A new project. A customer wondered about building a steel string baritone. Hmmmm… Kind of an interesting project, so I agreed. At the same time I’m building a steel string tenor, just for fun and investigation. This blog entry will be evolving (I’ll edit the end of it) to follow the steel string story. I’m going to try an include a fair amount of my thinking, not just what I will do, both since it might be of interest to you the reader and also to provide a record that I myself can refer back to in the future.

To start with, the requesting customer and I had to decide on what sorts of woods this little experiment would be built out of. Since this is a new thing for me, I’m taking some lessons from steel string guitars. I know that a ukulele is not a little guitar, they have their own considerations, but I figure a steel string guitar can provide some general direction. In all of the thinking the overriding consideration is the increased string tension of steel strings.

For the top we selected some old sitka spruce. Sitka is a classic steel string guitar top, so I figure it is well suited to deal with the string tension. I had some sitka that I got from a friend a while back, and he had been storing it for a very long time. It is at least 40 years old, feels very stiff, and is perfectly quarter-sawn.

For back and sides, from the steel string guitar perspective, one wants something stiff and dense, like the rosewoods. I have some casuarina that I cut some years ago that is baritone sized, and is very stiff and dense so we went with casuarina back and sides.

Bracing – The back bracing can be pretty close to what I normally use for a nylon string baritone as the back itself does not participate in the string tension.

The top bracing however is another story. The top is where the increased string tension is centered. I have been doing some reading, and thinking, and I decided to make up some brace stock where the braces were split out of a larger piece of wood, rather that sawn. I had some bigger pieces of nice fine grained sitka spruce. Splitting wood ensures that the wood fibers extend the length of the brace, rather than being on an angle, which is called ‘runout’. Under stress a brace with runout will tend to crack along the wood fiber lines, which are across the brace. If one looks at guitar repair videos where braces are being fixed, the crack is almost always across the brace, due to runout. After making them up the split braces feel particularly stiff. Maybe it is the piece of sitka spruce, maybe it is the fact that the wood fibers run from end-to-end.

A steel string instrument will have a pined bridge, like a guitar, where the ball end of the string is up against a bridge plate on the bottom of the top, and a tapered pin holds the string in place. There are two considerations here in my thinking. For one, the ball end of the strings produces a concentrated pressure in a small area of the bridge plate. Therefore I wanted something both hard and tough for the bridge plate. Secondly, the bridge plate, like all of the other components of the top, will have to vibrate to produce sound. My current thinking is that the lighter the bridge, and by extension the bridge plate, the less mass there is that needs to be vibrated, and thus the easier the whole top can be vibrated. I chose to use some Pennsylvania black locust as the bridge plate. Locust is quite hard and tough, but is not too heavy.

For both the reinforcement around the sound hole, and the bridge plate, I glue down somewhat thicker pieces and then run the top through the thickness sander to thin these pieces out. Gluing down thicker pieces is much easier than trying to glue down quite thin pieces. In this case I wanted the sound hole reinforcement thinner that the bridge plate, so I glued it down first, thinned it with the thickness sander, then glued the bridge plate down, and finally thinned it.

The bridge plate also needs to be larger that what I would use for a nylon stringed instrument, since the tapered pins will come through it and one does not what the pin holes too close to the edge of the bridge plate.  On steel string guitars the bridge plate is between the X braces, but I have been extending the bridge plate and notching the braces over the bridge plate on all my instruments. I think this adds to the stiffness because any rotation of the bridge/bridge plate will directly have to bend the braces.

For a bracing pattern I am using what is my normal nylon string bracing, an X braced system, without any transverse brace below the sound hole. This brace pattern has produced very nice results in the nylon string world, and I figure that vibrations are vibrations whether started with nylon strings or steel strings so it should work well in a steel string environment. Also, a whole lot of steel string guitars use an X brace pattern.

When making up the X one notches both of the the two braces half way through and then fits them together. This has the unfortunate side effect of basically cutting the brace in half in the middle. To make this X join point stronger I usually glue an extra little piece of wood across the top of the X to span the two sides of the brace where the cut faces up. From an engineering standpoint this will greatly strengthen the brace beam. In the case of this steel string baritone, as added stiffening/insurance, I per-glued the X together with epoxy. Epoxy will fill any little spaces where the two notched braces fit together and will weld the whole thing into one structure.

The braces are shaped into a triangular cross section since the stiffness of a beam (brace) is dependent much more on the height of the beam than the width. By shaving this off into a triangular cross section the brace is made lighter (easier to vibrate) without sacrificing much strength.

Gluing down the top

Once the top is glued down I shellac the interior.  A coat of shellac should slow down expansion and contraction due to humidity changes.  I have no proof that this is really necessary/beneficial, but a luthier who is much greater than I shellacs his interiors and it is a small thing, so why not.

The back is shellaced before it is glued down (while fitting the brace ends into the kerfed lining).  Glue on the back and this are all ‘boxed up’.

I am not a big believer in tap tones, at least as far a judging a top before it is glued onto the instrument.  I do tap things trying to learn some patterns (I bought some xylophone hard rubber mallets to give a consistent tap) but the sound changes enormously between the top with bracing by itself, and when it is glued down as part of the instrument.  Tapping on a closed box instrument I think gives some indication of what the eventual instrument will sound like.  One thing I can say about this instrument from tapping it is that it is LOUD!  Stiff spruce top, hard reflective back & sides (?) who knows but I think that the volume bodes well, since to me that is a sign of how well the box converts vibrational energy (from the top) into sonic energy.

