Tuesday, August 30, 2011

Coffee Anyone? (22)

Post 22 in the series. I've been describing the design and construction of a joined hardwood coffee table with a glass top. Previous posts are to be found in the blog archive to the right side of the page.

Lat post, I finished off the assembly and then showed a picture of the turned leveler feet. These were the last pieces to make for the table and were largely made on a wood lathe. Once they were shaped, and each one is slightly different due to the way they were made, I fitted a short piece of threaded rod to the leveler:

Previously I had fit T-nuts to the counter-bore at the bottom of the table legs. The leveler feet go straight in and screw down:


The leveler feet allow for about 1/4" of adjustment at each corner, which should be plenty for the coffee table given its relatively petite footprint.

During the design phase of the project, I had gone back and forth on whether or not to fit these small pads to the bottom of the legs. They are a feature I like on many piece of Ming furniture, and I drew several iterations. In the end, after a bit of humming and hawing, I chose to go with the feet, especially since they allowed for the leveling function. Now that they are in though, I am surprised just how much visual effect they add to the piece as a whole. They really make the bottom of the legs with their little 'stirrups' stand out and give the piece more of a light feel. I'm really glad I went with the leveler feet and it goes to show that 3D drawing cannot convey 100% of how a piece will look when constructed and placed in front of you.

With the construction phase at a close, I did another few rounds of finishing work, culminating in a steel wool rubbing, and a couple of coats of wax.Then I fitted the glass, which I then sealed at the edge of the frame with a bead of clear silicone. I think the client chose wisely in opting for a grey glass, as it goes very well with the wenge shelf panel. Here are some unofficial 'studio' photos of the piece, with delivery slated for tomorrow:

I hope you'll agree that the tiny leveler feet add more to the piece than one might have suspected.

Here's a shot from the design phase:

And the completed piece for comparison:

A closer in look at a corner of the table:

A pic from the early design phase, with legs more radially curved, without feet detailing, and a thinner shelf:

The later drawing:

And the completed piece:

That draws to a close the build of Client L's coffee table. I hope readers have had an enjoyable time following along.

I'll be getting some professional photography done soon and will post up the pics on my other blog at a later date.

Thanks for coming by the Carpentry Way. Comments always welcome.

Sunday, August 28, 2011

Enter the Octagon (II)

In my previous post in this series, I took a look at octagonal bay window bump outs and how they might be configured. I also mentioned the matter of octagonal construction in general and the popularization of the octagonal house in the US in the mid 1800's, following the publication of Orson Fowler's The Octagon House: A Home For All, or A New, Cheap, Convenient, and Superior Mode of Building - shortly thereafter reissued as A Home for All, or the Gravel Wall and Octagon Mode of Building. This work was re-issued by Dover in 1973 as The Octagon House, A Home for All, a copy of which I have just finished reading. The cover:

It seems that Fowler's considerable enthusiasm for octagonal building was quite infectious in its day, and more than 1000 octagonal homes were built in the US and Canada, along with octagonal barns, carriage houses, even churches and schoolhouses. Of the survivors, most of them exist in right here in the state of Massachusetts, and I plan to start checking them out when I can. In fact, there's one right here in the town where I live, though I hadn't noticed it so far on my travels.

Octagonal buildings were hardly new on the scene - for example, many octagonal one-room schoolhouses were constructed in Pennsylvania in the first half of the 19th century, like this one:

Thomas Jefferson's summer residence, Poplar Forest, near Lynchburg, Virginia is another example of an octagonal house that predates Fowler's work:

What Jefferson envisioned and created (he was in his 60's at the time) was a classical villa in the ideal of Palladio; the final result is a combination of Renaissance, Palladian, and eighteenth century French with a smattering of British and American elements. Such mish-mashes are not uncommon at all in US architectural history. I had noticed references to Fowler's book in various other places over the years, including most recently in the book Architecture, Men Women and Money in America 1600-1850, written by Roger Kennedy, the director of the Smithsonian's museum of American history. Unlike many other accounts, Kennedy did not poke fun at Fowler and suggested his work had something to offer. So I took a look, and I'm glad I did.

What surprised me the most, coming in with slightly low expectations, was how persuasive I found Fowler's argument for octagonal construction. I'm now a believer!

