I was asked to help with a Gage R&R that was done with a prototype gage.

The tolerance that they are measuring here is very tight. +/- .025mm

This process as far as I know has NEVER been stable, and they have always had to do a 100% to 200% inspection using Jo-Blocks for min and max.

They want to start doing SPC on this characteristic in hopes of measuring some sort of improvement, but needed to come up with a gage capable of measuring such a tight tolerance.

The prototype is a gage that uses lasers to measure the diameter of a part compared to a master, and gives a reading on how different each part is.

I am going to attach the Report from Minitab, I would appreciate anyones thoughts on this.

Most of the parts in the study look to be out of tolerance, but that isnt suprising based on what they have been telling me.
I know as a rule you are supposed to take parts that span normal process variation, and they should come from a stable process?

Is there any value in this study as an attempt to discern if this gage would help them at all?


6199

At a P/T Ratio of 230%, this gage is totally unacceptable for inspection to the tolerance.

If the part variation accurately reflects the actual process variation, the %GRR of 54.5% and an ndc = 2, also means that this gage is not acceptable for process control. The validity of these two measures depends on how well the part variation reflects the process variation.

The gage has problems in three areas: repeatability, reproducibility and an operator * part interaction. Without more information on the gage itself, including the parts and fixturing, we will not be able to provide much more assistance.

At a P/T Ratio of 230%, this gage is totally unacceptable for inspection to the tolerance.

If the part variation accurately reflects the actual process variation, the %GRR of 54.5% and an ndc = 2, also means that this gage is not acceptable for process control. The validity of these two measures depends on how well the part variation reflects the process variation.

The gage has problems in three areas: repeatability, reproducibility and an operator * part interaction. Without more information on the gage itself, including the parts and fixturing, we will not be able to provide much more assistance.

P/T of 230% means that 230% of the tolerance is consumed by measurement variation correct?

This gage basically sits at a 45 degree angle, with a V block slide that sits between two Posts. at the end of each post is a laser that shoots out,the Vblock is pushed up by hand into the laser which hits the outer diameter of each side of the part, and reflects back into a mirror.

The data is crunched by a control unit, and a measurement it spit out.

Trying to present the part into the laser in the exact same position is something they DEFINATELY need to improve. Just the way the Sliding V block was moving after doing 1 or 2 trials was causing issues.


I just wanted to verify with you folks who seem to have alot more experience with this what my conclusions were.

I told them that if they could devise a more effective way to present the part consistantly to the gage that they should try and re-run the parts this time with a part that comes from a known stable process.

The machine is able to store a master diameter from different parts so as long as the gauge is able to see the master they can use it to measure the same characteristic on different parts. I spoke to the gentleman who built the gage for them and he confirmed it.

Funny thing is that even he told me that he knew going in that the gage would probably have a really poor repeatablity.


Thanks for your input, If theres anything else I can add to this please let me know!

P/T of 230% means that 230% of the tolerance is consumed by measurement variation correct? Yes.

This gage basically sits at a 45 degree angle, with a V block slide that sits between two Posts. at the end of each post is a laser that shoots out,the Vblock is pushed up by hand into the laser which hits the outer diameter of each side of the part, and reflects back into a mirror.

The data is crunched by a control unit, and a measurement it spit out.

Trying to present the part into the laser in the exact same position is something they DEFINATELY need to improve. Just the way the Sliding V block was moving after doing 1 or 2 trials was causing issues. This could have an impact on all three areas (repeatability, reproducibility and the interaction).

I just wanted to verify with you folks who seem to have alot more experience with this what my conclusions were.

I told them that if they could devise a more effective way to present the part consistantly to the gage that they should try and re-run the parts this time with a part that comes from a known stable process.

The machine is able to store a master diameter from different parts so as long as the gauge is able to see the master they can use it to measure the same characteristic on different parts. I spoke to the gentleman who built the gage for them and he confirmed it.

Funny thing is that even he told me that he knew going in that the gage would probably have a really poor repeatablity.


Thanks for your input, If theres anything else I can add to this please let me know! I agree that the fixturing needs to be addresses as the first priority. Once that is done and the study repeated, it may highlight another area.

Gracias Seniõr Miner! :agree1: :thanx:

Hello, I just joined the forum so sorry for the late reply to this post. I have a little experiance with using laser gaging and performing gage R&R's on them so I can give you what little knowledge I have........ right or wrong.

Basic information you may already be aware of:
I guess the one question that I have is what type of laser is it, is the laser beam that shines across the part a band that will measure up to a certain diameter or is it a single direct (fine) beam. The reason that I ask is because if it is like a bench style laser mic (a laser band), the laser should always find the highest and lowest point on the diameter put in front of it. Laser mics are typically accurate and repeatable to about .000003" - .000005". So if your laser has a 3" laser band capacity you need to be sure that the part is placed within that band but if you are using a fine point laser you will need to ensure that when the part is introduced into the laser you are always stopping at the exact same place on the part (dead center to be accurate). So depending on what type of laser you are using, the fixturing may not be much of an issue, as long as you part diameter is not real close to the same size as the laser band width you should get a pretty repeatable value.

When doing your Gage R&R :
1. Make sure the parts you are measuring for the study have enough variation between them, you need to make sure the gage can tell the differance between .0001" & .0002", this was shown as an issue on the data you posted (it had only 2 Distinct Catagories). This was probably due to not enough part variation.

2. If the diameter you are measuring has form error (roundness (egg shaped or 3 lobed) this will cause havoc and skew your results. There are no round parts are perfectly round.You need to put some type of identification mark on the part and on the fixture so when each operator loads the parts they must line the 2 marks up together, this way everyone will be loading the part onto the fixture in the same orientation. This will help remove (lower) the operator to operator influence.

3. Contamination on the part when using a laser gaging system always seems to be an issue, the parts must be super clean. If one operator wipes the part off with some type of cloth they may add lint or something that may effect the reading. I typically have the operators blow each part off with compressed air before loading it.

4. Other things you can do if you cannot put identification marks on the part & fixture for alignment are: You may need to take 2 measurements on the diameter 90 degrees apart and take the average. Have the operators load part and then rotate it to find the high point (largest) reading and record that number.

5. You need to make sure all your operators are loading the part and taking the measurement exactly the same way, if needed, write a short SOP or procedure for them to follow.

I hope some of this information helps, a laser gaging system should have roughly a 10% Gage R&R when using a ± .001" tolerance.

Good Luck,
Trae1170