Basic Composite Fabrication

When I was building the BEHEMOTH substrate back in 1989, I needed a way to fabricate lightweight, rigid structures with arbitrarily complex (and sometimes curvy) shapes. In particular, both the bike’s trailer and the equipment enclosure behind the seat were fiendishly complex designs that required a lot of mounting points for equipment, and the thought of trying to do it with aluminum was disturbing. I asked David Berkstresser for advice, as he always knows about these things.

“Build it with cardboard,” he said. “I made a kayak out of cardboard once.” I had a good laugh, and then realized he was serious. Within a week I had scrounged a pile of clean corrugated cardboard and was building my trailer enclosure by cutting out pieces with a utility knife, sticking them together with a hot-glue gun, and fiddlling with the evolving sculpture until it was exactly right. At last, happy with the design, I lay fiberglass cloth on all surfaces, bonded it with stinky polyester resin, and built up multiple layers where there would be significant stress, hardware attachment points, or abrasion. I then “closed out” the edges with a Bondo-like paste, filled and sanded the surfaces smooth, and took the unit to a professional for an automotive-quality Imron paint job. The results were amazing: light, perfectly matched to the application, and even pretty:

Figure 1: The bicycle, BEHEMOTH, with enclosure behind seat and trailer fabricated using "CSPC" cardboard-core composite technique.

We dubbed the technique CSPC, or “Cellulose-Core, Silicon-Matrix, Polyester-Filled Composite,” and the technique is described here. (If you think I’m mad, consider that the BEHEMOTH enclosures only involved simple curves, and were thus trivial to form by making long parallel slits in one skin of the cardboard. David’s kayak, on the other hand, had compound curves…requiring a detailed pattern of little X-Acto knife cuts to “stretch” the surface.)

This may sound frivolous, but there was real engineering here. The skins of fiberglass cloth, once solidified in a matrix of resin, provided significant tensile strength. As they were held a constant distance apart by the corrugated cardboard core, any attempt to bend a panel would encounter strong resistance. You can see more sophisticated variations of this technique in boats, aircraft, bridges, and even some new roofing materials.

Cored Composites

The essential idea behind cored-composite structures is to maximize the stiffness-to-weight ratio. Since flexural stiffness of a beam is proportional to the cube of its thickness, it would at first seem desirable to make everything thick—but if the material is solid, that can become absurdly heavy. Fortunately, we can play a trick: use a lightweight core to add thickness. When you try to bend something made this way, the stresses resolve into three separate cases, so we can optimize the materials to handle them. Imagine pushing down on a piece of my cardboard-core bicycle trailer:

1. The upper fiberglass skin is in compression across its surface, focused at the point of contact
2. The lower skin is in tension
3. The core is loaded in “shear,” which tries to split it along the middle as the faces move like hands rubbing together.

Since everything is bonded nicely together, the characteristics of the long glass fibers work to our advantage (they are oriented at right angles in a normal cloth weave in this particular case, but there’s nothing sacred about that; in optimized structures, the majority of the fibers may be oriented along the lines of maximum anticipated loading). As long as the core maintains a constant thickness and doesn’t start to crush or split longitudinally, the whole assembly is resistant to bending and the skins don’t wrinkle or buckle; failures, when they do occur, are generally catastrophic. It’s analogous to an I-beam, where the skins correspond to the flanges and the core corresponds to the web. This approach also cuts weight, with the sum of the skins thinner and lighter than a “single skin” structure of equivalent stiffness (and it has better insulating properties).

Of course, cardboard is a wimpy core material, and would never be used in a real application (especially in a marine environment—one crack and the whole thing would turn to mush). Old boxes were a cheap option back in 1989 when I was bicycle-hacking, but I graduated to professional materials when moving on to boatbuilding; instead of cellulose, the Microship uses Divinycell, a closed cell rigid PVC foam with a density of 5 pounds per cubic foot. This is one of a wide range of materials that are made specifically for this purpose, varying in shear modulus, density, permeability, fabrication difficulty, and cost.

One familiar core material is end-grain balsa wood (from Baltek Corporation), oriented with the fibers running from skin to skin to resist shear loads and add compressive strength. Balsa is particularly good at handling attachment points without crushing, something that is more of a pain with foam. But as with the cardboard, damage to one of the skins (or even a hidden leak around a fastener) can eventually cause the structure to become waterlogged.

