Home           Road Map     Resources     Lutherie Tools

Weekend 4: March 26 - April 1

The Rosette from the LMI boxed materials set comes in three parts: two 35" strips of dark/light/dark veneer sandwich about 1/16" thick and 1/8" wide (the dark/light/dark shows on the 1/16" face), and a pre-formed circle of Martin-style herringbone purfling 1/4" wide and about 4" inside diameter. The ring is a little more than a full circle in its relaxed state, that is, it is not joined and the ends overlap, so it can be inlaid at that size with a little cut off the end, or pulled open to a larger size. The procedure is to rout a channel in the face of the top, surrounding the intended location of the soundhole, and to inlay the strips and the purfling ring into the channel. Well, that's how it came out. The intention was to rout a 1/8" channel for one double circle of veneer strip, then 1/8" further out, a 1/4" channel for the purfling, and then another gap and another 1/8" channel for veneer strips. 

I began by following the figures in Cumpiano and Natelson, starting with a 4" soundhole, 1/8" of wood, an 1/8" veneer strip ring, and so forth. When I routed the channel for the purfling ring, I naturally tried it for fit. It would not open out that wide. The most it would do was a 4-1/4" diameter, and I was asking it for another 1/8" beyond that, which was enough to leave a wide gap that would not be covered by the end of the fingerboard. I retrenched, and routed down the wood between the first and second channels, leaving a 1/2" channel into which I would inlay all three parts of the rosette. It came out fine; but next time I'll measure all my materials before working on any of them!

The process in the pictures, then, shows the start of a three-piece rosette ring that finishes as a one-piece. If you see anything odd, that probably explains it!

The veneer strips are too brittle to bend into a ring, so they need to be pre-shaped. Cumpiano and Natelson suggest soaking in hot water, then bending into a prepared mold with channels the same size as those in the top. I didn't have anything long enough to soak them in, and carrying hot water downstairs would be a problem anyway. I decided to try one of the specialist tools I got for this project: a bending iron (click here for the tool description). I soaked the first few inches of the strip in a bowl of cool water, then worked the strip against the iron. The iron was hot enough to sizzle the water, and the strip bent readily. As you can see here, I fed the strip in with my right hand and pulled it around the iron with my left. All it took was about two seconds sliding back and forth by about 1/2" at the contact point, and then moving over to the left another 1/2". 

Because I needed two circles on either side of the herringbone, I just continued curving the whole 35" strip, and then curved the other strip.

I had prepared a board by routing the channels I thought I was going to use, and used this as a mold to let the curved stripping set.

With the strips loaded in the mold, I placed another board over them to keep them from popping up, and put them aside overnight.

I used the template I had made by gluing the plan cutout to a board to trace the outline of the top onto the spruce sheet from the materials. I then cut the shape of the guitar top out of the sheet, leaving a 1/2" border all around. 

The materials come with the top wood in two "bookmatched" pieces; that is, they are two consecutive slices from the same log. To bookmatch the pieces, you place them in the orientation they had in the log, then open them like a book and glue the two near edges together. It yields a symmetrical grain pattern, and, for an instrument soundboard, a symmetrical wood structure. When you order a top, you can have LMI  sand them to a given thickness and glue the pieces together. These operations are tedious and they require either extreme skill with a plane or a heavy-duty thickness sander, so I opted to pay for the thicknessing and joining to be done for me. I also had them thickness and join the back, and thickness the sides.

This top is made of Sitka spruce, a wood native to the American Northwest. It has a very high strength to weight ratio and acoustic characteristics that make it the most popular wood for steel-string guitar soundboards. When I "tap test" the top wood by holding it near the edge and rapping it with a knuckle of my other hand, it has a bell-like sound. It is a beautiful piece of wood. I'm now about to take a chance at not wrecking it.

I agonized for a while about how to put the channels in the top for the soundhole rosette. I thought of grinding a special cutter for a fly cutter, but the one I have isn't big enough, and besides, if you look at it, it seems awfully risky. The thought of modifying my Porter-Cable router to cut circles as recommended in Cumpiano and Natelson  didn't appeal to me. That's a lot of horsepower against the delicate spruce of the top. LMI offers a specialized router harness for a Dremel tool that does all kinds of magic for lutherie...but it's 125 clams, and I've already spent a lot on this project. In addition to which, LMI takes forever to ship anything.

Then I noticed the lowly plastic router base I had gotten for my Dremel long ago. If you take away the fence from the router fence attachment, and leave the part that rides on the rails, you find that that part rides level with the bottom of the router base, and it contains a hole right through the center of it that perfectly accommodates a 1-1/2" finishing nail or brad. Swiveling on this nail, it becomes an excellent circle cutter. I bet it's even described in the instructions for the Dremel router base that I never read. I tried it out on the scrap board that became the mold for holding the veneer strips to set in their curved shape.

The rings come out of the mold on Sunday morning quite content to be rings. They are shown here with the LMI-supplied herringbone rosette. Using the Dremel tool circle cutter, I cut the first 1/8" channel and fit the smaller veneer strip rings into it, cutting the strip to make two separate rings so as not to have a bump where they overlap. Then the 1/4" channel for the purfling, when I notice that the purfling ring won't stretch to fit it. After shaving down the wood between the two rings, I have approximately a 1/2" channel. I rout down against the inner circumference of this channel, shaving about a 32nd of an inch each time, until all the rings fit into this channel.

