Sail Rocket 2 1:10.5 scale

Sail Rocket 2 set the current water speed record for wind powered craft in 2012 of 68.33kts peak and 65.45kts over 500m.   Paul Larsen was inspired by the design of Bernard Smith in 1963 to build a sailing craft like an aeroplane  that “comes down to grab the water”.  Bernard thought the main application of such a device was a submarine hunter as it could travel silently long distances at 40+kts without any need of fuel.

It essentially is an Ekranoplane that flies in ground effect just above water but powered only by the wind.   Unlike other sailing craft the stronger the wind the more stable it becomes…up until it breaks.  The colour was inspired by the Bell X-1 (first aircraft to break the sound barrier) as this vehicle was designed to smash through the 50kt barrier which existed at the time.

Short video on Sail Rocket 2, Paul Larsen on the record setting run and the build so far 

I was planning on a 1/5 scale model as Paul Larsen did prior to making the full scale version but decided to make a 10.5 scale to work out the basics of the design.  Why 10.5?, this is the absolute biggest scale that can 3D print the sail aerofoil in one piece on my Prusa i3 MK3S.  

You can view and rotate/explode the model here (note it takes a few minutes to load) which may help explain how it all comes together.    Use the explode tool or the animation to pull model apart.   You can drill down into the components to see where they go but please note all components are listed here including the old outdated ones https://a360.co/40eg5Ej 

Edit: Major revision: Apr 13th.   The original design tried to make two giant winch servo's built for ¼ scale sailing boats fit inside the fuselage.    This in turn led to lot of non ideal compromises and was difficult to print. (i.e. the Cockpit was two pieces with holes through it, the sheeting lines ran through the electronic bay lid, the lid didn’t have full length splash seal etc etc) It wasn't intended as a finished product as was just roughing out design for the larger 1/5 scale.  However I wasn't happy posting something with so many issues so I redesigned the fuselage to suit standard sized (40mmx20mmx40mm) winch servo's and fixed a bunch of errors found during the first build.  Also got rid of the S curve flap control to replace with \ micro servo (or linkage system).   The result is a much cleaner and stronger print and far more functional.   Way better for a scale model print at the very least.

Lots of words below but build video here to show how it comes together

Despite living in the third windiest city in the world there hasn't been a breath of wind here for weeks and no sign of any coming so can't get this bad boy out on the water….

There is also the additional of a compression system which puts the entire fuselage in compression which greatly increases strength.  This is important especially at nose and tail sections where the front and rear float attach as the whole model is around 1160mm long but standard carbon rods come in 1000mm sections.

Whilst I have finished the design and printing it may take a few days to upload and document all the changes so if you are making/or planning on making this, check back often.

Anything that has been changed or added now has a prefix V2.  For the fuselage, V2 parts are totally NOT compatible with the previous parts.    The foil fuselage(2) parts print horizontally but the rest of the fuselage prints vertically. Tail bits can print vertically or horizontally.  The main fuselage parts are designed to be printed in a certain orientation without supports (see the screen shot).  I have included print files for all the major components for i3 MK3S for PETG.  The print files all use the fastest possible settings i.e. 0.3mm layer height draft

Changes as follows, starting from the nose, working backwards:-

V2 Nose - now one part, 4 embedded nuts and includes holes for compression lines

V2 Front servo bay and lid, integral servo mount, better connecting bolt layout

V2 Cockpit, now one part.

V2 Electronic bay, clean and versatile layout and full lip for splash protection. 4ea 20x3x5mm rare earth magnets latch to Lid with same amount of magnets.   Addition of antenna hole.   A sheet line deflection mechanism bolts to the front which integrates into cockpit.  This routes the sheeting lines so can remove the lid easily.  (The alternative is to mount the winch servo's upside down with spacer plates).  There are also two sort of hooks at the rear, the intention of these is to hold rubber bands that will keep the sheeting lines in tension when lowering the sail for transport.

V2 Foil body bottom, no flap hole, simplified capstan/bollard system.  press fit nuts to bolt to Ebay, Made as long as can print on Prusa i3 MK3S horizontally.   The Foil body top now bolts on from underneath.

