Building the Walnut Case

Regarding a case for the project, one option which has been discussed on the Spare Time Gizmos Yahoo Forum has been custom frames available on eBay. While the particular link to an eBay vendor that I found in the forum is no longer valid, you can find lots of such offers by searching eBay for “custom made shadow box frames”. Please read the thread linked above for issues though. For example, assembled shadow boxes might have trapped glass. Also one side of the frame would probably need to be removable, to allow the electronics to be removed. You might need a custom kit.

Rather than try to order something, I decided to build my own case. That led to the design discussed here. This section lays out exactly how you can reproduce it, if you wish. I can’t say that it was easy but I’m quite pleased with the results. The enclosure consists of a walnut case and an aluminum back panel. The front panel of course, is part of the FP partial kit. The power connector, power switch and serial port are located on the back panel. Features of the case include:

• Solid walnut panels
• “Tung oil” type finish (without actual tung oil)
• Removable bottom and electronics system
• Eight-inch depth (but the exact depth could be more)
• Dado grooves to enclose the front and back panels
• Front and back panels are inset 1/2”. (Helps protect switches.)
• Front edge has a “picture-frame” profile.
• Rubber feet

Case Design
The overall dimensions of the case are controlled by the dimensions of the faceplate (15 x 9.5”), the depth of the dados and the thickness of the board. Our cuts will be measured to the outside edges of the sides

The Board

I needed a nice piece of walnut, 8 x 60 x 3/4 inches. (Actual outside circumference is about 51-inches.) That translates to “FAS” quality, surfaced 4-sides (S4S). Buying walnut lumber is, of course, different than buying commodity woods. The lumber supplier that I went to near Atlanta didn’t have such a wide board in the bin and instead brought out a palette of unplaned lumber. We chose a piece (which I couldn’t tell beans about) and they ran it through the planer. For some reason, they could only offer surfaced 3-sides, which left one side of the board rough. It also varied in width but always well exceeded 8-inches. As far as thickness, they offered “4/4” (four quarters), which nominally comes out as 13/16”. Of course, after sanding, it will be slightly less.

The first step was to rip (cut lengthwise along) the bad edge in the table saw, as seen above right. Due to the long length of the board, a “dead man” support was placed on the far side of the saw. On the foreground side of the saw, a camera tripod was used to provide similar support. To keep the tripod from tipping, one of its legs had to be secured to a concrete block. These arrangements were necessary for good control of the board, needed to get a straight rip.

Next, I inspected the board carefully, choosing the best side of the board as the outside of the case and the best edge of the board as the front side of the case. That also involved looking at the ends to see what bad spots would be cut off. I decided that the pieces would be taken sequentially along the board, to preserve the continuity of the grain over the visible part of the case (as seen above left). The order of the pieces would be: bottom, side, top and side. As it turned out, this was somewhat spoiled, as will be covered later.

Ripping the Front Profile

I liked the molding which was used for the case in the Spare Time Gizmos photo (at left). It’s reminiscent of the bezel that the PDP-8/e/f/m uses (at right). I thought about using separate molding at the front, maybe painting it white but felt that having it all a solid piece of walnut would work better and wouldn’t be hard to make on the table saw. The profile shown below right was chosen and I was pleased with the results. The angle is 45-degrees. It only takes two rips on the saw. After cutting the 0.15” corner off, I decided to cut slightly more, so I increased it to perhaps 0.2”.

Initial Sanding

I had decided to do all of the sanding before doing the crosscuts. It’s easier to sand one long board, final thickness would be established (ref “Calculating...” below) and there would be fewer concerns about sanding variations affecting the mitered corners. However, I did find that sanding on the tip of the nose of the front profile (at right) affected flatness along its length. This is important because it must serve as a reference edge for the crosscuts. The broader back edge of the board fared somewhat better but sanding it inevitably (for me anyway) rounded it slightly. That made it more difficult to align a square to the edge by feel.

So when should the sanding be done for this project? Here are my answers, though I’m hardly an expert:

• It’s certainly easier to sand before crosscutting, so we want to do that, where possible.
• In any case, there will need to be a final light sanding after assembly, to remove any blemishes occurring in the process.
• The large flat areas of the board can be sanded before crosscutting.
• Sanding the diagonal facet of the front profile is best done before crosscutting. Incidentally, I found that this benefitted from using the simple 45-degree sanding rig at right. It takes just three clamps and a few minutes to set it up.
• Sand the tip of the nose of the front profile after assembly.
• Sand the rear edge of the board after assembly.

