<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Pramod</title>
    <description>The latest articles on DEV Community by Pramod (@pramodvadali).</description>
    <link>https://dev.to/pramodvadali</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4145856%2Fa45308ec-6f8f-4ffb-96f7-ac83d0a6988d.png</url>
      <title>DEV Community: Pramod</title>
      <link>https://dev.to/pramodvadali</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/pramodvadali"/>
    <language>en</language>
    <item>
      <title>Oracle Fusion Quality: Sampling Challenge for WIP Inspections</title>
      <dc:creator>Pramod</dc:creator>
      <pubDate>Sun, 27 Sep 2026 18:28:09 +0000</pubDate>
      <link>https://dev.to/pramodvadali/oracle-fusion-quality-sampling-challenge-for-wip-inspections-220n</link>
      <guid>https://dev.to/pramodvadali/oracle-fusion-quality-sampling-challenge-for-wip-inspections-220n</guid>
      <description>&lt;p&gt;Setting a tight AQL on a small batch in Oracle Fusion Quality may result in a situation where the inspection screen returns numbers that look broken. Below is a situation I encountered recently and fixed:&lt;/p&gt;

&lt;h2&gt;
  
  
  The problem
&lt;/h2&gt;

&lt;p&gt;In-line WIP inspection configured in this manner:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;ANSI/ASQC Z1.4&lt;/li&gt;
&lt;li&gt;Single sampling plan, normal inspection&lt;/li&gt;
&lt;li&gt;General Inspection Level II&lt;/li&gt;
&lt;li&gt;AQL of 0.01&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;A process manufacturing work order for a 10 KG batch. The inspection triggers at the operation and shows:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;Accepted:  1,250
Pending:  -1,240
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;A negative pending quantity. It looks like a calculation error, so it usually goes straight into a support ticket.&lt;/p&gt;

&lt;h2&gt;
  
  
  The fix
&lt;/h2&gt;

&lt;p&gt;Two options. Pick based on what the business actually needs.&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;&lt;p&gt;Raise the AQL to something the batch size can support.** A ten unit batch can only demonstrate a quality level that a sample of ten or fewer can prove. That means a much looser AQL than 0.01.&lt;/p&gt;&lt;/li&gt;
&lt;li&gt;&lt;p&gt;Drop AQL sampling for small batches.** Use fixed count or percentage sampling instead. If you want all ten units inspected, configure a 100 percent plan and say so directly.&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  Root cause
&lt;/h2&gt;

&lt;p&gt;Sample size does not scale down with batch size the way people expect.&lt;/p&gt;

&lt;p&gt;Under Z1.4, the lot size and inspection level give you a sample size code letter. The code letter gives you a sample size. You then read across to your AQL column for the accept and reject numbers.&lt;/p&gt;

&lt;p&gt;The catch is what happens when the AQL is too tight for that sample to prove anything. A sample of 315 units cannot demonstrate a defect rate of one in ten thousand. There is no acceptance number that separates a good lot from a bad one at that level.&lt;/p&gt;

&lt;p&gt;Z1.4 handles this with an arrow in the table. Instead of accept and reject numbers, it sends you down to the first plan that can discriminate at your AQL, and you use that plan's sample size. At AQL 0.010 on single sampling normal inspection, that plan is 1,250 samples with accept 0, reject 1.&lt;/p&gt;

&lt;p&gt;That is where 1,250 comes from. Fusion is reading the standard correctly.&lt;/p&gt;

&lt;p&gt;The second number then follows:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;batch quantity  -  required sample  =  pending
      10        -       1,250       =   -1,240
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The batch made 10 units. The plan wants 1,250 inspected. Pending is the shortfall, negative because the requirement is larger than the entire lot.&lt;/p&gt;

&lt;p&gt;So the real answer is that an AQL of 0.01 is not achievable on a batch of ten. No sampling scheme can prove that quality level from a population that small. The screen is showing you a statistical impossibility, not a bug.&lt;/p&gt;

&lt;h2&gt;
  
  
  Additional tips
&lt;/h2&gt;

&lt;p&gt;Before you commit to an AQL, work out the sample size it implies and compare it against your normal lot size. If the sample is bigger than the lot, the plan will never run.&lt;/p&gt;

&lt;p&gt;Watch out for AQLs chosen because they sound rigorous. 0.01 reads well in a quality plan and almost nobody checks whether it is attainable at their batch sizes.&lt;/p&gt;

&lt;p&gt;The same symptom appears any time the required sample exceeds the lot, not just at 0.01. If you see a negative pending quantity, check the sampling plan against lot size first.&lt;/p&gt;

&lt;p&gt;If your tests are destructive or expensive, look at special inspection levels S-1 through S-4. They produce much smaller samples than the general levels and are designed for exactly that situation.&lt;/p&gt;

&lt;p&gt;Z1.4 exists to avoid inspecting everything when lots are large. On a ten unit batch there is nothing to save, so the whole mechanism works against you. Small batch process manufacturing and acceptance sampling are an awkward pairing, and it is worth deciding early which of your products should use sampling at all.&lt;/p&gt;

</description>
      <category>oracle</category>
      <category>codequality</category>
      <category>manufacturing</category>
      <category>erp</category>
    </item>
  </channel>
</rss>
