What is the quality control process at a TFT module factory?
Raw Material and Incoming Quality Control (IQC)
Every TFT module starts with glass substrates, polarizers, driver ICs, and backlight units. The IQC team receives these components with a sampling plan based on AQL (Acceptable Quality Level) standards, typically AQL 0.65 for critical defects and AQL 1.0 for major defects. For example, a 7-inch TFT panel might have a batch of 10,000 pieces; the inspector pulls 200 samples and checks for scratches, pinholes, and contamination under a 10x magnifying glass. Glass flatness is measured with a laser interferometer, with tolerance of ±0.1 mm for a 0.5 mm thick substrate. Polarizer adhesion is tested using a peel test machine at 90-degree angle, requiring a minimum of 0.5 N/cm force. Driver ICs undergo electrical testing on a dedicated IC tester, checking for open/short circuits and current leakage at 3.3V. Data from IQC is logged into an MES (Manufacturing Execution System), and any batch with more than 5% failure is rejected and sent back to the supplier.
In-Process Quality Control (IPQC) During Cell Assembly
The cell assembly process—where the TFT glass is aligned with the color filter and liquid crystal is injected—is the most critical stage. IPQC engineers monitor the cleanroom environment: Class 1000 or better (ISO 6), temperature at 22°C ± 2°C, and humidity below 40% RH. Every hour, a technician checks the cell gap uniformity using a spectrophotometer, targeting a cell gap of 3.5 µm ± 0.2 µm for a 5-inch module. The liquid crystal filling process is monitored by weight: a typical 5-inch module requires 0.15 g of LC material, with a tolerance of ±0.005 g. After vacuum filling, the cell is sealed with UV-cured epoxy, and a UV intensity meter ensures the curing lamp delivers 2000 mJ/cm². Any deviation triggers an immediate line stop and a root cause analysis. Data from the 10 most recent batches is plotted on a control chart, and if the CpK (process capability index) drops below 1.33, the process is adjusted.
Automated Optical Inspection (AOI) and Electrical Testing
After cell assembly, each panel goes through AOI. This is not a simple pass/fail—it's a high-resolution camera system that captures 20+ images per panel at 10x magnification. The AOI algorithm checks for mura (brightness non-uniformity), dead pixels, and line defects. A typical 7-inch module has 1,024 x 600 pixels; the AOI can detect a single dead pixel with a 99.5% capture rate. The threshold for mura is set at a contrast ratio of 1.05:1 or higher. Parallel to AOI, electrical testing is done on a dedicated jig that applies a 60 Hz refresh rate and checks for cross-talk, flicker, and response time. Flicker is measured using a photodiode and must be below -30 dB. Response time (from black to white) is measured with a luminance meter, targeting 25 ms or less for a typical TN panel. Any module failing these tests is marked with a red dot and sent to the repair station for rework.
Backlight Unit (BLU) and Module Assembly QC
The BLU—comprising LEDs, light guide plate, and diffuser films—is assembled separately. Each BLU is tested for luminance uniformity and color temperature. A 7-inch BLU uses 6 LEDs in series, with a total current of 120 mA. The luminance is measured at 9 points across the panel using a luminance meter, with a target of 300 cd/m² ± 10% and a uniformity of 80% or better (minimum luminance / maximum luminance). Color temperature is measured with a spectrometer, targeting 6500K ± 500K. After attaching the BLU to the TFT cell, the module goes through a final functional test: a 24-hour burn-in test at 50°C and 85% RH for industrial-grade modules. During burn-in, the module displays a scrolling checkerboard pattern, and any visual defect (like a flicker or line) is logged. For automotive modules, this burn-in is extended to 168 hours with a temperature cycling of -40°C to +85°C.