Note that only 1/2” of the inside of the case, at the front and back, is visible. Therefore, only the part near the edge of the inside plane of the board need be sanded. I sanded, proceeding through three grades of sandpaper: 100, 150 and 220 grit. However, I wasn’t quite as happy with the smoothness of the finish as I was with another walnut project, sanded to 400 grit. I had seen online comments which stated that wood like this could not benefit from finer than 220 and that’s why I stopped there. More recently though, I’ve read credible advice that, while they agree with that for film finishes, oil finishes like this can benefit from finer grit.

So if I had it to do again, I would probably choose a sequence that includes finer grades, perhaps 100/150/320 or 100/220/400. By the way, I highly recommend the sandpapers which claim 3X the performance of conventional products. They really made a difference in how fast the job went. It makes doing some hand sanding much more attractive. (You should be more careful though, that you don’t over-sand.) For most sanding, I just went with the same old cheap 1/3-sheet sander, which moves in small circles. The results look great to me. For some of the edges, I also did some hand sanding with a rubber, 1/4-sheet holder.

Cutting the Pieces of the Case

The cuts should be made with the good side facing up, for a cleaner cut. That means that the 45-degree miter cuts will alternate direction as seen at left. (Should be more space between cuts, though.) Hence, the board will need to be rotated 180-degrees between cuts. Notice that when the boards are assembled into a box, any error in the angle will result in a wedge-shaped gap, having twice that angle error.

It’s preferable that the error make the mitered corners sharper, rather than more obtuse. That’s because sharper (more acute) corners result in a gap which is on the inside of the joint, whereas obtuse corners result in a gap at the outside corner, which will be visible. I decided to offset the angle by about 0.5-degrees. Of course, one strives to keep the error small enough that the pressure from the clamps can bring the entire joint together, eliminating any gap. No way can you trust the tilt scale on the saw to set the 45-degrees. I went to considerable effort to set the angle of the blade, and found that name-brand drafting triangles were most accurate. It’s necessary to crank the blade to about max height to check it against a triangle. In spite of all the effort, I still was a little disappointed with the angle. I realized later that my once well-calibrated saw had drifted and the blade was slightly off in being parallel to the table.

At least for the first couple cuts, I found that pushing the initially 60” board through the saw was a bit unwieldy, requiring dead man support rollers at the side of the saw. Locking the board to to the miter gauge slider helped. With the first cut, you may want to remove some bad area at the end of the board. This is a tradeoff with the amount of bad area you can avoid at the other end.

Calculating the Length of the Pieces

Roughly, the idea is to wrap the case around the faceplate, which measured to be 15.0 x 8.5 x 0.19 inches. At right is a drawing of the case from the front, with the thickness of the sides greatly exaggerated for clarity. That defines the principal measurements, L, T and D. The dotted line shows the groove (dado) for the faceplate. For the moment, let’s imagine that there is no play between the case and the faceplate, so its edges are at the dotted line. We can calculate the Length of the top as: L = 15 + 2(T-D). (Sides will be 6.5” less.) So the length of the pieces we are going to cut depends on the thickness of the board and the depth of the groove. This is because we are effectively wrapping the part of the board between the groove and the outside, around the faceplate.

Adding Room for Play
Now we can add in some Play (P), giving L = 15 + P + 2(T-D). Notice that errors in thickness and dado depth are multiplied by a factor of two in the equation. If either is off by half of the play, the faceplate might not fit, so it’s critical that we have a good handle on those values. One issue that you don’t need to worry about, is the effect of sanding after the cuts, on the thickness. The thickness comes into play because we are measuring the position of the cut at the outside of the board, at the end of the mitered edge. Once the cuts have been made, the dimensions of the inside of the box are fixed. Those dimensions are what determine the fit of the faceplate. Sanding the outside of the box after the cuts will not affect the fit.

The amount of play needed depends on your skills at making precise cuts. Bear in mind that it is also burdened with allowances for errors in squareness of the box and changes in size due to clamping of the glue joints. I can generally hold an accuracy in cutting of 0.03”. However, I found that targeting a nominal play of 0.05 to 0.06” in this, uh, case, was hardly enough.

