Tips HeaderTips HeaderTips HeaderTips Header

logoHome Button  

Hardware Mounting  - (in wood)

One of the more challenging things of building a wood boat, is the mounting of hardware (typically strong metal parts), that take high, local loading. On most sailboats, these would typically be items such as chainplates, cleats and winches, but for some boats, such as small multihulls, there are other challenges such as parts of folding systems as well as eyestraps or track for high-load trampolines and other such items.
Even if sealants are used, if the mounted fitting is not mounted really solid, it will eventually start to move. Even if it cannot actually break-away, water can infiltrate the wood and rot will soon set in.

This somewhat lengthy article will discuss commonly used solutions as well as present some new ideas and procedures developed by the author. One may well conclude after reading this, that 'this can get damned complicated'! It may indeed seem that way, but if you need a really high-strength attachment, at least you know what steps to take. Leaving something out is always possible, but each step omitted slightly lowers the end result, so finally, it's your decision.

Typically, a metal fitting comes with a few holes for either screws or bolts and one assumes that they are sufficient in size and number. If the duty matches what the fitting was designed for, then that should be so—but if (as is the case for hinges used on the folding system of the W17) there's a change of load or direction, then one may need to reassess the attachments.
One also needs to consider whether the pull is parallel with the surface (shear action) or whether the pull is more directly away from the surface.

Should you use through-bolts or not?

While both loadings may typically justify through-bolts and backing plates, there are several reasons why this might be either not possible, or even, not advisable! So let's go briefly through these before looking at options.

  1. First of all, through-bolts need access to the nut side, for both removal and tightening. Generally, deck fittings and winch mountings will offer that but box beams generally do not. Whether any tightening will be required, can depend on the wood used and its moisture content at the time of build. If it dries out after the installation, then there's a good chance that tightening will be required at some point. And through-deck bolting often requires two available persons to execute.
  2. Wood tends to compress fairly easily across the grain—softwoods in particular and if the fitting gets slightly loose, water will run under and rot the wood. The fact that there's a through hole, always invites a leak.
  3. Fittings and equipment installed with countersunk heads will also convert typical shear loading into some axial load so this could also compress the wood it's mounted to (though less so with hardwood or plywood).
  4. Nuts and backing strips are not very attractive and in a cabin, generally need covering to avoid accidental head contact.
  5. Another factor that might be considered minor to some, is that the use of longer, full-diameter bolts with backing plates and nuts is certainly slightly heavier that using shorter machine or lag screws in an epoxy core.

OK, so those are 'some cons' for through-bolting—but surely, they must be stronger, no? Well, it's certainly very unlikely that backing plates will actually be pulled through the wood—yet, I cannot agree 'it's impossible'.
A short but true story is justified here.

I once owned a Buccaneer trimaran that had through-bolted tracks for the adjustable jib sheets. Unknown to me at the time, the high local loads had compressed the wood and water had continually run under the track into the wood. One gusty day, I suddenly lost all sheeting power and found I had 4 feet of alloy track flying around in the air with 6 mm bolts rattling around on it every 75 mm!! The whole track had ripped totally out of the wood rail—of which there was now virtually nothing left! The boat was only about 12 years old at the time.

After such an experience, I think I can be forgiven for at least asking the question, "Is there not now, another way that's equally strong?".

In the much referenced book on wood-epoxy boats, entitled "The Gougeon Brothers on Boat Construction", I think it is quite noteworthy that there's little mention of through-bolting of anything, other than for keels! Here, the Gougeons present a very useful chapter covering their tests and experience with epoxy cores for mountings, based on the use of their popular WEST System epoxies and even though the 4th Edition of this book is already over 25 years old, most of their solutions have stood the test of time. Such cores are not intended to go right through the wood they are fitted in, so for one thing, should never leak. Further, core pull-out tests have shown that IF the diam and length is sufficient for the type of wood used, the bolt should break before the core fails. So for starters, I would try to "buy, beg, borrow or steal" (well forget the last one ;-) a copy of their last chapter on this, as it's a valuable reference document.

So don't be surprised that I'm using very few through-bolts these days.
(and in case anyone wondered, NONE to attach hinges and latches to the beams of my personal W17. More later on this).

So as a main feature of this article, I'd like to take these basic epoxy cores a few steps further and show how to add even more strength, shear and pull-out resistance to them, as well as showing ways they can be a little shorter without pulling out any easier than the longer, straight core.

