Coleagues
Just another issue I´m unsuccesfull.
Trying to laminate a sandwich panel  with nonconstant thickness in 2 steps process. 1st step there is a cca 0.7mm thick panel based on UD carbon prepregs - 1x54+2x100+2x200gsm UD. On 2nd step apply the Rohacell core and inner skin based on twill 200gsm carbon prepreg, see the attchd. sketch. The curing process starts on vac-1bar@100°C for 4.5hours, cooling slow down to 30°C for 6hours. After demoulding the panel is going to distorsion about 10mm.
Any tips for process changes?
Thanks a lot, Rosta

Chris
1. Yes, the bottom skin was released from mould before.
2.The core chamfer is 45° all around the shape.
3. Yes, all the prepregs are cured with the same cycle
4. I have made an experiment where all the plies were cured in one shot, no surprise, the effect is totaly same.
This is my conclusion, you right, the laminate is TOTALY UNBALACED.
The picture below shows the basic experiment and the deformation results. If the sheet flatness is the crucial parametr, the laminate has to be strictly balanced and symetrical across the plies. Because the thermal expansion goes in fibers direction, on plain/twill ply goes 0°+90°, on UDs goes 0°. This fact generates the internal tension penpendicular to UD direction. On sample 3 the laminate is balanced and sheet is finaly flat.
On panel above there are  5 plies UD0° + 1 ply twill which generates the tension because contains fibers goes 90°to UDs direction and UDs plies has no 90°fibers...
That's the result I hope so...