Troubleshooting: Die Casting Dimensional Drift

What die casting dimensional drift actually looks like

Dimensional drift is not a single defect. It is a trend. You receive a lot of castings, and the parts are slightly larger or smaller than the previous lot. The next lot drifts a little further. Individually, each part may still pass. Together, they show a pattern that will fail downstream assembly or cause fit issues in production.

This pattern shows up in CMM reports, go-no-go gauge results, or in-field assembly complaints. The part may be 0.1mm too large at a boss, or 0.2mm short at a bearing seat. The shift is consistent. It is not random scatter. It is directional.

The problem is that dimensional drift hides in plain sight. First-piece inspections pass. Incoming inspection may pass if the acceptance limits are wide. The drift only becomes visible when you plot the data over time or when the part stops fitting in the assembly.

How to diagnose the trend before it becomes a quality failure

The first step is measurement. Do not rely on a single lot sample. Pull CMM data from the last three to five lots and plot the critical dimensions against the nominal value. You are looking for a slope, not just a point outside the limit.

If the data shows a consistent shift, the part is still in tolerance. If it shows a slope, you have a problem even if every point is green. The slope tells you the part will fail.

Common symptoms appear in specific ways:

• A bearing seat that grows 0.05mm per lot.
• A flange thickness that shrinks 0.1mm over a production run.
• A wall thickness that varies by 0.08mm between the start and end of a die set.
• A part that passes today but fails in the next production cycle.

The key is to separate process variation from mold variation. Process variation changes with operator settings, machine pressure, or melt temperature. Mold variation changes with wear, corrosion, or damage. Both can cause drift, but the correction differs.

Likely causes of consistent size changes

Symptom	Likely cause	What to do
Gradual increase in part size across lots	Mold wear at the parting line or core pins	Request a mold condition report. Review die maintenance logs. Check for scoring or corrosion at the parting surface.
Sudden shift in one dimension after a material change	Different alloy composition or melt behavior	Verify material certifications. Confirm the supplier used the specified alloy grade. Check melt temperature records for the affected lot.
Size change correlates with production volume	Die overheating or lubricant degradation	Request the process control chart for the affected run. Check machine oil temperature and die lubricant application records.
Variation between two dies in the same set	Uneven mold wear or uneven lubrication	Run a trial on both dies. Compare CMM data. Check for asymmetric wear patterns.
Dimensional change after a tool repair	Rebuild tolerance error or rework misalignment	Review the repair documentation. Verify the rebuild used the original drawing tolerances.
Slow drift over weeks	Progressive wear on a critical feature	Increase measurement frequency. Request a mold inspection. Plan a corrective action before the next lot.

The table above covers the most common patterns. In practice, the cause is often a combination of factors. A worn die plus a slightly high melt temperature will drift faster than either alone. The diagnostic work is to isolate the dominant variable.

How to correct the drift at the supplier level

Once you identify the cause, the correction depends on the supplier’s capability. A small zinc alloy shop may not have the same process controls as a large aluminum die casting plant. Do not assume they can fix it the same way.

For mold wear, the correction is physical. The supplier must inspect the die, resurface the worn areas, or replace the worn component. If the die is out of service, the supplier should tell you before you pull the next lot. A supplier that sends a drifted lot without warning has a communication failure, not just a process failure.

For process parameter issues, the correction is procedural. The supplier should adjust the machine settings, change the melt temperature, or adjust the lubricant cycle. They should document the change and confirm the next lot passes. If they cannot do this, the issue is capability, not effort.

For material variation, the correction is at the source. The supplier should verify the alloy batch, reject the affected material, or substitute a certified batch. If the supplier uses multiple foundries for the same alloy, the variation may come from the melt source. You may need to require a single supplier for the material.

The correction should not be a one-time fix. It should change how the supplier monitors the part going forward. If the drift happened because the CMM check was too infrequent, the supplier should increase the frequency. If it happened because the mold inspection schedule was too long, the supplier should shorten it.

How to prevent dimensional drift from recurring

Prevention is about making drift visible early and making correction automatic.

1. Define the critical dimensions in the purchase order. Do not just reference the drawing. List the features that must stay stable and the acceptance limits.
2. Require CMM reports with a minimum of five points per dimension. A single point is not enough to show a trend.
3. Ask the supplier to plot the last five lots on the quality report. If they do not do this, ask why. A supplier that tracks the data will catch drift before it becomes a problem.
4. Include mold maintenance records in the supplier quality package. The date of the last inspection and the next scheduled inspection should be visible.
5. Review the process control charts for the affected part. If the charts show a shift, the supplier should have an action plan. If the charts are flat but the part drifts, the measurement method is wrong.
6. Set a maximum allowable drift. For example, if a dimension moves more than 0.05mm over three lots, the supplier must submit a corrective action report. Without a threshold, drift is just noise.

The goal is not to make the supplier over-engineer the process. The goal is to make the supplier accountable for the data. If the supplier can show you the trend, the cause, and the action, they are managing the quality. If they cannot, you are flying blind.

How to talk to a supplier about dimensional drift

The conversation should be factual. Do not accuse the supplier of a mistake. State the data, ask for the cause, and ask for the action.

A good opening is to share the CMM plot. “We are seeing a 0.1mm increase on the bearing seat over the last three lots. Can you check the die condition and the process parameters for those runs?” This is not a complaint. It is a request for information.

The supplier should respond with one of three things: a confirmed cause, a corrective action, or a request for more data. If they give you a vague answer, push for specifics. “What is the current die condition?” “What is the melt temperature range for the affected lot?” “What change will you make before the next lot?”

If the supplier cannot answer, the issue is capability. You may need to qualify a different supplier or change the part design to be less sensitive to the process. Some dimensions are harder to hold than others. A thin wall that shrinks with mold wear is harder to control than a thick boss that grows with wear. If a dimension is consistently difficult to hold, the design may need to change.

The supplier should not be asked to fix a problem that is not their fault. If the drift is caused by a design change, a tooling rebuild error, or a measurement method issue, the correction is not on the supplier. But the supplier should still provide the data so you can identify the real cause.

How to measure whether the correction worked

A correction is not complete until the data stops drifting. Pull the CMM report for the next lot and compare it to the pre-correction baseline. If the dimension is back within the original range and the slope is flat, the correction worked.

If the dimension is back but the slope is still there, the cause is not fixed. The part will drift again. Do not accept a one-lot fix as a permanent solution. Require the supplier to monitor the next three lots before closing the issue.

Track the correction in your supplier quality system. Log the date, the part number, the dimension, the cause, the action, and the result. This record helps you identify recurring patterns. If the same supplier drifts the same dimension on three different parts, the issue is likely the mold, not the part. If the same supplier drifts on many parts, the issue is likely the process.

The data is your evidence. Without it, the correction is a promise. With it, the correction is a verified change.

When to escalate the issue

Some dimensional drift issues are not worth a long conversation. If the part fails in assembly, the supplier should be informed immediately. If the part is out of tolerance by a large margin, do not wait for the next lot. Hold the current inventory and request a disposition.

Escalation is not punishment. It is protection. If a supplier sends a lot that will fail in production, you are responsible for the downstream cost. The supplier is responsible for the corrective action. Both parties should agree on the next step before the next lot is released.

A supplier that hides drift or delays the correction is a bigger risk than a supplier that has a process problem. The first problem is fixable. The second is a relationship problem. If the data shows drift and the supplier does not act, the relationship is broken.

The goal is to keep the supplier aligned with your quality expectations. The data is the bridge. Share it, review it, and require action. That is how you stop dimensional drift before it becomes a cost.