Let us first look at:
Fastenings working in shear — with loads parallel to the surface.

As hinted earlier, one thing that has always bothered me, is the use of countersunk head screws for items being held in shear. The taper of the head will most certainly apply a force trying to lift the screw out, and this risks to compress the wood and loosen its mounting. Unfortunately, some fittings require a near-flush top—and the upper hinges of a W17 are just one such case, as they need to fold close to the touching point.

Where shear loads are really high, additional steps may be needed to keep things well anchored and when installing parts that must also have flush fastenings (and therefore countersunk head screws), it's well worth considering the extra steps outlined here.
Most of us already realize that a steel fastening screwed directly into typical boat-building woods, will not resist a side pull anywhere even close to its own strength. So the wood at least needs local reinforcement and the common way (as detailed in Gougeons' Book and elsewhere), is to surround the fastening with an annular epoxy core. This increases the side area on which the fastening pulls and greatly improves its shear resistance. Typically, this core or plug, is double the diameter of the fastening. Whether one can increase this slightly, will depend on the spacing between the fastenings, as a point can be reached, where there's not enough wood remaining wood between the plugs and complete block tear-out becomes a risk. Personally, I like to see at least 12 mm (½") between cores. But beyond the basic core, what more can be done?

Here are a number of additional actions that will each add shear resistance compared to a simple core mount in solid timber.

1 — Use of one or even two layers of marine plywood for the top wood surface.
2 — Add layer of fiberglass over that plywood—preferably bias or bi-directional weave.
3 — During fabrication, add an additional layer of fiberglass under that surface plywood.
4–7 — Properly prepare the underside of the metal surface to be bonded (see the 4 points below).
8 — Add a thin layer of 1oz random mat between the metal and the wood.
9 — Increase the cross section of the core plug at the top surface—particularly in way of any plywood. This can increase total shear resistance as much as 50%.
10 — Add small choppings of fiberglass or carbon fiber tow, to reinforce the epoxy used for the core.

For a ¼" diam machine screw or bolt, action 9 can be achieved by first using a ¾" or 78" diam spade bit to drill out the first 4‑5 mm and then continuing to drill out the main core with a ½" spade bit (see pic above right). In the event that you've already drilled the smaller ½" hole first, all is not lost, as by dropping a length of ½" hardwood dowel down the hole (with a 18" hole in the top for the drill tip), you can now use the larger spade bit to mill out the surface recess. (See pic on left)


A Specific Case

To illustrate these procedures, let's look at fitting of the top hinges for a simple folding system, as a good example of where these extra steps justify to be applied.

First, the shiny surface of a typical metal fitting base (including hinges), is not a good surface for epoxy to grip on, so special preparation is recommended.
Also make sure the holes are countersunk the right depth for the larger cs'k machine screws. If they do not need to be totally flush, at least relieve the upper edge of the screw on the grinder to form a moderate 'pan' head.

Then prepare the hinges with these 4 steps:
  1. Use a grinder to remove the polish of the underside.
  2. After that, drill 8 mm holes of about ½ depth in the underside of the hinge palms. These will fill with epoxy and fiberglass during installation, adding to the shear resistance.
  3. Then heat the hinge to about 150F and brush on some un-thickened epoxy. With a fairly coarse emery cloth, rub this into the surface with a circular motion (see pic above).
  4. On to this, sprinkle some 2–3 mm lengths of carbon fiber tow (or fiberglass roving) and stipple these short pieces into the wet epoxy. Then leave to cure.

Ideally, try to do this about 12 hours before you mount the hinge or fitting, so that the epoxy is hard but still not fully cured. At this point, the surface should look like this (photo above to right), and clearly offer a much better grip to the wood than the original polished surface.


Fitting the Machine Screws, Lag Screws or Bolts into the epoxy core

There are two suggested approaches here. Both first require that the equipment or fitting being mounted, first be positioned in final location and a shallow hole drilled in the surface, to show exact location. [In the case of the top hinges for the W17, it is VERY important that they be aligned exactly at right angles to the main central beam, in order for the folding system to work smoothly. So with a piece of flat, solid timber clamped to the beam side, place a square on this and align the hinge palm to this—see pic.   To help place the hinges on accurately, it helps to slide a roughly 2ft long 5/16" dia metal rod through the hinge knuckles, in order to exaggerate any potential misalignment].