V2 Tail, this is where the compression mech sits.   This works as follows.  Two 0.6mm piano wires run through the 3mm carbon tube from the nose to the compressor clamp.  (Note you have to splice the piano wires inside the carbon tube as they typically come in 80cm lengths…else head to your local piano tuner to get longer lengths)  possibly use guitar string instead.). From the compressor clamp a M3 bolt runs to the tail tip.  Tightening the bolt puts the piano wire (and bolt) in tension, thus compressing the entire model and mashing all the print layers together.  (This is a technique which really works I use for making 3D printed windsurf fins which vastly increases strength..see my other models).  Totally can skip this whole thing if don’t need the extra strength.

V2 Beam attachment, larger hole for running micro servo wires, slight change in sheeting hole arrangement, fixed hole size.

V2 Ball attachment, clean design, now just one hole for micro servo (or linkage)

V2 Wing Base Fillet, corrected errors where external edge didn't join in places.  Extended front strut which now interfaces to the V2 Curve fillet

V2 Curve Fillet (Solid print), now has micro servo bay for flap, recess for flap horn (so no need to cut Wing Ext), wire channel and connects to Wing base fillet with additional carbon tube (front, where the anti torque lines attach).  Designed for Emax ES 9051 water proof servo which fits.

V2 Wing Extension, remove errors where external edge didn't connect at places.  

V2 Wing tip, remove sheet out stuff and extend holes for carbon tubes and flap rod

Since I have ordered but not received standard winch servo's I made some dummy servo's to test mech.  Some that just want to make a scale model may not want to buy proper servos. 

Loose the original anti torque bar.  The real machine just uses ropes/wires which allows the sail to be lowered.

V2 rudder horn, this is slightly tapered to give clearance for the lid.   Use a standard servo horn filed down or shorter as stronger alternative

V2 Leeward Float support.  This now integrates the float strut and anti torquer stuff as was a weak point.

V2 Anti torque.  The Anti torque mech needs to keep the Wing extension, fillet and wing tip etc at 6 deg toe in to foil as far as real machine/Bernard Smith design.    Have made adjustable as model scale may be slightly different.   In order to lower the sail, the anti torque/mech needs to be flexible.   I went with 1mm stainless rope (since I had lots lying around).   The wire is held firmly in place inside the 3mm carbon tube with small piece 2mm x 0.45mm WT Al tube ( a drop of super glue to be sure).  A knot in the stainless rope is simpler but you loose a little of the carbon strength inside the curve fillet. Small tenth of 3.5mm brass tube stops the carbon splitting.    Wire then goes to the wire holders.  The adjustable screws need to be as short as possible so don't interfere with sail lowering.    Can adjust basic wire position by pulling the wire through wire holders that lock in place once in tension.

V2 wing support.  In order to quickly lower the sail I went with a 4mm clevis system that locks onto the wing support.  The main support is a 4mm stainless rod with M4 thread on one end that I salvaged from an old printer.  It fits onto the mast holder snugly and can adjust length with spacers inside.

To follow, I noticed the struts are for the 30% larger floats.   Need to adjust slightly for the scale floats but if I didn't tell you, you probably won't notice.

I noticed after reading Paul's Blog and looking at videos of real machine in detail the rear float includes a retractable skeg to assist getting machine moving initially.  However, I couldn't see this on any models so it may not be necessary.  If needed in future I plan to just duplicate the front float and include a locked rudder.

Included DXF files of Aluminium foil supports and Stainless Foil shape before bending.  

Note: due to number of components, highly likely I have forgotten to upload something so if appears anything missing, please let me know.

Disclaimer: This is not a sailing boat for lazy Sunday’s in your local pond.   It is a high wind specialised machine that only sails in one direction. Judging by other models, if it goes at all it is likely to go very fast and will be quite dangerous to anything in its path.   Please use common sense and don’t hold me accountable for any damage you may cause.  It sails a Starboard tack (wind coming from right) to match the real machine and the majority of speed sailing spots in Southern hemisphere: Walvis Bay (where the record was set), Luderitz, Lilacs, Chook Pond, Rsoles, Pink Lake etc

Update Apr 22.  Added line adjusters to change length of tie down string and sheeting strings easily.  Seems to work well with the 9Kg fishing line I am using.  Feed the fishing line in and out through all but one of the holes, create a loop then back through last hole and tie a surgeons knot.   Removed unneeded hole in tail bottom (V2.2), modified sheet in slightly (V6)

Apr 24: added Front servo bay + lid front tie down screw (50mm M3) to lock lid tight. V2.2

Apr 25: middle wing sheet control outside tie down hole is in wrong position.  Updated but seems need to shift the whole thing 1mm forwards…this version better than before but could be better.

