BVT (Build Verification Test)
BVT is the process from "code completion" to "testable state", focusing on whether the construction after compilation and linking can achieve the core functions. Specifically, after developers compile the code into executable files (such as APP installation packages and software clients), they will use BVT test cases to cover the "must-be-available" basic processes - for example, the "registration - adding to cart - placing an order" process in an e-commerce system and the "new - save - open" process in a tool software. These test cases do not go into details but only confirm "whether the most core functional chain works": if the registration button fails to respond when clicked, it indicates that there are problems with the compilation or dependencies, and immediate rollback and repair are required to avoid wasting resources in subsequent tests. In essence, BVT is a valve for "quickly filtering out basic errors", ensuring that subsequent tests are based on the premise that "core functions are available".
DVT (Design Verification Test)
DVT is a crucial step in hardware production to verify "whether the design is implemented". It covers full - dimensional testing from molds to electronic performance and directly determines whether the product meets the specification. Specifically, it includes three types of tests:
Mold testing: Check the dimensional accuracy of injection-molded/stamped parts (e.g., the deviation of the mobile phone middle frame ≤ 0.1mm), the smoothness of demolding (no mold jamming or deformation), and the integrity of the appearance (no burrs or shrinkage marks) to ensure that the parts can be assembled accurately.
Electronic performance testing: Verify the voltage stability of the circuit board, signal transmission efficiency (such as the download speed of a mobile phone's Wi-Fi), and anti-interference ability (such as no noise during a call) to ensure that the electronic components meet the design requirements.
Appearance test: Check the housing color (color difference with the design draft ≤ ΔE2), surface defects (no scratches or paint peeling), and logo position (logo offset ≤ 0.5mm) to ensure that the visual appearance meets the design expectations.
The goal of DVT is to "turn the design drawings into qualified mass - producible hardware". For example, in the DVT of a mobile phone, the touch sensitivity of the screen and the focusing speed of the camera will be tested to ensure that the design requirements are implemented 1:1.
PVT (Process Verification Test, small-batch process verification test)
PVT is the "mass production feasibility verification" in the small-scale trial production stage. The core is to confirm that "the mass production process can stably output qualified products". Usually, 100 - 500 units are produced in the trial production. The test contents include:
1. Full function verification: Cover all user-visible functions (such as calls, Bluetooth, and taking photos on a mobile phone) to ensure no omissions.
2. Stability test: Simulate long-term usage scenarios (such as continuous power-on for 72 hours and cyclic application switching) to check for problems like system crashes and freezes.
3. Reliability test: Verify the product's durability through environmental simulation (e.g., dropping from a height of 1.5 meters, IP68 waterproof immersion for 30 minutes).
The value of PVT is to discover "hidden problems in mass production" - for example, the internal resistance of the battery cells in a certain batch is too large, resulting in a shortened battery life, or the inconsistent assembly torque causes the screen to become loose - to avoid large-scale failures during mass production.
EVT (Engineer Verification Test, engineering sample verification test)
EVT is the "R & D level verification" in the engineering prototype stage, led by R & D engineers, focusing on "design feasibility". After the new product completes the prototype (such as a mobile phone mock - up or a functional prototype of household appliances), engineers will test the core functions (such as button sensitivity and screen display effect) and at the same time investigate design defects (such as whether the position of the heat sink affects cooling and whether the battery compartment can hold the battery). For example, if a loose charging interface is found during the EVT stage of a smartwatch, the mechanical structure will be adjusted immediately to ensure stable charging of subsequent samples. EVT is a crucial step in "transforming design ideas into a workable prototype" and lays the foundation for subsequent stages.
CTQ (Critical-To-Quality)
CTQ is the "core quality element" that directly determines customer satisfaction, and it represents the transformation from "general quality requirements" to "quantifiable control". These points are derived from customer feedback (e.g., milk tea customers complaining about "large fluctuations in sweetness"), design specifications (e.g., the touch sensitivity of a mobile phone screen ≥ 99%), or FMEA (high - risk items in failure analysis, such as the braking distance of a car). For example, the CTQs of a milk tea shop might be "sweetness deviation ≤ ±5%" and "boil pearls for 15 minutes" — these parameters directly influence the judgment of "taste good". The value of CTQ is to provide a clear goal for quality control and avoid "trying to do everything at once".