Using the small initial holes as a guide, now drill out the main holes for the epoxy core—ideally using a spade bit mounted in a drill press for accurate cutting and vertical alignment. Again, use a square to check that the beam is supported at 90 degrees to the drill.

[Drill ONLY as far as when the drill tip just breaks through the inner ply surface. This will prove important.]

From this point, there are two possible approaches. Both will require a mixture of un-thickened epoxy with added short cuttings of carbon fiber tow or roving:

  1. Fill these holes with the mixture and once cured, drill out pilot holes for the mounting screws or bolts and then use a tap to create the threads in the hardened core material, OR
  2. install the bolts or screws 'in the correct location' and pour the epoxy mix around them. The bolts/screws will have first been wiped with an oily rag, to prevent too solid a bond with the epoxy. (Oil should only be a VERY thin film—certainly nothing close to running—so it's best to wipe them off with a dry rag after applying oil to the threads.)

In my experience, the first system is simpler for alignment but less effective in providing critical strength. This is partly because the tap tends to tear up the hardened epoxy rather more than it would if cutting in metal.
This is probably due to the added 3 mm glass or CF clippings. So if planning to use a tap to cut threads, I would suggest replacing the short tow fiber cuttings by a hard density filler or metallic dust.
If tapped, you also have to be very careful not to overtighten the bolts or you risk to shear the threads. Should this happen, you will need to clamp that screw or bolt securely down in place with fresh epoxy around it and wait until it hardens around the threads.

Using the second 'pour' approach above has proven to allow a very tight joint and this is mainly because the epoxy is not disturbed once cured and the short lengths of carbon fiber are not cut by any tapping action, so remain fully effective. This is the system I personally recommend and use, but setting up for this approach takes special care though.

Again, here are two approaches to achieve this:
Both require that the bottom of the core hole be drilled to take the root diameter of the screw or bolt, so that it can be temporarily screwed in and stand vertical. For this method, it's also required to temporarily work with screws or bolts that are ½" (12 mm) longer than the final ones. The reason for this will become clear as we proceed.

Approach 1
If the holes for the core were correctly aligned for the fitting, then the screws will be close to vertical once threaded into the last 7–8 mm of wood at the base of the core hole. After lightly wiping the screws with an oily rag, screw just their tips down into the wood. You will now see the need for them to be longer than final, as this allows space under the head to pour the epoxy mix around the screw or bolt. Once this is done for all bolts, position the fitting or hinge over the fastenings to make sure that the head is exactly centered in the mounting hole. As this will not be readily visible with slotted head screws, only Philips or Robertson head screws should be used for this approach. (Should a screw or bolt just not want to stay in alignment, then use a thin strip of duct tape to pull it into line until the epoxy has hardened, as shown above for the 2nd screw.)

Approach 2, (that I more typically use) is to first fill the holes with epoxy and then pass the longer (oil wiped) bolts through the actual fitting or hinge and down through the wet epoxy and thread their tips into the wood below. To keep the hinge or fitting OFF the surface (to allow any surplus to overflow), 38" (9.5 mm) spacers are slipped under the edge of the hinge or fitting (see pic below). At least with this system, the screws can be tightened down to the hinge or fitting, enough to automatically center their heads to the exact location. Once cured, the spacers can be removed to give good access to the heads and these can then be easily cracked free of the epoxy with a small ratchet wrench or vice-grips (locking pliers). The long bolts are then removed completely, so that shorter bolts can be used for the final installation, once the mounting surface is ground or sanded flush.
This final mounting can use a very thin layer of fiberglass mat & epoxy for superior shear resistance and the fitting can be securely tightened in place with the ratchet and left to cure. Whether the threads are also wiped with epoxy or not, will depend on whether regular removal is planned. [Generally I'd recommend adding a touch of epoxy, as for the rare case of requiring removal for repair or replacement, this can still be achieved by applying heat.]

Such heating needs to be just local to each screw. This can be by an electric soldering iron placed against the screw head OR, by preheating a block of metal and letting it stand on the screw head until enough heat has spread down into the screw to permit removal. 150F will generally do it, with 200F a recommended maximum, so do not overdo it.

