Lost alignment

[[Da...]]

operator comes to me saying he had to do 3 manual alignments yesterday because when he went to run a second time the probe was using a z axis cp for a y axis ID!
He was quick enough to catch it before it crashed but it would have. After he did the manual alignment, it ran fine. It ran fine for quite a few parts then again, out of nowhere, it would go to a random spot. The spot doesn't line up with any of the known positions.

Any clues? we blocked edges and increased retract distance. He'll keep an eye on it.

[[Ju...]]

Performing a manual alignment incorrectly can flip the base alignment, causing clearance plane assignment issues.

viewtopic.php?p=43813

[[Ke...]]

A few things to try:
Restart the system, including the controller. The controller is a computer, and it is good practice to shut it down from time to time. I have seen some instances of base alignments going off being fixed by this.

Go to->Resources->Utilities->Set Base Alignment to Zero, then reteach the manual alignment. This will clear out any data or settings that Calypso may be holding onto for whatever reason.

With a blank program open, probe in cardinal directions on flat faces and compare the vector onscreen to the probing vector. In some instances a base alignment going off can be a sign of a probehead going bad, which can be seen in vectors being wildly off (like a +Z vector onscreen when probing in -Y). These vectors also show when doing a manual alignment, so keep an eye on them as well, which is good practice because as Justin said an incorrect manual alignment can cause bae alignment flip.

[[Da...]]

It is running fine for 3-4 groups of 3 parts, with several different parts running in between. We wouldn't be able to run at all with a twisted/flipped alignment.

After some time, it happened again. The machine had finished a rather large part and the large "t" probe finished above and to the left side of the machine. We loaded the problem part, changed probes and hit start. The probe went to a Zaxis position above the part, but in the center of the part. The first feature is an ID and the part sits in X axis, so this would crash.

So for a shot in the dark, we manually moved the machine to the right side of the machine and re-started from there.
It ran perfectly. This makes NO sense.