3D Printing surf-ski/canoe prototypes

Hi

I want to share one thing I was working on recently, hopefully somebody will find it interesting.

Since I got my first 3D printer almost 4 years ago (CR10) I always wanted to use it as a tool for composite making, at first I only considered the classic route of 3D printing patterns and taking moulds from them (after post-processing for smooth surface). While that worked fine, it was VERY slow process, first "normal sized" object I tried was small canoe for children, 2.2m long with ~50kg capacity, it took weeks of non-stop printing, sliced into many parts to fit on CR10 print-bed (30x30x40cm), I spent almost 10kg of filament on that thing.
Overall, it was not much cheaper and much slower process then cutting the pattern from foam on CNC 5axis milling machine, of course price of such 5axis CNC is a lot higher then $350 for CR10...
Another downside is that such pattern itself is not functional piece at all, even if we would skin it in fiberglass for strength, it would still be super-heavy and not comparable with end-product (moulded boat) at all, so any necessary changes to the design discovered during the prototype testing phase mean that all-this "3D print pattern -> make-mould -> mould prototype" cycle would need to be repeated.

Then I had an idea of printing the core for the composite sandwich directly, which can be hand-laminated (from both sides, or only one, depending on shape, thickness, etc.) and theoretically you could totally skip the "make mould -> mould prototype" phase and test your initial idea directly.
To check whether, its viable I first created custom 3D printing slicer which implements very efficient infill structures and enables very fast printing with no wasted movements (unlike normal 3D printing slicers which needs to be very generic and can't do things like that). 
Then I tested some samples for compressive/shear strength and compared it with high-performance foams/honeycombs normally used in composite building and it compared quite favourable -> 3D printing the core also have many additional advantages like varying thickness/density according to loads in the structure, which is much harder to implement when working with sheets of foam.
Then I tested the whole idea on the small 2.2m canoe again, this time with much faster custom printer with bigger build volume, and instead of weeks of printing 10kg of filament, it was 30h of printing, 1kg of filament, 800g or fiberglass/resin and end result was quite good  - needless to say it's very stiff and strong (the boat survived even being abused as a snowsled during winter, I was quite surprised) and I'm certain it will scale to adult size boat without problems.

My latest creation is 5.5 long very narrow and fin stabilised surf-ski/kayak prototype, here shown with reinforcing unidirectional carbon keel strip already laminated in place.

However in process of doing all this, I developed 3D printer which has quite good price/performance ratio, especially in bigger build sizes and I decided to dedicate myself to bring the 3D printer itself to market, since a lot of people asked for that and it's certainly quicker way to make money then canoe/surf-ski business. 

I still think it would be shame to not develop the concept further and I would be very happy to find somebody I can cooperate with - the software is open-source, I learned a lot about how to laminate different kind of plastic cores, what works and what doesn't, which adhesives are the best, etc.
If somebody would be eager to take the idea and try building boat in such non-traditional way, I would be happy to support it as well as I can and share everything I learned.
 

Very interesting and exciting post! I suspect it's too technical to find a lot of resonance here, though.
I'm interested in this stuff 'cause I'm tinkering with hydrofoils, e.g. see www.surfski.info/forum/35-surfski-innova...foiling-surfski.html and if I manage to get some foils to work (big IF) I'll eventually need a custom boat... I've not done any 3D printing so far 'cause I can't add another huge time-sink to my life at the moment :-). But if there's an overriding motivation...

I saw your threads on reddit, it seems that without real motivation, e.g. "I want to build THAT" it's difficult to generate interest. Bummer.

Some comments:

My understanding is that Don Kiesling 3D prints his rudders using reinforced polycarbonate filament. Maybe he'll chime in here. You didn't mention what type of filament you've been using. I believe his carbon rudders are laminated on top of the 3D print. You may want to contact him to exchange tips.

I'm currently stuck building hydrofoils. Real reason is too much other stuff going on. But I've tried to hot-wire cut foam cores and then laminate carbon over those, but with large high aspect ratio foils the hot wire is very tricky 'cause the cores become very, very thin. 3D printing the core would be awesome... Judging by youtube videos there seems to be quite some interest in DIY foils for kite/wing/wind-foiling, although that window may be closing given how many companies are now selling them. But that may be a place to find interest too...

