Friday, July 31, 2009

First Light XXIV

Moving right along with the barge boards and associated joinery here with this garden lantern prototype. Much of the time in this stage of the project I am making small adjustments, frequently refitting and checking, and cutting the joints stage by stage. It's more efficient of course to cut all the parts in one fell swoop ,and then connect all the bits together, however I am venturing into new ground in places here, and feel the need to tread carefully. The more hours one gets into the making of a given piece, the more is at stake, and I don't want to make any gross errors at this stage of the process. It can get a little nerve-wracking at times, and it's helpful to know when to take a breather here and there.

Last time I had completed the stub tenons, or mechi, on the ends of the wall plates and the ridgepole - here's the view from the backside:

Next step was to work on the mechanisms to draw the barge boards - hafū- tight to the beams. My idea here is to use a mechanism that draws the boards in horizontally to the beams. This starts with cutting some short sliding dovetail trenches in the lower ends of the hafū:

Look a little closer and you'll see them:

After these slots were cleaned out a bit, I refitted the hafū so I could transfer the lines from the dovetail slots to the nose of the ridge. Then I laid out the tapered mortises on the ridge pole ends which are to receive the wedges which are involved in the draw mechanism:

After most of the cut out of the wedge-mortises was complete:

Just a little more paring to clean them out remains - which will wait until I have have made the wedges.

Now for the draw bars - I returned to my pile of offcuts and dragged out another little piece of Bloodwood. This was a piece I needed to re-saw (re-taper in other words) on four sides to get the straight grain alignment I was after, followed by the usual jointing squaring and dimensioning by hand plane. Once at dimension, I cut them to length with a 240mm nokogiri, and then trimmed them by chisel, gang-style, to get them all the same length and with square ends:

A little fun with the router table later, the draw bars had their male dovetails cut, and were partially fitted into the barge board mortises cut earlier:

A closer view:

In the above view, the draw bar is only just started into position - another bit of trimming is required yet in the dovetail mortise to get it seated to final position.

A view from the end-grain side of a hafū board, showing the insertion of the draw bar:

Once the draw bards were fitted to a given pair of barge boards, I could check them in an assembled position to be sure that there was an adequate space between the two bars, one of which has been cut at the end to a half-lap (the other to follow shortly):

There's needs to be enough space for the bars to fit in their mortises, plus space allowance for them to move through the seasonal humidity swings.

Here's the slots in cut in the end of the ridge to receive the Bloodwood draw bars:

The draw bars need to be keyed to their slots so that the wedge cannot push them outward. This necessitated, as shown a couple of photos back, a half-lap on the end of each bar, which I processed with a saw cut and some chisel work:

Some scars in the end of my sawhorse cap will need to be dressed out when this project is done.

The slot mortises on the ends of the ridge needed mortises to receive the half-lap extensions of the draw bars. Blind mortises like these are not the easiest to cut cleanly - I start with a brad point drill bit:

It's handy to have a full set of long thin paring chisels sometimes - here my Tasai 6mm gets nosy:

Last one for today - a picture of the trial fitting of the barge board, the draw bar sliding close towards the half-lap mortise, and with the newly cut tapered wedge mortise also clearly visible:

I wonder if this will work?

Until next time then - please stay tuned and of course comments/questions/heckles are always welcome.

Thursday, July 30, 2009

First Light XXIII

More work unfolds on the 4 hafū boards for this Japanese garden lantern. The fitting of these boards are among the most complex aspects of the entire project and there are quite a lot of steps involved.

After the upper board support ledges were trimmed to fit against the ridgepole, I was able to fit up each board, to check position:

If things needed adjustment I did so, otherwise I marked the interface of the hafū and the end of the long side keta:

At the end of that day, I had the four hafū moved considerably along:

The next morning I set up a shooting board to trim the miters closer to the line:

Fit of the miter was looking clean at this point, though some adjustment may well be required further ahead depending upon how things sit once the assembly is on the keta and ridgepole:

A few steps omitted -- in the next photo I have the upper roof frame assembly, sans ceiling board, clamped to my sawhorse/workbench and I've begun work on the joints for the ends of the beams:

The keta have a central stub tenon, or mechi, which keeps the hafū from any tendency to slide downhill, along with a protrusion above which is meant to keep the piece from lifting up off the seat - here's a close-up:

Here's the end of the munagi after the initial cut lines have been defined (the knife lines are faint so I roughly ink just inside of them so as to not cut into the wrong spots):

The end of the munagi after the next stage of cut-out with the router, X-Acto knife and chisel. The vertical mechi are beveled to as to tighten in fit and drive the hafū inward as the hafū assembly is pulled inwards by the joint mechanism:

With the joinery on the beams moved along, I could fit the hafū and directly mark the location of the various mechi onto the hafū boards:

After the scribing, the cut out could proceed for the mechi mortises:

A rare opportunity to use the 1.5mm chisel to clean the 0.125" deep mortise:

Once the little mortises were cut, I could go back to trial fitting and adjusting the fit:

I have one side nearly complete, save for the cut out needed for the mechanisms that will draw the joints tight together:

This thread is ongoing, if not endless(!) and I will post again in a day or two - hope to see you then, at station 24 in the journey..