Foams are available in many flavors, ranging from cross-linked types like Divinycell (and Klegecell) to linear types that are more flexible (Airex). They can be had in a wide range of densities as indicated by the application—where flotation is an issue (as in surfboards) a very light core would be used, but that is of course more prone to impact damage if the skins are too thin. Always the trade-offs…

Figure 2: A sampling of core materials. Gray surface and scrap at upper-right are the Divinycell used throughout the Microship project; various honeycombs are also shown.

The most “high-tech” core material, usually seen in aerospace applications (or racing yachts) where there is the budget to pay for it, is honeycomb. This stuff looks like a beehive, with structural and weight characteristics nearly ideal for a cored composite. Naturally, it comes in many forms: Hexcel sells an aramid (Kevlar) material, Nida-Core has a relatively affordable polypropylene core, Tricel makes one out of paper, and you can even find it in aluminum or carbon fiber for special applications. But all these are harder to work with than foam—the close-outs (edges) are trickier, and when bonding skins you need to make sure that epoxy doesn’t flood the cells and turn the whole thing into a brick (Nida-Core solves this with a bondable scrim on both faces). Still, a properly made honeycomb panel is a thing of beauty, ultralight and extremely stiff.

Unfortunately, all cored composite structures introduce fabrication challenges. Imagine the simple problem of bolting on a piece of hardware: on a single-skinned panel, this is so trivial as to barely be worth mentioning. But with a cored panel, the compressive force of the bolt can be quite enough to crush the relatively soft foam or honeycomb, introducing a stress riser that can cause structural failure (or at least the eventual loosening of the bolt). To get around this, there is a whole body of techniques that involve replacing a section of the core with a rigid annular ring that resists crushing—in our case, usually accomplished by “hogging out” foam with a bent nail chucked into a drill, clearing the hole with compressed air and hemostats, injecting filled epoxy into the enlarged cavity, and then re-drilling for the fastener itself—or sometimes using a syringe to cast threads in place around a waxed bolt. It’s a nuisance that prevents the casual slapping-on of impromptu hardware, but taking the time to do it right is worth it.

Other fabrication problems include getting a perfect bond between skin and core—if we don’t, then bending stresses will tend to separate the two and lead to buckling. It’s not always obvious just how thick to make the skins and cores, what kind of cloth to use, the optimal orientation of the fibers, and how gradually to distribute localized loads across a wide area. And finally, remembering our rule that an infinite number of very light things becomes infinitely heavy, it is important to use just enough resin, but not TOO much. To this end, one of the more advanced fiberglass fabrication techniques is vacuum-bagging, which uses the weight of the atmosphere as a giant press to force excess resin out of a layup. Most of the time, fortunately, we can get by with “hand layups” and relatively simple techniques.

Indeed, the beauty of composite construction is that it’s pleasingly incremental: you sculpt chunks of foam, glue them together, add fillets to inside corners and round those outside, then bond on fiberglass skins of sufficient thickness to satisfy the structural requirements of your application. Sometimes you mold structures around the objects to which they will mate; other times you do freehand build-ups of thickened epoxy and then shape them with a Dremel tool. The result is a kind of ugly thing with the surface texture of hardened burlap, dangerously sharp bits where once-soft fibers now protrude like needles, and a patchwork of blah colors resulting from the types of epoxy and hardener used as well as the various fillers. To clean this up once all the engineering issues are satisfied, you sand down the rough edges without damaging any more fibers than necessary, blend microscopic glass balloons into a fresh pot of epoxy, butter the fabric texture and overlapping edges with squeegees and other tools, and wait. Once that cures (a function of temperature and epoxy chemistry), you can finish the resulting object to any level of perfection ranging from butt-ugly to showroom-perfect depending on how much of your life you want to devote to it, then get intimate with another suite of toxic chemicals to prep/primer/paint the surface. It’s all very messy, actually, but at every step in the process the thing you’re building is editable—time and again, we tore into “finished” work with fearsome implements of destruction to correct mistakes, add fixtures, and even change our fundamental design as evolving specifications rendered earlier assumptions fallacious. This is not even remotely like machining: once you get used to working with composites, there is liberation in knowing that anything can be undone or fixed; rarely is it necessary to start over. Errors are not fatal.