Another person you may want to contact is ExploreComposites on youtube: www.youtube.com/c/ExploreComposites He posted a really interesting series on very creative test panels but then abruptly stopped. Dunno what happened. But he may have interesting ideas and suggestions at the composites level.

WRT composites, the "holy grail" I've been pursuing is to get to a single lamination step around the core without further surface finishing work. Some of the RC airplane guys are really good at that. The key idea is to wrap the laminate with mylar foil so it comes out with a perfect surface. It's tricky because then you can't have a whole layer of blotting fabric to suck up the excess resin, so you need to do your lamination right from the get-go. Maybe the mylar is impossible on a boat given the complex curvature...

The other thought is resin infusion, which is really the best lamination method after prepreg (which can't be done here due to the required heat). I don't know whether you're familiar with that, but you need a special medium with channels that let the resin flow through the laminate (many examples on youtube). The special flow medium produces a rough surface, so one normally puts it on the back of a panel. In your case you have the core on the back, though. The solution would be to print the channels into the surface of the core . I believe there are honeycomb cores one can buy that have such channels.

All put together, you'd have a core with flow channels, dry laminate on top, then wrap with mylar and/or a vacuum bag that produces the best surface finish, and finally infuse the resin. Unmold when done and touch up imperfections. Not easy to perfect that, but hey, where's the fun otherwise?

Q: if I wanted to print a foil core, what would it take? The foil is about 1m by 20-30cm by 25mm. It can "easily" be cut at the center, the two sides are obviously mirror images of one another. I believe I want some dihedral, so the 25mm thickness may actually be 10cm in terms of print volume bounding box.
I gather you're in Europe? (I'm in California)

Thank you for awesome feedback !

I will try to answer all of your comments:

1. Yep, I was little bit sad that it didn't generate much interest, I thought that people would be more excited about the possibility of printing large items efficiently, but I fully acknowledge that my software is experimental, VERY hard to use for non-programmers, etc. so that's mostly on me, I need to work on interface to provide nice desktop/web GUI (not that hard at all, much less than actual inner workings)

2. I know about  rudders made by Don Kiesling, although I thought they are foam core - long time ago I created river rudder (small 10cm/4inch height, NACA0012 airfoil) for myself with 3D printed core (CF filled nylon) and fiberglass skins, I use it already for almost 4 years, it took quite a beating many times (dragging across river floor, etc) but it's still holding very well, despite having aluminium shaft (and being much, much lighter than for example stock Nelo 520 river rudder (which has very bad finish with prominent seam on the leading edge).
Of course premium product for sale would be made with carbon fiber shaft (or steel for more durability) and laminated with carbon instead of glass-fabric (spread tow would look amazing), also NACA0012 is a OK choice for hydrofoil (good range of acceptable CD/CL ratio, benign stall characteristic), but there are better airfoils out there designed from inception as hydrofoils with much larger area of laminar flow and better CD/CL ratio over similar AOE (Angle of attack) range.

3. Very good points, for now I stuck with simple hand-laminating of the 3D printed composite cores, but that of course means extensive post-processing to get that awesome mirror finish look - this is for example SUP paddle with carbon-petg filament core and 200g twill carbon fabric skin (2 layers + unidirectional strips on shaft for bending stiffness).
It's similar weight to production SUP carbon paddles (470g) and when I tested the shaft stiffness it was similar as well, but getting the finish to acceptable flatness was a major PITA and sanding/polishing that took by far the most work - if that would be production item, most of the price would need to cover that = not economic.
I experimented with Mylar foils but the classic approach is only fine with surface mostly curved in 2D and very little 3D curvature.
I have one idea in mind, which is basically vacuum forming mylar sheet over pattern to match the curvature perfectly, it should be possible with core 3D printed from high TG material (like PA12CF15, resisting 170C at 0.45MPA pressure), the caveat is that there is a thin line between sheet conforming enough and sheet conforming too well (3D printing imperfections and layer lines imprinted on the sheet itself).
I have no time for that now, but I will test it in the future.
It's not a big deal for boat prototypes though, for example that canoe is just laminated fiberglass with no post-processing and it's fine for its purpose as a functional prototype which you can throw on water, test its behaviour and quickly print/laminate a new one if you find out you need to make any changes.
The final iteration which you are satisfied with can be then painted/sanded/polished and you can take negative mould from that for future production.
Another possible direction is to target the tech for home-builders, current cedar-strip kits are very work intensive and basically mean weeks/months in the garage, with this tech you could sell just gcode (3D printing format) files and people could download/print/laminate them, providing they have good enough and large enough printer (which is my primary focus now, providing such printers).
Ff course I fully get the appeal of traditional wooden boats and I find wood-strip boats simply gorgeous, this would just provide a feasible alternative for quicker building of "modern" style/look boats.