Wednesday, July 29, 2009

First Light XXII

In today's post I will show how I took the hafū, which I rough-cut out and trimmed to shape last time, further into their more nearly complete form, short of the joinery. I've spent much time puzzling over the best way to attach the hafū boards to each other and, simultaneously to the ends of the keta and the munagi. There are several ways to do it - beyond the obvious method of spiking or screwing them into place, which are not under consideration here - or using threaded rods to pull things together (also not under consideration) there a couple (or three) all-wood joinery methods. These methods basically devolve into two types:

1) the hafū attach by dropping vertically into place on male dovetails on the keta.
2) the hafū attach by moving horizontally onto stub tenons on the keta and munagi, using various types of locking keys to pull them into place.

I'm going with my own version of method (2) above, with a few elaborations. We'll see how it turns out, as the prototyping continues.

Okay, after I had my four pieces of hafū roughed out, and the lines transferred to them from the templates, I proceeded to give the router a good workout, working out on my sawhorse set up in the yard (trying to keep the dust in the kitchen down a little bit).

First off, I processed some slots on the backside of each board, which define the position of the two roofing boards:

Here's a more close up look at the slots on the two boards on the right side of the above picture:

Next I processed a couple more slots on the backside of the boards, but these aren't actually slots as such, they merely define a raised platform which will abut the underside of the roof boards. This platform will help to guard against any tendency the quarter-sawn boards might have, in the sun, to cup. Wedges will help this process as will soon be apparent. This detail should make more sense as the pictures roll out in this build:

Then with a change to a shorter bit, I free-handed the router and removed a good swath of waste material from the back, taking the cut close to the edge so the router could remain supported while cutting:

Also note that the board support ledges on the lower side have been shortened to about 2" in length - I don't want these ledges to be visible at the end view of the boards.

A closer view:

And yet closer:

An electron-microscopic view of the fiber:

(just kidding!)

Next I turned to the opposite (face) side of the boards and processed the waste out, leaving me this pile of rubble:

Then I sawed the edge remains of the cut portions off, and out came the hand plane (and, oh yes, off with the earplugs!) to clean the edge up a bit:

Figuring was on to a good thing, I decided to trim the lower end of each hafū to the line:

The finish was acceptable:

Time for the always controversial show, a definite thrill-a-minute, yes, it's Sawing for Teens!:

As these boards are representing two separate parts, one atop the other, the cut is to define the lower board, the actual hafū-ita, which is set back a short distance from the end of the cap above.

Following the sawing, I did a little paring:

And this is how it looked after the chips had settled:

Now, on the uphill end of the boards, I needed to trim the board support piece back a bit to fit onto a notch which will be made on each side of the each end of the munagi (ridgepole). This I did by chopping and paring:

The result:

Well, that's the 15 pictures I wanted to post today - more to come! On to post 23

Tuesday, July 28, 2009

First Light XXI

In the previous post, I had detailed the fabrication of the dovetail sliding keys to help affix the lower ceiling boards to the wall plate, or keta, of the lantern housing. I try and restrain the number of pictures per post to a maximum of 15 or so, in the interest of keeping the download times for those with slower internet connections to a reasonable time frame. So last time I had to finish just short of showing the attachment of the lower roof boards to plate; therefore, without further ado, here's how that went:

The fit was quite snug, and I needed to use a mallet to help tap - not pound - each board into place:

They went in quite nicely I thought:

From underneath, here's a look at a keta nose-to-roof board intersection:

That should look a bit better when I get around to cleaning up the end grain cut of the nose, however I'm happy with it fit-wise.

And just to refresh the reader memory, and for those who just stumbled upon this page and are wondering what the heck I am making, here's a shot of the lantern housing with the pair of lower roof boards fitted up:

It's a pity I couldn't have come up with a mechanism of some sort to get the lower roof boards to flap like cuckoo wings, as the top of the lantern does look at this stage somewhat ready to fly off somewhere. Yes, I speak in jest.