Before leaving this section on fastenings that work in shear, I'd like to comment on the use of alternative bolt/screw heads.
Most of the above has considered countersunk heads, as these are generally required and/or supplied with equipment such as winches—and flush hinges. But as already mentioned, even if there's sometimes no option, I am not a great fan of cs'k heads. Yet some notables still recommend them—so just why IS that? Well, apart from the fact that most pieces of equipment are designed for cs'k heads anyway, it's the metal-to-metal contact that they offer—when securely tightened. Such contact, when present, should prohibit any lateral movement and therefore permit the important surface-to-surface bond to ADD measurably to the total shear resistance.
But is there another way to get good contact without the negative lifting action of a countersunk head?

Well first of all, a straight smooth-shank bolt will certainly not create any upward force on the core (unless overtightened), but it's quite possible that there would be enough slack in the mounting holes that there would only be metal-to-metal contact on a low percentage of the fastenings—again allowing the surface bond to overload and break down. Whether the bolts can stay tight enough to not permit such movement will depend on aggressive surface preparation and how solid the core is anchored to the wood*. But with a simple, straight parallel core, there will be a limit on that. One possibility to help here, is to apply a horizontal load on the fitting in the direction of future load, while the bolts are being tightened, but this is only a partial solution.

But there's another fastening that I like, that DOES offer some improvement on this. It's not perfect but it CAN certainly offer better metal-to- metal contact WITHOUT applying any significant vertical force on the fastening and core. This is the Lag Screw. Take a look at its shape (left) and compare it to a straight shank bolt. Note both its full diameter immediately under the head as well as the very slight taper. Certainly a LOT more material here than at the root of a machine screw thread and generally, even more than a plain, straight bolt. If the mounting holes are correctly sized, that fine taper can really bite into the upper surface of the fitting or hinge, and virtually eliminate any movement. The body is also interesting, as the threads are at least double spaced compared to a machine screw or bolt, and this allows a really solid chunk of reinforced epoxy core between them—less likely to get sheared off due to a high vertical load. Couple this with a good surface bond of the fitting itself (as detailed above) and I believe this promises an excellent attachment. But what does 'not quite perfect' mean? Well, the wide thread pitch compared to that of a typical bolt, takes less rotation to tighten it—and therefore, less to loosen it too. But as it's possible to bond these heads to provide resistance against rotation, I think overall, this is a very interesting option.
WARNING: One thing to watch out for though, is that some of these screws are cheaply made and the threads can be too rough for tightening in place, without risking to damage the epoxy core. Also, some lagscrews are now being made undersize and worse, some even without the important, lightly-tapered plain-area under the head, so be sure to examine or send for a sample before purchasing for any critical use.

When installing lag screws, remember to drill the hole for its core body diameter a little deeper, so that you can tighten it down on the fitting after the epoxy core has cured. Ideally, the slight taper under the head of the lag screw should fit the mounting holes very snugly. If shear loading is to be very high and space between holes permits, use the under-surface recess approach noted earlier (point 2, under hinge preparation), to further increase shear resistance.

I fairly recently used lag screws to attach some experimental latches that will be highly loaded. In this case, I drilled holes for 'anchored cores' (see pics) and after first wetting-out the core surfaces, poured in CF‑reinforced epoxy about 80% full. Then, with all the oil-wiped lag screws fitted tightly to the latch, I slowly lowered them into the liquid wells and held each screw-head down firmly with a clamp.

(I had a wax paper under the latch so that it would not bond to the surface prematurely.)
Once cured, I removed the latch but returned the lag screws back into their same threaded holes, so that I could add more epoxy around them, to fill the core flush to the surface or higher. Later I ground the surface off flush and reinstalled the latch permanently with a light mat under its surface. I'll be totally amazed if these ever move again.


Now let's look at:
Fastenings working with tensile pull — those being pulled straight out.

