Thread: Epoxy filler alternatives

I'd like to see you repeat the tests, though this time with appropriate and fair procedures. I've done quite a bit of testing and now don't recommend wood dust at all and wood flour (big difference in particle size), unless it's a non-oily hardwood (very finely ground).

As to the tests, some of the resin ran out, simply because the joints faying surfaces weren't "primed" before the addition of thickened goo. The test coupons should have been wetted out with straight resin first, so the wood wouldn't suck out the bonding resin from the thickened epoxy, when applied. This would have mitigated a lot of the "resin starvation" you noticed.

The problem with wood dust and to a lesser degree cooking flour (wheat, rice, corn, etc.) is the particulate size is too big. This is especially true of wood dust, like that from a sander or even as in your case, from the dust collection bag. What happens is the exotherm reaction, thins down the resin and it tends to run out of the joint (or to the bottom of the vertical portion of a fillet). This leaves you with a resin rich bottom and a resin starved top joint/fillet.

All glue, fillet, fair compound mixtures will have multiple combinations of materials. For example, West System 405 (a filleting mix) is silica, cotton flock, a pinch of talc and a pinch of wood flour. The wood flour is for color, the talc to smooth out the mix, the flock to add a very fibrous, interlocking material and silica to control viscosity.

There's no one material that can make a good whatever, though some can be used alone, it's usually best if used in combination with other materials. As a rule, silica is a thixotropic agent, which simply means it's used to control viscosity. Used alone, as many folks do, it is strong, stiff and tough, but also tends to make the cured matrix brittle and hasn't any "interlock" or interwoven fibers (like the rebar used in a concrete slab), to insure it can transmit loading forces, within the cured mass.

In a nutshell, there are three types of mix categories: a heavy structural, a light structural and a cosmetic mix. Both of the structural mixes will need a high fibrous content (milled fibers, flock, 'glass, polyester, stone, metal, etc.), maybe something to ease sanding effort (talc, etc.), though this isn't entirely necessary and of course some silica, so the mix will stay where you put it, depending on application. A cosmetic mix will have light materials (Q-cells, micro spheres, etc.) with a smoothing agent (talc) and again, enough silica to stiffen it up for the application.

With destructive testing I now use a load cell, but I used to simply use a fish scale. If the sample coupons are small enough, a small scale is all you'll need. Record the testing with video, so you can see the point of failure, which is used for comparative purposes. The key is to do them all the same, with standardized procedures. An example would be the amount of goo used in each joint. I use a notched plastic applicator, specially cut to make uniform 1/16" thick, 1" wide smears of goo. It's just a single 1/16" tall, 1" wide cut in the applicator, with a guide on each edge to hold the plastic off the surface, as I smear goo on the surface. This means each joint has the same amount of goo and the test isn't tossed off, because one joint had a significant difference in glue mass. I also cut the test coupons from the same piece of wood, which are all cut to the same size. The testing rig, holds each piece the same way, with the same "purchase" in the jaws, etc. I also do a "base line" test with each destruction run. This is how I discovered that the use of fillers in epoxy, actually weakens it to some degree. For example, a straight silica fillet has a slight decrease in compressive strength, a modest increase in modulus, but a significant decrease in tension and peel resistance, compaired to a straight resin application. In other words, filler materials alter the physical properties, of straight resin to some degree. Some add a little of something, while detracting in other areas. I found this to be surprising, but it makes sense from a physical/chemical standpoint.

Hi PAR, great read thanks for the reply. I haven't been ignoring you, rather pondering whether to redo the test. I was in a hurry to see if the wood dust and or flour would work well enough to hold gunwales on, which so far all's well, though we'll have to wait over time to see if there's any long term failure. I've used the dust/flour mix to glue on inwale spacers yesterday and will do the same for the inwales and knees later today but all framing and chinelogs were with the powder. I should add, cost was not a factor in choosing an alternative, it was purely lack of ready supply of the commercial stuff and time running out before I have to go back to work. The major structural stuff has been done with powder and it was just these few items to go before gluing's finished. The 4Litre bucket of powder could've lasted for the whole build but I've wasted so much, which is to be expected when I'm fairly new to this I suppose.

Regarding priming, I agree, though the thought never occurred until it was too late . Remembering advice about precoating endgrain I did precoat the bow end of gunwales that meet but never thought to wet out bare ply prior to gluing. The ply I've got, although sanded, isn't a finish sand of course and well you can see how resin was sucked away in canyons, so to speak. I wondered whether I'd be left with glue starved or dry joints. For my gunwales at least there's just so much surface area glued I assume it'll be ok. I've come a long way in trusting epoxy. I put screws everywhere in my canoe but none in the GIS, so far.

I wish I did a straight resin test against the others and have some sort of scale to test against. I've read similar to what you wrote regarding mixes and compromises plus straight epoxy, additions that add features to the thickened epoxy and so forth and how straight epoxy is strong. It does all make sense. I would love to try some microspheres on cosmetic fillets for it's ease in sanding, but also demonstrate on a scale why they shouldn't be used for gluing.

I'm not sure if I want to put effort into doing another test under stricter conditions,yet, I wont say never. I'm still curious and I'm not sure anyone's shown pictures or demonstrated differences before visually. None that I could find a few days ago. Although I do remember watching in the past a Matthias Wandel youtube video or two where he performed great demonstrations on a home made test rig on adhesives namely different PVA glues, super glues versus fasteners and I'm sure he threw some straight epoxy in one as well. Just a shame he's not into boat building.

I'm a big fan of Matthias Wandel's efforts, though some of his conclusions about PVA's and epoxy aren't consistent with known testing, he's usually anal enough to get things right.

Testing is pretty common with boat builders, because we need light, yet strong structures, which requires dependance on glued joint performance. This assumes you'll be doing a lot of "goo work". For one boat build, you can just muddle through, likely with few issues, but if you're afflicted with the illness, you'll need to know the realities. The testing permits you to trust goo, but more importantly allows you to setup procedures that you can have faith in, with repeatable results. This is probably the biggest thing with epoxy in general and glues specifically. How tall to make a fillet and what materials to use in it, for a particular application, will help you know how much goo to mix up, how to mix it and how to apply it so it does what it needs to do. For example, after testing (many years worth) I now have premixed jars of fillers; a light structural mix, a heavy structural mix, a cosmedic mix and also mixes specific for different substrates (wood to wood, metal to wood, metal to metal, metal to 'glass, etc.). This way I can just grab the one I need, dump some in the goo and get started, usually with only a pinch of silica to control viscosity for the particular application.