<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>M9V.com — Journal</title>
    <link>https://m9v.com/en/blog</link>
    <description>First-principles process intelligence for high-volume manufacturing.</description>
    <language>en</language>
    <lastBuildDate>Mon, 24 Aug 2026 00:00:00 GMT</lastBuildDate>
    <atom:link href="https://m9v.com/feed.en.xml" rel="self" type="application/rss+xml"/>
    <item>
      <title>Tolerance stacks: over-specified in three places, loose in one</title>
      <link>https://m9v.com/en/blog/tolerance-stack-up-over-specified</link>
      <guid isPermaLink="true">https://m9v.com/en/blog/tolerance-stack-up-over-specified</guid>
      <pubDate>Mon, 24 Aug 2026 00:00:00 GMT</pubDate>
      <description>Most production tolerance stacks are inherited from a first drawing and never re-derived against measured variation. Worst-case arithmetic weights every contributor equally, so it offers no ranking, and engineers tighten whatever is cheapest rather than whatever dominates the output variance. The signature is three features held far tighter than needed and one, often a locating datum or an unbudgeted thermal effect, carrying most of the variation. Reallocating the budget from measured standard deviations cuts scrap and inspection cost without changing design intent.</description>
      <category>tolerance stack</category>
      <category>GD&amp;T</category>
      <category>metrology</category>
      <category>process capability</category>
      <category>scrap reduction</category>
    </item>
  </channel>
</rss>