The Ingredients

It seems, at the outset, like this should require only a minimal inventory: a roll of fiberglass cloth, a few gallons of resin and hardener, and something to use as a thickener for fillets and fixtures. But this is only the beginning, as every aspect of composite fabrication involves a wide range of choices. Let’s set the stage for getting our hands dirty by running through what it takes to set up a fiberglass shop.

Figure 3: The Zone of Goo

A Note on Safety

First, I should mention the problem of being a delicate biological organism in contact with profoundly nasty industrial chemicals. In addition to the obvious physical hazards of skin laceration and itching, there are two classes of things that you don’t want in your body. First, many of these chemicals are evil: the hardeners used with epoxy resin are sensitizers that lead to severe allergic reactions, the MEKP used as a catalyst for esters will cause blindness on eye contact, and various essential solvents are highly toxic (many are also carcinogenic). Sanding a fresh layup generates fine, chemically active dust that is optimized for inhalation poisoning, and working for hours with volatile organic compounds in an enclosed space will push the limits of any safety gear as you sweat profusely and roll around in epoxy drippings. Inhalation, ingestion, and skin contact are equally dangerous, and it is essential that people working with this stuff take protection seriously and use high quality masks, gloves, and eye protection. Unfortunately, there is a sort of macho mentality among some builders (I actually heard one guy say, “this stuff doesn’t affect me”), but it’s only a matter of time and has nothing to do with masculinity—the toxic effects are cumulative and eventually sensitization will occur. At this point, the project ends. If you are patient, you can find good deals on partially built boats and kit airplanes for exactly this reason, though it seems a rather sad way to take on a project.

The other hazard is more “mechanical”—the non-chemical but equally damaging effects of things that can be inhaled but never eliminated. Some filler materials, most notably colloidal silica and other sharp-edged fibrous substances light enough to hang suspended in the air for some time, can become a permanent fixture in your lungs. Those nuisance-dust masks from the paint section of Home Depot are not enough; it takes good filters and careful attention to detail (not shaving my beard constitutes risky behavior; it’s hard to get a good seal around my 3M activated-charcoal dual-cartridge mask).

When you think about it from a health perspective, this whole pursuit is insane. If you decide to do this, please spend some time planning around the safety issues: install dust control facilities like a shop air cleaner with HEPA filter, ventilate well when using solvents, understand and label all chemicals, don’t scrimp on personal safety gear (we’ve gone through over a thousand latex gloves), wear your mask when sweeping, use a positive pressure respirator of you’re crazy enough to do your own spray painting, use wet sanding techniques to minimize dust whenever possible, use hearing protection when grinding, treat the table saw with Deep Respect, don’t eat in the shop, and don’t smoke at all (it makes your lungs sticky). Read and take to heart all the literature on the subject, and spend some time hanging around professional boat shops to see how people with legal exposure are covering their assets. And don’t forget that poisons don’t magically disappear when placed in the trash; some of this stuff is classifiable as toxic waste and does not belong in aquifers, landfills, or waterways (the correct way to dispose of trace amounts of hardener, for example, is to use it to catalyze a bit of resin and then discard the resulting solid object).

OK, assuming that we still want to risk shortening our lives with this stuff, what do we need to stock?

The Essential Goo

One of the first issues is the choice of the resin that will be used to solidify an otherwise soft layer of cloth. The broad choice is between epoxy and the esters (polyester or vinylester). The latter are cheaper and more suited to volume production and “chopper guns,” but are structurally weaker. And they stink. For hand layups and one-off custom projects, epoxies are the only rational choice; the extra cost is more than worth it.

Epoxies are available in a variety of systems from different manufacturers, and consist of resin and hardener that must be mixed in a precise ratio (by weight or volume). The Microship project mostly uses West System products, though for the solar panels we consumed a few gallons of a special low-viscosity resin from Composite Materials, optimized for high-temperature applications. West is perhaps best known in the marine marketplace, with the other big name being System Three; there are ardent devotees of each. (I have a stack of very informative System Three how-to books; paypal $6.00 to wordy@microship.com if you'd like one; I need to reduce excess!)

