HMS Surprise by irxum - PLASTIC - Experimental 3D printed hybrid scratchbuild + first build

[irxum]

Hi all. This might more properly belong in the scratchbuilding build log forum - I hope to turn it into a reasonable build log - but I thought it might be a bit presumptuous. I guess the moderators can decide.

 

INTRODUCTION:

I see a lot of people here building these beautiful ship models out of wood. What I think they don’t realise is that they and their skills have been rendered totally redundant by the swift march of technology. Do you hear that whirring? It’s a 3D printer. That’s right - the way you have fun is obsolete. Machines can do all this better and quicker than you old-timers with your skills, care, precision, and attention to detail. What’s the point of carefully planking a hull by hand when with a bit of clicking this and clacking that, and some swift beep-boop, a machine can make one that’s not only harder, but better, faster, and stronger too?

Of course, literally everything I wrote in the previous paragraph is complete nonsense. 3D printing can’t and, in the sensibly-foreseeable future, most likely will never be able to reproduce a quarter of the work done by hand here -- and even if it could, so what? It’s not my aim or opinion that it should.

Sadly, though, I don’t have nearly the quantities of free time necessary to get as far into the hobby as I’d like. Nor can I sustain the interest - I can focus for 16 hours a day on something, while it captures my imagination, but after a few months I will lose interest and move onto one of my other interests, possibly for as long as a couple of years (for example, my 2 year old Venetian polacca kit). If that seems pathetic to you, then fair enough. I envy your persistence of interest. Anyway, these factors combine to point to ship modelling not being the ideal hobby for me. But I nonetheless want to build ship models and I want to have the finished models in my house; so, I turn to technology.

I have poked around in the CAD/3D printing section of the website, and seen that people here have of course looked at 3D printing and arrived at the sensible conclusion that only stereolithography prints are up to the requisite standards for most applications - figureheads, cannons, decorative reliefs, etc etc., but that that technology was mostly too expensive to own at home and to print large volumes. What occurred to me -- and, indeed, user @RKurczewski had actually done and suggested -- was to print hulls in sections using a home FDM printer and process them afterwards into the finished model. Unfortunately, although RKurczewski looks to be more-or-less still around (hello?), they didn’t follow up on that idea with a meaty build-log since they posited the idea in 2014, so I’m going to see if I can bring the idea to life a bit.

There are several ways to proceed with regards to post-processing the models, and each have different demands at the CAD/3D modelling stage, as well as different pros/cons, so let’s examine them a bit:

1. Leave it as is:
    - Probably want to print decks, deck furniture, whales, masts, transom, beakhead, etc. etc. in place - all will turn out low-resolution and/or with pronounced layer artifacts
    - Joins between sections will be ugly
    - Hard edges of model will be ugly
    + Simplest and quickest
2. Do some post-processing, e.g. sanding, painting:
    - Detail problem as above
    - Bits that aren’t painted will look rubbish
    + Could hide/minimise hard edges
    + Can hide joints
    + Moderately simple and quick
3. Post-process as above and detail the model with wooden parts:
    - Takes longer
    - Bits that aren’t painted will look rubbish
    + Hide joints & hard edges
    + Detailed structures and furniture
3b. Print in wood-fill plastic, post-process and detail the model with wooden furniture:
    as above, but
    + effect of unpainted plastic will hopefully be mitigated
4. Post-process, then completely plank over the plastic elements
    - Takes ages

    - I don't have the materials
    + Few to no 3D printing-related drawbacks?

Actually, many other alternatives exist - e.g., print frames complete with camber for a POF model. I'd like to try this at some point.

So, bearing in mind this hopelessly incomplete list of pros and cons, I’m going to shoot for 3, and then 3b if I meet any success/can sustain my interest - woodfill is quite expensive stuff.

So, onto the build log.

METHODS:

I tried to make a 3D model of a ship from plans. It turned out like this:

Pretty useless. The middle is just about OK but I can’t seem to find high-resolution scans of the ships I'm interested in, or interpret ship plans at all well, nor can I work out how to make a good-looking transom or bow.

So, I thought, how can I take a shortcut? Can I piggy-back on someone else’s hard work and jump straight in at the preparing for 3D printing stage? And the answer was, yes!

I turned to my dedicated friends over at the Pirates Ahoy! community. These guys have been maintaining and developing a massive, awesome, comprehensive community patch/mod for the 2004 Sea Dogs game by Akella that was acquired and released by Disney close to the end of its development cycle as a tie-in for the film, Pirates of the Caribbean. They also are developing their own game to escape the limitations of PotC, called Hearts of Oak, and I did a bit of volunteer work on this game a while ago. With the tools they use to work on the PotC engine, you can get access to the models in their mod. However, the export tool is a bit janky w.r.t. the topology of the models, and they are designed for use in a gen-before-last game, so they are low-poly. I did press ahead with one of those models, HMS Surprise (not HMS Rose), but one of the most prolific ship modellers at PA!, Armada, also offered to let me use his WIP ship for HoO, HMS Bellona, which is a beautiful model and for which I thank him profusely. I'll be cross-posting this over there so NRG might get some visitors.