AQL (Acceptable Quality Level, Acceptance Qualification Standard)
AQL is the "core parameter" of sampling inspection, not the quality standard itself. It represents the "acceptable average quality level" - for example, AQL = 1.0 means that 1 defect is allowed for every 100 products. However, this is the basis for "judging whether to accept the entire batch", not the "standard that the products must meet" (the standard is a clear rule such as "no scratches on the surface"). For instance, during incoming inspection, if AQL = 0.65 and 13 samples are drawn, 2 defects are acceptable: if the number of defective samples ≤ 2, the entire batch is accepted; otherwise, it is rejected. The function of AQL is to balance the "quality requirements" and the "inspection cost", and to avoid over - inspection or missed inspection.
COPQ (Cost Of Poor Quality)
COPQ is the "total economic loss" caused by defects, which can be divided into two categories:
Internal failure costs: Losses before product delivery (e.g., labor/materials for rework, costs of scrapped parts);
External failure costs: Losses after the products leave the factory (such as return logistics fees, claim amounts, and sales lost due to brand damage).
For example, a mobile phone manufacturer recalled 100,000 units due to screen light leakage. The COPQ includes: recall logistics (50 yuan per unit), screen replacement cost (100 yuan per unit), and loss from customer churn (200 yuan per unit), totaling 35 million yuan. The significance of COPQ is to drive improvement through "economic accounting" - for instance, a factory reduced COPQ from 5% to 2% of its revenue, directly increasing the net profit by 3%.
TQC (Total Quality Control, Total Quality Management)Total Quality Management Total Quality Control
TQC is a quality management concept of "full participation", emphasizing that quality is not "the business of the quality inspection department", but the responsibility of the entire process from the CEO to front-line workers and from design to sales. The core logic is that "quality problems stem from process loopholes and require the collaboration of all employees to solve" - for example, Toyota's "completion of one's own process": each worker checks their own process (such as whether the screw torque meets the standard), and immediately stops the production line upon finding a problem to prevent defects from flowing to the next process. The essence of TQC is to "make quality the underlying ability of an enterprise through systematic methods" rather than relying on post-event inspections.
LRR and VLRR (Online Defective Rate and Confirmed Online Defective Rate)
LRR (Line Reject Rate) is the real-time defective rate on the production line. The data comes from inspection stations at each process (such as post-station inspection on the assembly line). For example, on a certain mobile phone production line, 5 out of every 100 units are rejected due to misaligned screen pasting, so LRR = 5%, which reflects "immediate defects during the production process".
VLRR (Verified Line Reject Rate) is the rechecked version of LRR and needs to be confirmed by engineering and OQC (Outgoing Quality Control). For example, for the issue of "misaligned screen sticker", it is necessary to check whether it is within the allowable deviation (e.g., ≤0.5mm is acceptable) to avoid misjudgment by workers. Therefore, VLRR is more accurate. For instance, if LRR = 5% and it is confirmed that 1% is a misjudgment, then VLRR = 4%. Both of them jointly monitor the stability of the production line, and the data sources are all from production, engineering, and OQC.
MRB (Material Review Board)
MRB is a cross - departmental mechanism for handling "uncertain defects" and is applicable to "ambiguous issues" in all inspection processes: incoming materials (raw material dimensional deviation), production process (part cracks), shipment (finished product stains), and customer returns (unknown failures).
Specific process:
1. Convener: Initiated by the position where the problem is discovered (for example, if the IQC discovers an abnormality in raw materials, it shall convene a meeting for discussion).
2. Participating roles: Project manager (coordinating resources), Purchasing (liaising with suppliers), Process engineer (analyzing processes), Planning (adjusting production), and the Quality/Production manager makes the final decision when necessary.
3. Conclusion output: Quickly determine 4 types of results -
- Continue to use (the deviation is within the allowable range);
- Use after rework (e.g., deburring and polishing);
- Scrap (cracked parts);
- Return to the supplier (raw materials do not meet the specifications).
The value of MRB is "to quickly convey problems and avoid stagnation". For example, when a batch of plastic parts cannot be judged due to burrs, the MRB decides to "use them after grinding", which not only reduces scrap but also does not affect production.