Friday, June 8, 2012

Five-Course Mandolin Part 5

Binding is fussy. In an effort to stay away from plastic I've decided to use holly, which is a very pale, rather softish wood. A single black-white purfling strip (made from black-dyed pear and maple veneers) will suffice, and it does away with the necessity of routing two separate channels.
Here's my high-tech binding jig. The undulating nature of plates on a mandolin make it almost impossible to hold a router plumb to the surface. This little Makita laminate trimmer is fixed to a plywood carriage that slides up and down on two sturdy drawer slides. The device is bolted to the bench and instrument is presented to it in some form of carrier that keeps the sides perpendicular to the bench surface. I used the carving cradle with some blocks screwed down to keep things from shifting.

Here's the exciting end. A small rabbet cutting bit is surrounded by a disc of slippery plastic (a furniture glide).  The router only rests on a tiny area of the top plate about 3/16" (4mm) wide and thus does not get thrown out of plumb by the rising surface of the arch.  The ball-bearing guide on the bottom of the bit has been modified with the addition of clear plastic tape. I know it's ridiculous, but it lets me dial in the exact depth of cut without having to fumble through a huge and expensive stack of bearings in umpteen different diameters. Care must be taken to rout in the correct direction and I normally use a couple of light passes to reach full depth.  Obviously this is an inherently dangerous operation and hands must be kept well away from the bit and all safety precautions addressed.
Ah, if they were all this simple!  The gentle curve and supple binding materials make for an easy ride. I used regular masking tape to hold things in place. Having the binding channel correctly sized takes a lot of stress out of the procedure. If the binding is higher than the channel it can have a tendency to tip inward slightly, causing gaps.  One has to be careful removing the tape. Sometimes slivers of grain want to come up with it.

The back gets the same treatment with the addition of a decorative line at the top. The heel cap will mate up with this and complete the curve, hopefully in a graceful manner.

Thursday, April 26, 2012

Five-Course Mandolin Part 4

I apologize for the delay between posts. I've been enjoying an experience with kidney stones.  Anyhow, I have been fitting the interior tone bars. To do this I've prepared a couple of sticks of high-quality spruce and held them in the appropriate positions while scribing the interior undulation using a small washer to hold the pencil in position against the brace. I trimmed down to the lines with a knife and you can see I've come pretty close on this brace. The fit is important. I want perfect contact along the entire length and width without having to use pressure to close the gap. The bar in violin family instruments is occasionally "sprung", being glued under tension. This can have the effect of kick-starting the instrument's response, hot-rodding it as it were. The effect is temporary in bowed instruments and it's not generally seen as beneficial for plucked ones like the mandolin or archtop guitar.
To refine the fit I've placed some 120 grit garnet paper in position and will sand judiciously. The movement is very small. Broad sweeping strokes will act to average out the inconsistencies in the curve rather than mate the two surfaces.
Traditionally chalk was dusted on the plate surface and the brace pressed into place. I'm using carbon paper instead. It's slightly less messy and makes reading the transferred areas easy. You can see I've carved the f-holes in a previous session I neglected to document.

Here's a high spot. It's only thousandths of an inch proud so I'll use my carving knife held perpendicular to the surface as a scraper. Soon the entire surface is grey, and a final gentle scrape leaves the surface ready for gluing.
To glue these bars down I relied on a heavy strip of cork, 1/4" (6mm) thick on the outside surface of the plate to absorb the clamping pressure from several C-clamps. Traditional violin construction sometimes involves casting a negative mold using plaster of Paris, especially if the bar is to be sprung. These braces fit perfectly and the hot hide glue pulls the pieces together as it dries. 
To finish things off the ends of the braces are tapered off and feathered into the surrounding surface. The bars are trimmed to final height and rounded.  You can see that the bars aren't symmetrically placed. The bass-side bar is slightly shorter. The treble bar is narrower, and is at a more acute angle to the centre-line of the plate.

Saturday, April 7, 2012

Five-Course Mandolin Part 3

Carving the plates requires a certain kind of determination. This is the back, and I have used a router to remove a series of rings of material at varying depths to provide a guideline for carving. It's like a topographic map in maple. I'll use a small plane to take down the pointed areas of the terraces followed by a scraper.
You can see traces of the stepped profile remain as I further refine the curve. Still a lot of lumps and bumps. I've also started to carve in an indication of the "recurve", the area next to the edge of the plate that dips before rising again to resolve at the binding.

More scraping. It's easy to start chasing the lumps around. They disappear in one area only to pop up again in another.

Ready for hollowing!  I'll hold off on carving the recurve to its final depth until after the mandolin is assembled. The recurve of the back has a major effect on the tone of a carved instrument and it's best to sneak up on the optimal flexibility.

I embedded a dowel in the surface of my drill table that is centred on the bit and marked out some depth notations on the inner surface of the plate. By setting the drill stop at various depths I will give myself a visual aid in the hollowing process. To make measuring between bit and dowel easier I made up a little wedge of hardwood with increments of 1/2mm marked on it.  I left a generous amount for waste.

This is a quick-and-dirty carving cradle I whipped up. It's made from 3/4" (19mm) ply that is sawn out to a profile that supports the plate around its edge. I lined it with some scrap foam I had sitting around. This allows me to clamp the work in position. Carving involves some muscle and the movements are repetitive. It's especially hard work in maple, though spruce furnishes its own set of challenges. It's much easier to tear the grain of softwoods. In either case most of the rough carving is done across the grain. 
Like so.  I'm using a small convex-soled plane and various shallow gouges. The thickness profile was taken from a smashed 1950's Gibson I ran into quite a long time ago.  It's thickest right in the center and thins toward the edges, but not in a strictly concentric way. The area near the neck end is quite thick and it extends forward .  When I reach the bottom of the guide holes I start to pay very close attention to what is happening, making frequent checks with my thickness caliper.
I even got out my baby Ibex plane!  Isn't this ridiculous? It works well though, even if it is hard to hold on to. Curved scrapers finished the graduation and then a quick sanding with a random orbit sander.

