Slipstream Table (Experiments in Aluminum and Plywood Lamination)

The Slipstream Table is the culmination of a number of side projects and experiments in CNC profile cutting, lamination, sheet good manipulation and traditional woodworking. The project is the result of the economy and efficiency of cnc fabrication, as well as the realities of the hand finishing required after these products are processed by these high-tech robots.

Inspired by the aesthetics of mid century modernism, the name 'Slipstream' is inspired by the tapered aluminum points at the ends of the top and legs of the table. For me these features evoke the imagery of Airstream trailers, early fighter planes and motorcycle fairings all crafted out of aluminum in a fight against the forces of air resistance in the endless pursuit of aerodynamics.

The table is a finished product that took numerous ad hoc twists and turns throughout its conception and creation. It is a result of the process of serial profiles of birch plywood being cut and being laminated together to create a objects that is unexpected and visually dense. This Instructable outlines the process by which this table was created, as well as failures and triumphs that were discovered throughout the nearly three week build time.

Before embarking on the full scale table fabrication, I wanted to test the fabrication and design qualities of this process through a smaller project. In this case this project was a small end table that was built using the same construction processes. I will not go into any detail in this step, other than to say, when beginning a large project that will require large amounts of time and materials, tests are always worth their time and effort in allowing you to feel confident and comfortable moving onto the primary project.

The design of this table was governed by a number of factors, chief among them: aesthetics, machine processes, minimal waste strategies, precision, and structural stability. In the context of aesthetics, I looked towards precedents from the modern movement. These pieces of furniture with tapering features and subtle angularity were crafted with hardwoods and traditional joinery, but the lines and figure of the pieces could be distilled, reproduced and carried over into the world of digital fabrication.

The physical elements of this project were fundamentally simple, there were no complex shapes or cuts, and so it was important for me to make sure that the lines and design of the object were specific and satisfying. The stacking language of the extruded construction method allows for a pure object, and with this clarity, the quality of the lines is easily transmitted.

I toyed around designing the project in both 2d and 3d. Between sketching, drafting and modeling, the language of the smaller table was carried through to the larger full-scale project. These details included tapering edges, angled and tapered legs and the ridge between legs. These details allowed me to create a table with subtle features, that could be immediately understood. But also an object that had more depth upon further investigation.

As mentioned earlier, material efficiency quickly became a focus of the project. Upon looking online, I found a great number of inspirational and intriguing examples of laminated plywood furniture and sculpture. Something that continued to nag me when looking at these projects was how their laminations came at the price of efficiency. Creating swooping forms that occupied countless sheets of ply was indeed romantic as a product - but the lack of nest-ability and the implicit material waste associate with these forms made me shy away from any form that could not be built efficiently.

The concept behind the nesting and material use of this table was as follows: the tables large and cumbersome end pieces used the majority of a sheet of plywood, and did not allow the left over material to fit other pieces of the table. However, in this leftover space, there was room to fit pieces of a smaller piece of furniture. In this context the idea became to nest a bench in the off-cuts of the cumbersome end pieces.

This model did end up working on the first sheet, though eventually I realized I could break the leg pieces into segments and re-attach them with glue and clamps after they were out of the cutter. This separation and re-attachment model lead to even more efficient nesting of cuts.

Here at Pier 9, we have the luxury of having access to a number of large format profile cutting tools, including an Omax Water Jet, Coherent MetaBeam Laser and Shopbot CNC Router. For the purposes of the wood portion of this project, I opted to use the Shopbot for the following reasons:

Firstly, the shopbot uses a router bit that allowed me to specify a specific and accurate hole size that I drilled into my layers of material. This specific bit size and corresponding hole size (in this case 3/8") allowed me to effectively thread the layers of material along a finish dowel, making precise indexing much simpler in the assembly stage of fabrication.

Secondly, the shopbot cuts with a blade and not a laser or water beam, this means that the cut you get off of the shopbot is clean, unmarred and uniform in color. This is particularly important in the context of this project, where the cut quality defines the quality of the surface, so it was vital that the tool with the highest quality of cut was selected to ensure the most uniform surface quality.

The wooden core of the table is constructed from 42 individual laminations of 3/4" inch thick void-less birch plywood. Each of these laminations are cut out on the Shopbot 3 axis CNC router. Starting with 4'x8' sheets of plywood, the components were cut out with the aforementioned nesting strategy. This cutting process was both interesting and frustrating for a number of specific reasons.