Although my home built test rig has a limit of just over 2000 lbs, at least I could test the smaller bolt sizes and weaker woods. By then cross-referencing to other published tests, I was able to extrapolate results for slightly larger bolt sizes. High accuracy of measurement appears not essential, as there's a fairly large range of results for what is seemingly the same size and material, say ±30%, so it would not be smart to work with a safety factor of less than 2 anyway, and I'd recommend using 3 on anything critical.
Let's first look at the regular epoxy core insert for this duty. In the sketch at the end of this article, I'm suggesting that the epoxy core be 2 times the bolt diameter and that the depth be 2 times the core diameter. In a hardwood like mahogany, this will give a pull out force that is pretty close to the strength of the bolt—presumed to be weakened under its head by its thread. Increasing the core diameter to say 2.25 times the bolt diameter will add 15% but one needs to be careful when getting to the larger sizes, as damage from excessive exothermic heat (caused by chemical reaction) can significantly affect the strength. In the accompanying table, to cover the low end of things, I show the force one might expect for pull-out from one of the weakest woods used structurally on a boat. The same insert in decent mahogany, gives about 50% more holding power. Adding depth to the core will also add shear area in direct proportion, but sometimes, the wood is not thick enough to deepen the hole. So here below, is another effective way to add pullout resistance.

Both the possible vertical core displacement from using cs'k heads with a horizontal shear load and of course the straight pull-out, can be helped by using something I've devised that I call a 'core-lock' or 'anchored core'.
This requires a simply-made tool to undercut the basic core hole so that at the lower part of it, a rim is formed that truly anchors the core deep into the heart of the wood in which it is poured and formed. (see pics below)
I personally believe this 30 second extra preparation to be well worthwhile for all countersunk machine screws and also for any attachment that will be loaded axially—ie: for things like screw eyes and even eye-straps that have to take heavy pull-out loads from trampolines and the like. Also eyes for shrouds where the eyebolt end cannot be accessed the other side to receive a backing washer and nut—in fact, any location where the 'pull-out' axial load is high on the core itself.

Here is how to make it and what the revised bit looks like. All you need is an old spade bit (½" will be good when using ¼" bolts or screws). First grind off the bottom point and round off the surface to that it has no cutting edge at all.
Then, grind the upper sides of the spade—also rounded with no cutting edge—right down to the last ¼" of the bit. That's the ONLY part you need sharp.
Now, after you've drilled your regular hole for the standard epoxy core, pass this bit down inside and while keeping it down in the bottom of the hole, slowly go around the periphery of the core hole and this little blade will dig out a recess deep down inside the core hole. Then, when you fill this with FG or CF‑reinforced epoxy, it will be solidly anchored against any vertical dislodgement (see pic below). Just remember to FIRST clear out the dust and prime the hole and recess with unthickened epoxy, so that the wood grain is filled before the final core is created. Mounting bolts or screws in this core is exactly the same as for a parallel core. The only fastenings for which an 'anchored-core' might be overkill, could be a straight bolt or lag screws working in shear. Both give little upload.

Naturally, for all cores and fastenings subjected to vertical load, an 'anchored core' would seem most beneficial and worth the small extra effort. It can also help when attaching hardware to relatively thin wood, though a through-bolt with a backing plate is the recommended mounting here if there's access to the other side and the nut does not get in the way.

So how much more resistance can we expect from such a ring? Tests on epoxy seem to indicate one can reliably get about 1500 psi in shear, so for a ¼" undercut and ring of 2 × bolt diameter, we can expect an added shear area of about 2d × π × 0.25" — or 1.57d. This gives 1.57 × 1500, or 2355 lbs × bolt diameter, ie: about 590 lbs extra pull out resistance for a ¼" bolt—adding 25–35% over a straight plug.

There's no reason one cannot combine the two 'extra-diameter rims' that I've proposed above, and use both for a particulary demanding attachment.
They really take very little extra time to cut and as long as the epoxy used is reinforced with short fibers, such cores will be nigh impossible to break out. The only thing to keep in mind is that epoxy starts to soften with high temperatures and also needs protection from direct sunlight.


Added 2019:  In case anyone thinks I am totally against thro-bolted hardware, I am not.  But this article is to show, its not the only way and not always the preferred way. Through bolting requires access to the nut and also a core of high density that will not readily compress.  Otherwise, a compression-resisting-tube is required. (as proposed for bolting the W17 beams down to the ama deckpad).   The metal surface of a through-bolted fitting also needs sealing with its wood base to prevent water ingress, and a non-hardening polyurethane is one recommendation.

 The author welcomes comments plus any further input and experience on this critical subject. Please use the standard questionnaire with 'Hardware Mounting' as the subject. Thanks!


Read more Construction Tips & Techniques.


"New articles, comments and references will be added periodically as new questions are answered and other info comes in relative to this subject, so you're invited to revisit and participate." —webmaster


"See the Copyright Information & Legal Disclaimer page for copyright info and use of ANY part of this text or article"