Home

Furniture Gallery

Individually Styled and Crafted

Fine Hardwood Furniture

by

Andrew Pitts ~ FurnitureMaker

Work in Progress
File Cabinet

Article completed 22 March 2008

Disclaimer: This discussion chronicles the making of a piece of furniture in my workshop. My intention is to bring the reader virtually into my shop to generally see how I do the work and share in my thought processes. Although I try to point out how the tools work and the applicable safety considerations, this discussion is not intended to be a text on how to work with tools, nor how to execute operations with the tools shown. I am not providing instruction in woodworking methods. Woodworkers attempting to imitate my methods do so at their own risk.

Click on the images to enlarge them and see the detail

My workshop has a room in which I do my finishing, and since it is nearly sawdust free it makes a great place for my desk, as well. But the cleanliness of the space belies the clutter of my desk! I needed a file cabinet! Now, earlier in the year I had made a rolltop desk with nice file drawers, and later on I made my Chest sans Sides which won three awards, so why not combine the ideas into a file cabinet. And while I'm at it, why not build a second cabinet that I can show in galleries? So off to the computer and my CAD program, and this is what I came up with.

The cabinet has two file drawers on full extension ball bearing guides, plus a four inch drawer in the center for saving paper and such and a cubby hole just below the top (a nice place to keep the phone book). As in Chest sans Sides, this piece also has no sides .... the drawer sides, themselves are the sides of the cabinet, making it easy to see my dovetails! And of course, the legs are curved to keep the viewer off balance just a little.

The first part of the cabinet I tackled was to build the legs. Since I use DesignCAD as the tool to design all my pieces, it was easy to print out the leg profiles full size (taping the pieces of paper from the printer together) to make templates. Then I cut out the profiles and transferred the profiles to poster board to provide some beef to my templates. Using some 4" x 4" red oak, I laid out the profiles, seen at right, so that I could later bandsaw the shapes, seen at left. In this design, the curves of the front legs are pretty radical, so 4" stock was not enough to make the entire leg. In fact, I had to cut another 4" block and glue one to each end of the basic leg stock to make the last eight inches or so of the stock actually 4' x 8". The resulting blanks looked a little like a block letter "C", if you get my drift. For this straightforward gluing I used my old standby, Franklin Titebond Original. That is a great glue for work that will stay dry and when a long setup time is not needed, plus it's pretty inexpensive. It's just the good ol' yellow glue with which most

woodworkers are familiar. One operation had to be done before the bandsawing, though. If you look closely at the rendering of the design you will notice that there are horizontal pieces of wood about centerline of each drawer that are let into the leg insides. These are the "guide rail housings", or the piece of wood that I will screw my drawer extension slides into. Since they are let into the legs, I had to rout out slots for them while the leg blanks were still square. Had I waited till after bandsawing the curves, it would have been very hard to rout accurate slots. So, I laid out the locations for the slots and then devised a little template from a scrap of 3/4" Medium Density Fiberboard (MDF- a building board material made of compressed and glued random wood fiber - I use it for templates and jigs). To the right you can see me routing the slots, and to the left you can see the template clamped in place over the slot after removing the router.

After the slots for the guide rail housings were routed and the legs were bandsawn to rough shape, I had to smooth them to a uniform surface because I would be gluing cherry faces on opposite sides (the fronts and backs of the legs) to create a laminated look. The cherry would darken with age and contrast nicely to the red oak, which hardly darkens at all if left natural. I used a spokeshave for much of this work, since the odd shapes of the legs did not lend themselves to any mechanical approach. The photo to the right shows me working one of the faces of a leg with the shave.

The cherry faces on the legs were to be about 3/8" thick, and at first I had hoped that the curves of the legs would not be so radical as to prevent bending that thickness of cherry. But my first attempt at gluing faces to the oak blanks resulted in a failed glue joint. Now, and this gets interesting, the failure was not entirely due to the thickness of the cherry bending around a curve. In fact, the leg that failed had a rather easy curve. I was using Unibond 800 glue, a very stiff two part resin glue used in applications where "creep" is undesirable. Creep is where a glue joint slowly slides one way or another over time, and glues like Titebond (remember the old shop standby?) do exhibit creep. When making a curved lamination, you don't want creep. Anyway, back to the story. Unibond 800 is a very good glue, so I was puzzled why the