So, I’m pressing ahead with the two ships, HMS Surprise, by now-sadly-inactive PA! contributor pgargon, and HMS Bellona, by Armada. Working on other people’s models means that I’m not free to distribute the files - kind of a faux pas in the 3D printing world, but sorry.

Here’s what HMS Surprise looked like when I first opened it (all views are from the software I’m most used to modelling in, Wings3D):


And here’s HMS Bellona:


As you can see, HMS Bellona will need much less processing.

HMS Surprise first.

There is a lot of polygon wrangling involved. One problem you face here is that models for games do not need to be manifold; that is, topologically closed, >2D, and not self-intersecting. To save processing power, they are often made of surfaces that are not only 2 dimensional, but probably only have a visible surface in one direction. They don’t necessarily join on to surfaces that look contiguous (which is also I think exacerbated by a quirk of the .gm to .obj conversion software). However, the model that I print does have to be manifold, so all the planes need to join up into a 3 dimensional object, and should definitely avoid intersecting with itself, and also ideally with anything else, or becoming inverted. So, disconnected sheets of polygons have to be joined up, thickness added to them, quirks of the mesh need to be mended, and, importantly, polygons need to be added. Game engines smooth the boundaries between polygons by rendering the edges between them as ‘soft,’ but 3D printers do no such thing. Therefore, to avoid too many hard and obvious edges, the hull had to be smoothed. I daresay any 3D modelling program can do this. Beforehand I recommend getting as many tris up to quads, though, because it looks less awful.

We go from this, which will be conspicuously facetted when printed:


To this, which will be a bit less:


As you can see here, the geometry around the stern gets kind of weird:


I’m hoping that it will come out in the wash, by which I mean I’m going to take a rotary tool with a sanding drum to it until it stops looking weird. The thickness of the hull was eyeballed so so it would turn out 2-3 layers. Fortunately that worked.

I decided not to put any of the wales on HMS Surprise because I think it will be possible to get a better finish on the hull if it has no surface details. For this reason I also didn’t add the keel into the hull geometry - at the moment there’s just a groove where hopefully I can find the wood to add one later.

I expanded the gunports so they can be lined in wood (maybe a bad idea? does rely on the upper whales being planked for good effect), added guides to locate and correctly angle the mast and bowsprit:

Played around with plans for stern framing:


and added rails/guides to help locate the decks:


I’m not sure the guides for the weather deck will print properly, but we’ll see. I also added little tiny keys to help locate the parts against one another:

I wanted the print to be about as big as possible. The 3D printer I’ll be using, my brother’s Turnigy Fabrikator, has a build volume of 130x140x90 mms. For the first run the scaling turned out to be a little arbitrary - I haven’t yet calculated what scale it turned out.

Finally, the ship was cut into 9 parts and printed in 5 sessions, each 5-7 hours.

HMS Bellona:

Unfortunately I didn’t have time to process Bellona at all before going home, and by the time I had finished Surprise, I was getting so close to the end of my time with the 3D printer (Xmas holiday at the family home) that I had to rush it rather. I fear that I haven’t really done it justice - sorry, Armada - but I will definitely revisit it when I get my own 3D printer in the near future. I took the hull model straight from Armada’s files, cut it into 4 pieces, and printed it as is - keel and whales in place, and without any deck guides. Unfortunately I was a bit of an idiot and scaled it so it was too tall for the print bed. We actually discovered in the process that the printer could do 10 cm high pieces, but a bit of gcode at the end of prints moved the hotend through the print and messed it up a little quite a lot. I didn't have time to reprint so the remaining two pieces were halved.

I have now printed the parts out (sorry if the tenses/perspectives get a bit screwy throughout the log, I wrote half of this before I printed and half after) but I haven’t got any photos - to follow in the next post. Please don't expect this build log to move with any degree of haste, as I'm unfortunately pretty busy.

Edited January 11, 2017 by irxum

[EJ_L]

Interesting to watch you design and print this model. I think this is going to be a rather fun and informative build. Looking forward to seeing your progress unfold.

[HIPEXEC]

Looking forward to see this build

[PeteB]

I've been occupying myself lately going through Solidworks  and Fusion 360 Tutorials just to try and get up to speed with the changes from my old AutoCAD Lt days. Your log is just what I've been waiting for so I would like to tag along with you on this one. Cheers Pete

[Hias]

Hey have you ended the project? 
 

could you send me the 3d files for a tip?

 

cheers