故事 · 摩托罗拉为什么发明 DFSS
Origin Story · Why Motorola Invented DFSS
1980 年代摩托罗拉用 DMAIC 把工厂良率一路推高,可工程师很快发现一个扎心的事实:
很多缺陷在产品被画到图纸上的那一刻就注定了 —— 公差太紧、零件太多、装配太复杂,再怎么改善产线也救不回来。
于是有人提出:与其事后救火,不如一开始就把质量设计进去。这就是 DFSS(Design For Six Sigma)的起点 ——
它不修旧流程,它盖一座生来就达标的新房子。一句话:DMAIC 让坏的变好,DFSS 让新的生来就好。
In the 1980s, Motorola was using DMAIC to drive factory yields up — and engineers ran into a painful truth:
many defects were locked in the moment a product was put on paper. Tolerances too tight, too many parts, assembly too complex — no amount of line improvement could rescue it.
So someone made the leap: instead of fighting fires after the fact, design the quality in from the start. That is the origin of DFSS (Design For Six Sigma) —
it does not renovate an old process, it builds a new house that is born up to spec. In one line: DMAIC makes the bad good, DFSS makes the new good by birth.
1 同一个规格窗口,两种逼近方式 One Spec Window, Two Ways to Close In
DMAIC 改善中2 改善的边际递减 vs 重设计的台阶跃迁 Diminishing Returns vs the Step Jump of Redesign
横轴是投入(改善轮次 / 重设计),纵轴是 Cpk。蓝线是 DMAIC:每轮上涨越来越少,逼近一条天花板就再也上不去;紫色台阶是 DFSS:一次重设计直接跳到新平台。 X-axis: investment (improvement cycles / redesign). Y-axis: Cpk. The blue curve is DMAIC — each cycle yields less than the last, asymptoting to a ceiling it cannot cross. The purple step is DFSS — one redesign leaps to a new plateau.
3 现实里何时该选 DFSS When to Reach for DFSS in the Real World
全新产品:还没有流程可改善(从零开发新机型、新软件、新服务),只能用 DFSS 把质量设计进去。
Brand-new product: there is no process to improve yet — a new model, a new piece of software, a new service. DFSS is the only way to design quality in.
撞到天花板:流程已逼近能力极限,DMAIC 再怎么改 Cpk 也卡住不动,必须重新设计架构。
Hitting the ceiling: the process is at its capability limit. No matter how many DMAIC cycles you run, Cpk stops moving — only an architectural redesign can break through.
成本前置:设计阶段就决定了 70–80% 的产品成本与质量,事后改善只能在剩下的两成里挣扎。
Cost is front-loaded: 70–80% of product cost and quality are locked in at the design stage. Post-launch improvement only fights over the remaining 20%.
目标 6σ:客户对缺陷零容忍(医疗、航空、芯片),与其指望产线救火,不如一开始就设计到 6σ。
Targeting 6σ: when customers have zero tolerance for defects — medical, aerospace, semiconductors — designing to 6σ up front beats firefighting on the line.
一句话In One Line
DMAIC 与 DFSS 不是对手,而是互补的两把工具:有现成流程要变好,用 DMAIC;要从无到有造一个生来就好的东西,用 DFSS。
关键判断是「改善还救得回来吗」 —— 如果分布已贴着天花板、再挤也挤不出空间,继续 DMAIC 就是把钱扔进无底洞。
记住那条最贵的曲线:设计阶段锁定了 70–80% 的成本,越早把质量设计进去,越省。
DMAIC and DFSS are not rivals — they are complementary tools: improve a process that already exists with DMAIC; build something new that is good by birth with DFSS.
The deciding question is "can improvement still save it?" If the bell is already glued to the ceiling and there is no room left to squeeze, more DMAIC is throwing money into a hole.
Remember the most expensive curve in the room: the design phase locks in 70–80% of total cost. The earlier you design quality in, the cheaper it gets.
常见误用Common Mistakes
对全新产品硬套 DMAIC。还没有流程可测量改善时,应走 DFSS(DMADV / CDOC)从设计入手。
Forcing DMAIC onto a brand-new product. When there is no process yet to measure or improve, start from design with DFSS (DMADV / CDOC).
撞到天花板还死磕改善。Cpk 卡住不动就是信号,该重新设计架构而非继续打补丁。
Doubling down on improvement after hitting the ceiling. When Cpk stops moving, that is the signal: redesign the architecture instead of patching it further.
以为 DFSS 能取代 DMAIC。两者互补:DFSS 盖新房,DMAIC 持续维护与改善存量流程。
Assuming DFSS replaces DMAIC. They are complementary: DFSS builds the new house, DMAIC keeps maintaining and improving the existing ones.