Mast Support Beam Replacement

This page was last updated on 26 November 2003.

IM009393.JPG (151135 bytes)Given my worries about the mast support in general, and the repairs required to the deck, it was only natural that I decided to inspect--and remove--the existing mast support beam, which was original to the boat.  Removal involved removing several pieces of wooden trim, which were screwed in place, and then the glued-on  fabric over the beam.  It was obvious that there had been movement in the center portion, especially in the forward add-on piece.  There was a gap between the beam and the bottom skin, and there was still room to hike the thing upwards.  The beam itself, however, was still in decent shape--the same as before, so I wasn't sure initially how I was going to handle is:  reinforce, or build new.  My initial thoughts were to reinforce the beam with some metal plates, as this would avoid the need to delve too deeply into the finished surfaces surrounding the beam.  On further inspection, though, it looked like the beam would come out easily, if I could removed two small sections of tabbing and cut the heads off a series of bolts passing through the forward bulkhead.  

IM009396.JPG (155545 bytes)

The next day, I attacked the beam removal.  It all started out well enough.  First, I installed a spiral bit in my Dremel tool, and easily cut through the tabbing on each end of the beam.  Then, with a cutting wheel in my 4-1/2" grinder, I buzzed off the nuts securing the bolts running through the beam; the heads were hidden in the vee berth behind the Formica bulkhead covering, so I didn't want to try and re move them traditionally, as I hoped to not disrupt the Formica.  Cutting the nuts off released the add-on section of the beam in the forward part of the passageway to the vee berth, so I removed that section.   Then I made attempts to pry the beam out of position, which should have been easy.  

IM009411.JPG (171844 bytes)No luck--it barely budged, though it moved enough that it seemed there was nothing serious holding it  in position.  Thinking logically that the wood was still gripping the4 or 5 bolts passing through the beam (even with the nuts removed), I used a drill to ream the wood out around the bolts, which should have released the beam.  Still, I couldn't get the beam to come out.  I tried several approaches to no avail, and the job rapidly became infuriating.  There was no reason why the beam shouldn't have come out, or so I thought.  After all, I had managed to pry it a short distance away from the bulkhead on each side, so I figured there must not be anything really holding it.  My mood was made no better by an unfortunate lack of control with my pry bar at one point; as nice as can be, the wild end caught the starboard head porthole and shattered a hole in the glass in the blink of an eye.  I became somewhat more unhappy at this point, and might have spoken a few harsh words.  I can't remember; perhaps you should ask my neighbors what they heard.

IM009407.JPG (152927 bytes)After much struggle, accompanied by a not-so-nice-something-is-being-ripped-apart sound, I finally determined that there were two secret screws (driven from the forward side of the bulkhead back into the beam, so they were invisible behind the Formica) holding the beam in place.  The ripping sound I had heard was actually the plywood bulkhead being torn apart as I tried, unbeknownst to me, to pull the screw the wrong way through the material.  I was less than pleased at this new development, but at least now it was a cinch to cut the screws with my Sawz-All and release the beam.  It came out easily after that.  What a stupid hour I had spent.  But the beam was out, never to return:  I planned to build a new, laminated beam to go in its place.

IM009408.JPG (181387 bytes)

IM009409.JPG (136273 bytes)After cleaning up the mess in the cabin--dust, splintered wood, and the like--and removing the broken port, I felt much better, as if things had returned to some semblance of control.  Mass destruction of a finished boat is less than my favorite thing.  With the beam now on my bench, I inspected it and found that it was in generally good condition, with only minor--and expected--checks.  Still, the sawn beam is inherently prone to some flattening of the arch when under load, so I felt good about my decision to build new.  I put the beam aside for a week or so while I concentrated on completing the deck-repair portion of the mast step project.

IM009412.JPG (172527 bytes)

Building a New Beam
After concentrating on the mast step reinforcement and deck repairs for a number of days, as well as waiting for delivery of some lumber with which to fabricate the new beam, I eventually turned my attention to the issue of constructing a new beam.  I chose to cold mold the new beam out of epoxy-laminated strips of white oak, which would form a very strong beam resistant to the forces placed upon it by the mast compression.

IM009489.JPG (165809 bytes)The white oak lumber I ordered came in the rough state, with a nominal 4/4 dimension.  I received two boards, each about 9' in length and between 7" and 9" wide.  I thought I could get all the material I needed out of the larger of the two boards, so I began with that.  The first step was to plane the board smooth on each side, which reduced its overall thickness to about 3/4".  Knowing how much rough lumber I would be using in the future as I complete projects on Glissando and also for the construction of the Triton Daysailor, I decided to invest in a small thickness planer rather than fight with other means of smoothing the boards.  I purchased a 12" Delta planer and, in the days before beginning work on the mast beam, I experimented with the tool to gain some knowledge of its use.

