Flipped axes

[[Pa...]]

I am the supervisor of a well established CMM room. I have some Calypso but have never been the guy spending hours at the control. Both of my experts are on vacation, and we have a production part we cannot run.

We've made and measured a million of these. We've been making them for all of the 15 years I've been here. The feature to be inspected is a 0° helix involute spline. We use GearPro. The spline is a 10+ inch diameter located at the end of a shaft 10 feet long. I'm in an Accura 2 with the long axis of the part oriented parallel to the CMM Y axis.

Opened the program, clear existing results, default is already set at the "named" alignment. Pull the trigger and the CMM comes down several feet away from the part.

Kill it. Start again, this time using a manual alignment. Alignment is a circle near one end (circle is in X/Z, a circle near the other end (same orientation), and a plane on the Y- end face of the shaft.

We've got a really good set of adjustable tables upon which we perform all inspections. Operator setup instructions control the location of the workpiece inside the measuring volume to within .050" in any direction.

So now I've read about flipped alignments in resulting from GearPro's inability to evaluate splines parallel to any axis but Z (Why is this, btw? All we make are huge cylinders) but do not know how to proceed. Given this is aerospace R&D, I cannot simply make changes to the program.

Guess what I'm asking is how to proceed. I know to look at the blocked edges. If I change those, what other factors needs consideration?

We lost all of yesterday and are most of the way through today now....

[[Ri...]]

I would try and run from the Base Alignment as if it is a new setup, this will over write the saved Alignment you are using.

[[Pa...]]

I wondered what I had left out of the OP.

So we did try to run from the new base alignment. It was a new setup, but as stated we know that the piece is reasonably close to the same location and orientation as the last one we ran. As such we did not (and typically do not) use a start alignment after prove out.

In total we tried 3 times with the current program. We even dug an obsolete revision out of the program out and tried a few times; same results.

Unfortunately it would seem that there's some critical piece of info that I've forgotten...

[[DW...]]

In the "Start Measurement" window, under base alignment, select the program named alignment(CNC). There are three options for a base alignment - Manual Alignment (you manually take the alignment, easy enough). The next is "Program Name" Alignment. This base alignment is the last known manual alignment. Then there is "Program Name(CNC)". This is the last known completed base alignment in CNC mode.

Is this information helpful?

Also, I am guessing the cylinder axis created from the two circles you measured is your spatial orientation in your base alignment.. is it oriented in the Y axis?

[[Je...]]

So it sounds like you are using a 3d-Line axis as a spatial for your alignment being built off of center points of circles. Doing this in a Base Alignment without a Start Alignment can result in these types of troubles as the features can be probed in a manner that causes the calculations to have the wrong origins/directions. I am going to assume that you do not have a planar rotation set as you are machining cylindrical parts. This also presents another condition that can contribute to part "flipping"

If this is the case, the alignment may not be calculating properly due to the perceived orientation of the part. There are a few ways to correct the issue you are seeing (if I am interpreting correctly).

First off, if you are going to use a manual alignment to reconstruct a location after a problem with navigation, before probing your alignment go to Resources > Utilities > Set Base Alignment to Zero

Go to your inspections directory and delete (or cut/paste to a different location) the Base Alignment file named for that program if it still exists under C:\Users\Public\Documents\Zeiss\CALYPSO\workarea\basesystems (this is double duty on the same task but if there is corrupted data, this step is necessary)

Attempt your manual alignment again. Be sure to probe any 2d lines in the direction the alignment specifies or any features used in constructions in the proper order and direction as shown in the feature itself. Doing this improperly will "flip" the feature.



Edit 1 - temporary fix method - Create a new Base Alignment, keep all things the same except change the spatial feature to any plane that is perpendicular to your cylindrical axis (an XZ plane at the -Y is usually easiest but the largest of the planes is typically most robust in terms of projection error.) Select OK. Perform a manual alignment and allow the machine to measure the Base Alignment in CNC mode (I am not referring to alignment *named*(CNC)). Once it completes, create a new Base Alignment again and construct it as it was prior. Run the program using a named alignment. It should run as normal.

Note - A proper start alignment set up in this manner will prevent this from happening in the future.

Edit 2 - Probe the granite surface of the CMM to create a 2d-Line in an -X - +X direction roughly the width of your X clearance cube size. Use this 2d-Line feature as your planar rotation in Base Alignment. This will prevent CALYPSO from "flipping" your part about its Y axis in relation to the actual CMM

NOTE: DO NOT EDIT your alignment. Only CREATE NEW. This prevents the recalculation of the Base Alignment coordinates that changes your feature nominals. You are not intending to change nominal data.

[[Ow...]]

Quick read but if it's going to nowhere land when you fist start the program, I bet the alignment is flipped, which you should see in the menu so, I think Jeffery has you going in the right direction to fix this issue.
If it's going to nowhere land once you go into gear-pro, then it's the secondary alignment (with a special rotation) used to flip the alignment for gear-pro to run horizontally.
The pic below shows how I had to set one up to run on the side.

[[Dy...]]

Make sure your using the "active" stylus when taking your manual points? Not sure what your probe set up looks like but taking manual points with the wrong probe can also cause this issue

[[Ka...]]

It's probably that the program uses a "start alignment" as others have said. Ideally, the programmer has included setup instructions in the program (not in front of software, but I think it's the "user information" button in the start screen).