IfcSite LocalPlacement Values not correct?

[JSN]

I encountered a strange bug (?)

1 Project - 3 Part-Models - absolut identical Setup (from AC29)

Indeed the IFC Model Setup used to be absolutely fine …
Here is an overview:

But, now I got the new models and suddenly …

1 Model has a different IfcSite Origin

Here’s where the numbers come from … position of IfcSite

Author swears he has not changed anything … and I believe him because this kind of Setup is afaik not possible to achieve because no matter what’s the export settings …

Prove: (Not relevant here but just to be sure) the Survey Point is still the same

And also the geometric model positon fits together … nothing has changed.

A look into the file confirms this surprisingly

There is no hint in the original file where the -105,314 / -84,940 comes from?

What kind of black magic is making Solibri showing those values for only this file every time?
Am I missing something or is ther an easter egg showing those random numbers due to a hidden trigger?

Update

So it seems like it’s because geometry was classified as IfcSite and therefore Solibri is confused and it mixes information and merges the “real” IfcSite attributes with the local placement (?or maybe the bounding box?) information of that specific element together in one … wild.