Planing was uneventful.  Given the hardness of the oak,  and the novelty of the planer to me, I was conservative in how much material I removed with each pass, so it took a number of passes before the wood was smooth, flat, and at the proper thickness.  I was pleased with the performance of the planer on the tough white oak.

Next, I crosscut the board to make it easier to handle, and to reduce its length to that which I wanted to use for the new beam.  After measuring the existing beam, and allowing plenty of extra, I cut the board off at 6' in length.  Conveniently, this left more than enough length in the board to use the short cutoff as the stock for the additional reinforcing section of the beam that extends forward into the vee berth from the main beam.

With that done, I turned to the next task:  ripping the board into strips of the proper width and thickness.  The edges of the board were curved, so I set up a long straightedge and trimmed one edge with a circular saw, being careful to remove only the minimum amount of wood possible.  With one straight edge, I could begin ripping the board into strips of the proper width for the new mast beam.  The old beam was just about 2" wide, so I decided to go just over 2" for the new one.  At that width, I managed to rip four strips out of the original board, with only a narrow piece of waste left over.

Now I had to resaw each strip to reduce the thickness of the boards that I would use for the lamination.  I had originally hoped to use boards of about 1/4" in thickness, which would have given me 12 laminations in the 3" height of the beam and all but eliminated any appreciable springback after gluing.  However, the original oak board was not thick enough to get three 1/4" pieces from, and I didn't wish to waste so much material in order to only get two boards.  Instead, I decided to resaw each strip in half, which gave me new boards of about 3/8" in thickness with no  material left over.  I used the table saw to resaw the boards, and ended up with 8 boards, giving me a total height of the desired 3".  Using a formula for roughly calculating springback, I determined that with 8 laminations, I could be looking at nearly 1/4" of potential springback after the wood was glued.  I would have to take this into account when constructing the mold.

With the 8 main pieces milled to the proper dimensions, I continued by milling the shorter lengths into similarly-dimensioned pieces for later use as the beam reinforcement as mentioned above.

IM009490.JPG (147060 bytes)

Computing springback in a curved laminate

When a strip of wood, a ply, is bent to a curve and released, it springs back to its original shape. A stack of loose plies will do the same. When these plies are glued together while clamped to a curve, they tend to stay curved, but will straighten out or spring back a little. When building the mold or form for the laminate its common to make the curve a little tighter to compensate for the springback. The amount of springback depends on the number of plies in the laminate, not in the ply material or the ply thickness. A simple formula can help predict the amount of springback you can expect:

y = x/n2

y = the amount of springback
x = the amount of deflection
n = the number of plies

Thus with four plies, the springback (y) will be 1/16" if the amount of deflection (x) is 1". By using plies half the thickness you double the number of plies for a given thickness of laminate and reduce springback four times-in this case to 1/64".

Tip courtesy of West System website.


I decided to laminate the beam on a horizontal mold, rather than a vertical one.  Having read others' experiences with laminating over a saddle-type vertical mold, I thought it would be more convenient and easier to have the whole mold built on a flat platform, with the strips oriented on edge.  This would help keep the edges of the strips aligned, would prevent side slippage, and would prevent epoxy spillout from pouring into the clamps--a common problem with the other method.

I began with a sheet of particle board that I had lying around.  I ripped it in half and secured the two pieces together with screws and 2x4 bracing beneath across the seam, giving me a flat platform of an appropriate width and length.  Having the 2x4s beneath had the added benefit of raising the platform off the bench, which I thought might come in handy should I need to clamp the beam in a downward direction during lamination.  I laid the old beam down on the platform and traced its profile directly on the particle board.  Because the old beam was slightly taller on its aft-facing edge (since the top was beveled to approximate the angle of the cabin trunk), I chose to trace it in that direction, which would give me the tallest beam necessary and replicate the largest arc.  I marked a straight baseline and determined the beam centerline, and made the appropriate marks. With the outline traced, I covered the whole platform in some 6 mil clear plastic, which would protect it from epoxy and would also prevent the new beam from sticking once it was glued up.

lam1.JPG (153308 bytes)To take the potential springback into consideration, I made a new mark at the centerline of the beam arc, but 1/4" higher.  Just beneath this mark, I screwed a short piece of 2x4.  Then, at each end of the beam tracing, I screwed in another short piece of 2x4, but kept those edges at the original marks.  With the center of the new shape 1/4" higher, the theory is that the ends will spring up enough when the material is glued and cured, and will therefore ensure that the beam ends up with approximately the correct shape and arc.  

