Fellow Quality Professionals,
I am trying to compare two measuring equipment, and I would like to confirm that the method I am using is the most appropriate.
The Facts;
1- What is being measured is the width of a line
2- The line is a Au plated circuit printed on a clear material
3- The measuring equipment is an optical inspection system that scans the edge of each side of the line and provides the width
4- The measuring system acts with almost no human interaction.
5- Lights, magnification, and position of the parts in the equipment are the same as both systems work with the same software
The test method;
1- 20 pieces with the same feature were selected.
3- All pieces were fabricated at the same time & under similar conditions
4- One feature is measured within each piece once.
5- All 20 pieces are measured in each equipment for a total of N=20
6- A pared t test is used to verify whether u1 = m2
7- The variance will be tested also to be the same between both pieces of equipment
Please let me know whether you would consider a different method should be applied.
Thank you,
Fernando
Six Sigma and ISO 9000: How To Choose
Hi Fernando,
I read about several articles and I found one article that is very informative and useful tool. And I think that these articles would help you alot.
If you need some more reliable information please don't hesitate to ask.
Six Sigma and ISO 9000: How To Choose
Here's more: Process Change? Think ISO 9000!
QmFriendlyGhost
Comparison of Two measurement Systems
Fernando,
Yes, you are conducting this study correctly by using a paired t test and a variance check. The variance check should be a t test and not an F test due to the small sample size. It is recommended that you use at least 25 samples so that you have enough data to make sound decisions. I would also create an individuals and moving range chart (x and MR chart) of the delta's to make sure the differences between the two systems is in the state of statistical control. I wrote a paper on this subject and presented it at the Minnesota Quality conference in oct of 2007. If you still have questions please research Nicholas R. Farnum's book "Modern Statistical Quality Control and Improvement", page 286. ISBN # 0-534-20304-3 or leave me another question on this forum. Good Luck. OH - I highly recommend that you first test each of the gages individually to make sure that are stable and consistant within them selves.