The problem arises during the "slicing" stage, when it processes the build plate!
It generated a shape on the build plate beneath the object, causing the support attached to the object to warp!
However, as shown by the slicing of the newly created composite—using the exact same base objects—it doesn't produce that discrepancy.
Ok. I think I get an understanding of the issue. Just summarizing in my own words to make it easier to spot misunderstandings:
Own design with a challenging print feature: The lower holding flange needs to be printed directly on a support structure before growing to join the main body higher up.
=> For this challenge I usually use a dissimilar material interface. PETG for PLA and vice versa. This allows the use of a 0mm top z-distance in support settings. Both materials need to be absolutely dry (i.e. self-dryed, especially (!!!) if new).
The support failed. This is not surprising with anything but a 0mm top z-distance. Warping/curling of the support may be an issue, but drying the filament and reducing speed (by half) helps a lot. Also, due to the long lever arm before the flange joins the main body, use 2 walls for the support trees.
The second set of pics still show the needed support. It has just moved from the build plate to the part. In my experience, it is better to place supports on the build plate, use 2 walls for them and use dry filament at slower speeds in case of issues. The pri ting of “islands” in the air requires a 0mm top z-distance unless “islands” are tiny and quickly join the main body.
PS: Apologies if I misunderstood or jumped over a few explanations. This is quite a challenging print. You coul also use thr cut tool to cut it vertically and then join it post printing. This would avoid the need for supports and make it much stronger, but with a noticeable line down the middle.
I wrote that you need to use a dissimilar interface if you need a 0mm top z-distance.
I also wrote that the model geometry has a big, long lever arm while sitting only on the support. If that long arm is neither small nor attached quickly to the main model, it’ll wobble on a support with a weak interface during printing until failure.
The partial screenshots indicate the difference between supports on build plate only enabled and disabled.
But Studio is not bug free. Most bugs I have seen relate to memory issues so a restart (mainly Studio but sometimes the PC/Laptop) usually resolves them.
Either a bug or changed settings. Even a small change elsewhere can change the result of an optimization algorithm like tree support generation.
I never claimed it was. It is a challenging print. Good luck.