Monday, February 20, 2012

Hexagonal Table Made on the 5-Axis CNC Router

Here is a hexagonal table I designed and built as a test of 5-axis CNC wood joinery. It was cut in the Digital Fabrication Lab at the Taubman College of Architecture at the University of Michigan.

My idea for this project was to use a joinery technique that is difficult to make using conventional woodworking tools (table saw, sliding compound miter saw, etc). I choose to use all compound angles (non-right angles cut in two directions at once) with loose mortise and tenon joints. I'm using the 6mm x 40mm "biscuits" normally for use with the Festool Domino. The top is joined to the legs with dowels. There is a spline joint used between each side of the legs.

Here's the design (see-through so you can see the joinery):

Here are a few parts showing the joinery - on the top is a side, on the bottom is half a leg:

The wood I'm using is quarter-sawn white oak. It's stable when cut - and the figure evokes mission furniture which is a precursor to the look I'm after.

The top of the table sits in a rabbet in the sides. It is not secured to accommodate wood movement.

Here's the top, dry-fit. You can see the rabbet and storage available beneath the top. 

Here are some photos of the C.R. Onsrud router. This machine has a 12 h.p. motor.

In order to cut these pieces on the router, fixtures were required. These are simple CNC cut Medium Density Fiberboard (MDF) pieces, with grooves cut to allow the vacuum pressure form the table to be distributed under the workpiece thus holding it in place. The grove at the edge of each fixture holds a rubber gasket (removed in this image). It prevents the vacuum pressure from escaping the fixture.

The smallest piece - the stretcher parts between the legs - required some pins to hold the work from rotating. On work that small the vacuum pressure is not sufficient. For all other parts the vacuum from the table below was fine. 

Here are some parts, loosely dry-fit, before glue-up. You can see the knife edge on the leg parts. The spline joint was a huge help during the glue-up of these parts: 

Here's a video of the router in action cutting one of the six sides. You can see the vacuum table, the pods which lift the fixture and workpiece up to a sufficient height and transfer the suction to hold the board securely.

Sunday, February 12, 2012

Wolff Sleigh Bed Fabrication Details

Below are the fabrication details of a sleigh bed I designed and built for the Wolff family. The bed is made of Cherry and Tiger Maple.

I did most of the work at the Digital Fabrication Lab of the Taubman College of Architecture at the University of Michigan. I am currently a student there working on my Master of Science in Digital Technologies degree (to graduate in Spring 2012).

First a word of thanks - especially to Rodger and Patti Wolff who were fantastic clients. Also to everyone at Taubman College who were so helpful to me - Wes McGee, Maciej Kaczynski, and Ben Hagenhofer-Daniell. And a special thanks to my good friend, Bob Ongaro who was enormously helpful through the entire process!

The Wolff's wanted a clean design where the wood was the main feature. They also wanted to use two contrasting woods. We looked at a lot of different combinations before choosing Cherry and Maple. Here's the original computer model after several design iterations.

Pieces of Cherry were carefully chosen for grain match-up, cut to length, then glued up to form the blank for the sides.

The blank is secured to a 4'x8' sheet of plywood and placed on the 3-axis CNC router table.

Various cutters carve out the form. Here you can see the initial roughing done by a 1/2" wide straight cutting bit. This remove the bulk of the material.

Next a 1" diameter ball bit makes passes over the surface to smooth it out. Each pass steps over 0.07" from the previous pass.

The machine goes around and around until the entire surface has been smoothed.

A side fully routed. A small chamfer was added with a hand held router on the back side to soften the edge. That was inaccessible to a 3-axis CNC. The large chamfer on the outer face was cut by the CNC.

The next step is making the headboard and footboard from Maple. Each piece is curved so in order to fabricate them I needed to build curved fixtures - four in all - one for each face to be cut. Here you can see the medium density fiberboard blanks being glued up in the vacuum table. This needs to be shaped smooth.

Here the router does a similar job as was done to the sides. First the blank is roughed then a ball bit it smooths the surface.

Here's the fixture after have been smoothed.

Next a grid is cut to distribute vacuum pressure over the entire face of the blank.

This is the fixture for cutting the convex side. Holes are drill through the entire fixture to allow the vacuum pressure to come from the table below and into the grid.

A rubber gasket is inserted into the outer groove. This prevents the vacuum pressure from escaping. Also, paint is used to seal the edges of the fixture - again to ensure the vacuum is only applied to the part.

Next the flat pieces of 6/4 Maple are glued up, edge to edge, in the same way a barrel is made. Below the individual boards have been cut on the table saw to the appropriate angle and jointed. Here they are shown before before any clamp pressure is applied.

Back to the router... First a bit begins to cuts out the overall shape. Then the ball bit does the smoothing.

Finally the headboard is trimmed on its edges. These sit flush against the side pieces.

Here's the footboard being milled. The shape has been cut and the ball bit is smoothing the surface.

The finished piece.

This results in a lot of shavings. On top the Cherry, below it the Maple.


Here are the pieces during scraping and sanding.Eventually the parts were sanded to 400 grit between coats of finish.

The next task was to fabricate the hardware to join the sides to the headboard and footboard. An 1/8" piece of stainless steel is secured to the abrasive waterjet cutter.

Here's it is in action. Steel is no match for water and garnet abrasive under intense pressure.

Here's the finished part. The missing piece in the upper right was removed during the cutting to test the fit - it is one of the hooks and slides into the grooves cut into each piece.

These then need to be released from the blank and countersunk so screws will be recessed. The hooks then get welded into the plates.

The hardware is recessed into the headboard and footboard sides using a Dremel tool with a teeeeeny router bit and router base. The corners are cleaned up with a skew chisel.

All done while standing on a step ladder to reach the edges and to work horizontally. (In the background of the picture you can see the Fab Lab's 7-axis robot).

Here you can see the hooks of the plates. These latch into matching plates mortised into the sides. Note that the angle of the hooks have to be perfectly vertical in the final orientation in the bed. Otherwise the parts would not be able to slide together.  

And set up in the Wolff bedroom.