Perpendicularity Glitch

[[Ji...]]

We are having a problem where Perpendicularity is checking good then all at once it is checking .900". I have noticed that when this happens it is exactly the same as the projection number. Friday, I had to re-program the pocket, or it gave me the projection number for an actual every time. Today I ran 30 pieces and 4 of them were .900" scattered throughout the run. Is there a fix for this glitch? We are scanning the pocket at the top and .900" for the bottom. It has .200" clearance, so it is not hitting anything. I am at a loss for what could be causing this. When I run the parts again, they are all good.

[[Ji...]]

Since all of the smart people on this forum don't know, looks like it is time to turn in a ticket to Zeiss. Hopefully this is something they have a patch for.

 

[[Ma...]]

I think you gave people here a small amount of time to respond. I am not currently at machine until tuesday.

I would suggest to observe your actuals for feature - especially vectors/axis.

[[Ri...]]

Not a lot of information here to work with. 

Is this the feature a Cylinder/Plane/3d-Line? What do the actual values show for the A1/A2 for the feature? What do the actual values show for the A1/A2 for the datum? What is your measurement strategy? Can you post screenshots showing data?

Are you using the new GD&T engine or the legacy? What fitting algorithms are you using? Filtering? Outlier Elimination?

[[Ma...]]

Whenever I've seen this occur. the cause has been that the feature is calculated to be parallel to the datum axis rather than perpendicular to it. Usual cause is order of manually taken points changing the axis direction for a datum feature.

[[Ji...]]

I can give you every detail that some of you are asking for. The problem is not in any of those details. I have ran these programs for years without issue. All at once it starts and is intermittent. When it does this right for the pockets all of the x and y data is correct. All of the angles for A1 and A2 are near zero. When the glitch occurs, it will have crazy numbers in each box. X actual may say 4032.7456 Y actual 8256.2565 A1 405.876 A2 45.098. When I have ran these for years and never had it before now and can re-do the program for the feature involved and go back to running good parts. This leads me to believe it is something in Calypso having an issue. We do 30 piece runs using these programs all the time, so why has the problem just started? Since this started, we have tried many of the different ways to get perpendicularity and still get the same crazy numbers intermittently. It is a plane to a pocket. We have tried scanning a line, cylinders, 3D line and scanning multiple lines for a cylinder. All will check good with all nominals correct then get 1 or 2 that gives us the projection length for the actual. This has been a major headache for the last 2 weeks. I hope I have answered all of the questions everyone had and somebody has an answer. I have been doing this for 20+ years and never seen it before. All help is much appreciated. 

 

[[Ri...]]

Unless you are able to share the measurement plan, or answer the questions that I asked, I'm not certain we will be able to assist you. Feel free to reach out directly for assistance as well. Thanks. 

[[Ji...]]

Richard I gave you the fact that the nominals are where they should be. When it does the stupid the actuals have extremely large numbers. The filters and outliers are the standard from Zeiss. We rarely use anything different. Have not heard of the new GD&T engine. When nothing has changed with the program in years, it just makes zero sense why it just started. I also stated that I have tried changing the cylinders to a line scan and a 3D line. All give the same answer within a tenth when it doesn't go crazy. All still go crazy too.

[[Ma...]]

What you haven't given us among other things is what is a "pocket"  or how are you determining the axis of the pocket.

Probably best for you to call Zeiss and let someone get on your system with TeamViewer so they can see what is going on.

[[Ji...]]

Please sign in to view this quote.

The pockets are an arc. I have tried recalling feature points in a cylinder. I have tried scanning a 2-D line and recall into a 3-D line. I have tried doing a cylinder straight up with scanning the arc top and bottom.  None of what everyone is saying I am not giving you has anything to do with the problem we are having. Our company has let the SMA expire and have no plan of renewing due to cost. Therefore, I cannot get help straight from Zeiss. Our management has dug in on this. I can't really blame them with the outrage's prices Zeiss charge now. I understand pay to play, but the management team is not about that.

[[Ma...]]

I would start with restricting of radius - this should help with finding axis for cylinder

[[Ji...]]

Please sign in to view this quote.

Thank you, Martin. This is the first response that has offered something to try. I will try it and let you know. I am not allowed to share the program or anything that would be helpful. Corporate policy are very strict here.