Apr 26: Middle wing sheet control had sheeting hole in wrong spot, corrected with V2.3. Thinned mast holder slightly to give better clearance.  Correct some weird offsets in model..still some remaining..doesn't effect print

Apr 30: Added storage/transport clamp that holds the wing securely in lowered position so can fit in vehicle/hang out of the way.  It uses the wing support with its clevis to lock the wing and so you don't loose it.   It is 3 parts Beam, Front and rear clamp held together with 4e 10mm M3 bolt

May 27 update.  Still waiting for wind.  Contact cemented 10mm polystyrene on bottom of each float to give more buoyancy.   Painted edges black so not noticeable.  Seems to give right amount of flotation.     Added GoPro mount to fit between the two cockpits.  Made this is two pieces but turned out flexible enough to make in one piece.  Camera leash goes to the mast/beam tie down point.   Took craft to Wallis Bay look alike (Lake Brown in honor of Steve ‘Reg’ Brown) for maiden voyage in massive 7knots.   Unfortunately water depth not enough and rudder bouncing on the hard salt bottom.

End update, original post follows:-

Only realised a few days ago there was competition so uploading the design as is.  Currently has a few minor errors but main issue is all the servo's I used in this design are big heavy duty ones I have for the 1/5 scale.  The winch servo's I have are massive...they still just fit in the 10.5 scale but at last minute I realised some errors in design that doesn't allow cockpit lid to fit with these big servos installed.   You could totally use much lighter/smaller servos and just redesign the servo mounts.  Hopefully for obvious reasons you need to use water proof servos unless you are just making a display model.

I scaled the floats up 30% to increase the bouyancy but have included scale floats.  The reason for this is the wing with the carbon mast and rods weighs 637g but the scale floats each only have displacement of 421mL. The scaled floats have 1.15L displacement so far more likely to float.  I filled them with expanding foam which makes them strong and waterproof.  The big floats I put webbing in but ended cutting all that out using snips.

The sail support uses 1.5mm x 20mm Al flat bar cut to 10mm wide except near the mast.  Bend it to suit the skin.  The front and rear stiffeners holding the foil are made from 3mm x 32mm Al flat bar.   The longitudinal carbon rods through the fuselage and sail are 3.4mm dia but if you print horizontally like I did better to go with 3mm or get yourself a long drill.  The mast is 9mm carbon rod.  The beam is centred around 8mm square carbon.   It is important that the screw from the ball for the mast support goes through the 8mm carbon square for strength.   Easiest is to build from wing back to fuselage as can trim the carbon square to size easier.   

The carbon rods through the fuselage go through the carbon Beam square AND the 3mm Al front and back stiffeners.   So install the Al stiffeners and carbon Beam square then drill from the back up the fuselage holes to mark position, then remove, drill, then replace.  This is where all the load is on the model so important everything intersects nicely.   You may have to file a little on the top of the front Al stiffener where the sheet in and out control strings go as tolerance is very tight.

The curved section of wing was too tricky to do in vase mode so is printed normally but with 5% or less infill.  

The mast connects to the beam using a standard ball/eye joint you can get at any hobby shop. The eye bolts to the mast using the mast plug which you jam/glue into the mast.   You need to Dremel a little section of the wing fillet base directly above the solid curved section to enable to fit to the ball.

The foil section I made with 1.5mm stainless steel.   Fusion 360 Hobby mode won't allow PDF drawing of flattened profile so I've included the STL which hopefully helps.  The section in the water is 105mm long.   The foil leading edge just like the real thing is razor sharp and the trailing edge is 90deg.  This is because (in theory) at speed the foil is supercavitating and needs to travel like a bullet.   The foil eye threaded bar and 2mm piano wire pivot are sandwiched in the middle foil skin by the foil and the foil sandwich which I used 3mm Al and bolted together with 3mm x 10mm nuts and bolts

The local hobby shop was selling proper metal model boat rudders for $2 so I used that. 

