2026 版 AIAG-VDA 合并版 SPC 手册正式发布，对比 2005 旧版，改动不算多，但每一条都戳在旧版的软肋上：哪些特性该上 SPC、特殊特性怎么管、设备验收报什么指数、上 SPC 之前先做什么。这篇把四个变化点理清，再给你车间里立即可做的三件事。
▌ 变化一：风险分级，不是所有特性都要上 SPC
手册 5.2.2 条款说得很明白：企业不一定有资源对全部特性实施 SPC。资源怎么分配？按风险分级：
- 高风险特性（与安全、法规、装配、功能相关）→ 优先部署 SPC，用控制图持续监控；
- 低风险特性（外观、非关键尺寸）→ 采用抽样检验等低成本方式，不必强行上图。
一句话：SPC 是管风险的工具，不是铺满所有特性的面子工程。
▌ 变化二：特殊特性，三条硬性要求
对识别出的特殊特性，新版把要求钉死了：
- 必须获取过程能力证据，不能用经验代替数据；
- 必须持续 SPC 监控，不能做一次 Cpk 就收工；
- 过程能力不达标时，强制 100% 检验。
识别路径也随之明确：FMEA 识别 → 图纸标注 → 控制计划落地，三步缺一不可。
▌ 变化三：设备验收换了新标尺——Pm/Pmk
旧 AIAG 手册没有设备能力指数，德系惯用的 Cm/Cmk 又没进正文，供应商报告各写各的。2026 版统一为 Pm/Pmk，定位设备验收：验收时把人、料、法、环全部固定，让波动主要来自设备本身。
名字里还藏着一处关键修正：短时采样无法证明长期稳定，所以叫"性能"（Performance）而不是"能力"（Capability）。新设备验收单上再写 Cm/Cmk，就是没跟上版本。
▌ 变化四：SPC 有了前置流程
新版把上 SPC 的路径排清楚了：公差设计 → 风险分析筛选特殊特性 → MSA 验证 → 过程表征 → SPC 监控。顺序里最关键的是 MSA：测量系统没验证，控制图画得再规范也不可信——你监控的可能只是测量误差。
▌ 车间三件事，现在就能做
- 拿控制计划对照 FMEA，把特殊特性清单补全，高风险特性优先部署 SPC；
- 新设备验收报告改写 Pm/Pmk，别再用 Cm/Cmk；
- 先验证 MSA，再上 SPC，顺序不能反。
记一句：新版改动不多，但每一条都戳在旧版软肋上——SPC 管的是风险，不是表格。
想系统落地 2026 版 SPC 手册，把特殊特性管理和设备验收一次理顺？欢迎联系续创信息技术咨询。
The combined 2026 AIAG-VDA SPC manual is officially released. Compared with the 2005 edition, the changes are few — but every one of them hits a weak spot of the old version: which characteristics deserve SPC, how special characteristics must be managed, which index to report for machine acceptance, and what has to happen before SPC even starts. This article sorts out the four changes, then gives you three actions your shop floor can take right now.
▌ Change 1: Risk Grading — Not Every Characteristic Needs SPC
Clause 5.2.2 states it plainly: an organisation may not have the resources to apply SPC to every characteristic. How should resources be allocated? By risk:
- High-risk characteristics (related to safety, regulations, assembly or function) → deploy SPC first, with ongoing control-chart monitoring;
- Low-risk characteristics (appearance, non-critical dimensions) → use lower-cost methods such as sampling inspection instead of forcing them onto charts.
In short: SPC is a tool for managing risk, not a vanity project that covers every characteristic.
▌ Change 2: Special Characteristics — Three Hard Rules
For identified special characteristics, the new edition nails down the requirements:
- You must obtain process capability evidence — experience is no substitute for data;
- You must run continuous SPC monitoring — a one-off Cpk study is not enough;
- When capability falls short, 100% inspection is mandatory.
The identification path is now explicit too: FMEA identification → drawing designation → control plan deployment. None of the three steps can be skipped.
▌ Change 3: A New Ruler for Machine Acceptance — Pm/Pmk
The old AIAG manual had no machine capability index, and the German Cm/Cmk never made it into the main text, so supplier reports were written in every dialect. The 2026 edition standardises on Pm/Pmk for machine acceptance: during the study, people, material, method and environment are all held fixed, so the variation comes mainly from the machine itself.
The naming carries a key correction: short-term sampling cannot prove long-term stability, so it is called performance, not capability. If your new-machine acceptance sheets still say Cm/Cmk, you have not caught up with the version.
▌ Change 4: SPC Now Has a Prerequisite Flow
The new edition lays out the path to SPC: tolerance design → risk analysis to filter special characteristics → MSA validation → process characterisation → SPC monitoring. The most critical step in the sequence is MSA: if the measurement system is not validated, even the most carefully drawn control chart is untrustworthy — you may just be monitoring measurement error.
▌ Three Shop-Floor Actions You Can Take Now
- Cross-check the control plan against the FMEA, complete the special-characteristics list, and deploy SPC on high-risk characteristics first;
- Write Pm/Pmk into new-machine acceptance reports — stop using Cm/Cmk;
- Validate MSA before launching SPC. The order is not negotiable.
One line to remember: the new edition changes little, but every change hits an old weak spot — SPC manages risk, not paperwork.
Want to implement the 2026 SPC manual systematically and straighten out special-characteristic management and machine acceptance in one go? Contact CINO Information Technology Consulting.