other iSixSigma newsletter subscribers:
Font Size


  • SPC on the inputs is the simple answer.

  • As @rbutler has hinted….fitting distributions can be tricky. How does one really know the parameters for the distribution.

    A “quick and dirty” approach is to just calculate the ppm defective. If the project/problem is solved, then the ppm defective will be drastically reduced and success can be assumed.

    My two cents.

  • Katie Barry replied to the topic Six Sigma Magazine in the forum Welcome 1 week ago

    @bmac A reply was sent to you on 11/3. I just forwarded that initial reply. Please be sure to check your spam folder. You can reach me directly at editor (at) isixsigma (dot) com.

  • I meant dictionary on the site

  • See the library on this site or google or … Maybe I don’t know the intent of your question?

  • Succinctly and nicely written.

  • Chris Seider replied to the topic Green Belt Dropout Rate? in the forum Training 2 weeks ago

    Let’s remember the adult learning model. If someone doesn’t practice what they are taught, they won’t begin to master or actually learn the tools.

    If a manager can’t complete a project that takes up to 10% of his/her time and has a real team to help divvy out the tasks, maybe it’s a problem with scoping or commitment to process improvement.

    If…[Read more]

  • ppm defective is the easiest approach.

    some folks will convert ppm defective to other quality measures but only if someone says they need to see one…

  • Your continuous example may be misleading. If you pick 3 different “SKU’s” and test the measurement system across such a broad range, the % error of the MSA relative to the process will be mistakenly thought to […]

  • Torque is a very odd measurement. First it is a destructive test if you are doing a MSA and you should. The way you handle the wrench makes a huge amount of difference. There are hard and soft joints and there is breakaway and running torque. If you want to talk through some of these issues please email me at

  • check out Magnatag

  • @mike-carnell is familiar with this also.

    If you take a cap off and retorque, you have no confirmation it’s to the same torque.

    Consider a bottling line, if you had 6 fillers, take 2 or 3 consecutive bottles off the same filler head. E.g. filler head #1, after the process lines out, consider taking bottle #1, #7, #13, and #20 and test with your…[Read more]

  • I’d consider doing a gage R&R with both set of “devices” and then doing a paired t-test to see if there’s a statistically significant difference between the two technologies.

    Remember, most things can’t be retorqued again so you have to assume that consecutive parts off a “machine” are similar enough which can lead to an overstated gage R&R but…[Read more]

  • Nicely written….a measure of acceptability would be a good addition.

  • Yes you can but don’t fall in the trap of just putting things in a neat and tidy place to sustain. Look at the processes and see what can be improved by the team(s) and 5S along the way.

  • @Enewport Did you check our 5S section of content?

  • Chris Seider and Profile picture of Joel MasonJoel Mason are now friends 4 weeks ago

  • I’m not as close to it but if you generate a “curve” every time with one set of run, then consider a poor man’s solution. If you can generate best fit coefficients for each run, then you can have a model that affects the curve from your DOE. This would require the same curve (e.g. Ax3 + Bx2 + Cx + D) and you’d create four columns for A-D and see…[Read more]

  • Think of process maps used to solve problems and VSM to help identify problems/projects.

  • Why not OEE?

  • Load More

Login Form