Whatever the system, the chemistry is similar: when the resin and hardener are thoroughly mixed, an exothermic polymerization reaction results. Long molecular chains start to become cross-linked, and eventually the pot of goo will get warm and gel, which is your cue to stop being so finicky with that one persistent fiberglass cowlick and finish quickly. You can push the envelope quite a bit if you’re working with a thick filler mixture, but if you’re trying to wet out cloth the time limit is more severe. This is strongly affected by the shop temperature (up here in Washington we get longer pot life than we did in California), as well as the particular choice of hardener—West provides type 205 (slow) and 206 (fast), which can be mixed in any ratio before in turn being mixed 1:5 with the type 105 resin (changing this ratio is not allowed, and will result in either a brittle layup or a horrible uncured mess). Depending on the mix, you have 15-30 minutes of working time before it gels (and about 9 hours to a sandable partial cure; overnight is generally assumed unless you get up early, skip the morning email, and keep the woodstove stoked). Measuring the ingredients is typically done by volume using pumps, but when it’s cold this is physically difficult; we switched years ago to an Ohaus triple-beam balance and carefully monitor the tare weights of empty containers to compensate for spills and the recycling of mixing pots.

The “completeness” of the mix is critically important, and the standard method is to use tongue depressors with one end sanded square to get into the corners of the unwaxed paper or plastic mixing cups. Any little bits of resin that are not commingled perfectly with hardener will be forever gooey and require removal, a most unpleasant task.

Let’s see, we’re already up to a stock of resin, two flavors of hardener, cups, pumps, sticks, gloves, and a scale… plus acetone, methyl ethyl ketone, and isopropyl alcohol for various cleanup jobs. That’s just the beginning…

Filler Materials

When you whip up a pot of pure epoxy (which we have come to call neat, as in a glass of whiskey), you have a low-viscosity material with the approximate consistency of corn syrup. This is suitable for all the basic layup requirements: bonding fiberglass cloth to foam, adding layers to an existing structure, or vacuum bagging. But quite often, you want the goo to have some body of its own. For this, you use fillers, and there are a variety of choices depending on the application.

Earlier, I mentioned micro-balloons. These are tiny, non-structural bubbles that are used as a thickener for filling low spots or getting a more complete bond to a textured surface such as foam. This filler is available in phenolic or glass form, though it’s a bit of a pain to blend as it wants to blow all over the place and is so light that it runs away from the tongue depressor like a strange fluid. There’s also something called “microlight” from West that is somewhat easier to use—but any of these are good for fairing (smoothing a surface) and other applications where you’re trying to achieve beauty, not add strength.

One of the more annoying characteristics of epoxy, even filled with a thickener, is that it tends to sag on vertical surfaces and run out of places where you really, really want it to stay put until it hardens. This can cause some amusing scenes in the shop as someone tries to compensate for poor planning by frantically shoveling goo back where it belongs, unable to get away long enough to conjure an impromptu dam or to refixture the project in a way that renders gravity less traumatic. A useful additive to reduce this tendency is colloidal silica (or Cab-o-Sil), a synthetic silicon dioxide powder that magically renders epoxy thixotropic. We often mix this with other filler materials to change the flow characteristics without otherwise having much effect on consistency.

For general bonding and filleting, we turn to fine cotton fibers. This stuff is light and easy to work with, and since it’s fibrous, it has fairly decent structural characteristics once hardened into an epoxy matrix (compared to, say, little glass bubbles). As more and more of this is added to a pot of fresh mix, the consistency progresses from syrup to honey, thence to mayonnaise and eventually peanut butter. You can take it all the way to bread dough, if you want… this allows thick buildups for things like mounting pads without a tendency to droop or run away.

When the objective is maximizing compressive strength, there are various high-density fillers that lend themselves to hardware bonding and similar applications. The choice of materials becomes a judgment call, based on experience with the material and a bit of destructive testing; inevitably, despite our simplistic opening assumptions when stocking the shop, we ended up with an inventory of pretty much everything in the book (and had to re-order almost all of it at one time or another).

In practice, we found ourselves doing custom blends for nearly every job, factoring in the characteristics of different filler materials to optimize the mix for a given application while not losing sight of secondary issues like ease of sanding and weight. On big jobs early in the project, a whole pot of goo would typically be used for one task (like forming a long fat fillet between two foam panels glued at right angles), but when it came to dealing with countless little bonding jobs we would start with a runny mix carrying only a few fibers, inject some of it with a syringe to pot threads, add a bit of colloidal silica and more microfibers to form a buildup for a mounting pad, then blend a load of micro-balloons into the dangerously warming leftovers to use up the remainder on some always-available surface filling job in the endless quest for a decent finish. It becomes a sort of dance, choreographed by a list of tasks on the blackboard categorized into type of mix, all crammed into about 30 frenetic minutes.