moldready.JPG (191879 bytes)To ensure a fair curve and the proper placement of additional clamping blocks, I bent one of the oak plies around the three main blocks (center and two ends), and traced its arc on the plastic.  Then, I installed further blocking beneath this line, with blocks located every couple inches.  To complete the mold, I added small pieces of plastic sheeting over the end of each 2x4 block to prevent the beam lamination from sticking irrevocably to the mold.  I added a final block on each end well outside the actual end of the beam shape; the arc must continue past the actual stopping point to ensure  that the curve is fair and smooth.

dryfit1.JPG (174629 bytes)Not wanting any unpleasant surprises while the wood strips were covered in gooey epoxy, I decided to perform a dry run of the lamination to see if any issues arose.  I clamped all eight strips securely on the centerline, and began pulling them in towards the mold on each side.  I found that it was helpful to place a squeeze clamp (such as a Quick Grip) at the end location and pull the plies in from the end, clamping the sections closer to the centerline as I went.  I soon realized that I would need some sort of hold-down clamping setup to keep the plies in line, as they tended to slip in a vertical direction even with no epoxy in place yet.  To that end, I clamped a couple scraps of wood across the plies and pulled them tightly down, which nicely aligned the plies.  This was the sort of issue I had hoped to uncover during the dry fit.

Another problem that surfaced was that, when placed under clamping pressure, one or two of the 2x4s split around the screw holes--a common problem with junk timber, especially older ones that were dry.  It was a snap to replace the broken ones, but I was glad to have discovered the weakness before the actual glueup.

clampsready.JPG (164913 bytes)Happy that it looked like all was going to work fine, I unclamped the dry-fit beam and made final preparations for the actual glue-up.  I covered one of my benches in spare plastic, so that I could keep the surface clean while wetting out and spreading the epoxy on the wood plies, and laid out my clamps and other equipment near the mold location on the other free-standing bench.  I gathered extra clamps and set them nearby, just in case, and collected spare pieces of plastic and scrap wood, along with my deadblow mallet, trying to prepare for any contingency.  While there is always more time than expected to work with the epoxy when used in this sort of method, I wanted to take no chances.

readyforwetout.JPG (162633 bytes)After laying out the eight plies in the order I wanted them (I chose a smooth, nice face for the ply that would end up forming the bottom--and most visible--portion of the new beam), the first step was to wet out one side of each ply with unthickened epoxy.  The purpose of this was to ensure that all areas of every bonding surface would be coated with epoxy, but without creating excess waste from the inevitable squeeze out.  I spread the epoxy on the upturned side of each ply, ensuring that they were evenly wet out.

With that step complete, I took one last look around to ensure I was ready for the main lamination, and went ahead and mixed up a batch of epoxy resin thickened to an adhesive texture with cabosil.  I left the mix quite loose to allow it to spread easily and flow well once pressure was applied during clamping.  The idea in laminated glue-ups is to provide an even, consistent layer of adhesive between each strip, without squeezing out all the glue and without creating thick pockets of adhesive that are difficult to even out under pressure.  I thought I mixed more than enough, but of course halfway through the process I ran out and had to mix up another batch.  Beginning with the lowest layer, I spread the adhesive on, then stacked the next ply on top--with the resin-wetted side facing the just-applied thickened epoxy--and then repeated the process till all eight boards were stacked on the bench, oozing epoxy.

lam3.JPG (178268 bytes)Then, I picked up the stack and placed it on the mold, aligning the ends of the boards and lining the centerline up with the marked line on the mold.  I installed one clamp in the center, fairly tightly, and then placed one of my squeeze clamps at each end to begin pulling the plies into the mold.  I drew it in a bit, then installed more clamps from the center outward towards the end, gradually pulling the whole mess is towards the mold blocks.  I didn't worry at this stage about unevenness in the top edges of the boards; I'd take care of that later.

lam4.JPG (166637 bytes)Once all the main clamps were on, I used my deadblow to  roughly even out the tops of the strips--wearing safety glasses in case the hammer blows splattered epoxy spillout everywhere.  Then, when I was satisfied with the pressure on all the clamps, I installed a pair of hold-downs, one at each end, which I had practiced with during the dry fit.  Each was made from a piece of scrap ash, which I rested on top of the laminated plies with a piece of plastic to prevent epoxy adhesion, and I installed a clamp on each end to pull the scraps down, which had the effect of evening out the plies.