Here's the result, messed up with glue squeeze-out. The gap in the herringbone ring will be covered by the end of the fingerboard. 

It begins to look like part of a guitar, no?

Now the glue has dried, and the rings must be scraped level with the soundboard. I clamp the top to the workboard -- backwards, so as to have the soundhole close to me -- and set up a light so I can see the high spots as I work. I'll use the scraper rather than a plane or an abrasive, because the scraper is easier to control than the plane, and I want to scrape the rings without too much scraping the top, and an abrasive might drive dark-colored dust from the ring material into the light top wood.

The scraper takes fine shavings from the ring material. Soon the rings are level with the top surface. This took about ten minutes of scraping. After scraping, the area is smoothed with 120-grit sandpaper on a block, and then lightly scraped again to remove any dark material from the spruce.

If you use a fly cutter to take out the soundhole, don't be alarmed when the soundhole climbs up the drill bit as it cuts through!

The next step in Cumpiano and Natelson  is to do the top braces, but I have the Dremel-router out and I know I eventually have to inlay the decorative back strip. Here is the gorgeous East Indian Rosewood back material. I really don't want to mess this up. You can see the results of bookmatching on this piece: the light, slightly curved figure streak about 1/6 of the way from the edge repeats in mirror image on the other side. The material for the sides is closely matched to the back in color and figure. Luthier's Mercantile Inc. promises first grade materials in the boxed set, and in my admittedly inexperienced opinion, they certainly deliver.

LMI offers two grades of boxed materials for steel string guitar, differing mainly in one having mahogany back and sides and the other rosewood. The other differences are in the ritzier decorative materials in the rosewood set. The rosewood set comes with Martin-28-style herringbone purfling and rosette, and the back strip features an elongated checker pattern of maple and rosewood pieces. Picture a little further down!

The Dremel router base comes through again, with one flat edge offering the perfect bearing surface for routing a long straight groove by following a long straightedge. I'll retake this picture another time -- it's a little shaky. I find that not using the flash I get better colors, but I really have to steady the camera with a tripod or some other support...I skipped that this time

Routing a long straight groove to take the decorative strip that disguises the joint between the two halves of the back. The strip is about 3/16" wide. In order to get a snug fit, I start with a 1/8" router bit.

The initial try at routing to full depth sends up smoke, so I back off the depth a bit, and take about .02" at a time. Any more than that burns badly. This rosewood is dense, which makes it a stiff and resonant reflector of sound. After getting to about .07 inches (the material is .11) I need to increase the width of the groove. This I do by adding successive layers of tape to the bearing edge of the router base. This adds less than 1/64" at a time, so I am sure I will not go over and get a sloppy fit.

Here's what I meant by "elongated checkerboard". The strip is glued into the groove and I'm scraping it down. You know you've learned to sharpen a scraper when you get curls like this. I saved a bowlful of them to make people guess what they are. The discoloration on either side of the strip is from acetone that I wiped into the groove (and slopped over). I did that on the advice of a book called "A Guitarmaker's Manual" by Jim Williams. Seems the oils in rosewood can keep glue from adhering properly. Wiping with a paper towel lightly moistened with acetone certainly brings a lot of gunk out of the wood, and a bright purple color. Cumpiano and Natelson  don't mention this. I worry about the discoloration, but it scrapes away once I get the strip down.

I had to get in my weekly macro photo of a curl of wood. This came from the scraper, not the plane, and it shows the rosewood having no trouble adhering to the glue that sticks it to the back strip.

And here's the result.

And now for something completely different. The wood for making the braces that help to strengthen and stiffen the underside of the top comes in a billet split right from a log. As you can see, it is split along the radial lines of the fibers of the wood. To make braces that will have the best acoustic properties and greatest strength, you should have wood that has no run-out of the grain, i.e., the fibers should continue through the piece to the greatest extent possible given the shape of the piece. The way to get that is to split the material along the grain down to the approximate dimensions of the piece, and then saw and plane to the final dimensions.

The normal way to do this would be with a froe, a specialized tool for splitting wood. It has a long blade with a handle mounted perpendicular to it. You drive the blade into the end grain of the billet, then pull the handle to the side to turn the blade and widen the crack. I found Sunday afternoon not quite done when I glued in the back strip, so I went over to the Woodworkers' Club to see if they had a froe. They carry two models, an out-in-the-woods he-man one and a small one for splitting basket strips. I figured on the small one. Well, they didn't have either one, so I settled on the method pictured. I started a crack about where I wanted it by driving a chisel deep into the side of the billet near the end, then walked the crack down the billet by leapfrogging two chisels down the crack. Apart from having whacked my thumb four  or five times, it worked like a charm.

Having accomplished this glorious split, I decided to call it a day. Actually, I went back and scraped the back strip down as described above, now that the glue was set, but that's telling the tale out of turn. Next week, I'll lay out the bracing pattern and make the braces and maybe get them glued onto the soundboard.

Previous     Next

Home           Road Map     Resources     Lutherie Tools

Copyright 2001 Stephen Miklos