During those times, we don’t answer the phone.

A Suite of Fibers

During this whole discussion of polymers and powders, I’ve been alluding to fiberglass—which, one could argue, is the central component in all this. Naturally, there are a lot of variables here as well (Figure 4).

Our basic building material was a 100-pound roll of 48-inch wide, 10-ounce (per square yard) cloth generously donated by David Berkstresser and completely consumed in the construction of Wordplay and Songline. The roll was mounted on a pipe fixtured above one of the shop windows, and directly below it was a dedicated 4x4-foot cutting table (thus appearing to visitors as if we were making boats out of our curtains). Known as bi-directional cloth, this material has half the fibers parallel to the selvage edge and the other half perpendicular. We used a Sharpie, a yardstick, and some good scissors to cut all the pieces, and kept a bag of scrap that often served up just what was needed for patches or other small jobs (sometimes so small that we would tease out individual fibers and apply resin with a pipe cleaner or artist’s brush).

Fiber orientation in most places on a boat is not terribly critical, but it comes up in two ways. First, sometimes, like around the mast step, you care deeply about stress distribution, and the final structure can be significantly lighter if you think it through and ensure that the applied forces run “with the grain.” Second, these thin strands of glass are quite flexible, but they still don’t like sharp corners. Orienting the cloth on the bias so the fibers approach sharp bends at a 45-degree angle solves the problem.

But this is just the vanilla cloth. Sometimes all the stress is along one obvious axis, and accommodating anything else is just dead weight (along with the epoxy that fills it, representing at least the weight of the cloth itself, and that’s if you’re very good at doing a proper layup). This kind of situation calls for a more specialized cloth known as unidirectional, or uni in shop parlance, which has 95% of its fibers running one way and only 5% used to keep the rest together. The result is maximum strength along the long axis with a minimum of overhead.

Another common fiberglass material (which we avoided completely) is mat—nonwoven fibers oriented randomly and sold as a thick pad-like material that shreds easily. This is useful in some large repair jobs, but seems most associated with projects whose weight and elegance are not critical (hot tubs, shower stalls, and mass-market powerboats). Be careful with this stuff, for in addition to causing very heavy layups, most of it (including the sexy-looking “X-Mat” that somebody shoved my way before I knew better) contains a polyester-soluble binder that is incompatible with epoxy. The resulting layup might look like a good bond, but the glass fibers are captured in the plastic matrix more mechanically than chemically unless you switch to vinylester resin.

Figure 4: Four flavors of fiberglass cloth (contrast-enhanced for weave clarity). From left to right: 10-ounce bidirectional, 4-ounce bidirectional, unidirectional, and X-mat.

All of the above are available in various weights, widths, weaves, and thicknesses—some bi-directional cloth is so thin that it virtually disappears when wet out, and is ideal for creating a tough abrasion-resistant waterproof seal over a wood surface. This material is commonly known as deck cloth, and one product is only 1.45 ounces per square yard. Having spent months trying to get all the boat surfaces smooth enough to look good with glossy polyurethane paint, I can attest to the desirability of getting as close as possible with the basic layup if you possibly can. In our case, we used lightweight cloth occasionally when needing more fiber flexibility for detailed structures, but didn’t really take advantage of this approach since there are so few clear expanses of deck.

If you want to buy some fiberglass cloth, the current eBay listings are at the bottom of this page. Ain't technology wonderful?

So far we have only discussed glass fibers (yes, it’s really glass). This is great stuff, and even reasonably affordable at about $5/yard in the standard 4-foot width. But for more demanding applications, one can achieve much better performance.

Kevlar (aramid fiber) has a much higher strength-to-weight ratio, with a tensile strength that makes it the material of choice for bulletproof vests. For a given thickness, it’s about half the weight of fiberglass, but 2.5 times stronger. But this comes at a cost: cutting it is a major pain (it eats scissors, requiring special $60 models to do the job), and sanding a surface turns it to fuzz instead of making it smoother since no mere sandpaper is going to abrade the fibers. Some kayak builders have developed a very high-performance sandwich that consists of an inside layer of Kevlar bonded to an outside sacrificial layer of conventional fiberglass that’s easy to repair—the former has the tensile strength to resist holing; the latter behaves better when dragged across a beach of fractured oyster shells. Kevlar costs 4-6 times as much as glass and would be overkill for most of our work, although the original canoe hulls are made of it and are thus much more likely to survive the occasional clumsy encounter with an underwater rock than they would be otherwise.