lam5.JPG (166974 bytes)     lam10close.JPG (167933 bytes)

lam8.JPG (161775 bytes)I spent some time ensuring that all the clamps were properly situated, and added a third flattening clamp setup on the centerline, where there was some tendency for the beam to spring up.  Then I used a plastic squeeze to clean up as much of the squeezeout from the top of the beam and around the edges as possible, to make sanding and surfacing easier later (cured epoxy and cabosil makes for very difficult sanding).  Then, just to be sure, I laid the old beam on top of the clamps and compared the curves.  I had done this several times earlier in the process, of course, but it never hurts to check again.  The curves matched closely, allowing for the additional 1/4" of height on the new beam to account for springback.

lam11close.JPG (174619 bytes)     lamendclose.JPG (173651 bytes)

lamabove.JPG (152308 bytes)
     oldandnew1.JPG (161436 bytes)

smallbeam1.jpg (51124 bytes)I left the beam to sit in the mold and clamps for nearly 36 hours, to ensure that the epoxy would reach an advanced stage of cure, before continuing.  Then, I unclamped it and, with a small amount of effort, broke it free from the plastic-covered mold.  It took some minor persuasion in one or two spots where the epoxy had oozed out more heavily, but in moments it snapped free, leaving behind a nearly clean mold ready for another use, which I shortly put to work:  using the same techniques described above, I laminated up the smaller forward section of the beam, using the offcuts of the same material.  I left the new piece to cure in the mold for a day or two, and turned my attention once more to the large beam.

The surface of the beam that had been facing upward was relatively free of epoxy, but there was some slight unevenness in the plies.  The other side, the one that had been firmly pressed into the plastic, was covered with a thin, smooth, even layer of epoxy spillout, with additional minor unevenness to match the other side.  The beam overall was straight, true, and free from twists or other problems--not that I was worried, but it was still nice to see a new project turn out as expected.  Springback was minimal when I unclamped the piece--perhaps 1/8", but certainly less than the projected 1/4".  I was very pleased with the result.

My next step was to remove the thick globs of hardened epoxy from the edges, and then to sand the beam smooth, particularly on the aft-facing side, which was to be varnished and exposed in the cabin.  The forward side would be invisible, so all I had to do was smooth it to ensure it was flat and even.  To clean up the beam, I briefly considered trying to run the whole thing through the planer, but decided that it was prudent to at least try sanding first, lest I ruin the beam, the planer knives, or both in the process.  To sand the beam, I used a belt sander with 60 grit paper, running the sander perpendicularly across the plies to even out any irregularities and to remove excess epoxy.  Once the whole face was smooth, I ran the sander with the grain to remove cross-grain marks caused by the flattening process.  Then, I used a DA sander to sand the top and bottom edges of the arc, just enough to remove the epoxy oozeout and handprints.  The sanding wasn't particularly difficult at all, and did a fine job cleaning up the beam.  More final sanding will occur later in the process before the beam is varnished and installed for good.

Next, I began the process of sizing and fitting the beam.  First, I laid the old beam on top, and marked the ends on the new beam.  Then, to double check how the old beam fit, I took it up to the boat and put it in place, and noted that it didn't extend all the way to the sides of the cabin.  Therefore, I decided to cut the new beam slightly longer, which would not only give me a fudge factor for the final fitting, but also might allow me to run the beam all the way to the edges.  With my checking and double checking done, I cut the new beam roughly to length on the chop saw.  The clean cut highlighted the even, thin cutposts.jpg (38815 bytes) glue lines between the plies, indicating that the clamping pressure appeared to have been even and appropriate.

With the beam cut to length, I tried fitting it in place.  Because of a slightly different arc built into the new beam, I found I had to remove a small amount of material from each of the two vertical support posts on the port side so that the new beam could slide over the top.  I cut these with a hand saw, which was made overly difficult and time-consuming because of the fine teeth on the one decent backsaw I had.

My first attempt at a dry fit revealed that I had to cut the ends at a slight angle relative to the face of the beam, to accommodate the fact that the cabin trunk narrows as it runs forward.  A compound miter box would have been nice, but since I don't have one, I worked out a way to use my fixed miter box to make the cuts at the appropriate angle (about 11.5 degrees).  After a few more trial fits and slight modifications to the ends and the top surface of the beam (slight stock removal in a few areas necessary), the beam slid tightly into place--just the fit I wanted.  I still needed some minor persuasion to get the beam snugly into its home, but it wasn't so tight that it was being forced.  

portbeamdryfit.jpg (27144 bytes)     stbbeamdry.jpg (32582 bytes)

newbeamdryfit.jpg (21923 bytes)With the beam fit in place, I marked the extent of the opening in the bulkhead on the forward side of the beam, and prepared to cut the smaller laminated sub-beam to fit.  After measuring, cutting overly long on purpose, and some additional trial fitting, I had the second beam section properly sized.  I discovered that, because of the forward-sloping top of the coachroof, the small beam section protruded below the main beam by 1/4" or so--had I though it through, I could have simply eliminated 1 or 2 of the laminations during the build process.  Now, however, I needed a way to trim the bottom of the small beam down to size.  

