Received my copy of the standard about a month ago. Seems straightforward and gets away from the percentage nightmare. Mark Twain said it best with "There are 3 kinds of lies in this world - lies, damn lies and statistics!!"
Now I have some nagging questions and would like to find solutions, suggestions, perceptions from the community at large.
The standard (in its simplest form) has the DPMO Index based on the output for the factory, which I read to mean only the quantity of product processed that is completed to Finished Goods.
How do I take into consideration major process steps? For example-
SMT completes 100 units PTH completes 50 units Test completes 30 units Final QA completes 30 units
They are all working on the same assemblies, each group adding their respective value. Do I calculate the DPMO index for the quantity passed by Final QA? Can I include the output for each department? (I realize there is overlap) Do I hold each department only to the actual Terminations, Components, and Placements they are responsible for? making the Final output a cumulative answer?
What about rework? For instance, if a SMT defect is found late in the process, do I charge the origin or impune the finding party?
Test typically does not add parts - what portion are they liable for? likewise for Final QA.
This can get nasty and become a logistical nightmare trying to maintain spreadsheets for multiple products running simultaneously in a CM environment.
DPMO Index calculation - The total number of defects (Dc + Dp + Dt) on a completed Printed Circuit Board divided by the total number of defect opportunities (Oc + Op + Ot) for that PBA multiplied by 1 million. Your calculation deals only with the Component Defects, Placement Defects and Termination Defects. Test and Final QA are considered as gates that detects the above defects. I think thats how it should be done... That's how I understand it I maybe wrong mate...