Troubleshooting: Supplier Inconsistency in Die Cast Dimensions

Why Dimensional Drift Happens in Die Casting

A single out-of-tolerance part is an inconvenience. A pattern of small deviations across multiple lots is a system failure. When a die casting vendor delivers parts that drift outside the specified range, the problem is rarely a single bad shipment. It usually points to a breakdown in machine stability, process control, or communication between your engineering team and the supplier.

Dimensional inconsistency in die casting is a cost center. It triggers rework, scrap, line stoppages, and supplier disputes. For a quality engineer, the goal is not just to reject the bad parts. The goal is to find the root cause and remove it before the next lot arrives.

Consider a hydraulic pump housing. If the bore diameter drifts by 0.05 mm over three months, the seals will not seat correctly. The part may look fine on a caliper, but the assembly will leak. This is not a quality issue in the abstract. It is a functional failure that reaches the customer.

The root cause of drift is often invisible. It hides in the thermal history of the die, the viscosity of the metal at the moment of injection, or the wear of a single guide post. You cannot see wear with the naked eye. You can only see its effect in the data.

Identifying the Pattern

Before you escalate the issue, map the data. Look at the last five to ten lots from the supplier. Plot the critical dimensions against the specification limits.

Check for a specific pattern:

1. Is the drift linear? The part gets larger or smaller with every batch. This usually indicates thermal expansion issues or machine wear.
2. Is the drift cyclical? The parts are good for a week, then bad for a week, then good again. This suggests a change in material lot, a shift in operator practices, or a scheduled maintenance window.
3. Is the variation random? The parts scatter around the mean without a clear trend. This points to poor process control or unstable melt temperature.

Random variation is the hardest to fix. It requires tightening the process capability. A stable Cpk of 1.33 or higher is a standard target for many industries, but your specific tolerance requirements will dictate the needed capability. If the supplier is running a Cpk below your minimum requirement, the process itself is not capable of holding the tolerance, regardless of how many good parts are in the current lot.

Imagine a supplier delivering a series of small brackets. The first lot measures 10.0 mm. The second lot measures 10.1 mm. The third measures 10.2 mm. This is linear drift. The die is likely heating up more than expected, or the mold is wearing out on the parting line. The solution is not to reject the third lot. The solution is to audit the mold and check the melt temperature logs.

If the data shows a cyclical pattern, look at the calendar. Did the supplier change their die casting machine? Did they switch to a different supplier for the aluminum ingots? If the parts are good on Monday and bad on Wednesday, the issue might be operator-dependent. One operator might run the machine at 500 degrees Celsius, while another runs it at 450. Both are within the broad range, but only one is in the sweet spot for your part.

Common Symptoms and Corrective Actions

The following table lists common symptoms of supplier inconsistency and the likely causes behind them. Use this as a starting point for your root cause analysis.

Symptom	Likely cause	What to do
Gradual increase in part weight	Melt temperature too high or mold wear	Request melt temperature logs. Audit mold condition.
Sudden drop in dimensions	Melt temperature too low or gating change	Verify melt temperature. Check gating and feeders.
High variation in one dimension	Unstable mold clamping or machine wear	Inspect machine guides and clamps. Check Cpk data.
Inconsistent flash	High injection pressure or worn mold	Adjust injection pressure. Replace worn mold inserts.
Porosity in thin sections	Gas entrapment or poor degassing	Review degassing process. Check vacuum systems.
Surface defects on critical faces	Mold corrosion or poor polishing	Polish mold surfaces. Apply mold release agents.

If the supplier cannot provide the specific data requested in the “What to do” column, that is a red flag. A competent die casting vendor maintains records of melt temperature, injection pressure, and machine status. If they do not, they are not controlling the process. They are just manufacturing.

Take the symptom of high variation in one dimension. Suppose the length of a motor mounting flange varies by 0.1 mm from shot to shot. The supplier claims the machine is new. You should not accept that. Request the Cpk data for that specific feature. If the Cpk is low, the machine is not stable. If the Cpk is high but the parts are still out of tolerance, the drawing is too tight. You need to clarify the tolerance, not blame the machine.

The Role of the Drawing and Specification

A lot of supplier inconsistency comes from ambiguity. If your drawing does not clearly define the critical dimensions, the supplier will make assumptions. Those assumptions will vary by operator, shift, and machine.

Review your drawings for these common gaps:

• Tolerance stack-up: Does the drawing account for the cumulative effect of multiple dimensions? A part might be in tolerance on each individual feature, but the assembly might not fit.
• GD&T: Are geometric tolerances defined where dimensional tolerances are not enough? Flatness and perpendicularity often matter more than a simple length measurement.
• Material properties: Does the drawing specify the minimum tensile strength or hardness? Sometimes a part is in tolerance but lacks the mechanical properties needed for the application.
• Inspection method: How should the part be measured? Is a caliper sufficient, or does the part require a CMM? If the supplier measures with a micrometer and your drawing implies a CMM check, you have a mismatch.

Clarify these points in a drawing review. If you are sourcing from a second-source die casting vendor, this step is even more critical. New vendors need a clear baseline to understand your expectations.

Consider a part with two parallel holes. The drawing specifies the center-to-center distance as 50 mm +/- 0.1. It does not specify the flatness of the face the holes are drilled into. The supplier drills the holes. The distance is 50.05 mm. The part fits. But the face is wavy. When you bolt the part to the mating component, the holes do not align. The part is in tolerance on paper. It fails in assembly. The drawing was missing a geometric control.

