Don't use the collapse transformations/continuously rasterize switch.
If you need to have that layer crisper when scaled, precompose it, tick that continuously rasterize box, and scale it up. Use that precomp in your main comp.
Ok. so the individual layers which need to be continuosly rasterized should be enabled individually within the comp or the PreComp itself should be used in composition with Collapse Transformations enabled?
If this comp is nested further on in other comps should all of them be enabled with Collapse Transformation?
All of this really depends on what you're doing. The only reason to enable the continuously rasterize button is if you need AE to continuously rasterize a vector image.
The only reason you would need that is if you were planning to scale it larger (or move a camera closer to it). This is taken care of by scaling in a precomp. That way any effects that get messed up by the continuously rasterize/collapse transformation switch won't matter.
I'd suggest reading up in the AE help on the use of this switch in working with vector graphics.
I read the help, it states that Collapse Transormation buttons treats the PreComp as seprated layers, which preserved 3D treatiment for casting shadows and does not cut off shapes by PreComp size.
It seems to work for me that I check continuosly rasterize on each individual illustrator imported layer (inside the PreComp1), so when I enlarge the shapes they stay crisp. Then I keep on checking Collapse transformations on all consequtive PreComps and everything stays sharp.
My question is at what stage I would better apply Transform effect, which doesn't seem to work. Should it be on the PreComp1 layer with all the Continuoiusly rasterize enabled or on final Render Out PreComp where "Collapse Transformations" is checked off?
In eather way because of movement of these illustrator layers, Transform position seems to be fixe on screen, but is not relevant to shape movement whithin the PreComp1.
What previously worked for me was that I applied Transformation on each individual illustrator layer, however I did not enable "Collapse Transformations" in and consequtive PreComps. As soon as this feature is enabled everything goes awry....
The most common reason things go awry is that users fail to realize that when you collapse transformations on a 3-D layer in a pre-comp that the camera in the main comp is the one that will be used. This really fouls things up if you have a camera in the pre-comp or you are not looking at the active camera view.