Choosing Dado Depth (D)
Since the depth of the dado is a factor here, let’s consider what it should be. My rough target was to make it about 1/4”. At that depth, the right-angle power connector is starting to get close to the side of the case though, so it can’t be much more than that. I decided that it would be nice to have it at least 3/16,” to handle variations in measurements, play and to provide plenty of overlap to hold the panel. It’s good to target less dado depth than max, to allow it to be increased to deal with problems after the cuts (but before the dados are done). For example, if you need more play, increasing D can provide it, while holding L constant. Another case might be that a piece has come out short and you would like to shorten the one on the other side of the box to match, so that squareness is maintained. Changing to a larger D allows a smaller L, without changing the play. I had started with a target D=0.27 and ended up pushing it to 0.31, when I had to shorten L to adjust miter angle. As a result, the power connector is quite close to the bottom of the case. With the minimum at 0.188” and the desired max at 0.25”, I recommend an initial target D=0.2”.

Target Measurements
The thickness of my board averaged about T=0.79”. If we allow play, P=0.06”, we get the length of the top (and bottom), L = 15 + 0.06 + 2(0.79-0.2) = 16.24”. The length of the side pieces (Ls) is 6.5” shorter, Ls = 16.24 - 6.5 = 9.74”. Note that if the thickness of your board is different, the cut lengths (or D) should be adjusted. Now you are ready to make the cuts: As mentioned above, you may need to position the first cut to eliminate some bad area at the end of the board. Then cut the bottom, left side, top, right side and top, as discussed. As each is cut, label it on the inside surface of the piece. By the way, there is no way that we could target the cuts, so accurately, that the faceplate and back panels would be held rigidly in place. Later, I’ll address how the panels are secured.

Cutting the Dado Slots

My first thought was to use the router table to cut the slots that will hold the faceplate and back panel. In fact, I even bought a slot cutter for the back panel slot. However, the crude router table which I had built has sagged in the center and made me worry about being able to control depth well. Since the depth affects play on a two-for-one basis, I decided to use the table saw for the dados, instead. It’s a good thing too, because I quickly found out that it was much easier to setup and run the cuts on the table saw, anyway. Advantages include: a nice, big, solid table; easy to adjust depth precisely; large, sturdy, well-made fence; results easily and precisely repeatable; dado width variable by moving fence; good blade cuts clean slot bottoms and clean edges. I used a 60-tooth, carbide tipped, 10” blade made by Kobalt and purchased at a home improvement store.

You can see the slots that I cut at right. The kerf (cutting) width of this blade is 0.11”, which is fine for the slot holding the 0.07” gauge aluminum that I used for the back panel. Of course, it would be very difficult to size the slots to hold the panels rigidly in place. The faceplate thickness measured 0.19” so it was necessary to cut that slot in three passes. For each pass, the fence is moved slightly. I targeted the slot width at 0.24”, leaving 0.05” clearance. This is needed because the sides of the box might not align perfectly. Also, I expected some irregularity in the cut and possible glue encroachment.

To make the 0.24” slot with the 0.11” kerf blade meant that I would need to move the fence a total of 0.13”. It was very tempting to try doing that in one step, hoping that the 0.02” lack of overlap would simply be swept out on the second pass. I tried that on a piece of scrap and ended up with the cutest little dado, having a paper-thin wall in the middle :) Okay, so much for the short cut. I moved the fence 0.06” for the second pass and 0.07” for the third pass. Worked great. By the way, if you’re thinking that it would be easier to do the dado cuts before the crosscuts, I would advise against it. Doing a long rip like that causes trouble in supporting the piece and keeping it steady. It was very easy to run the set of shorter boards though, once the fence was set. The slots came out very well aligned.

No doubt, you will want to mockup the box before attempting to glue it. Don’t make the mistake that I did, though. I setup the box on the table saw as at right, only without the tape. It seemed pretty steady but it wasn’t. The top casually fell over. Wouldn’t have been any big deal, except that it fell onto the teeth of the saw blade. It made the little holes seen at left (after finishing). No way could reasonable sanding take those out. Wood filler would no doubt show marks, with the fine oil finish. The only thing to do was to designate the bottom as the new top and live with the grain discontinuity on one side and the less desirable figuring of that part of the wood. Oh well; The best laid schemes o’ mice an’ men, Gang aft a-gley, An’ lea’e us nought but grief an’ pain, For promised joy. :)

Gluing the Top and Sides

We will be fastening the bottom only with screws to provide access to the electronics. Thus, only the two joints at the top side are glued. However, the bottom must be included in the clamped assembly during gluing, just as if it too were being glued. Prior to gluing, I used masking tape to protect visible areas near joints from glue seepage, as seen at left. Since I am using an oil-type finish on the natural walnut, any sealing action caused by excess glue might very well mar the finish. The tape extends along the full length of the joint on the other side of the boards.

