Dear all
Just working on the project based on carbon prepreg honeycomb structure. The procedure below is OK, the final rigidity is excelent, but there is a problem on visible surface goes to be non flat after vacuum bagging. See the procedure:
1. Hit the first  layer, carbon prepreg 210gsm, 1 layer, vacuum bagging, autoclave 100°C@-0.98bar/+8bar. Perfect results, excelent flat visible surface, no pinholes, final thickness cca 0,25-0.28mm.
2. Apply the epoxy glue RG (5min.) for better adhesive properties, put on the nomex honeycomb OX4.8, thickness 3mm.
3. Apply the outer layers on honeycomb side, 3x carbon prepreg 210gsm.
4. Vacuum bagging, autoclave 100°C@-0.98bar, no overpressure.
Final structure is fine, but on the mold side - the visible side of structure, there are after vacuuming around all the surface where the honeycomb was applied..how to say that..hollows? The surface is not absolutely flat like in step 1, but deffected by inner honeycomb core. See the picture. Any ideas what's wrong?
I'm convinced, the problem is vacuuming. If I hit the outer and inner side of prepregs separately in autoclave and glue the sandwich together in vacuum, the defect is quite same, if the thickness of outer visible side is 0.7mm instead 0.25mm, the defect is quite same.
Many thanks for support.
Rosta

Its hard to tell from the picture but could it be some pre-release of the mold-side skin from the cooling off after the first cook?  I've had that happen esp. on aluminum tooling.  If that happens the golf ball thing is more likely - even with 3mm cell nomex - not sure what your cell size is...

F1rob mentions vented core and he may be onto something - that .28mm skin isn't much - guessing its woven if just one ply.  I am a big fan of slicing the top of the cells (nomex) or vented core (aluminum) and in either case perforating the top skin every 50mm or so to release pressure as the cook ramps up.  Can you get vented nomex?

Can you cook the whole thing in one shot?  You might need to warm debulk the first skin to get a nice surface but it is totally possible to get a really nice surface finish from a one-cook cored panel.  It's corners and core fitting that make multiple cooks necessary most of the time... unless you're doing something crazy complex.

Lester's theory about the glue I agree with, especially if you are using an ambient cured epoxy to bed the Nomex and it gets up near/above it's Tg during your top skin cook... Can you get some glue-film that's compatible with your resin system?

Good luck - I hope you figure it out!