4. I'm familiar with resin infusion and I think I have a pretty good setup for that - this is for example resin infused flat panel and it turned out very nice. The idea with flow channels printed directly in core is brilliant, I already tested how airtight you can make the print (so resin will stay in channels and won't get sucked into hollow spaces in core), it was very promising, certainly doable.
With higher temperature polymers (PA12CF15 or even more exotic PPS/PEEK with TG above 200C) even pre-preg lamination is fully feasible.

5. Yes, I'm located in Europe (right in the middle), but I would be happy to print that airfoil and send it to you just for the shipping costs. Do you have it already designed/tested in XFOIL or similar program ? If not, my custom designer/slicer incorporates airfoils namespace/library where any airfoil shaped surface can be easily defined, airfoil cross-section, planform (tapered/elliptical) etc.



 

This is very very impressive. As a 3d printing enthusiast I can appreciate the challenges and the sheer bulk of work here.
Do you have your slicer somewhere on Github or similar?
Also quite curious about the custom printer you mentioned so casually

Thanks !
Yes, the slicer code is hosted on github, it's in very experimental stage and GUI is lacking, meaning you have to interact with slicer through programming language API, but it's quite easy and there are examples how to do it - chisel

The printer is a high speed delta concept with very rigid frame (spaceframe with fully triangulated sides) and ultra-lightweight motion system (mostly carbon fiber), this is the production machine in "normal" 200x200 build volume size which I will start shipping in 2-3 weeks - spacedelta
The concept is very scalable and my biggest printer is actually more then 2M tall and has a build volume of 0.5x1.5m while still being quite fast, here is a reddit topic about the printer  - needless to say this was the first prototype with spaceframe chassis and is much less refined (frame, motion system) than the smaller delta shown on instagram post - all new printers will be based on the latest design regardless of size.

My latest big object printed on the 0.5x1.5m machine is this very fast fin stabilised paddle craft I call hammerhead, this is bottom/side rendering of the hull (bare hull without side mounted fins & rudder), it's 5.5m tall and only 26cm wide at the widest point (slender ratio of more the 21), deck will be fully flat with just depression for feet, with seat on top of the hull:

This was for example printed in 5 pieces, each 1.1M tall and took less then 3 full days of printing (4kg of filament), here is the printed hull , this is already glued together , with smoothed joints and unidirectional carbon strip already laminated in place.

I still need to laminate the outer skin (due to the very efficient almost tubular cross section of this boat, it's not sandwich construction but just single outer skin with internal 3D printed stiffening rib structure), create/bond deck + fittings (seat, rudder, stabilisation fins) and test it out, unfortunately I don't have time for that now as I'm fully occupied with the 3D printer business, but hopefully later

Helldiver , this is beyond awesome. I own two FLSun deltas, QQSP and SR. The QQSP is converted to direct drive, but after a couple of simple upgrades the SR is my favourite even in the Bowden form. 
I am quite intrigued by your large printer design. Would love to have plans for that if they are available. I have been mulling around 3d printing a Greenland paddle, just not too exited about gluing the parts together. 

What's the best way to follow your work? FB, etc?

Thanks @SpaceSputnik !
My first delta printer was FLSun QQS-Pro, good machine for the money after few upgrades like rod stiffeners, tight tolerance capricorn xs bowden tube and Klipper FW with correct tunning - in its prime, it was printing at 3000mm/s2 acc/  20mm/s scv (~jerk) with good quality. 
My biggest complain about that machine was poor design of rod-ends with very small open metal-to-metal rod-joints, which meant that play developed quite soon, running it with significantly higher acc/jerk than stock didn't help as well.
The printer design I'm about to launch will have dedicated page with order form, technical details, print gallery etc., it's almost ready, professional quality photos were done couple of days ago, in the meantime you can check out my instagram for some info, I also try to be active in the FLSun fb group, helping people with all kind of issues, open-sourcing helpful little bits like those bowden couplings superior to stock pneumatic fittings, etc.