I was surprised actually at how rigid the connection was between the plate and lower roof boards using just a pair of keys each side. It wasn't as if I could do chin-ups off the board edge or whallop it with a 2olb. sledgehammer, but it is quite a strong connection nonetheless and in concert with the other means I will employ to put the roof parts together should result in a solid construction. We'll see, I guess.

I could have moved on to a number of other tasks at that point, and I chose to proceed with working upon the gable end boards. In traditional Japanese construction, there is a fair amount of careful detailing an elaboration when it comes to finishing off the edge of a gable (or, kiru-zuma) roof. There is a barge board which forms a vertical fascia at the edge of the gable - called a hafū-ita (破風板). Then, along the top edge of the hafū-ita runs a horizontally-oriented cap board termed a nobori-yodo (登り淀). More elaborate roofs might have two or even three nobori-yodo, each offset from the one below. In this lantern housing I am employing a certain simplification and exaggeration of the respective roof components to achieve a traditional look. This is what the Japanese do as well, from what I have observed in studying these lanterns. The nobori-yodo in this lantern roof is relatively chunky compared to one in a full-size building, and thus suggests not just the board itself but the ridge of tiles that runs down the slope at the gable verge (or close to the verge, I should say, as the edge of the gable is also normally rolled over as well, a treatment called the minoko). I also have decided, for simplicity's sake, believe it or not, to make the nobori-yodo and hafū-ita out of one piece of wood instead of two. I make this decision partly out of practicality/efficiency, and partly as a result of my decision from the outset of this build to eschew the use of metal fasteners or glue, if at all possible, in the construction. Making the gable end boards one piece allow me to make a better connection that will prove stronger, or at least that is how I am planning it!

So, the first step in making the hafū, as I shall term this amalgamation of the two components into one, is to make a full size template. I will use this template not only to mark out the cut sections from board blanks, but to use as a guide for my router, with which I will do much of the curvilinear cut-out and shaping.

After the layout was done on a piece of 0.5" MDF, I used a jigsaw to cut the template to rough form:

Then I clamped the template to the sawhorse, my 'work bench', and used a spokeshave to shape the rough cut curves smoothly, fair and to the line:

Once I had the template cut to shape, I stuck another piece of MDF to it using double-sided carpet tape and used a bearing-guided pattern bit in my router table to create a duplicate directly:

After cut out, I transferred layout lines from the original template to its doppelganger.

Then I took the template sandwich and used it to mark out the finished shape on the four pieces of wood which would become the hafū. These pieces of course had already gone through the usual round of sawing, planing and jointing to get them to the dimension I wanted:

Most ideally I would have used wood that was both edge-grain on the face and curvilinear (from a curved tree) however that option was not readily available, as they say, so I had to 'settle' for rift grain pieces that have a slight amount of curve in their lines, hopefully not so much that the grain lines will 'fight' the line of the pieces, and certainly there will be no issue with short grain in this case at least.

Next I ran the pieces through my deluxe Italian Agazzani band saw - whoops! Wait a minute! I don't have a band saw, at least not on this end of the continent. Oh well. Out came the jigsaw again and I trimmed the boards to their layout, leaving the cuts about 2~3mm wide of the lines:

After rough cut-out the piece was starting to look like something, perhaps a blank for a field hockey stick:

And then there were four hafū, all ready for the chopping and hacking to come:

The next step was to use the double-sided carpet tape, a godsend for the woodworker if there ever was one, and attach each piece of hafū in turn to the template sandwich, which I then took to my router table and trimmed the boards to the template. For this I used a combination of three bits, a template bit with a 1/16" (0.0625") offset between bearing and cutter diameter, for the initial rough pass; then either a top-bearing mounted Whiteside bit, solid spiral carbide, or a bottom-bearing mounted flush cutting template bit. I had to alternate between cutters, sometimes cutting with the board on top, sometimes with the template on top, due to my desire to cut with the grain, not against it. The use of the additional offset bit as the initial step eased the work for the other two cutters which followed and meant that I was always using the router cutters as they should be: to trim, not to hog waste. Here's the piece after the trimming is complete, still stuck to the template sandwich:

A little while later, I had my stack of 4 boards all trimmed to shape:

The next stage (which I completed before I separated each board from the template) was to transfer layout lines from the template to each piece at their edges, and then connect the lines across the faces:

Next installment in this series (post 22) should see the hafū through to completion. I hope you'll enjoin with me then, enjoyin' the joining.