Besides the octagon form itself, Fowler argues for what were at the time unusual amenities, such as dumbwaiters, speaking tubes, central heating and water closets -- also known as indoor toilets. Fowler believed that houses should be well ventilated and let in lots of natural light, for the physical and emotional health of their inhabitants. I doubt most of those features, or Fowler's philosophy of building in that regard would incur argument from anyone today. And I for one wish that dumbwaiters would make a comeback - that, and laundry chutes!

I'd like to summarize though the essence of Fowler's advocation for the octagonal form, an argument which emphasizes its economy. When you get down to it, the primary considerations in residential construction, after site, style, and materials, are the costs of construction on a square foot basis. Ideally, one would think, one would like to obtain a given squre footage with a minimum expenditure of materials, or, for a given amount of material obtain the most square footage - two sides of the same coin.

The intent of Fowler's book is to demonstrate a way of building for all, that is, a form of construction which was more affordable. The reasons why an octagonal form of structure might be more affordable may not be apparent at first, so I'd like to take a closer look.

We can build houses in a wide variety of plans, however, in North America at least, the most prevalent form is the rectangle. A typical suburban house, might be, say, 24' on one side, by 44' on another:

Two obvious facts: the square footage you get for that is 24 x 44, or 1056 square feet. The wall length required is (2 x 24) + (2 x 44) = 48 + 88 = 136 linear feet. 136 linear feet of wall, sill, plate, sheathing, plaster, baseboard, etc., looked at simplistically (ignoring the thickness of the wall, interior partitioning, etc.).

The interesting point is that the more pronouncedly the house becomes rectangular, the longer the wall required to enclose the same amount of interior square footage. Let's say, for example, the house was only 12' wide on the short end of the building, which would make for a long side measuring 88'. Same square footage (12 x 88 = 1056), but the length of wall required is now (12 + 12) + (88 + 88) = (24 +176) = 200' 

If dollars and cents, yen, or pesos, etc., mean something to you, it would appear that designing long rectangular houses, or houses which stretch out by means of wings and staggered jogs, is not a good direction to go.

Let's consider that same amount of square footage, 1056,  going the other direction of plan, namely in the case of a house which is perfectly square. To obtain the measure of the side of the house, we take the square root of 1056, which I will approximate to 32.5:

Same 1056 square feet obtained, however the total length of wall, foundation, exterior sheathing etc., requires (4 x 32'-6"), 130 linear feet. So, just in shifting around the plan configuration, I have saved some 6' total in wall.

That may not seem particularly dramatic, so let's look at it another way. Given the 136 linear feet of wall we obtained with the 24' x 48' rectangular plan, we could configure a square plan with the same total length of wall, namely 136 / 4 = 34' per wall section: 

If we look to see what sort of square footage we obtain, we find 34 x 34 = 1156 sq.ft., a gain of 100 square feet, just by changing the plan configuration. A 100 sq.ft. of floor area is quite a gain, more or less equal to adding an extra room in the house if one so desired.

Now let's take a look at an octagon plan and compare. Given a perimeter of 136 linear feet, and dividing by 8 sides, we obtain a figure with sides measuring 17':

How many square feet do we get for the same 136 linear feet of wall? The formula, which I won't spend time looking at in detail in this post is:

The side, s,  measures 17, and 17 x17  = 289. Therefore:

Area = 2(289)(2.414213) = 1395.4

So, an octagonal configuration of the same wall length increases interior square footage from 1056 to 1395 or so, some 1.3 times greater! Compared to a square plan, which yielded 1156 square feet, the octagon gives about 20% more square footage. That's significant!

Now I'm not arguing for building larger homes (as Fowler in fact does), but if one can obtain a given desired square footage and economize with the same exact materials one would be using otherwise, then I can't see any disadvantage to that.

One might wonder what the consequences of other shapes of floor plans might be. It turned out that the most efficient shape for maximizing area with a minimum measure of perimeter is the circle. It's no surprise that nature chooses this form, and the spheroid forms which develop from it for so many storage devices, like seed pods, eggs and so forth. Another aspect of the circular form is that the minimum of material can cover the greatest amount of square footage - hence the reason why the circular pizza makes the most use of a given amount of dough, and allows the fewest toppings to go the furthest as well.

To obtain a square footage of 1056 sq.ft. in a circular plan, we would need a radius of 18.3 feet:

The total length of wall required is only 115.19, quite a savings over the 136 feet required in the original 24' x 48' rectangular example we started with.

