those materials respond differently to the same heat input. Send your alloy, thickness, and joint design to [email protected] and we will confirm the tool and force envelope.
Tool and Fixture Conditions That Reinforce the Fix
Tool condition is the third check after travel speed and downward force, but it becomes the first check when a previously stable recipe starts producing tunnels. The common failure pattern is not a sudden tool break. It is a slow loss of pin feature height over several hundred parts. As the feature wears, the tool transports less material per revolution, and the shoulder has to contain a smaller volume of plasticized metal. The machine record often shows spindle torque dropping, which is the signal to inspect the tool instead of adding more heat.
For butt welds, measure pin length before installation. In aluminum sheet between 2 and 5 mm, a pin that is 0.1 to 0.2 mm shorter than expected can leave an unfilled root channel while the surface looks completely normal. On our machines, we record pin length and shoulder diameter before each production run because a tunnel that appears after a tool change is usually a tool setup issue, not a parameter drift.
Tool geometry should be reviewed once the weld record rules out a straightforward parameter drift. <AEE provides customer with optional features FSW tool> covers how optional pin features and shoulder profiles are matched to material flow requirements, which is a useful check before committing to a new tool for a persistent tunnel.

Fixture behavior can mimic a tool problem. If the tunnel appears on one side only in a long profile, check clamping before changing the recipe again. A gap between sheet edges opens a path for softened material to escape, and the backing bar may not be making full contact. Add local clamping or adjust the backing bar only after measuring the actual gap and the current clamp force.
Correcting FSW Tunnel Defects Without Repeating Scrap
Repeated tunnel defects are expensive because the same seam is often cut, reworked, or scrapped before the root cause is clear. The lower cost path is to validate the parameter correction on a start tab and a runout tab before welding the production part. A corrected setting should produce a smooth surface, a stable torque trace, and no visible exit void on the tab. If those conditions hold, move to the production seam and still record the first part in full.

When the tunnel persists after parameter, tool, and fixture checks, the issue is usually a machine capability boundary. In those cases, we review force capacity, spindle torque, and tool shoulder design together. If you are working on a project where the tunnel defect resists standard adjustments, send your part number, alloy, joint thickness, and current parameters to [email protected] or call +86 18325808715. We will work through whether the problem belongs in the recipe, the tool, or the machine.
Common Questions About FSW Tunnel Defects
Can a tunnel defect be fixed by increasing spindle speed only?
In some cases yes, but only if the spindle speed increase is paired with enough shoulder containment. Tunnel defects often close when more rotation adds heat and material transport. If the increase is too large, the weld surface will show flash or the joint will thin, which replaces the tunnel with a different defect. For 6xxx aluminum in the 3 to 6 mm range, I first try a 50 to 100 rpm increase and watch the torque trace. If torque rises but the void does not close, keep the speed change and move to downward force.
Will a visual inspection catch a tunnel defect?
A clean cap surface does not rule out a tunnel. Tunnel and wormhole defects sit below the surface, so visual inspection catches only the exit opening or secondary surface distortion. In many cases, the first reliable indication comes from radiography, phased array ultrasonic testing, or a sectioned coupon. If you only check the weld face, you can pass a structurally weak seam. That is why we qualify a new parameter window with X-ray or ultrasonic inspection before releasing it to production.
How do I know whether the tunnel is caused by parameters or by tool wear?
It depends on when the tunnel appears. If a previously stable recipe starts producing voids after a tool change, measure the pin length and inspect the pin features before adjusting the recipe. If the tunnel appears on a new part number or alloy, check the parameter window first. A tunnel that grows slowly over a production run is usually tool wear, because material transport per revolution drops while the commanded parameters stay the same. Confirm by comparing spindle torque and the weld surface between the first part and the latest part.
Which FSW process record signals should I review first?
In our process reviews, I start with spindle torque and Z-axis force because those two channels show tool load and shoulder engagement. A dropping torque trace with a stable Z position often means the tool features have worn. A falling Z force in force control can mean the shoulder is lifting as the sheet deflects. The weld record tells you which variable moved first, and that is more useful than the average value alone. If the record shows a pattern you cannot match to tool or fixture behavior, share the log and weld setup with us at [email protected] and we will confirm whether it points to a parameter, tool, or machine issue.
If you’re interested, check out these related articles:
AEE provides customer with optional features FSW tool
Customized Dissimilar Material Welding Manufacturers in China
Refill Friction Stir Spot Welding (RFSSW)