(This paper was presented to 1992 Marine Applications of Composite Materials Conference.  I see that all the graphics and diagrams have fallen out during up-grades of Word versions.  They will be found and re-inserted soon.)





Kurt Hughes, Kurt Hughes Sailing Designs





Composite one off multihulls have tremendous amounts surface area for a given volume.  Consequently considerable time is spent in molding a composite multi.  Builder's attempts at reducing the molding time are usually spent fairing the work later on.  Any strategies that can reduce multihull molding time and keep the quality of the work high are desired.






Mold building is a significant time expense of any composite project.  With a multihull, the problem is particularly acute as it will have twice the number of molds, and a larger total mold surface area than an equivalent sized single hulled boat.

Beginning in the Fall of 1988 I began experiments with possible strategies to mold composite multihulls without molds or with minimal molds.  The resulting project, a 40' trimaran, was built with two of the following strategies and will be launched whenever I get time away from the office. 

Since that time several other multihulls have been built using these ideas.



The fastest possible way to build a one-off multihull would probably be stereolithography.  Stereolithography is where X, Y, and Z-axis lasers, driven directly from the CADD software, beam into a vat of polymer to form a part.  For several reasons however, this is not yet the answer.  One has to keep looking to find faster ways to rapidly build composite multihulls.



Building a composite boat is outside of the nature of materials.  Unlike plywood for example, the amorphous blobs of resin and fabric must be molded to be useful.  One of the things I do as a designer is look for industrial processes or the results of industrial processes that already do what I need. 

One example is a plate glass molding table used for molding large powerboats.  Large boatyards already use the plate glass molding tables to create large fair compos site panels.


Both of my experiments were very successful.  The molding times of the hulls were trivial, less than the time to wax, buff, and pva the mold itself, not counting a couple of diversions.



The first molding system used thin tortured plywood to develop the hull shape.  Next core was vacuum bagged onto the inside.  Finally fabric was bagged both inside and outside.  While not true molding, it did rapidly define a proper fair surface, and the composites "froze" that shape.



The second system used Cylinder Molding as a female mold.  First, thin sheets of 3mm plywood had a smooth molding surface put on them.  These sheets were then vacuum bagged together using the Cylinder Molded technology.  The plywood sheets with smoothed surfaces inside were bagged to another layer, giving more thickness and strength.  The two sides of the mold-to-be were wired together along the keel and stem with copper wire.  The keel where the two mold sides were sewn together got a cove of structural "bog" placed in it.  Bulkheads were then glued to the mold on the outside.  The result was a rather traditional 40' female mold in very little time.



The final strategy will use formica or thick mylar surface spanning former stations on perhaps 16" centers.  The fabric will first be laid up flat on the formica or mylar, then the whole thing moved onto the forming stations.  The port and starboard halves will be attached using traditional stitch and glue keels. 

I did see multihull hulls and panels being built in a similar way to this in Britain on my last trip there.

Those hulls were laid up with no rocker, like a uniform U shaped formica section.  To achieve rocker the appropriate bits were cut away from the forward and aft ends of the keel area.  To force the lot to come together, transverse slits or cuts were made all along the bottom 15% of so of the hull.  The result was a totally unacceptable level of fairness and repeatedly chopped full-length fibers.  Of course it is well known that full length properly oriented fibers are the keys to strength and stiffness.

I see perimeter stiffness as the key to making this work.  The keel edges must be kept from buckling as they are sewn together. 






The main hull on the project trimaran was built using this strategy.  It is a system using no mold, unlike the next two, which are rapidly built molds.  Since there was only one main hull but two amas used in the project, it did make more sense to use this system for a main hull. 

In this process a thin plywood hull is built using traditional stitch and glue systems.  One stitch and glue system is Cylinder Molding, where two thick nesses of thin plywood are vacuum bagged together.  The resulting panels are full length and have a transverse curve in them.  The other system that can be used is StressformTM or tortured plywood.  In both systems the identical port and starboard panels are cut to shape, wired together, spread out and an epoxy-roving glass and filler keel is poured in.  After the keel cures, the spread sheers are brought back in to final shape with a deck flange.  This will be the final hull shape.  Like a large three-dimensional batten it will be fair and smooth.  A composite stem is next glassed in. 

The hull panels are built using traditional CM or Stressform TM techniques.  Refer to the Cylinder Mold Construction for these techniques.

First verify that there are no holes in the panels. See that the scarphs do not have holes or gaps in them.  Any nail holes or rough scarphs must be filled with epoxy and bog.    Sand these filled areas smooth.  Any glossy epoxy found on the panel must also be sanded.

Naturally the keel area has many holes.  These are dealt with during the keel pour.