Considered another way, if we wanted to work with the same 136 lineal feet of wall in a circular plan, we would obtain a circle with a diameter of (136 / π) = 43.29' or so. Taking half this measure (21.645') as the radius, and using the good ole' "pi x r-squared" formula from grade school, we obtain an area of π (468.509) = 1471.86 square feet. That's some 40% greater interior space for the same wall length as compared to the rectangle we looked at first. So, clearly, the circular form is the most bang for the buck in terms of wringing square footage out of a plan.

That said, the circular form is not terribly practical for building unless you are talking about tipis, yurts, or igloos, etc. Working with timbers generally tends toward working with the grain in the material, which generally runs straight in the timbers we obtain building materials from (the softwood species). Cutting circular sections from straight timbers is obviously wasteful and trying to obtain trees bowed in the right radii and then working with them involves other challenges in the way. Or one could laminate up circular wall sills and plates, cut from plywood and glue up, etc. Those challenges generally add up to dollars and cents in the end, especially if you are trying to work with commercially available materials. Then, add to that the issue of windows and doors, which are not easy to make to fit curved walls, along with plastering, finishing, and furnishing the interior of a circular dwelling with modern appurtenances (none of which are round), and you can see further hassles ahead. It's not a practical form for wooden construction, generally speaking.

With polygons, the more sides you have the closer one gets to a circular outcome, however when you look at the choices, the octagon really is the most convenient. All carpenters are familiar with working with 45˚ and 22.5˚ angles, for one thing, so the lay out and cutting is not a significant challenge, excepting the roof work. A hexagon is not as efficient at converting wall measure to square footage as compared to an octagons, but the 30˚-60˚-90˚ triangle associated to a hexagon is straightforward to employ. A heptagon, one of my favorite polygonal forms, has an odd angular value integral to it (360 / 7 = 51.428571..., with the .428571... repeating infinitely), and the nonagon with 40˚ and 140˚ angles, offers no significant advantages. Same for the decagon. Polygons with more facets than an octagon, while becoming more efficient in producing interior space, trade back the advantage in the greater number of angled and mitered cuts of all the wall and roof materials to obtain the result. A zero-sum game it would seem.

In the next post in this series I'll take a look at octagonal buildings in more detail, east and west, and consider some of the aesthetic and practical issues which associate to these forms.

Thanks for coming by the Carpentry Way. Comments always welcome.

Friday, August 26, 2011

Coffee Anyone? (21)

Post 21 in the line up of episodes for this build series on a coffee table. You can find previous entries in the archive located to the right side of the page.

In the previous post I had pre-assembled a leg and drawbar to a pair of top frame rails, one long side and one short side. I made up two such pairs of assemblies.

The next step was to attach the remaining pair of legs to the stretched octagonal frame of the table shelf, one on each corner, diagonally-opposed to each other. The rubber mallet helped seat them most of the way:

Time for the grand assembly  - the shelf with its two attached legs is positioned in between the pairs of the leg &top frame rail assemblies:

Here's the leg which connects to the top frame rails starting to engage to the mortise on the shelf frame rail:

Meanwhile, the leg attached to the shelf frame is positioned between the top frame rails, like a sandwich:

Again, the clamping cauls come into play:

When the joint got close to being drawn up, I decided, after some consideration (but mostly paranoia), to add a dab of glue to the cross-wise floating stub tenon (not the drawbar, to be clear). I reasoned that in this case the form of joint would be more difficult to disassemble once the pins were driven in, and that the use of a little hide glue satisfied my paranoia about 'the worst that could happen', and could still be steamed and reversed later on if a repair were ever necessary (and the repair-person could figure out how the joint disassembled!). So, I put a dab of glue on a popsicle stick and applied a smear to the floating stub tenon (one side of which was already glued in with hide glue):

One of the guys upstairs of course took his opportunity to razz me a bit about using some glue, about the 'slippery slope' I was embarking upon, and yes, I actually felt slightly chastened for a few moments. But in the end, my goal has always been use glue minimally if at all, and in this case it is indeed minimally, so I feel fine about the 'compromise'.

The cauls were then drawn in to close the joint some more, and I alternated my attention from one end of the table to the other as this process went on:

While the cauls were tightened in alternate fashion, I also gave the legs connecting to the shelf frame a tap with the mallet as they required - this one is now about 1/8" (3mm) from closing:

I repeated the process until everything was drawn tight, then fitted and drive in the 4 shachi sen on each of the remaining two corner joins, just as I showed in the previous post.

