Wood/Epoxy Keel and Keelson

The problem with maple, although it’s very hard, it’s not particularly stiff, in fact for its weight (which is considerable) it’s far less stiff than almost any of the oaks.

Red Oak in not at all suited for marine environments, as it caries its own fungus just waiting for a little moister to start its action. However the tried and true marine structural wood of choice, for several centuries, has been white oak.

Maple does have one advantage over oak, it’s a little more impact resistant, due to its flexibility, and less grain run off, but if you were looking for a hardwood with high impact resistance hickory would actually be a better choice.

As far as the beam, as was stated earlier, an I-beam over a span is every bit as strong as a solid one, where it lacks is in both point loads, and impact loads.

You can get around this by making it solid at the connection to the keel and in the bow where loads of both kinds are expected. You needn’t worry about finding a large piece of wood for the solid parts; laminated ones are straighter, easier to shape, stronger, and dimensionally more stable.

Another thing to remember about wooden composite structures, is point loading in general, great care must be taken to insure that your hard points are sufficiently strong to dissipate some truly outrageous loads.


Good luck with your project.

I’m currently building a 30 foot catamaran using spruce strip planking core under a carbon skin; the difference in stiffness is compensated by running the unidirectional carbon skins a right angels to the longitudinal spruce strips.

The strips act as cores in loads more perpendicular to the keel and as the major structural member in loads more parallel to it.

 

Ha hah! You've spotted the pattern. I am indeed choosing local woods over imports even if sometimes imports cost cheaper (like Khaya). Sustainability and environmental friendliness is the goal. After all, if the woods are available and can do the job (with the help of epoxy), why not put them to use. White pine is very close in characteristics to Cedar. Maple on the other hand is kind of different...

Asleep Helmsman:

Interesting info. Are you referring to a particular kind of maple?

 

Canuck....

The simplest inner keel/apron is a rectangular block, typical size for a 40' boat might be 3.5" by 8" up to 4.5" by 12-15". It is just very easy to cut a square notch in your forms to the correct depth and lay up the entire keel (pieces properly scarfed) in one go.

Major function of the keel is to resist the longitudinal bending loads imposed by the rig (or hydrodynamic loads in powerboats). Thus you want lots of material at the bottom and top (deck) of your hull girder. Some extra weight in the keel is not a bad thing when something goes bump in the night. I would not be comfortable with only 2" by 5" of (continuous) longitudinal wood in a 40' boat. I would also be very uncomfortable with a composite (carbon/Kevlar/wood) structure unless I had performed some real world (your shop, materials, and methods) tests of those bonds.

The bonds in your keel are loaded in rolling sheer, some woods (doug fir, mahogany) have been tested and perform well under this type of load. Others woods are weak in this condition, I would have a serious look at the question before committing vast resources to this.

A good reason for using the I-beam keel is a high stiffness requirement vs lack of vertical space. If your hull form (sections) is not tangent at centerline, the floors become vital in tying the two halves together. In a marginal headroon situation you can get the floors across the centerline and max longitudinal stiffness in the keel using the I-beam. In a racing boat it's considered okay to stumble over structure, thus the keelson can be higher than the sole.

By blocking I meant the short pieces between the floors, (the web part of your I-beam).

 

Tad: thank you very much. This is the information I was looking for.

 

Looking at the USDA Encyclopedia of wood, I find the Modulus of Elasticity of Sugar (hard) Maple to be 12,600 MPa with a density of .63. For White Oak it is 12,300 MPa for a density of .68. That means that the Oak is 1% less stiff than Maple. Also, the white oak's density is higher. Hard maple is the heaviest but not stiffest of Maples (in fact, average oak gravity is .665 whereas avg maple is .538 according to USDA EoW).

So I'm not sure where you're getting your data. Do you have a different reference or any reason to believe the USDA data is inaccurate?

When you speak of the tested and true properties of stock like Mahogany and Douglas Fir, are you referring to ultimate strength, fatigue behaviour, stiffness... all of the above?
By tested, I'm assuming you mean by ages of boat building experience as opposed to actual engineering testing. Do you yourself conduct tests in these situations or instead refer to a particular body for testing. I notice that the afore mentioned USDA handbook only provides shear parallel to grain which is not rolling shear. It does however provide some information on modulus of rigidity by direction.

I'd like to thank everyone for the time dedicated to this thread. It has been very useful and informative.

 

Rolling sheer testing of wood and epoxy bonds was done by the Gougeon Brothers a number of years ago in regard to windmill blades that they were building. Michael Zuteck did the research and came up with many of the techniques. Talk to the tech department at Gougeon's for a copy of the paper.

Before committing to building an entire boat of unusual materials, I would do a great deal of testing. I urge people to do tests using the actual materials they will use to build the boat, and to create samples in their own shop using their equipment. Testing done under some other conditions in some laboratory are not real world, or at least not your real world. Hopefully your own tests just substantiate those done by others. But wood has highly variable properties depending on all kinds of factors.

 

Here's a link to a pdf where 6 oz. of glass is mentioned enough times that I consider it a good number.