Nice. It took some doing. The curved profile of the hollow extends all the way out to where the lining meets the back.  The top was carved using the same method.

Wednesday, March 21, 2012

Five-course Mandolin, part 2

 I've tried several different ways of attaching necks to bodies.  I decided to go with a straight dovetail for this project owing to the low-profile design I have in mind for the heel. The original Gibson mandolins had a complicated curved pattern obviously machined on a shaper with custom knives. Though not quite beyond my skill level, the amount of time needed to cut it by hand seemed wasteful.

This is a jig for my router table originally made to cut straight spline joints for classical necks.  It works for dovetails too. There is a bearing on the shaft of the bit which rides against the walls of the slot

The neck sits at a five degree angle and so I had to make up a wedge to keep it registered. You can see that I will nibble away at the sides of the heel and gradually move the neck over. As it turned out, removing material made the neck progressively more prone to tipping so I left a generous amount of waste to be refined by hand.

 Cleanup was accomplished with a chisel and this handy sanding stick that has the 14 degree angle cut into it. I aim for a very slight taper, so the fit gets tighter as the neck is pressed home into the slot.

Like so. The neck actually stands proud of the block by about 1/2" (12mm). This will accommodate the thickness of the top and an added fingerboard support extension.

I'm experimenting with solid-surface countertop material for a headstock inlay. Mother of pearl and the pretty shells of various bi-valves are now extremely hard to move across borders due to the U.S.A passing a piece of legislation called the Lacey Act. This is such a complicated issue.  In short, there are a great many natural materials which are legal to own, but sometimes illegal to sell, modify or move across state lines or into the country.  This Corian blank is about .100" thick (2.5mm). I've surrounded it with some hardboard to provide a stable surface for my miniature router and I glued on a slip of paper with some guidelines.
A tiny little carbide burr will be used to cut nearly all the way through the material and then the outlines will be smoothed up by filing. There will be some wiggles and eccentricities because I am not a laser engraving machine! A cavity in the headstock will be routed to fit and black epoxy employed to hold things in place and fill the gaps.  
I don't often put my name of the headstock of instruments, but this is small and subtle and will be partly obscured by the strings.  As it turns out, the radius I used on the end of the headstock is at the outer limits of what one can bend wood binding to. I had a big pot of boiling water and a little form to pin it to but it fought me every step of the way.

Friday, March 9, 2012

Five Course Mandolin with Fanned Frets

Shawn Trotter is a versatile local instrumentalist. I met him at a coffeehouse where he was displaying his dazzling acoustic fingerstyle guitar and mandolin chops. Soon after I found myself singing in a vocal ensemble with his partner Nikki and it was at one of our performances that he gave me a nod and mentioned that he'd been messing around with some Bach on mandolin and he was thinking about an instrument with an extra course.  "Interesting", I think I replied.    He had me at Bach, really.  Play one of the cello suites on slide whistle and I'll sit in rapturous admiration.  It seemed like a fun challenge and I had some nice old wood just waiting for something like this to come along.

"What about fanned frets?"
Shawn had done his homework. The low C strings would be pretty floppy, otherwise. I was aware of the concept of fanned frets.  The Orpharion was a wire-strung, lute-like instrument from the 16th and 17th centuries that employed the multi-scale idea.  Essentially, each string has its own scale length, thus the basses can be made a little longer to offset intonation and tension problems.  More recently Ralph Novak patented a fanned fret system coupled with some innovative bridge designs for use in his electric guitars.

The idea is deceptively simple. Just plot out two scale lengths, decide on a common fret and join them together.  Easier said than done. Accuracy is always of utmost importance when it comes to frets and not having them perpendicular to the center line adds some challenges!  I decided to keep the treble scale close to what Shawn was familiar with on his Gibson F-9.  After drawing a theoretical fretboard I came to the conclusion that an additional 1-5/8" (39mm) was as much as I felt comfortable adding to the bass. So it will be 13-7/8", sloping to 15-1/2". I have an expensive machinist's rule marked out in 64ths that I use to plot the scales. Using a magnifying lens I can split those increments.

I couldn't use my table saw to cut these slots. It's ancient and fine for parallel slots in a standard board but this called for hand work.  I came up with a little scratch knife to make things easier.  My fretting saw cuts a kerf .022" wide. I took a .023" feeler gauge and glued it into a handle made from ebony offcuts and ground the edge into a little hook.  It's basically a tiny scraper that initiates a groove for the saw to sit in.

I used some double-sided tape to keep the board secure and overlap the end of my bench. There's a knack to setting up the bevel gauge. One has to get things in position, gently snug it up, move it off the board before cinching it tight and then test to make sure nothing moved. I was able to center the scratch tool right on my marks. This seems like an ideal system for people without mechanized fretting equipment - certainly more accurate than trying to hold the saw in place!

Don't be offended by the rippled surface from the jointer. I'll be sanding a 16" radius into the board and that will go away soon enough. The original A-5 had a flat board but a little curve just feels better.

Here's the other tricky bit. The board will be bound in white-black-white. I decided the best way to do this was to plane the board to its full dimension, cut off the area of the binding, glue on the purfling strip, radius the board, re-cut the slots through the purfling, then finally glue on the binding. This would take care of any chipout on the ends of the slots and reduce the amount of filthy ebony dust ground into the pores of my nice white holly binding.

I set my marking gauge very carefully and did a test-run before touching the board. I'll use a sharp plane to remove this excess material before continuing.