Then, a little trimming followed:

Here's a look at the top of one of the corner joins - you can see a little rub mark or two from the cauls on the front edge of the profile:

A view of the inside of the same corner - a little attention is needed in that blobby area on top of the curve of the leg where the frame rails meet:

The joins of the legs to the frame rails came up tightly as well:

Another corner (with apologies to Adam Macer for this run of photos):

Panning back a little:

It's alive!:

A look at the join of the shelf to the leg, and the shelf's mitered locking joint:

A glimpse underneath, with a little re-oiling soon to come:

Final step in assembly was to drive in four pegs, komi-sen, to fix the four legs and their attached drawbar tenons to the shelf frame, using a bubinga scrap as a drift pin:

After the first pin went in a bit slowly and with more force than I would like, I decided to lubricate the next three pins with camellia oil to facilitate their fitting. That worked like a charm. Once the pins were in, I trimmed them flush and cleaned up the surface:

Another corner miter:

Any paranoia I had been experiencing about the integrity of the corner joints had dissipated completely as a result of driving the four shachi sen into place at the corners - the force of the joint being drawn tight was sufficient to force any residual hide glue up and out the top of the joints. They're very strongly connected, completely tight all around, and I really could have dispensed with the glue dab. Live and learn.

Did I mention? - It's alive!:

And next on my rapidly shortening 'to-do' list were the leveler feet. These I fabricated on a wood lathe and with aid of a sliding chop saw:

In tomorrow's post I'll show the installation of these pads and some more photos of the nearly-completed piece. I'm very happy with the way this table has turned out!

Thanks for coming by.

Thursday, August 25, 2011

Coffee Anyone? (20)

Twentieth post in a series. Welcome back to this description of the design and build of a bubinga framed coffee table. Previous posts may be found in the blog archive at the right side of the page.

Getting quite close to the end of the build phase. The next step in the 'to-do' list was to fit my maker's mark to the wenge shelf panel. I stated off by mortising the panel in the Rockwell Radial Ram drill press, which has no issues drilling in the center of quite wide panels:

I held off drilling full depth, as I didn't want the forstner bit's pilot to pop through the 1/2" thick panel. My next move was to apply some painter's tape to the area:

With the tape down, I then used a small top-bearing router bit to take the mortise to full depth - 3/8":

Then the holly flower could be fitted to the socket and the outline of the flower scribed onto the taped surface:

With the scribing complete, I removed the flower and pulled the tape off to reveal the housing to be cut:

Then I used a combination of two spiral carbide router bits to freehand cut the waste out:

I was well conscious of the risk of the cutting, as one slip could ruin the piece, and dis-assembly and possible replacement of the vertical grain wenge panel was simply not on the plan list. The cutting came out well however:

I checked the fit of the flower and made a few tweaks to the fit, then on goes the yellow glue:

Down goes the flower and I used a hammer and block to seat the carving:

Then I grabbed a clamp to squeeze any excess glue out from the middle:

Lastly I cleaned the glue off and let it set up.

With the inlay done, I turned my attention to assembly. The assembly sequence for the table parts is on a diagonal, and the first order of business was to fit pairs of table top frame rails together to their respective legs. In order to facillitate pulling the frame rails together around the leg's twin tenon and drawbar assembly, I decided to make a couple of clamping cauls - you can see one fitted on the right side of the picture:

Then I tapped one of the leg/drawbar assemblies sideways into place on the long rail:

With the leg assembly all the way in, I then brought the short rail into position, sitting atop its clamping caul:

I pushed the two together as far as I could, and then it was time to clamp, or cramp, as the English prefer to say:

The clamping cauls drew things together quite well:

Now it's time to fit the shachi-sen into place:

I drive the pins in, all four in quick sequence:

I had some apprehensions during the design phase about how tightly I might be able to seat the pins, thinking I might be risking some short grain issues, but these concerns evaporated as I was able to drive the pins in quite tightly and with no issue whatsoever. The joint drew up very well and the miter is completely tight, as I will show in the next post.

Once the pins were fully in, I used the Miyano to trim them flush:

Pegs trimmed, with some smoothing of the finish yet to come:

Here's a look at how the top of the joint looks, the normal viewing position:

I was quite happy with the way this joint came together, and repeated the process with the other pair of table top rails and their associated leg/drawbar assembly.

More to come, stay tuned. Thanks for dropping by the Carpentry Way. --> on to post 21