I have to admit that this did help a great deal with the excess glue problem. However, I did experience the issue that the Titebond III wood glue seemed to react with the adhesive at the edge of the tape. As a result, it took some effort to get all of the tape residue off. I imagine that another type of masking tape (perhaps the blue stuff) would perform better.

Before attempting the gluing operation, I setup the clamps on the box, as seen below. One set of corner clamps was used to insure that the box would be square and two band clamps were used to apply pressure to the joints. I laid out the top and two sides end-to-end and applied a coat of glue to both sides of each joint, using a small brush. Following advice I had seen online I went back over it, applying a second coat. I was hoping that this would give me more open time, with which to setup the complicated clamping system. No such luck.

I was barely able to get it all together before the glue started getting pretty “grabby.” It did go together satisfactorily but the accuracy of alignment wasn’t perfect. I scrambled to wipe glue out of the dados, using foam swabs. Part of the delay was due to fumbling with the metal corner pieces while trying to tighten the band. I recommend preparing tape on the corner pieces before gluing, to hold them.

Holes for the Bottom Screws

After allowing the glue to dry for 12-hours, you can drill the holes for the screws in the bottom of the case. It’s best not to remove the band clamps until after the screws are inserted but go ahead and remove the corner clamps. Also remove the masking tape.

Locate the four holes on the bottom, as shown at left. Positions are 3/8” from the sides, 1.5” from the back edge and 2” from the front edge. Drill the holes to accom­modate #8 x 1.75” flathead screws (below left). Counter­sink the holes as at right. You can use a 1/8” bit for the pilot hole and a 5/16” bit to countersink. Also drill out clearance holes in just the bottom board, using a 11/64” bit. I recommend that you apply silicone (or other) lubricant to the screws. Insert them but don’t over-tighten. When the glue has dried for 24-hours, it’s time to take the wraps off! The fabricated box is seen below right.

Final Sanding

Assuming that you followed the sanding recommendations discussed in “Initial Sanding” above, the box is ready for these final sanding steps:

• Sand the tip of the nose of the front profile. This needs to go through all three grades of grit.
• Sand the rear edge of the box. I just used one or two grades of grit for this.
• Do a final hand sanding with the finest grit, on the front 45-degree facet and the 1/2” wide strips of interior, which will be visible.
• Motorized final sanding of the outside of the box, using the finest grit. Be very careful at the corners. You do NOT want to dull the corner. Rather, sand each side separately, allowing only a small part of the sander to go past the edge. That should take off any overhangs.

The Oil Finish

I chose an oil finish for these reasons:

• Easy and foolproof to apply. Applicator doesn’t matter.
• Dries quickly.
• Only takes one or two coats.
• No clear coat needed.
• Brings out the beauty of walnut.
• Can be renewed if necessary, without stripping.

Choosing an Oil Finish Product
The first question is, which oil finish to use? I once used a Tung Oil Finish I was happy with, made by Behr but that is no longer made. Note that in spite of its name, it contained little tung oil. The product that I found to replace it is Watco Danish Oil (above right), which is widely available. Such oil finish products are typically a combination of oil, varnish, solvent and dryers. They don’t dry to a hard film like varnish and don’t build up thicker and thicker as more coats are applied. Instead, they penetrate the wood and seal it, providing protection against moisture and dirt. Since there is no hard shell present, there isn’t quite as much protection against abuse but it compensates by being easy to reapply, if needed. Since it hardens in the wood, I suppose that there is some fortification of the surface. You can get Watco Danish Oil with colors added but I’ve found that applying the Natural color version to walnut results in a rich, fairly dark color anyway, as seen below.

Cleaning the Box
Before applying a finish, we need to remove all of the sanding dust. For that, I first wipe off as much as possible with a shop rag. Then, in lieu of the elusive tack rag, I use a clean rag, moistened (but not dripping) with turpentine. Don’t make the same mistake I did, by trying to substitute other solvents (like alcohol). The key here is to use something wet which won’t raise the grain, which we have painstakingly sanded smooth. Turpentine for sure fills that bill. After drying, I blow it off with canned air (wishing I had bought a compressor :)

Applying the Finish
With the box assembled, apply it to the bottom first. I used the handy plastic painter’s pyramids (below) to support the box while applying the rest. It’s possible to align the points with the Phillips bottom screws but I found that it wasn’t really needed. I just followed the directions on the can, using a cheap foam brush:

• Apply a heavy coat of Watco Danish Oil and let it soak-in for 30min.
• Apply another coat and let it soak for 15min.
• Wipe off all product. Dries in 8-10hr. You could apply another coat but I didn’t see much point in it.