glue joint failed, so when in doubt, I went back and reread the instructions. There, in bold letters was the statement, "DO NOT USE UNIBOND 800 BELOW 65 DEGREES AS PROPER CURING WILL NOT TAKE PLACE". Well, duh! It was winter, and to save fuel I was cutting the temperature back to 60 degrees at night, and I had made the glue up late in the day. So, the glue had not been allowed to cure properly. My solution was two-fold; first I kept the thermostat at 70 degrees through all the gluing, plus I made the 3/8" thick laminate out of two 3/16" layers. Two thin layers bend easier than one thick layer, and you get twice the glue surface to keep the bent laminate stiff. Those two changes did the trick and I was back in business. The photo to the left shows two of the legs being glued - lots and lots of clamps were required!

After the legs were all glued and cured, the excess cherry was cut off flush with the oak using a ball bearing flush cut bit in the router table. My bit is carbide tipped, like most router bits, today, and that is a good thing, since a glue line will quickly cut a notch in a steel cutting edge, but carbide will hold its own. At right you can see how this cut is done. With the excess cherry laminate cut off, I used the spokeshave to relieve the sharp edges, used planes and a card scraper to smooth the faces, and sanded all around to 320 grit. Later, after all the fitting of parts is complete, but before gluing the cabinet together, I will sand to 600 grit and apply shellac polish. At left you can see how the spokeshave was used on the edges.

With the legs completed I next focused on the panels that would serve as top, sub-top, dust panels, and bottom. These were made as simple frame

and panel assemblies, using sycamore for the frames and red cedar for the panels. There is a story behind the sycamore I used here. We had to have a sycamore tree cut down at my Church because it was dying and was located right along the road and had electrical wires running through the branches. If the tree had fallen in a storm, and there are lots of storms in Eastern Virginia, it would have been quite disasterous, so down came the tree. Now, since I have a sawmill I decided that furniture would be a nicer end for the tree than firewood, so I carted it home and milled it. That's a story in itself -- I pulled at least seven big 20d nails out of it. I think might have served as the community sign post for years, located right along the road as it was. Anyway, the wood looked great off the mill. The lacy pattern was spectalar! But, when I dried the lumber, most of the boards split along the annual rings! OUCH! It seems that the tree was weaker than we expected, possibly due to a cement truck hitting it some years ago (or so the story goes). All I got from the tree was short pieces, but they were ideal for the frames in this piece, so the file cabinet gets the spectacular sycamore! Back to the building: the advantage of using frame and panel construction with solid

lumber is that the panel insert (the red cedar) can expand and contract with seasonal moisture changes without splitting, as the panels float in grooves cut in the inside edges of the frame. I often use open mortise and tenon joints in the frames, as I like to see the joinery. The photo at left shows the cutting of a mortise using the table saw and a tenon jig. Cutting the tenons is similar, except that the jig must be adjusted a bit so the blade cuts off the tenon cheeks. The panels that fit in the frames are cut to the opening size plus an allowance for the rabbet to be cut all around the panel so that it can fit in a groove to be cut inside the frame. Once all that cutting is completed, the frame and panel assembly is glued up, as shown to the right. These assemblies are very straightforward to make, and they serve the purpose well!

These frame and panel assemblies had to all be made a bit different in size, depending on their locations in the cabinet. Also, the frame thicknesses varied depending on the panel. Before attaching the panels to the legs, I had to make the guide rail housings on which the ball bearing drawer extension slides would be attached. These pieces also served to hold the front and back legs of each side together to simplify the final assembly. Once they were made and attached, and with the side leg

assemblies together, it was a much simpler matter to align the panels to the correct heights so they could be attached to the legs with dowels. First, I had to figure out where the dowel holes needed to go, so I used blocks to hold the panels at a certain height, then marked the location of the legs on the panels and similarly the location of the panels on the legs. Then I drilled holes in the panels for the dowels. I used metal dowel points to mark the corresponding locations on the legs for the holes. A dowel point is simply a piece of metal shaped like a dowel with a pointy end. The dowel point is placed in a dowel hole in the panel, then the panel is held up to the correct location on the leg and the leg is pushed into the point, which then makes a mark where the dowel hole needs to be drilled in the leg. This works well if you are very careful with the fit up. With the bottom and the two dust panels attached in this manner, the sub-top could be screwed to the tops of the legs, then the spacers (shaped like extensions of the legs) between the sub-top and the top could be attached with screws, and then the top could be screwed to the spacers. The photo to the left shows the completed assembly.