Printing and laminating the paddle will work very well when done, correctly, this is SUP paddle I did, the part I most like is the seamless transition from shaft to paddle:

Shaft/Paddle Core is 3D printed from PA12CF15 filament and hand laminated with 2 layers of ~200gsm twill carbon fabric for blade, and unidirectional carbon fabric strips + 300gsm weaved carbon "sock".
Weight is 470g and stiffness comparable to production paddles, fiber volume fraction is of course a lot worse than resin-infusion/pre-preg, but superior core which is optimally oriented partially makes up for that (compared to foam core).

Gluing that together (I think 4 shaft sections + 2 blade sections) was very easy and not problematic at all, laminating was relatively easy as well, surface finishing was by far the most time consuming and difficult .

If you already nailed down dimensions for your greenland paddle, you can share them with me and I can generate gcode files for your QQS-Pro/SR so you can print them yourself, both printers you own are perfectly suited for that.
Or in the case you are cautious about printing gcode from other people (there are good reasons for that), I can send you models/print-specs and you can generate them yourself with the open-source Chisel tool

Cool, thanks, appreciate it. Haven't gotten too serious about the paddle yet, but I should.
Yeah, qqsp's arms are kinda sloppy. I think I have a bit of play in one. SR certainly removed that problem along with a few others.

Also, check your direct messages.

Hi,

All your work and experiences sound very interesting.
I do myself try to sort something out of 3d printed core with prepreg materials and still have quite an issue with core/laminate interface adherence.
The material I use is a poly lactic compound so not the best companion for epoxy but properties allow a process over 100°C under vacuum.
I have in mind that CF filled PA would be a better candidate but is also twice the price.
What is your experience on this particular issue ?

Thank you for your help.

Hi Benbourg,

I'm quite surprised that PLA is able to withstand anything at 100C - usually, it's TG is way less, 55-60C, unless it's annealed in controlled conditions while being stabilised to retain shape (pieces are often immersed in ground salt or another fine powder), such annealed PLA has very high TG of ~150-170C, eq right in the exotic, expensive high-performance polymer territory (PPS, PEEK, etc.).
PLA can be surface activated with acetone, bonding is significantly easier after that.
Unfilled PA is quite difficult to bond to anything, what helps immensely is flame-treatment (that's true with almost all difficult to bond polymers with low surface energy like PA, PP, POM). Luckily, CF/GF filled PA is quite porous when 3D printed and has a rough "sandpaper" surface, so no special treatment is necessary and epoxy creates good enough "mechanical" bond.
I wish PA (even unfilled one) was only twice the price of PLA, I can buy usable PLA/PLA+ filament for 8e/kg, while PA12 is at least 6-7x more expensive...
PA6 is less expensive but unusable for mechanically loaded 3D printed parts IMO, while it's even stiffer than PA12 when dry, it's very hygroscopic load creeps like crazy (turns into expensive plasticine) once moisture is absorbed - You can probably work around that in 3D printed core by airtight sealing of the core by epoxy/cloth skins, but I wouldn't want to rely on that.

Thanks for your quick reply.
Still very interesting information.
Regarding the material I use, this is not pla exactly but some derivate which has a higher tg. Thermomechanical behaviour is maintained at 115°C continuously and indicated 160°C peak.
This is why it is more expensive than simple pla but still easy to print. Additionally, I use the carbon filled type (10%) which make the parts a bit lighter for similar characteristics. I may give you the reference in PM. I also anneal my parts also to avoid differential shrinkage (still quite an issue)
Regarding PA, I only tried to print pa6 with glass and carbon but did not make the process through laminating yet.

As you recommend, I will try acetone activation on my next parts.
As I mentioned too, I would be interested in tips regarding that shrinkage. I think I should anneal at higher temperature to keep next step shrinkage at a lower level.

Benbourg

I'm very interested in the PLA based material you mentioned, so I would be glad for reference in PM.
Shrinkage is usually much better in fiber filled plastics, I'm mostly using 15% CF loaded PA12 and shrinkage is non issue (and there is no need/advantage of annealing that material).
When annealing PLA, it's usually best to keep temperature relatively low and make up for that by long cycle time, but I don't have much practical experience in that, since all my 3D printed PLA cores were always laminated either by hand or by vacuum-bagging/resin-infusion, never prepreg with raised cure temperature.
Of course exothermic reaction of epoxy resin raises the temperature as well even without externally heating the part, but that was never problem for PLA.