At first, I thought I would try to make the small beam flush with the main beam.  Since I don't have a bandsaw, my tool of choice was my portable jigsaw.  Unfortunately, the saw made a terrible cut because of blade wander through the thick, dense material.  While the visible cutline on top of the beam seemed OK while I was making it, when I finished I realized that the cut was completely unacceptable:  wavy, uneven, and just awful.  I put the piece aside for the night while I thought about other courses of action.

In the end, I decided to use my belt sander to remove the material necessary, since none of the saws I had were up to the task.  Using the main beam as a guide, I marked a smooth arc on each side of the small beam with a pencil, clamped the beam upside down in a Workmate, and fired up the belt sander.  The coarsest sanding belt I had was 80 grit, and I went through three belts as I removed nearly 1/4" of white oak.  It only took about 10 minutes, and I was pleased with the end result.  Since a flush fit was impossible (after my botched initial cut), I chose to recess the forward piece by about 1/4".

With that done, and a final trial fit, the beams were complete.  I removed both pieces and performed some finish milling.  I routed a 3/4" roundover in the area where the beams span the passageway, and a 1/4" roundover on the remaining edge of the main beam, and then sanded both sections smooth through 220 grit.  The beams cleaned up extremely well, and the large-radius rounded edges just screamed out to be touched.

mastbeam110903.jpg (8457 bytes)

I applied a coat of sealer varnish (50% varnish, 50% thinner) to the beams to prevent any staining or discoloration before continuing with the final fitting.  

I decided to emulate the original installation technique, which had certainly seemed adequately strong and suitable for the task.  The original beams were bolted through the main bulkhead and to each other, and featured no other means of attachment except for some lightweight tabbing at the beam ends.  The load from the mast is transferred directly downwards through the beam arch and into the compression posts secured to the bulkhead, with some load transfer also occurring through the bolts and into the bulkhead.

I ordered some 5/16" silicon bronze carriage bolts from Jamestown (expensive!) in two lengths--one set for bolting the main beam to the bulkhead (3-1/2"), and some longer ones for bolting the two sections of beam together in the center (4-1/2").  I went with bronze for strength, longevity, and, mostly, looks, since the bolts would be visible in the final product.

mastbeam111303.jpg (29072 bytes)With the beams in place, I marked exactly where the forward section of beam fell against the main beam, and then removed both sections one last time.  The marks were important because I needed to mock up the setup on the bench so that I could drill holes for the bolts, and so that the boltholes could be located in an aesthetically pleasing manner.

I marked for three holes towards the center of the beam, evenly spaced through the smaller forward beam section, and four additional holes--two per side--for attaching the main beam to the bulkhead.  Then, I drilled all 7 holes with a 21/64" bit on my drill press.

With the holes drilled, I lightly sanded the beams and applied the first of several more coats of varnish before installing the beams in the boat.  I finished up with a coat of a floor polyurethane, with a satin sheen finish, that looks simply awesome on oak.

fwdbeam111303.jpg (48223 bytes)

mbaftcomplete.jpg (29797 bytes)When the
varnish was complete, I installed the beam in the boat.  To secure it, I used silicone bronze carriage bolts through the main bulkhead, with the rounded heads facing the (suitably) head, and the nuts and washers in the forward cabin.  I attached the sistered forward portion of the beam with three similar, though longer, bolts.  I decided that the bolts were more than sufficient to secure the beam.  Fortunately, the new beam fit closely enough to the overhead, especially in the crucial center span beneath the mast step, that no further reinforcement was needed.

mbfwdcomplete.jpg (38350 bytes)I trimmed out the edges of the mast beam with narrow mahogany strips, many of which I had used in the original setup.  I also had to cut a few new pieces of trim here and there to replace some of the old pieces that didn't fit properly anymore.  

The mast beam replacement was complete!

Glissando, Pearson  Triton #381

We recommend viewing this site with your screen resolution set to 1024 x 768 or larger.  Problems?  Email the webmaster.

1999-2014 by Timothy C. Lackey.  All rights reserved.  No duplication of any portion of this website allowed without express permission.  Permission may be obtained by emailing the webmaster.