Wax Job
Since an oil finish doesn’t provide a hard shell of protection like polyurethane does, a final wax finish adds a significant level of protection. Also, it will add some luster to what might otherwise be a matte finish. I would advise applying two to four coats of wax. Before applying wax though, make sure that the oil finish is completely dry. I’d wait at least 24-hours for that. I used the widely available brand at left. Just follow the directions on the can.

Congratulations — that completes the case!

Back Panel Fabrication

The back panel is shown at right. The design concept is to make the power switch easily accessible at the top of the rear panel items, centered vertically and close to the right side of the box. That makes it simple to reach around and operate. The serial connector and power connector come below, separated by one-inch, so plugging them without turning the case around is feasible. To make this, you will need a piece of aluminum sheet metal. I used some that I had obtained from a sheet metal shop long ago, having a thickness of 0.07”. The fabrication steps are:

• Cut a 15.0 x 8.5” rectangle from 70mil aluminum using a cutoff wheel on the table saw. Before cutting, you might want to check the actual inside dimensions at the the back of the box. If you made the dado depth D=0.2”, then the dimensions of the back panel should be 0.34” greater than the measurements. (That allows 60mils of play.)
• Clean up the edges with a grinder, file and/or sand paper. The cutoff wheel leaves a crust, spreading upward from the cut as seen at right but it is readily removed with a grinder, leaving a decent edge. The other edge shown is the cleaned-up cut.
• I have provided a template here, which you can download and affix to the left side of the aluminum panel as shown below right. Cut the left side of the template along the vertical line. Aligning that cut with the panel will place the items one-inch from the side of the panel. With the dado, the items will be 0.8” from the case.
• Drill the holes at the sizes marked on the template, except that the switch at the top may need a different size, depending on which switch you choose.
• The DB9 connector D-hole will need to be completed. You can use a jig saw to cut out the trapezoid, as shown immediate right. Note that the blade must be released to thread it through one of the holes in the D-area. After cutting the outline, a Dremel tool helps in tidying up the edges. The result is below.
• Sand the visible side of the panel with 220 grit to get a uniform surface. You should wear a respirator for this because it’s not good to get aluminum dust in your lungs.
• Clean the good side with alcohol, apply primer and black paint. I used the Rustoleum primer (far right) and textured black listed in the SBC-FP Enclosure Parts List. Primer is important here, since paint has a hard time sticking to aluminum.
      To speed the drying process, I bake in the oven at 150F for several hours. (Uh, don’t let your wife catch you doing that :) By the way, for some reason, when I first began to spray the primer, it formed bubbles in a small area, as you can see in the upper left corner of the primer picture. I should have wiped it off and started over but I guess I hoped they would pop and level out. Negative. That area remains lumpy. Woof. Life’s hard.
• For the DB9 connector, we’ll use the one which is on the AD2 adapter, specified on the SBC-FP Parts List as shown at right. Unfortunately, the connector needs to mount on the outside and the IDC-10 connector, on the other end of the cable, won’t fit through the DB9 hole. Take off the hex standoffs which hold the plastic shell to the DB9 and pull the shell back. Make note of the connections (which are only tack-soldered) and desolder them. Mount the DB9 on the outside using the hex standoffs and nuts from the shell. Then resolder the cable (after discarding the shell).
• (Refer to the pictures above.) Mount the power connector and power switch. The tip is positive on the connector. Wire the switch in series with the positive lead from the power connector. Twelve-inch red and black wires go from the switch and the negative lug of the chassis power connector. Solder the right-angle female power connector (P1 on the SBC-FP Parts List) to the end of the red and black wires.

That completes the back panel assembly.

Summary of Enclosure Materials and Measurements

Here is the SBC-FP Enclosure Parts List in PDF format and Excel 2002 format.

* Depends on the thickness of the board (0.79” here) and the depth of the dado slot (0.2” here). See text.

<<< Return to Part I - Overview, Documentation and Parts

<<< Return to Part II - Circuit Assembly, Software and Testing

Now at Part III - Building the Enclosure

Continue to Part IV - Final Assembly and Support Software >>>

Copyright © 2013 by Stephen H. Lafferty. All rights reserved.

_