There are some errors in this design like blocked holes for push rods but nothing a drill/dremel can't fix.  This was a first pass build.   I may upload a corrected version but probably will work on the larger scale as priority.   The flap control is a bit dodgy and needs serious silicon lube to work.  Probably replace with a micro servo or linkage system or just loose it.

I used kite string in this build but would use fishing line in future as kite string goes all weird when trying to thread through the holes.

If you are interested in more on Sail Rocket you can check Paul Larsen's YouTube channel on the Sail Rocket https://www.youtube.com/@paularsen1 or the main SailRocket 2 website http://www.sailrocket.com

I designed the model importing drawings from Sail Rocket 2 website and Wired magazine article into Fusion 360 https://www.wired.com/2013/01/ff-paul-larsen-sailrocket/  Knowing that the craft had to fit inside a 40 ft container it was 12.2m long.   For small details did the same using photos like cross sections of wing from the sail rocket site.   Since all the views have the shape of the sail at an angle had to project the shape back to a 30 deg plane to get the correct profile.  Then used airfoiltools.com using a Clark Y foil at 11.7% which was the closest to the about 10% chord width that Sailrocket seems to use.  Then lofted the foil using the projected profile as guide rails.  The curved wing was hardest to get to work in vase mode but the basic technique came from based on Tom Stanton's video below on his sub channel.    The sail and beam sections are printed in vase mode with 0.2mm layer height.   

The Antitorque bar is made from 3mmx10mm Al bar and holds the Wing fillet, Wing ext etc in place and opposes all the sheet in/out forces.  Can probably use string instead.   Note the original wing fillet base upload doesn't include the strengthening strut to hold the Al bar so use the Wing fillet base2.  You need to cut /dremel a tiny slot to slide the Al in. In the real machine it appears to have two of these, one going to nose of base fillet and one to about mid way, luckily there is already a carbon strut there to attach this too.  Note I didn't realise the importance of this in my original print so is not shown in the photos.  Without it the wing fillet will pivot around the mast freely and also be pulled by the sheet out strings.   In the real machine there is also a pole sticking 90 from mast directly forwards.  Not sure if this is also to oppose sheeting forces or is used to sheet out.

I've included working Gcode for PETG for a Prusa i3 MK3S for all sections that need to be printed in vase mode.  This is fastest I could get to print correctly but you may be able to tweak settings further.  Note for the lower wing and middle wing bottom you have to disable the skirt otherwise print will be outside allowed print area (and it takes some time to align it to fit).   If you are slicing yourself you can see big chunks missing in wing sections if the vase settings are wrong so it should like screenshot attached.

Note pretty much all the STL's are straight from design so require rotating down to the print bed before slicing as didn't have time to do this. 

This is meant to be FPV which is why there is carbon rod sticking out of mast tip where the camera will mount but this isn't done yet.  Another small error I only found after printing is the angle of the ball support doesn't allow the wing to fully rotate horizontally for towing/storage.

After printing I noticed that the original ball holder won't allow the mast to rotate fully horizontal for ‘tow mode’ or storage/transportation, so have added a quick fix ball mount 2 that allows wing to fully rotate to horizontal.

I added an improved sheet out control that is more like the real deal and stops having to route the string through the wing extension. There is modified ‘Lower sheet control’ and ‘Middle sheet control’ that pin to the Lower and middle sails with the 3.4mm carbon rods near the trailing edge.  There is a small hole on each at the leading edge of these ‘controls’ that allows for a 2mm carbon tube and piano wire inner ending in an eyelet.  The length of these carbon rods should place the eyelet where the sheet out strings go an identical distance from the mast as the carbon rod where the sheet in strings are tied.    Hopefully the attached picture explains it better.  It also makes tying the sheet in strings to the sail much easier for the middle sail as don't have to drill the sail.

I named my craft ‘Spirit of Reg’ after a sailing buddy who suddenly and unexpectedly died whilst windsurfing just a few weeks ago.