4.1 Alignment Problems

[gcunniff]

First of all, I'm coming in behind someone who already setup about a dozen programs using two different fixtures. For the sake of continuity, I followed their lead. Here is the situation:

There are two fixtures that check about a dozen or so parts. An separate alignment program was written to rough locate the fixtures on the plate. Basically two threaded holes in the fixture plate and a surface point on the plate.

Next, these rough alignments are recalled into part programs as external alignments. Next, the guy just open up the auto features window and sampled features on the part itself, not any features off the fixture. There is no correlation between the external alignment and the auto features that are developed off of it, and the part nominals. The features are simply measured to what ever alignment was created off the external alignment program. When sampling part features with the external alignment, the machine is in DCC, so basically you end up with a DCC program that samples features and then creates a part alignment.

Now the problems begin. After creating a part alignment, I brought in a part CAD file and did a CAD = Part. I wrote a program for the part using the CAD model without any problems. When I tried to run the program from the beginning however, the nominals for the DCC alignment features changed. I believe this was a result of the CAD = Part command. Furthermore, after resampling the DCC alignment features to the external alignment again, I found that the rest of the features are no longer linked to the CAD file and now they want to check out in space somewhere.

In the past, I have always written these automatic alignment files in a separate program and then called that external alignment into the part program. So basically, you would have a manual alignment program that would locate the fixture on the plate, a separate DCC alignment program off of that fixture or the part and finally a part program that recalled the external DCC alignment. This way, if the fixture didn't move on the plate, you would just run the auto alignment and go.

What am I doing wrong here? Is there something going on with this external alignment? The DCC alignment inside the program (with nominals that don't relate to anything except to rough fixture alignment to the plate) or is there something going on with the CAD = Part command? It looks like all of the programs that were written previously used the method described and in fact, I was able to write a program with this method until my problems tonight.

Also, why is CAD = Part sometimes in the Alignment menu and other times, it is not? Am I pulling the CAD into the program at the wrong time? Any help with this would be appreciated. I am completely flummoxed with this **** thing!

[Don Ruggieri]

whoa

You have several questions there, let me try my hand at answering some of them.

Yes, it is true, sometime the CAD=PART offering is not listed in the alignment men u. This is by design. It is found elsewhere in the menu structure. I don't remember the details on why. I suppose it has to do with recommended procedures. Just know that you are not going crazy here.

It is also true that CAD=PART modifies the feature coordinates of the commands at the start of your program. This is also by design. It is not intended to be used for full DCC programs, at least not in the mtehod you are using. The reason is that when you measure those first few features, you are using an un-related coordinate system. There are several ways to handle this.

One such method might be to re-learn those "front-end" features in a new program and then paste them into this program, replacing the original features. This will allow those first few to be relative to the fixture alignment, and the rest to be relative to your CAD model.

Let me stop there and see if it helps so far. . . .

[gcunniff]

Ok, I will give this a try. One other thing that I noticed is this - when I imported the CAD into the program and it drew up on the graphics display window, the points that I had measured to create the part alignment were not matched up on the CAD file until I selected CAD = Part. I would expect that to be the case, since the features were measured to a different alignment, unrelated to the CAD file. However, when I attempted to create a new alignment that would place the part into CAD orientation, I could not orient the alignment in such a manner that these points would line up with the CAD file without selecting CAD = PART. Shouldn't I have at least seen these points moving from one side a an axis, depending on how I rotate a line around the level axis? The axis of the trihedron (or what ever the **** you call it) changes in the graphics display window, but shouldn't the orientation of my points change also?

[gcunniff]

What would happen with the following scenario - write the DCC alignment in a separate program, which recalls the fixture to the plate alignment. Turn this into a sub routine. Call the sub into the part program to run the DCC alignment. Would CAD = PART modify the feature coordinates in the subroutine the same way it would if those points were inside the part program?