The keel pour will be done nearly the same way as a stitch and glue hull using no core.  Since these thin ply hulls will usually be much larger hulls than the ply thickness would suggest, the keel fabric must be carried farther onto the hull than one might expect.  That puts the turn of the bilge out farther.  The keel fabric edge transition must also be more gradual than with a thicker plywood hull.  Both these items are covered in the plans or revision sets.   The keel fabric must be placed with peel ply on top.  The peel ply surface and any bumps must be roughened and knocked back with a small grinder.


The sheer timber will be installed exactly as on a traditional stitch and glue hull.


The deck flange will be built and installed exactly as on a traditional stitch and glue hull.


Typically thin plywood hulls using core will not have stringers.


I am assuming that the core will be bagged into the inside of the hull.  While core could be bagged to the outside, the fairing job would be much bigger and since contour core will probably always be used, the scrimm would be on the wrong side.

Once the thin plywood hull is compounded to exactly the right shape, the core materials will be cut to shape and dry fitted. 

The core to be used must first be cut to proper shapes if necessary, all pieces fitted, oriented and each piece numbered. A dry fit should have very piece to be used in place.  A typical core piece will have the number, orientation and side of the hull noted.

The same steps are done to the fabric.

On my project the main hull plywood was 3 mm Asian luan.  The hull was built using the tortured plywood method.  In fact the folded-up hull was so delicate and fragile that a fan could almost make it flutter. 



I had expected two problems that did not occur.  I worried that the vacuum bagging might distort the folded up hull.  It did seem that the vacuum pressure should press equally on both faces, but I recall suggestions advising a very robust mold for vacuum bagging.  In fact there was no distortion at all. 

I also was not sure that a tortured plywood hull this large would have port side exactly match the starboard one.  The largest tortured plywood hull that I knew about was also 40' long but had very low freeboard, unlike the 6' wide panel that I was using.  Outside of the standard barely visible recurve in the topsides, there were no problems controlling the fold up hull, nor in symmetry.  The standard requirements of the opposing panels being exactly the same species and density still apply.

The molding portion of this hull, building a 3mm thick hull, took about 15 man-hours to do. 



On this project I used 1/2" thick contour core balsa, with the Al600 coating.  As the core shear of the balsa is between 235 and 360 psi,(1) I worried that the plywood skin would be vulnerable to rolling shear failure and be the weak link in the layup.  The core shear of the plywood was in fact tested as 713 psi(2) and should not govern. 

The fabric used on this project was Heinsco Rovelock unidirectional E glass.  Each layer was 9-ounce weight and about .009" thick.  It is a very high strength glass using





polymer fibrils rather than "slug tracks" or stitching to bind the strands.  We were rather beta testing the product and did have some problems.  The polymer holding the glass together had some sort of memory and was very difficult to squeege down.  Every square foot of glass layup had perhaps a half dozen little moguls that could not be worked down.  I thought the vacuum pressure would help but I discovered that the fabric arrangement is so dense that air would not permeate through the fabric, not even under vacuum.  I did finally have to use a

dremel tool to take out the hundreds of memory bubbles in the layup.

The layup was a single layer in the 0 degree direction and one in +/-45 degree directions.  My calculations showed that it was overly strong(3) in the +/- 45 directions.  The same layup was used inside, vacuum bagged over the core, and outside, without bagging. 

The two serious concerns that I did have about this method of non molding composite boats turned out to be groundless.  I consider it to be a very rapid and successful way to build composite one off multihulls.




                                     (Figure 2)


In the next experiment I attempted to take the tremendous time and fairness advantages that Cylinder Molding gives and apply them to traditional female molding.  The amas of the 40' trimaran were built this way, and this Spring a 48' foam/glass catamaran on Kauai in Hawaii will be built this way.

Twenty sheets of 3mm plywood were used to make the Cylinder molded female mold.  These were vacuum bagged into a mold two sheets thick.  The first step in this molding process was to smooth finish the faces of the sheets of plywood which would become the inside surface of the mold.  I flow coated epoxy and then sanded it smooth, but I'm sure there are faster ways of doing that job.  A plate glass table to lay the epoxied ply onto might work well, or even some formica might have worked as an inside surface.  This coating and sanding the flat sheets, plus putting scarphs in the ends of each sheet took 14 hours.  Again, with a better facing system, that could be a much lower figure.  I used standard Cylinder Molding technology to build the panels.  Each full length panel took three hours to build in all.  A flat surface profile was cut into each panel and these were wired together along the keel in the regular stitch and glue process.  An epoxy "bog" keel was poured into the wired and spread assembly.  The "bog" keel was hand shaped which meant that it had to be sanded later.  A better choice would have been to form the keel pour with a large diameter pvc pipe or some other smooth round pressed into the bogged keel area.  That sanding, which I should have been smart enough to avoid, took two days of brutal labor to do.  The sheer spread and keel pour took three hours.  A previously built deck flange defined the deck shape of the foldup.