The other high-tech fiber that is available for composites fabrication is carbon—ultra light and strong, made of extremely fine filaments. Where weight is critical, this is the material of choice; it’s particularly good for kayak paddles, windsurfer masts, and other things that either have to be manipulated by human muscle or represent significant weight aloft (where the moment of inertia multiplies the effects of mass at the end of a long arm). I have a sample honeycomb-core panel with tri-axial carbon skins, and it’s so light that it almost wants to float out of my hands and hang in the air…

As you can see, there are a lot of choices to make, and a composites fabrication shop quickly becomes a rather large investment (and we haven’t even talked about tools, abrasives, molding supplies, fixtures, hardware inventory, paint, metal preparation chemistry, and all the other essential stuff that makes a 3,000 square-foot building seem not so big after all). Let’s take a quick tour of the facilities and the tools I consider essential for a project of this nature (after a brief commercial for relevant books).

Project Facilities

It seems, at the outset, to be a rather simple problem. After all, people have been building boats in backyards for decades, and it can be argued that motivation is much more important than having the ultimate shop. But a large-scale composites project pushes the envelope of complexity, and some of the funkier shade-tree construction methods would, alas, be inadequate.

Our lab is a 40x56-foot heated metal building with integral benches, enclosed dust-control areas, electronics lab, fiberglass fabrication area, roll-up door, machine shop, inventory shelving, and an upstairs office and video/publications facility. The boats themselves, two little 19-foot micro-trimarans, are end-to-end down the middle.

I have seen some amazing boat and airplane projects come together in rickety facilities, but it really does make a difference in attitude if you start with adequate workspace and a good set of tools. Ours are listed below; your mileage may vary...

Tools

The Microship project called for a substantial upgrade to my battered collection of tools, some of which dated back to the ‘70s. Fiberglass jobs require a prodigious quantity of abrasives and specialized cutting implements.

Here’s our inventory of shop tools, all of which have been used on this construction marathon, some heavily enough to be nearing their “last legs”:

• a copy of the essential book, Gougeon Brothers on Boat Construction
  
• Recirculating industrial air filter, 2-stage (Grizzly)
• 16-speed floor-mount drill press with cross-slide vise, 1 HP, 5/8" (Foremost)
• Sheet-metal brake/shear/roll(Grizzly)
• Belt sander/grinder (Rutland)
• belt (4") and disc (6") sander (Delta)
• Swiveling bench vise (Wilton)
• Air compressor, portable tank, and 100' hose (Campbell-Hausfeld)
• Bead blasting cabinet (Grizzly)
• Table saw, 10” (Foremost)
• Makita 4" Angle Grinder
• Laminate trimmer, basically a small router (Ryobi)
• Palm finish sander (Makita)
• Sawzall (Milwaukee)
• Jigsaw (Ryobi)
• Makita 1 1/8in. x 21in. Variable Speed Belt Sander, Model# 9031
• Random orbit sander, 5” disc (Bosch)
• Cordless drill (Makita)
• Electric drill (Ryobi)
• mini disc sander (Dayton)
• Walking-foot heavy-duty sewing machine (Sailrite)
• Dremel tool (we're on our third one)
  
• Hot glue gun (we're on our second one)
  
• Industrial strip heaters for foam bending
• Refrigerator/freezer to suspend epoxy cure and extend brush life
• Oxygen-Acetylene portable welding outfit
• Vacuum-bagging facility (pump, table, epoxy trap, consumables)
• Homebrew anodizing tank and power supply for quick jobs, (although alodining is easier)
• 5 "rolly carts" of different types for material handling
• Hydraulic jacks, block and tackle assemblies, ratchet straps, and other gadgets for manipulating heavy things
• 2 homebrew boat workstands on casters
• Huge pile of spring clamps and other fixturing devices
• The usual sprawl of miscellaneous hand tools, bits, blades, abrasives, etc.