Tad, thanks again. Very much obliged.

 

Canuck thank you for that info

 

Well I guess my information came from years of building architectural millwork. Maple (hard maple) feels heavier than oak and when you carry it, the rhythm or your gate makes it oscillate more.

Not science I agree.

Also we found that maple (hard once again) will flex better than oak for tight radiuses.

So after you challenged my long standing gut feeling that oak is stiffer than maple I decided to look into it. So I asked other wood workers.

We all had the same impression. (based I’m sure on the fact that hard maple is used more extensively in millwork than soft)


So next I decided to look it up (same as you) and found some interesting figures.

First I would treat soft maple and hard maple as two distinctive wood types, there are more differences between hard maple and soft maple than there are between red oak and hard maple.

Next I have considered that white oak was generally used in ship building as high load members, based more or less on its ability to resist decay. In an encapsulated modern boat this is obviously not a consideration, so let’s just concentrate on Red Oak.

According to the American Hardwood Export Council there are two different versions of red oak, Southern and Northern

The Northern has an MOE of 12,549 and the Southern has 15,721. However the Southern turns out to be heavier (denser) than the northern by .05, which is .04 denser than hard maple, but on has a MOE of 12,618

Soft maple has a specific gravity of .54 and modulus of elasticity of 11,308

The stiffness to weight ratios are as follows: (larger number represents stiffer verses density)

Species.....................Stiffness...............Density..........................Ratio
Southern Red Oak........15721_____________0.68________________23119.12
Soft Maple.................11308_____________0.54________________20940.74
Hard Maple.................12618_____________0.63________________20028.57
Northern Red Oak........12549_____________0.63________________19919.05



Take a look at this web site
http://www.ahec-europe.org/ahectec/SpeciesGuide.aspx?lang=en

The other issue with hard woods in general is the fiber length. Softwoods tend to have longer fibers making them more suited to spanned structural members like wooden I beams.

But as was mentioned earlier I would also experiment with epoxy adhesion to any wood you choose if sufficient data are not available.

The Gougeon Brothers are always willing (and able) to provide good data.

 

AH:

Thanks for that write up.

I've started doing some testing around the shop on maple as Tad suggested (especially regarding rolling shear). I did call Gougeon Brothers like you mentioned and they sent me some info on a test they performed for rolling shear.

I've added two pics. The 2D pic denotes where pressure is applied. The 3D figure has the end grain of the wood marked with a cross. The two brown pieces are the wood being tested.

GB recommend using a hard wood as the support blocks through which you exert pressure (the gray blocks) and, interestingly enough, I think they used Maple.

I did a quick prelim test using Mahogany as the test wood, and maple as the studs, and it performed pretty much as I would expect. The purpose of the quick test was to go to failure and see if the test worked as expected. It did: the mahogany split along its (barely visible) growth rings.
I also did the same test with the equal volume of Maple (hard).

However, trying the same test with maple as test sample and frame didn't work so well. I thought that even though the test harness and the material being tested was made of the same wood, it wouldn't be a big problem since shear is so much lower than flex/tense strength. Anyways, long story short, I didn't manage to fail it in the same way: the epoxy bonds broke. So I'm doing another test batch to at least try and make it fail along growth rings as I expect it will be.

By the way, I am not worried that the epoxy didn't bond well: this was a quick test, and I barely did any surface preparation. The maple was planed and was shiny smooth. I've done other adhesion tests on maple where the bond was strong enough to break the wood instead of the bond.

As a side note: I used a simple vise to manually apply the pressure, and let me tell you, while the mahogany went pretty quickly and easily, I put pretty much all of my weight on the maple. I was actually worried I might damage my vise. So just from the feeling, I went way way higher in pressure than with the mahogany.

I will be doing a test with load cell and measurements tomorrow.

I've been thinking about the upsides and downsides of using maple on the keel. Originally, my sole comparison point was the scantling used on Golden Daisy (mahogany keel). So I was basically considering Maple versus Mahogany. Now that you've put in Oak in the mix, I'm having an easier time actually...

I have several factors that I'm considering: price, strength, decay resistance, and weight. Decay resistance is still nagging me to no end, but the more I hear the possibilities where something other than Mahogany was used, the more I feel comfortable letting go of that kind of ideal wood.

Weight is actually not an issue at all. This is the keel after all, and if I'm going to have something heavy in the boat, this is a pretty good place to have it. But even at that, the whole keel structure barely weighs 200kg (with maple as wood). The difference between .4 and .6 density will be negligible.

Regarding strength: these tests are making me more and more convinced that Maple is indeed as strong as the data indicates. Yes, there might be subtle differences in strength, but if I consider using the same volume (not same weight) of maple/mahogany/oak, then maple and oak are going to be substantially stronger.

And price: I've found Maple (hard) for $2.75/bf whereas White oak is $3.50/bf. For a keel that's 4"x10"x37', that's a 90 dollar difference.

That is such a small difference that I guess the only remaining thing that bugs me still... is the decay resistence.

Thanks Asleep Helmsman and Tad, I will seriously consider Oak instead of Maple.