Another common gap is the material specification. You specify “A380” on the drawing. But you do not specify the minimum tensile strength. The supplier buys cheap ingot that is on the low end of the A380 range. The part casts fine. But it cracks when you bend it. The material property was not controlled. You need to specify the mechanical properties, not just the alloy designation.

Process Control and Capability

A stable process is the best defense against dimensional issues. Ask the supplier to show you their process control plan. Look for these elements:

1. Control charts: Are they plotting key dimensions over time?
2. Reaction plans: What do they do when a point falls outside the control limits?
3. Maintenance logs: How often do they service the die casting machine?
4. Material certification: Do they verify the chemical composition of each aluminum or magnesium ingot batch?

If the supplier relies solely on 100% inspection, they are treating the symptom, not the cause. 100% inspection is expensive and inefficient. It does not prevent the part from being out of tolerance. It only finds it after the fact. A capable process produces parts that are consistently in tolerance.

Look at the control charts. A good chart shows a stable mean with variation within the control limits. If the mean shifts, the supplier should stop the process and investigate. If they do not stop the process, they are producing bad parts on purpose. The cost of stopping is low. The cost of shipping bad parts is high.

Maintenance logs are often ignored. A die casting machine is a complex hydraulic and thermal system. The guides wear. The pistons leak. The mold heats up. If the supplier does not log these events, they cannot predict drift. They are flying blind. You need to see the maintenance schedule. How often do they replace the seal? How often do they calibrate the temperature probe?

Audit and Verification

Do not take the supplier’s word for it. Schedule a periodic audit of the die casting vendor’s facility. Focus on the following:

• Machine condition: Look for worn guides, leaks, and proper lubrication.
• Operator training: Are the operators trained on the specific process? Can they explain the control limits?
• Laboratory capability: Do they have the necessary instruments for material testing?
• Documentation: Are the process control plans current and accessible?

An audit is not a punishment. It is a relationship-building tool. It shows the supplier that you are serious about quality. It also gives you the opportunity to see the process firsthand.

During an audit, walk the floor. Watch a shot being cast. Observe the operator. Do they check the temperature before starting a new batch? Do they clean the mold between shifts? Do they have a logbook on the machine? If the logbook is filled out, is it filled out by the operator, or by the quality manager at the end of the week?

Check the laboratory. Do they have a spectrometer? If they do, do they have a certificate of calibration? If they do not have a spectrometer, how do they verify the material? They might be buying the wrong alloy. They might be buying the right alloy, but with the wrong impurity levels. You need to see the test report.

Corrective Action Request

When a problem occurs, issue a formal Corrective Action Request (CAR). Do not rely on email threads. A CAR document forces the supplier to follow a structured process.

A strong CAR includes:

1. Problem statement: Specific lot numbers, part numbers, and dimensions.
2. Containment action: How are you handling the current bad material?
3. Root cause analysis: Use a 5-Why or fishbone diagram. Do not accept “human error” as a root cause.
4. Corrective action: What specific change will prevent the problem from recurring?
5. Verification: How will you verify that the fix works?

Give the supplier a defined timeframe for response. If they miss the deadline or provide a vague answer, escalate the issue. Consider moving volume to a different source if the problem is systemic.

“Human error” is not a root cause. If an operator made a mistake, ask why. Why did they make the mistake? Was the procedure unclear? Was the machine difficult to operate? Was the training insufficient? If the answer is “the operator was not paying attention,” you have not found the root cause. You have found a symptom.

When you issue a CAR, be specific. Do not say “The parts are out of tolerance.” Say “Lot 2310-045, part number 123-456, bore diameter measured 10.15 mm on 12 of 100 samples. Spec is 10.00 +/- 0.10 mm.” This leaves no room for debate. The supplier knows exactly what you are talking about.

Long-Term Supplier Management

Dimensional consistency is a long-term game. Build a relationship with the supplier based on data and continuous improvement.

• Share your data: If you find a trend in your incoming inspection, share it with the supplier.
• Review performance: Meet quarterly to review PPM, Cpk trends, and corrective actions.
• Invest in the supplier: If a supplier is good but not great, help them improve. Fund a mold upgrade or a new CMM.

A good supplier will respond to this approach. A bad supplier will get defensive. Choose your partners carefully.

If you find a trend in your incoming inspection, send a graph to the supplier. Do not send an angry email. Send the data. Show them the drift. Ask for their input. They might not have seen the problem because they are measuring a different sample size or using a different method. Sharing the data opens a dialogue.

Quarterly reviews are not just for checking in. They are for problem-solving. Bring the data. Bring the issues. Bring the opportunities. If the supplier is struggling with a new product, offer to help. If they are struggling with a machine, offer to share your maintenance logs. If they are struggling with a process, offer to share your process control plan.

Investing in the supplier is a business decision, not a charity. If you help them improve, they will deliver better parts. They will have fewer defects. They will have lower scrap. They will be more reliable. You will save money in the long run.

Conclusion

Dimensional inconsistency from a die casting vendor is a solvable problem. It requires data, clear communication, and a focus on process capability. Do not react to the next bad lot. React to the pattern. Fix the system, not just the part.