Final Quality Control (FQC) and Outgoing QC (OQC)
FQC is a 100% visual inspection under a darkroom environment with a D65 light source. The inspector checks for cosmetic defects (scratches, bubbles, dust) and electrical defects (dead pixels, line defects). The acceptable defect count per module is zero for critical defects and up to 2 minor cosmetic defects (like a 0.1 mm scratch) per module. After FQC, OQC uses a random sampling plan: for a batch of 5,000 modules, 125 samples are drawn and tested for all parameters—luminance, contrast ratio, viewing angle, and response time. The contrast ratio is measured with a spectrophotometer, targeting 1000:1 for a typical IPS module. Viewing angle is tested at 80 degrees horizontal and 70 degrees vertical, with a contrast ratio of 10:1 or better. Any sample failure triggers a 100% re-inspection of the entire batch. Data from OQC is compiled into a monthly report, and the factory's overall yield rate is tracked. For a mature TFT module factory, the yield rate is typically 95% to 98% for standard products, and 90% to 95% for custom designs.
Reliability and Environmental Testing
Beyond production line checks, a TFT module factory runs a reliability test on every new product or major process change. This includes a temperature cycling test: -20°C to +70°C for 100 cycles (2 hours per cycle), with a dwell time of 30 minutes at each extreme. After cycling, the module is tested for electrical performance and visual defects. A humidity test is done at 60°C and 90% RH for 240 hours, and the module must show no corrosion or delamination. A vibration test is performed on a shaker table at 10-200 Hz, 1.5G acceleration, for 30 minutes per axis. For automotive modules, these tests are more stringent: temperature cycling from -40°C to +105°C for 500 cycles, and a salt spray test for 48 hours. The factory also conducts a drop test: a 1-meter drop onto a concrete floor, repeated 5 times, and the module must still function. Data from these tests is used to update the FMEA (Failure Mode and Effects Analysis) and control plans.
Data Management and Traceability
Every module gets a unique serial number, printed as a QR code on the flex cable or back cover. The MES records every test result, operator ID, and timestamp for that serial number. If a field failure occurs, the factory can trace back to the exact batch of glass, driver IC, and polarizer used. For example, if a module fails after 6 months of use, the factory retrieves the MES data and checks if the AOI missed a defect or if the BLU current was out of spec. This traceability is critical for automotive and medical applications, where a recall can cost millions. The factory also maintains a database of SPC charts for critical parameters like cell gap, luminance, and response time. If the trend shows a drift, the process engineer adjusts the equipment before defects occur. This proactive approach reduces scrap and rework costs.
Supplier Quality Management (SQM)
A TFT module factory relies on multiple suppliers for glass, polarizers, driver ICs, and LEDs. The SQM team audits each supplier annually, using a scorecard that covers quality (defect rate), delivery (on-time delivery rate), and cost. For critical components like driver ICs, the factory requires a PPAP (Production Part Approval Process) submission. The supplier must provide a control plan, FMEA, and measurement system analysis (MSA) for each new part. The factory also runs a incoming inspection on every batch, and if the defect rate exceeds 1%, the supplier is put on probation. For example, a polarizer supplier with a 0.5% defect rate is considered acceptable, but if it rises to 1.2%, the factory sends a corrective action request. The factory also maintains a list of approved suppliers, and any new supplier must go through a 3-month trial period with 100% inspection of their parts.
Continuous Improvement and Six Sigma
The factory runs a continuous improvement program, with a dedicated team of Six Sigma Black Belts. They analyze yield data and identify top defects. For example, a common defect is "mura" caused by uneven cell gap. The team uses a DMAIC (Define, Measure, Analyze, Improve, Control) approach: they measure the cell gap across the panel using a white light interferometer, find that the gap varies by 0.3 µm, and adjust the spacer ball density. After the improvement, the gap variation drops to 0.1 µm, and the mura defect rate falls from 2% to 0.5%. The factory also runs a Kaizen event every month, where operators suggest improvements. One such suggestion was to add a pre-heat step before the polarizer lamination, reducing bubble defects by 30%. The factory tracks these improvements in a dashboard, and the annual cost savings from quality improvements are typically 5% to 10% of the total manufacturing cost.
Capture at the speed of your stack.
93MB installer. Three-second cold start. The recorder engineers actually keep open.