Meshroom - Void method with the crossed polarisation

I have been trying to use Meshroom to create full 3D models for some time now. I’m relying on materials available on the internet, mainly the videos on YT:
Grzegorz Baran - Photogrammetry Setup for Indoor 3D Prop Scanning
Erik Christensen - Scanning in the Void - A short video on good turntable setups
and on all the materials uploaded by Eugene Liscio on the 3D Forensics

Based on these materials (all of’em are great!), I have been able to get models good enough for archaeological finds:
Model 1

However, I often run into a problem when scanning longer objects, such as fragments of vessels or bowls, that simply don’t fold in SFM. I don’t have this problem with the simple long objects like below:
Model 4
So, I ask if anyone has had similar problems and found a solution?