It could go on quite a bit further, believe me, but we had to draw the line somewhere. Despite chronic techno-lust that keeps me drooling over machine tool catalogs, trips to Grizzly in Bellingham, and the tool departments of local homeowner meccas, at some point we had to stop building the lab and start building boats. As such, we depend quite a bit on nearby resources for TIG welding aluminum and stainless, serious lathe and mill work, electric discharge machining, anodizing/sealing of aluminum parts, panel silkscreening, and other “big stuff” that requires serious equipment and/or skills far beyond my own. Continuous shop enhancement is a dangerous and expensive obsession…but I would consider the above list a fairly minimalist collection of necessary tools for a composites-based project of this scale. (Note that a milling machine and lathe are conspicuously absent; I miss trusty old "Cecil be da Mill" from the bikelab, but have managed to get by with crude methods, hacks, and help from experts… it was downright humbling to watch what Bob Stuart could do with the table saw, a laminate trimmer, and a bit of creative jigging.)

This is a must-have tool for complex fabrication projects... I had never used one of these before, but now can't imagine working without it: a long, skinny belt sander that goes where no sander has gone before. In addition to the obvious fiberglass sculpting applications, I frequently chuck it in the bench vise and use either a standard sanding belt or one made of 3M non-woven abrasive (from McMaster-Carr) to do aluminum finishing with much more agility than the various fixed sanders in the shop. It's an expensive machine, but well worth it... click on the photo to pick one up from Northern Tool:

Stock Materials

Of course, that’s just the start. Every job involves not just a tool, but something to whack with it; as Microship fabrication progressed, we amassed quite an inventory of parts and raw materials. To help you plan cabinetry and shelving for your project, here’s a quick look at the major supply categories:

Massive inventory of stainless-steel fasteners (many thousands of parts, but never enough ) Stock bins of aluminum, machinable plastics, stainless steel rod, scrap, etc. A variety of woods, foam core materials, veneers, and other sheet stock 100-pound roll of 10-oz fiberglass cloth; smaller inventory of other weights, unidirectional fabric, and so on. Many, many gallons of epoxies and hardeners A half-dozen flavors of filler materials Countless paints, adhesives, lubricants, anti-sieze compounds, sealants, and other “goo” Stainless-steel wire rope, thimbles, sleeves, anhydrous lanolin, heat-shrink, and swaging tool Shelves full of marine supplies, hinges, mounts, blocks, cleats, widgets, gizmos, and framuses (frami?)

One of the most frustrating problems in a complex one-off project is finding the random odd bits of metal needed... we spent entirely too much time prowling scrap yards for sheets and extrusions various. There is now an excellent online solution:

When we moved to an island in the Pacific Northwest from the heart of Silicon Valley, my primary concern was the loss of ready access to goodies—what would I do without Halted Specialties, ham radio stores, local tech-savvy vendors, and yes, even Fry’s? There’s a lot of good stuff in Seattle, but it’s a pain to leave the island, so we just hang out here in the forest and work on projects. But wherefrom cometh the goodies?

Perhaps the greatest single discovery was McMaster-Carr, with a distribution center somewhere down around Los Angeles. Their huge catalog, with all pages available as PDFs on their website, has been our central source of hardware, materials, and tools for this project—and they are fast, often managing next-day delivery for ground UPS rates (they ship so much stuff to Boeing and its contractors that there is a daily container load transported via air courier to Seattle!).

For nautical supplies, we depend quite a bit on West Marine as well as other chandleries in the area, the most interesting of which is the venerable Marine Supply and Hardware in Anacortes. One of the nice things about being in the Pacific Northwest is that we’re not the only people crazy enough to build boats around here, and there are a number of resources nearby for quality woods, fiberglassing materials, polyurethane paints, and other stuff that would be hard to find locally in, say, Nebraska. But between the Web and a few high-profile mail-order suppliers (like the two biggies in the kit airplane business, Wicks Aircraft Supply and Aircraft Spruce & Specialty), it’s not too difficult to track down just about anything in the composites department.

All this to just outfit a canoe for a river and coastal expedition! Lewis and Clark would be scandalized, and we haven’t even peeked at the electronics side of the lab yet (where things get truly out of hand). But hey, part of what drives this is a strange obsession that would render a journey aboard a non-geeky boat somehow unfulfilling.

And if you've read this far about immersing yourself in toxic goo and itchy fibers, you know the feeling. I hope this has been a useful introduction... may all your layups solidify... but not too soon!