MrSteve Cool! Will be interesting to see the results, as I've not tried annealing any PET-CF yet. To try and address your questions:
- I'd stick with Bambu if you are using a Bambu printer, otherwise just use Prusa.
- With lowers the outer perimeters is what effects the shrinkage. Infill only makes an impact if it's pretty thick or has large flat sections.
- So I'd not worry about the masked areas.
- Annealing is pretty much just a scaling issue, though warping could be an issue, and there is little on the slicing end of things that will fix this.
- Probably not worth the trouble, but could be fun.
I think it's best to anneal quickly to eliminate any effects water might have, but with PET-CF this is probably a non issue. Definitely leave the supports on.
To address number two from above a bit better, you need to look at how the part shrinks. There are two types of shrinkage. The first is absolute shrinkage, the actual print lines becoming shorter. The other is line to line shrinkage, the layers and lines shrinking together. Generally the line to line shrinkage is the significant one, though absolute shrinkage can be high with some unfilled Nylons and PLA.
Print a long sample with aligned infill and use that to determine absolute shrinkage. If you make the sample wide enough you can also use it for line to line shrinkage on the horizontal axis. If you measure it in the middle the perimeters at the ends will have no effect. I'd print a second sample for the Z axis.
An interesting part of these two modes of shrinkage is that many infills below 100% will shrink at the absolute level, possibly very little. Mostly it's the Z axis and parallel lines compacting.
Would be interesting to get some theories on why the line to line is typically greater then the absolute.