The largest hurdle was a function of the nesting strategy itself. Because the cuts were so densely arranged, the typical 'tabbing' strategy of securing the cut pieces to the existing material was not possible. The cuts could not be secured to the wood itself because there simply was not enough left over material. In order to secure the wood then, a vacuum table would have been ideal. Unfortunately this was not a feature we had on our Shopbot. Instead a system of work-holding was devised that involved the use of a brass screw and plastic washer holding the pieces to the table, all inserted into the existing holes drilled for the indexing. The holes were drilled first, then the screws and washers were fixed, then the ultimate profile cutting happened around the screws.

This fixturing and cutting process was time consuming, and there was some trial and error involved. The main issue was that I originally bought 1-1/2" screws for the job, these ultimately proved to be too short, as they only grabbed into the mdf base 1/2". On multiple occasions these screws were ripped out of the mdf and sent unsecured pieces flying into the router bit, cutting into the pieces and destroying them. After the first sheet the screws were exchanged for 2" screws and this issue was resolved. One issue with the 2" screws was that they were driven into the plywood backing sheet, with an impact driver these screws tended to shear off leaving a useless nub of brass that had to be removed with pliers. Not as elegant solution in practice as it was in theory.

After the shopbot cutting was complete all of the pieces needed to be cleaned up before they were laid-up in order to ensure that there would not be abnormality in the profiles. Uniformity would lead to void-less surfaces without high-spots.

The pieces were first taken to the router table and cleaned up with a flush-trim router bit, then any other high spots were hit with 120 grit sand paper.

Before heading in with wood glue and clamps the first step in construction was trying the dowel and hole method dry. I got about 1/3 through the table before I realized there was a problem. Between the imperfect straightness of the finish dowel and the unfortunate reality of the inaccuracy of the shopbot, the stack was not lining up in the perfectly robotic error-free, so this is the time when I must admit an unfortunate truth about the project.

So I have to admit something here: my master plan of drilling, screwing, doweling... was a flop.

In order to ensure the most precise layup, the pieces were referenced against the face of my work table and ultimately glued, brad nailed and clamped. The nails were almost surely overkill, but in order to register each piece to the next without clamping and laying up in sections, it fastened the pieces together during glue up.

The body of the table was completed and glued up. While it dried, the leg assemblies were also glued together. Each leg assembly is comprised of three layers of wood. These three layer assemblies were then attached to the table using compressed dowels, glue and a whole bunch of clamps.

Once the core was fully assembled it was a misaligned stack of plywood. I was pleased with the structural stability of the table, but the finish was really quite terrible. I had to come to terms with the fact that the robot had failed me yet again. With this realization, I armed myself with an arsenal of belt, orbit and Fein sanders and got to work.

Sanding... It's not a quick process. Using 80 grit paper on a belt sander I must admit I spent over a dozen hours getting the surface uniform and all brought down to the lowest point. This end-grain surface is full of glue, and requires a ton of work to sand down.

The finishing was done with 320 grit paper on a random orbit sander.

The aluminum caps were incorporated for a number of reasons. Firstly, the use of the bright metal helped to break the monotony of the plywood-on-plywood aesthetic of the table. Secondly, the face of birch-ply is soft wood and would not hold up well against years of use, so the aluminum serves as an advantage in terms of durability. Thirdly, this project was undertaken in the context of the pier, and so it was interested in taking as much advantage of the tools available to me as possible. In this light it made good sense to add a waterjet-cut metal component for the 'because we can' element.

Using the Omax waterjet cutter and 1/8" aluminum, the exterior profiles of the table, as well as the interior leg profiles were cut.

There were a few potential ideas for fixing the aluminum to the wood. Fixturing with screws or bolts was an original idea, but the interruption of the surface felt like a misstep. Ultimately gluing became the answer. It was decided that contact cement would be the most viable adhesive for the metal-wood connection.

The metal was roughed up with a pneumatic sander and cement was painted on both sides of the connection. You only get one shot with the contact cement, it has to be laid down perfectly in place, then fixed in place with pressure, in this case a rubber hammer. After 30 minutes of curing, the layup was preformed and luckily went off without a hitch.

The reality of all the sanding of the wood was that the entire profile of the table lost about 1/16" on all sides, since the aluminum cutting was based on the original dimension the aluminum needed to be routed to fit the final profile exactly.

Using a trim router and a flush trim bit the work was fairly simple. The only part of the process that was precarious was routing the ends of the table and legs where the router did not have a sold base to reference off of, and so there was the danger of digging into the material and a more than 90 degree angle. In taking my time I avoided these gouges.

I decided to hold off on finishing the wood until after the aluminum was applied in order to not mar the finish while routing and applying the aluminum.

Once all the pieces were attached and trimmed flush, I applied five coats of polyurethane semi-gloss and finished by wet sanding by hand with 600 grit paper.

Thanks for taking a look!