Klein Peenemunde update:
The pic above shows that I've now printed out a full set of parts for the Rev 0.5 bird. The following pic shows most of the items that have been addressed in this version:
These parts are just test-fitted together -- no gluing, so the joints aren't as tight as they'd be if they were glued. You can see the steam vent printed into the bottom of the body, the new fin-body mechanical interface, the new, more realistic cross-section of the bottom of the fins and the printed-in rudder-actuator chain fairing.
There are a number of important details that can be see in this pic. The first is something that, when I was younger, would have caused me to throw a screaming fit when I noticed it (and probably put the project aside for a long time, or forever, out of frustration and anger). Fortunately, with age comes more patience. The rudder-actuator chain fairing is wrong, wrong, wrong. It's about 30% too short at the top. I spent a couple of hours frantically going back over the V-2 blueprints to confirm that I got this wrong and figuring out how I could make such a stooopid mistake. One wrong choice of a "key-point" on the blueprint to use as the guide for the length of the fairing ... Given that each fin takes almost three hours to print when I add up all three parts that make it up, and that I was well along in surface finishing when I noticed this, this one stooopid error accounts for almost a week of "wasted" work ...
But it wasn't all wasted, because I've been using the time to continue experiments in surface finishing. One experiment involved dipping parts in a bath of acetone. The idea with this method is to slightly melt the outer surface of the part to cause the striations created by the layering of the print process to even out. FAIL. I can see how those who have recommended this method got somewhat good results, given what they were trying to achieve, but the loss of surface detail and precision is too high for a scale model, and the lack of control over the process (at least with the equipment I have) is too great.
I also did a number of experiments with the auto paint "filling primer" suggested by Pickelhaube. This showed significant promise in a number of experiments, but seemed like it would not address the whole problem.
This lead me to work up a whole series of experiments combining various methods in series. The goal was minimizing sanding time and improving final surface finish. The current best result, which I'll be pursuing with the fin set seen in the pics above, consists of the following:
Step 1: Initial heavy sanding. Sanding the printed piece with very heavy grit sandpaper (100 grit). This takes out the largest imperfections and decreases the over-all relief of the striations SOMEWHAT.
Step 2: Component joinder. Major assemblies glued together with CA.
Step 3. Heavy puttying with acrylic modeling paste (the same filler I used for the entire finishing process on the Rev. 0 bird). This fills the largest joinder gaps and all of the layering striations.
Step 4. Heavy sanding. The acrylic paste is brought down to the top plastic striation layers, again with 100-120 grit sandpaper. This leaves some significant pits caused by air bubbles in the initial paste coat.
Step 5. Acryllic paste touch-up. The largest remaining pits are filled with acryllic paste and sanded again with heavy grit paper.
Step 6. Filler primer first coat. This is sanded with heavy, and then medium paper and provides a "macro-smooth" finish. At the end of this stage some of the surface below the filler primer is visible and the surface shape of the component is as good as it gets and only sanding artifacts remain.
Step 7. Filler primer second coat. This is sanded with increasingly fine sandpaper grades all the way to super-duper-fine polishing paper. At the end of this stage, only the primer is visible.
Step 8. Painting with lacqer. Just like a normal project.
That may seem like a LOT (and it is) but, believe it or not, it actually involves about half the sanding time that I put into equivalent surfaces with the Rev. 0 bird, and produces a glass-smooth finish.
I'm going to follow this procedure for the first iteration of the whole Rev 0.5 bird. What won't be there when I finish will be any fine surface detail -- panel lines or fasteners. I haven't even begun to consider seriously how I'd address that, although I've had some discussion about it off-forum with folks who are interested. (And note -- I'll continue that discussion soon.)
Finally, the third thing that's apparent to me from the second pic is that I definitely went the right way with the new fin-body mechanical interface, because it's much more conducive to the iterative rapid protoyping 3d printing methodology. With this design, I don't have to chunk the whole bird because I screwed up one design detail, and I can continue to improve various major components on a single development bird. Thus this Rev 0.5 rocket will likely remain the on-going development article for some months, until I'm ready to declare a true "Rev 1."