This was and continues to be a very informative discussion. Thanks for keeping it rolling.

 

Final post, test results etc...

I am posting the results of rolling shear test of Hard Maple (aka Sugar Maple) for posterity. The test was rough, but did expose an approximate order of magnitude of things.

The testing jig used was the one detailed in the previous post, concept graciously provided by Gougeon Brothers at WestEpoxy.

The measurement device was a tension based load-cell. Attached is the figure of the testing contraption. A steel pipe was rigged perpendicular to a wall. A metal shim was used to distribute the load evenly on the sample wood. Load was measured from bottom wire which was at roughly 30˚ from horizontal.

Maple Failure (2) shows how the failure occured. All pieces are hard maple, and as can be seen from the darker shade due to increased epoxy absorbtion, the tested pieces had their end grain facing the camera.

Test pieces were 3 3/4" long by 3/4" wide (bonding surface) by 3/4" thick, for a total surface of 2.8 sqin.

Epoxy used was 105 & 206 with 405 filler at syrup consistency. All pieces were coated with an initial coat of unthickened epoxy before the thickened mixture was applied for the bond. Maple surface was sanded with 80 grit until soft to the touch.
The same preparation was used to bond a test sample of Khaya Mahogany which failed, as expected, along growth rings.

Results:

Failure occured primarily (~80%) in epoxy bond despite having been thoroughly prepared. A small part of the grain ripped off the testing stud (longitudinally aligned fibers). See "Maple Failure" for illustration. The pieces being tested were entirely intact.

Peak tension read by the load-cell was at 3200 lbs. Angles as described in the figure were roughly 60˚ and 45˚ respectively. According to my calculations, that means that the load on the test subject was about 1400 lbs.

This would indicate a surface shear pressure of ~500psi. Which is in the same magnitude as official numbers of shear parallel to grain.

There was no damage visible due to rolling shear to the wood being tested. No cracking, checking, buckling, or sounds. Only a slight smell of burnt material on failure.

Metal shim holding the sample in place was deformed as can be seen in images. This piece was a 3"x1" by 1/8" thick steel beam (unknown grade of steel but used for rigging).

The test was officially a failure, since the rolling shear properties were not measured directly. However, I did draw the conclusion that Hard Maple can withstand at least 500 psi of rolling shear without any complaints. I am not sure about the bonding anymore, but given that mahogany doesn't exceed this number (500psi), I wonder if it is relevant whether or not the bond was hard enough to break the wood or not.

Once again, thanks all for time and input. If you have comments or questions, feel free to ask.

 

Hi Canuck Guy
Thanks for taking the time to publish your very interesting study.

Unfortunately, I am a bit thick, and I didnt cotton on to the latest test method.

Could you elaborarate on the last method with the metal bar a little bit more for the mentally challenged ?

 

The test was basically meant to be performed using a vise. I've drawn up a picture meant to describe the various forces being created. The sample wood's end grain is facing you.

The metal bar was because I had a load-cell that measured loads only in tension, and not in compression. So I rigged something that looks like a mast. Basically, picture the image so that the metal bar is pointing vertically up, and it's the exact same dynamics as a mast, headstay and backstay. Given the tension of the headstay and the various angles, I can roughly determine what the tension on the backstack and mast foot are. For details on that: I basically break down the tension force (which is aligned with the wire) into horizontal and vertical components. Both of these are known on the side of the load cell. Then I find the tension of the wire on the other side by saying that their horizontal components are equal since the device is at equilibrium. And I finally deduce the vertical component of the other wire, and add the two vertical components to get the final pressure on the mast foot.

I stand to point out once again that these measurements were quite inaccurate. For one, I guesstimated all the angles. I simply published what I ended up doing so that the thread doesn't turn into one of those dead end google searches where a question is asked, but no answer is ever given.

So the only real conclusion I came to was that Hard Maple was able to take at least as much punishment as mahogany. Everything else was just a matter of experience that I will probably use in further tests down the road.

Hope this helps.

 

Nice work Canuck!

It would appear to me that while it was fairly easy (in your shop) to get a bond strength equal to or greater than wood strength in mahogany, this was not possible (using your test methods) in the maple? As a general rule the wood should fail before the bond.

I don't recommend oak for work with epoxy because of unpredictable reliability of the bonds. They usually give up eventually, not always, but it's a question I don't want to consider in the dark when the wind is howling.

That's the next question for the maple bonds, long term cyclic loading? I realize you have no way to test this. You might throw those test bonds in a bucket of water, take them out occasionally and try crushing them in the vice.

 

Yes, that's definitely my next area to tackle. I want to be able to bond the wood better.

The other thing is that the test harness was made of the same wood as the tested sample. Which isn't really ideal. This is at least somewhat significant since the maple did break along the grain a bit.

I'm kind of clueless regarding weather/decay testing and cyclic testing. Any methods I can think of involve industrial solutions, something which I can't really afford or achieve on my own.

I'm not really out to change the world, but my assumption is simply that maple wasn't considered before since there was no epoxy. I'm trying to see if, with epoxy, the rules of the game change...