Now it was time to make the drawers. These drawers would be unique for three reasons. First, the file drawers would have the ball bearing extension slides embedded into grooves in the sides of the drawers so that they would be hidden when the drawers were closed. Remember, this is a cabinet without sides, so I had to take positive steps to hide the less attractive, but necessary parts of the cabinet. Second, the fronts of the drawers had to have a curve reflecting the curve on the front of the legs. And finally, since the cabinet has no sides, the drawers would have to be made very attractive - of course,

I always do that, which is the reason I designed pieces without sides in the first place! Now, the sides of the drawers would have to be quite thick, like an inch, to have enough meat for cutting the 1/2" deep groove for the extension slides. By now you are probably wondering what I am talking about, so let me give you a sneak preview of a completed and mounted file drawer and I think a picture will be worth at least a thousand words! The two photos at left show how the ball bearing extension slide fits into a recess in the drawer side as well as how it fits onto the inside of the guide rail housing. You should definitely click on the photos to enlarge them, since the thumbnails just don't show the detail.

In order to keep the drawers as light as possible I chose a light weight wood for the sides. I already knew that the fronts and backs of the drawers would be cherry,

to go with the legs, so I chose tulip (yellow poplar) for the sides. This wood gives a nice contrast to the cherry, making the dovetails stand out, and it has a nice heartwood that changes from greenish to brown with age. These pieces of wood had to be about 12 inches wide, and since I did not have cherry available in the required width and thickness, I glued up the fronts from two pieces, and did the same for the backs. I happened to have some wide tulip, so no gluing was necessary. After rough milling the stock, I let it stand for a good long time while I worked other projects. That let all the wood pieces equalize in moisture content to reduce movement in the pieces after assembling the drawers. The photo at right shows the drawer stock with the glue up of the fronts in progress.

When I was finally ready to construct the drawers, I finish milled all the pieces to exact size and then set up my dovetail cutting jig. At left you can see me flattening the backs of the large drawer fronts, making them ready for cutting dovetails. Now, I'm not a purist. I will cut dovetails in a manner befitting the piece at hand. If I were cutting only a couple of dovetails, or cutting them in curved or other unusually shaped pieces, I would cut them by hand. However, in these two cabinets there were to be a lot of dovetails, and large ones at that, so logic held that router cutting them made sense. I contend that if, in the old days, the masters had machines to do the work, they would have chosen machines over hand methods if the quality was not compromised. That's my story and I'm sticking to it!

After cutting the dovetails I constructed the drawer bottoms. In this project I used red cedar, as I often do, for the bottoms. In fact, for the large

file cabinet drawers the bottoms are actually two wide slats with an air gap on either side of the pieces. That made it so the slats could change dimension seasonally very easily, and a file drawer does not need a solid bottom, anyway. The center drawer, however, needed a continuous bottom so I glued up a couple of pieces of cedar to make the wide sheets needed. Then, with everything fit properly, I glued up the drawers, shown at left. After the glue dried, I continued shaping the drawers to fit. At right you can see me using a special wood carving tool on an angle grinder to remove a lot of meat fast to make the curve in the drawer fronts. After using that tool, I used

increasingly finer grits of sandpaper in my circular sander to make the curve gentle. I also had to fit the sides so that they would move nicely in the carcase and so the extension slides would fit the file drawers. The center drawer is guided by a 3/4" x 3/8" x 18" center guide rail attached to

the underside of the dust divider above the drawer. The drawer back has a notch cut in the center top to ride on the guide rail, allowing the drawer to close without hitting the sides of the back legs. The photo at left shows the rough fit of the drawers in the carcase. Finally, I fit rails to the insides of the drawers for the file folders to hang from. The photo at right shows how I constructed the rails. I made red oak holders into which I glued 3/16" steel rods and screwed the assemblies to the insides of the drawers.

With the drawers fit, I did the finish sanding of all the drawers and applied the shellac polish to the drawer exteriors. Of course, I had already finished the interiors before I glued up the drawers. The question of drawer pulls plagued me, as I had some pretty wild ideas of how to do them, but I finally settled on a modest red oak cleat type design that was doweled to the fronts of the drawers. And that was that. I quickly put my file cabinet to use in the office, and I am happy to say that it does the job nicely and with style. The other cabinet is ready to go to the gallery and eventually become part of a very progressive looking office!