The Sample Approval Form Is Not a Manufacturing Procedure
Quote from chief_editor on June 5, 2026, 3:00 amApproving a sample from a Chinese supplier establishes that one instance of the product was acceptable. It does not establish that the production process is controlled to repeat that instance. The gap between those two things is where production quality problems begin.
A manufacturer of water treatment systems in France approved a sample of a critical filter housing component from a Chinese precision machining supplier in Jiangsu. The sample approval was thorough: dimensional measurement by CMM, surface finish measurement, material certification review, and a pressure test at rated conditions. The sample passed all criteria. A production order for 500 units was placed.
When the first production batch of 120 units was inspected at delivery, 34 units failed the pressure test at a leak rate that had been absent in the sample. Investigation established that the sample had been machined on the supplier's newest CNC turning center with a specific tooling setup optimized for the part. The production batch had been machined on three different machines using different tooling, with operator-dependent setup practices that produced dimensional variation in the sealing land geometry that exceeded the sample's precision.
The sample was a perfect part. The production process was not capable of reproducing the perfect part consistently.
The Conceptual Error in Sample-Based Procurement
The assumption embedded in sample approval processes is that a supplier who can make one acceptable part has a process that will consistently make acceptable parts. This assumption is not inherently wrong -- it is the basis of all sample-based qualification. It becomes wrong when the sample is produced under conditions that differ materially from the production conditions.
In Chinese precision manufacturing, the production of first article samples and the production of volume orders frequently involve different equipment, different operator attention levels, and different process validation rigor. The first article sample is the manufacturer's best demonstration of what they can achieve. The production batch is what their production system achieves under normal operating conditions, including machine variation, operator variation, tooling wear, and the economics of volume production.
The gap between sample capability and production capability is most visible in three categories of manufacturing situations. First, when the part geometry has features that are at the edge of the supplier's machine tool capability -- the sealing land geometry in the French filter housing was a feature where a small diameter deviation significantly affected leak performance. Second, when the part requires operator-dependent setup that varies between operators and shifts. Third, when the part's critical features are sensitive to tooling wear, which increases predictably across a production batch but not across a single sample.
A supplier whose sample approval process involves producing the sample on their best equipment under carefully controlled conditions -- which is rational and expected -- has demonstrated peak capability. Whether their production system achieves peak capability consistently requires evidence beyond the sample. That evidence is statistical process control data from production batches.
What Changes When Process Capability Is Verified Before Production
The additional verification that converts sample approval from peak capability demonstration to production capability confirmation is a production capability study conducted before the full production order is placed. The study involves producing a defined quantity of parts -- typically thirty to fifty -- under normal production conditions, including normal machine selection, normal operator rotation, and normal tooling conditions, and measuring the critical features with statistical analysis.
The statistical analysis computes the process capability index (Cp and Cpk) for each critical feature. A Cpk value above 1.33 indicates that the process is capable of producing parts within tolerance with a low probability of defects. A Cpk value below 1.0 indicates a process that will produce defects at a rate that is significant for production volumes. A sample of one does not provide this data.
For Chinese precision machining suppliers, requesting a process capability study before a production order is placed is a specific and quantifiable requirement that capable suppliers can satisfy and incapable suppliers cannot. A supplier whose response is that their sample passed inspection is telling you they have not measured their process capability. A supplier who provides capability data showing Cpk values above 1.33 for the critical features has demonstrated something qualitatively different from sample conformance.
The sample approval the French water treatment manufacturer conducted was technically correct within its scope. The scope was not sufficient to confirm what the buyer needed to know: that 500 production units would meet the same standard as the one approved sample. That confirmation required process capability data that was not requested. Requesting it before the production order adds two to three weeks to the supplier qualification timeline. The alternative timeline is the one where thirty-four rejected units arrive in a production batch.
Approving a sample from a Chinese supplier establishes that one instance of the product was acceptable. It does not establish that the production process is controlled to repeat that instance. The gap between those two things is where production quality problems begin.
A manufacturer of water treatment systems in France approved a sample of a critical filter housing component from a Chinese precision machining supplier in Jiangsu. The sample approval was thorough: dimensional measurement by CMM, surface finish measurement, material certification review, and a pressure test at rated conditions. The sample passed all criteria. A production order for 500 units was placed.
When the first production batch of 120 units was inspected at delivery, 34 units failed the pressure test at a leak rate that had been absent in the sample. Investigation established that the sample had been machined on the supplier's newest CNC turning center with a specific tooling setup optimized for the part. The production batch had been machined on three different machines using different tooling, with operator-dependent setup practices that produced dimensional variation in the sealing land geometry that exceeded the sample's precision.
The sample was a perfect part. The production process was not capable of reproducing the perfect part consistently.
The Conceptual Error in Sample-Based Procurement
The assumption embedded in sample approval processes is that a supplier who can make one acceptable part has a process that will consistently make acceptable parts. This assumption is not inherently wrong -- it is the basis of all sample-based qualification. It becomes wrong when the sample is produced under conditions that differ materially from the production conditions.
In Chinese precision manufacturing, the production of first article samples and the production of volume orders frequently involve different equipment, different operator attention levels, and different process validation rigor. The first article sample is the manufacturer's best demonstration of what they can achieve. The production batch is what their production system achieves under normal operating conditions, including machine variation, operator variation, tooling wear, and the economics of volume production.
The gap between sample capability and production capability is most visible in three categories of manufacturing situations. First, when the part geometry has features that are at the edge of the supplier's machine tool capability -- the sealing land geometry in the French filter housing was a feature where a small diameter deviation significantly affected leak performance. Second, when the part requires operator-dependent setup that varies between operators and shifts. Third, when the part's critical features are sensitive to tooling wear, which increases predictably across a production batch but not across a single sample.
A supplier whose sample approval process involves producing the sample on their best equipment under carefully controlled conditions -- which is rational and expected -- has demonstrated peak capability. Whether their production system achieves peak capability consistently requires evidence beyond the sample. That evidence is statistical process control data from production batches.
What Changes When Process Capability Is Verified Before Production
The additional verification that converts sample approval from peak capability demonstration to production capability confirmation is a production capability study conducted before the full production order is placed. The study involves producing a defined quantity of parts -- typically thirty to fifty -- under normal production conditions, including normal machine selection, normal operator rotation, and normal tooling conditions, and measuring the critical features with statistical analysis.
The statistical analysis computes the process capability index (Cp and Cpk) for each critical feature. A Cpk value above 1.33 indicates that the process is capable of producing parts within tolerance with a low probability of defects. A Cpk value below 1.0 indicates a process that will produce defects at a rate that is significant for production volumes. A sample of one does not provide this data.
For Chinese precision machining suppliers, requesting a process capability study before a production order is placed is a specific and quantifiable requirement that capable suppliers can satisfy and incapable suppliers cannot. A supplier whose response is that their sample passed inspection is telling you they have not measured their process capability. A supplier who provides capability data showing Cpk values above 1.33 for the critical features has demonstrated something qualitatively different from sample conformance.
The sample approval the French water treatment manufacturer conducted was technically correct within its scope. The scope was not sufficient to confirm what the buyer needed to know: that 500 production units would meet the same standard as the one approved sample. That confirmation required process capability data that was not requested. Requesting it before the production order adds two to three weeks to the supplier qualification timeline. The alternative timeline is the one where thirty-four rejected units arrive in a production batch.