The scarphs joints every eight feet were filled and sanded if needed.  The female mold shape was kept intact by bogging on external bulkheads every two feet, in addition to the deck flange.  These bulkheads were very rough and took two of us two hours to do.



Once the mold was waxed, buffed and pva was sprayed on, I realized that these release agent steps actually took longer than building the mold itself, ignoring the keel sanding.



This mold stayed joined at the keel.  It could have been cut along the keel for molding where part of the deck is also molded at the same time.  This mold could have been used for a one off or a small production run.



The hull was laid up with a single layer of the 9oz Heinsco unidirectional in the +/-45 and the 0 degree directions in the outer skin.  A core of 1/2" balsa was bagged on while the fabric was still wet.  I did spread a thin 1/8" layer of epoxy bog "gel coat" onto the mold before the glass was laid down.  That proved to be a serious mistake as the impermeable fabric trapped air bubbles in the bog.  Again, the vacuum would not pull the air out through the fabric.  While the fibers were not affected by this, the entire bog layer had to be ground off later.  With this fabric it would have been better to have laid the fabric directly on the mold.

After the outer layer and core cured, the scrimm was removed, and an inner layer of an identical layup was vacuum bagged on.  The nice heavily rounded sheers that we all like were outside of the ability of this mold.  The sheers were laid up separately.  These were half length, laid up in a half round sonotube.  These parts had the 0 degree glass only on them.  They were wired to the hulls and +/-45 glass laid inside partly bonded them to the hulls.  Finally a premolded deck was set in place.  The deck, sheers, and hull were finally and fully joined with a +/-45 layup tying them all together.



These 38' long amas ended up weighing 312 lb and 318lb each, including connective and rigging reinforcements, and a solid foam crash bow.  I consider these amas very robust.



                                       (Figure 3)


The final rapid building strategy involves Cylinder Molding again, but using no plywood at all for boat nor mold surface.  At the time of this writing I have not had time to build a hull using this system.  When I do, it will again be a female mold.  I will use computer plotted mold former stations.  Each plywood station will have a slightly different concave curve.  Into the series of curves stations a surface, probably 12mil mylar, will be placed.  Before the mylar is placed on the mold stations, the fabric will be wetted out and squeeged on a flat surface.  As the profiles of a hull must

mirror each other, the order of the station formers will be reversed to mold the opposite hand panels.

After these all epoxy/glass panels are cured, the panels will be trimmed into reflected profiles and wired together.  Again, the sheers will be spread, a keel poured, and the whole assembly

brought together after it cures, with a deck flange.  Also again, the core and inner layer of fabric will be vacuum bagged onto the inside of the hull.  After the hull is "frozen" with the core and inner fabric, the keel wires will be removed.  The glass edges will be ground back if needed, and the outside of the keel sheathed in a final layer of fabric.



I see possible difficulties with this system if the glass fabric layup is too thin.  It would possibly then buckle before compounding as much as needed to develop a proper hull shape.  The best prevention for that problem is some sort of stiffened perimeter.  This strategy will probably require a couple of hulls to be built using it to sufficiently understand what is needed.






Each of these rapid molding systems has certain constraints and disadvantges.  Also, they all share certain constraints.  None of these systems will work with hulls that need to be fuller than 10/1 on the waterline, or about 6/1 on the deck.  Since few reasonable modern multihull have hulls fuller than ten to one, that is not a serious problem. 


None of these systems will work easily with complex hulls that need flares or other complications in section.


The first system, with the thin plywood as part of the hull, probably should have carbon fiber instead of E glass in the

high load 0 degree direction as the plywood and carbon have similar elongations to failure.  Laminate testing should be done to determine the best layups to resist multihull loads.  As this system has no real molds, the outside glass layer will need to be filled and smoothed.  Also, this system only really applies to one off multihulls.






Each of these different rapid molding systems shows promise in reducing the number of hours spent mold building for multihulls. The first system needs more study in laminate analysis for the optimum layups.

The other two systems can probably be improved by input from experienced composite multihull builders.

The last yet untried system naturally needs some hulls built using it to see if it is as promising as it appears to be.






(1)        Baltek Corp. Reference Book  "Mechanical Properties of Belcobalsa R" P-2


(2)        Comtex Development Corp. 2/16/'89 "USCG Certification Testing"


(3)            Reichard, Ronnal P., International Conference Marine Applications of Composite Materials, Structural Design of Multihull Sailboats, 1986, P-1-P-9.