As a professional liquid silicone rubber (LSR) products manufacturer, Xiangchu (Hubei) Rubber Co., Ltd. still adheres to manual inspection even with an annual output of 5 million LSR parts. A return incident prompted us to re-examine our quality inspection process for LSR products. This article shares our thinking on quality control of LSR parts to ensure stable delivery quality.
Liquid Silicone Rubber (LSR) manufacturing has scaled rapidly over the past decade, driven by booming demand from medical devices, consumer electronics, automotive, and baby product industries. At橡楚（湖北）橡胶有限公司, we hit our annual production target of 5 million LSR parts two years ago, a milestone that we celebrated as proof of our capacity growth and manufacturing optimization. But what we didn’t anticipate was a 12% batch rejection from a European medical client that forced us to step back and re-evaluate our entire quality inspection process. The question we’ve been asked repeatedly since that incident is simple: with an annual output of 5 million LSR parts, why do we still rely heavily on manual inspection instead of full automation? This article shares our technical assessment, process review, and the balanced approach we’ve landed on after that costly learning experience. We’ll break down the unique challenges of LSR part inspection, compare automated and manual inspection capabilities, and explain why a hybrid model centered on skilled manual inspection remains the most reliable choice for our production at 橡楚（湖北）橡胶有限公司.
LSR is a high-performance material with properties that set it apart from thermoplastics, conventional solid silicone, and other elastomers. These same properties that make LSR ideal for end-use applications also create unique inspection challenges that do not exist in other mass-produced rubber or plastic parts. To understand why manual inspection remains core to our process, we first need to break down these inherent challenges.
Unlike thermoplastic injection molding, which relies on consistent melt viscosity and fast cooling to produce uniform parts, LSR injection molding involves a chemical curing process that introduces subtle variability even with the most well-calibrated equipment. Liquid silicone rubber is a two-part material (Part A polymer, Part B cross-linker) that requires precise mixing, temperature control, and pressure hold during cure. Even small deviations in mixing ratio (±0.5% from the target) or mold temperature (±5°C across different cavities) can lead to visible surface defects that impact performance or aesthetics.
Common defects that occur in mass LSR production include:
Beyond material-related defects, most of the 5 million LSR parts we produce annually are small, complex geometry parts for medical and consumer electronics applications. 72% of our annual output has at least one undercut, internal cavity, or flexible thin-wall section (thinner than 1mm) that is difficult for automated vision systems to capture. For example, the medical silicone gaskets we produce have a 0.3mm sealing lip that runs along an irregular internal contour: subtle edge curling on this lip will cause leakage, but the curvature and low contrast between the lip and the rest of the part make it hard for standard automated vision systems to detect.
A large share of the 5 million LSR parts we produce annually are for regulated industries: 42% go to medical device and baby product manufacturers, which require strict defect traceability and zero critical defects. For medical parts, even a 0.1mm embedded bubble in a sealing component can lead to product recall, which imposes significant liability on both the end manufacturer and our factory.
These requirements create two inspection constraints that are hard to address with full automation:
When we scaled to 5 million parts annually, we first invested in two automated vision inspection systems to replace 80% of our manual inspection work, based on vendor claims of 99.7% detection accuracy. The system failed to deliver on that promise, and the 12% rejection batch that triggered our process review was a direct result of over-reliance on this automated inspection. Below we break down the key limitations we encountered, based on two years of real-world testing with our LSR production.
Automated vision inspection (AVI) systems work best for high-volume parts with consistent geometry, uniform color, and well-defined defect criteria on accessible surfaces. For LSR parts, we found three persistent technical issues that could not be resolved even after custom calibration by the vendor:
To quantify our test results, we ran a side-by-side comparison of automated inspection and manual inspection on a 10,000-part batch of 0.3mm lip medical gaskets, one of our most common high-volume parts. The results are summarized in the table below:
As the data shows, while automation has a throughput advantage, it also has a much higher rate of missed critical defects, which is unacceptable for our products. Even with automated pre-sorting, we still need manual inspection to catch defects that automation misses.
Beyond technical limitations, full automation of inspection for 5 million LSR parts is currently not economically viable for our operation, for three key reasons:
After the 12% batch rejection incident, we did a full review of our inspection process, talked to other LSR manufacturers, and tested multiple inspection models. We ultimately landed on a hybrid process that leverages automation for high-volume, low-complexity sorting, and keeps experienced manual inspection as the final verification step for all 5 million parts we produce annually.
We tier our inspection process based on part complexity, regulatory requirement, and defect risk, to balance throughput, cost, and quality. The three tiers are:
These are large, flat, non-critical parts (e.g., large gaskets for industrial equipment) with high contrast and simple geometry. We use automated pre-sorting to remove obvious defects like short shots and large bubbles, followed by a 10% random manual inspection check per batch. This balances throughput and quality for these lower-risk parts.
These include baby product components, consumer electronics keypads, and automotive seals, which require full inspection for aesthetic and functional defects. We use automated pre-sorting to remove obvious defects, which reduces the number of parts that inspectors need to check by 30-40%, then every part goes through a full manual inspection by a trained quality inspector.
These are medical device components, implant-grade LSR parts, and baby skin-contact products that require zero critical defects. For these parts, 100% of the inspection is done by experienced manual inspectors, with a second 100% check by a senior quality technician for critical dimensions and defects. We do not rely on automation for these parts, because the risk of missed defects is too high.
Instead of investing millions in full automation, we invested in training and retaining skilled manual inspection teams. We found that an inspector with 2+ years of LSR inspection experience can detect 99.6% of critical defects, which is far higher than any automated system we tested. To support our team, we have implemented the following processes:
We also recognized that inspection is a downstream check, and the best way to improve quality is to reduce defects in the molding process itself. To reduce the burden on our inspection team while maintaining quality for 5 million annual parts, we implemented the following upstream improvements:
These upstream changes have reduced our overall defect rate from 4.2% to 1.8% over the past two years, which reduces the workload on our inspection team and keeps our costs competitive even with manual inspection.
Scaling LSR production to 5 million parts annually creates a lot of pressure to automate every step of the process, including quality inspection. But after our costly batch rejection and extensive testing, we’ve learned that full automation of inspection does not deliver the quality our clients need, especially for the complex, regulated LSR parts we produce at橡楚（湖北）橡胶有限公司. The unique properties of LSR, the complexity of the parts we manufacture, and the strict regulatory requirements of our key markets mean that skilled manual inspection still outperforms full automation on critical defect detection, consistency, and compliance.
That doesn’t mean we reject automation entirely: our hybrid approach leverages automated sorting to remove obvious defects and reduce inspector workload, which keeps throughput high and costs competitive. We also continue to test new automated inspection technologies as they emerge, and we will update our process if new technologies can match the defect detection accuracy of manual inspection at a reasonable cost. But for now, our commitment to manual final inspection is not a sign of backwardness — it’s a deliberate choice to prioritize quality over maximum throughput and cost reduction, especially for clients that rely on us for zero-defect LSR parts.
At橡楚（湖北）橡胶有限公司, we hold ISO 9001 certification, and we operate from our facility at 湖北省鄂州市鄂城区经济开发区凡口街道内河巷54号, producing 5 million high-quality LSR parts annually for export clients around the world. If you have any questions about our quality process or are looking for a reliable LSR manufacturing partner, contact us at 18071171144 or email us at churubber@163.com.