The back and sides are made of casuarina. Here in south Florida there are actually three closely related species of casuarina, C. equisetifolia, C. glauca, and  C. cunninghamiana. All considered invasive, with C. glauca being the worst. One tells them apart with a very careful examination of the ‘leaves’, which look like pine needles, hence the local name of “Australian Pine”. They are similar but I see some differences in the wood of at least two, if not all three species (I just recently acquired some C. cunninghamiana).  The casuarina I have used in the past has all been C. equisetifolia, quite hard and dense. The back and sides here are C. glauca. This is the first time I have worked with this species of casuaarina.  This was cut a couple of years ago from a big blow-down after a glancing hurricane blow.

Trees of course differe even among the same species, but this C. glauca I have found to be really really hard. I started rough sanding the body of the baritone and this casuarina really dulls sandpaper quickly.  Even with frequent sandpaper changes progress is slow.  I think this is the hardest wood I have ever used for back & sides.

– – – – –

The binding is done.  The actual binding wood along the edge has a couple of purposes.  It covers the end grain of the top and back, and it provides a bit of edge bump protection.  Of course, almost any wood will do to to meet these purposes.  The choice of wood is about the look.  The customer with whom I am working on this little project is not afraid of a little bling, and one does not get much more bling-ie than properly cut sycamore for binding.  The purfling, that decorative strip inside the binding, is just that, purly decorative.  Again, we went with a pretty bling-ie purfling.

Neck

The neck of a steel string instrument will be subject to more tension because of the strings than the usual nylon stringed instrument. On all my instruments I insert a carbon fiber beam into the neck for added stiffness. My usual beam is 1/8 x 3/8 inches. Early on when preparing a neck blank a 1/8 inch slot is routed in the neck. This slot becomes the true center of the neck, regardless of where the edges of the blank may be. I have a number of jigs used when making and shaping the neck, all of which have 1/8” pins sticking out of them to index to that 1/8” slot.

For added stiffness I realized that once all of the neck shaping that involved indexing to the 1/8 inch slot were finished I could re-route a wider channel. I made the channel ¼ inch wide so I could double the width of the carbon fiber beam.

The beams fit tight into the slot (they have to be tapped in with a small hammer) and are then flooded with thin CA glue to weld everything together.

Intonation

The position of the bridge in relation to the frets is absolutely critical if the instrument is to play in tune. On a theoretical basis the 12’th fret should be exactly ½ the distance from the nut to the saddle, and play exactly an octave above the string open. However, because the strings are above the frets, when a string is pressed down and fretted, it is stretched slightly which makes the note play a bit higher in pitch. To ‘compensate’ for this one adds a bit of compensation by moving the bridge/saddle slightly away from the nut so the fretted note is a little lower in pitch. The amount of this lengthening of the string, the compensation, is dependent on the type of string and how the instrument is set up. (A high action means the string will be stretched more, which will require more compensation.)

Since these steel strings are a whole new thing for me I utilized a setup I came up with a while back for setting the compensation. I have a false tail-piece and stretch the strings over the bridge/saddle to the tuners. I can tune the strings up (I do only the 1’st and 4’th strings) and move the bridge/saddle back and forth so that the note open and fretted at the 12’th fret are exactly an octave apart. I bought a high quality stroboscopic tuner so get this really correct. Here is an older picture of the setup, not this steel string instrument but the setup is the same.

Finished

The instrument is finished. How does it sound? Great, very loud and sustain that goes on forever. Looks great too, all the parts seemed to come together to make a whole that is greater than the sum of those parts.

It has been a week since I strung it up and I have been monitoring the top to see if there is any distortion from the tension of the strings. The instrument is not over-braced as judged by the sound production. String tension will tend to rotate the bridge, which creates a hollow in front of the bridge and a bit of a hill behind the bridge. So far, things are looking good. There is no discernible distortion in the top which means that the bracing seems to be adequate, though it will bear watching over time.

Here in Florida I have run into wood turners who have used what they called “Florida rosewood”.  Next door neighbor came up with a big tree and kindly enough gave me a big basal section.  The tree is apparently a true rosewood, geneus Dalbergia.  The stuff that has been planted around Florida seems to be Dalbergia sissoo, known commonly as North Indian rosewood or shisham, not to be confused with the widely used (for guitars) ‘East Indian Rosewood, or Dalbergia latifolia.  Got some nice billets cut, big enough for tenors, and maybe a baritone or two.  After re-sawing and a couple of years drying of course.

To strengthen my necks, and to prevent future warping I install a carbon fiber beam down the center of the neck.  The beam I normally use is 1/8 x 3/8 inches.  I route a 1/8 inch channel for the beam when I make up the neck blank.  This 1/8 inch slot becomes a reference for the center of the neck on a number of subsequent jigs that have 1/8 inch pins on them.  The slot fits over the pins and allow the neck to slide back and forth to fit the jig.  The carbon fiber is glued in only rather late in the neck making process, when all the jig-based cutting is done.

For baritones which have longer necks, and parlor guitars, I was thinking that I would like to add a bit larger carbon fiber beam, but that 1/8 inch slot was integral to my build process.  I realized that I could cut the 1/8 inch slot and use it as the jig reference right up to the point that I would normally install the carbon.  Right at that point I could put the neck back on the router table and widen the slot.  For a baritone I got some carbon fiber which is 3/16 thick, a bit more robust that the 1/8 inch.  For a parlor guitar I could make the slot 1/4 inch and just double the 1/8 inch carbon.

The carbon fits tightly into the slots, I have to tap it into place.  I used to glue them in with epoxy but epoxy is messy, and the carbon fits so tightly most of the epoxy got squeezed out making more of a mess.  Now I just tap in the carbon and flood it with thin CA glue which wicks in and glues everything together with no surface mess to clean up.

I have found that a side sound port makes a real difference in the sound of an instrument.  It kind of ‘opens up’ the instrument so it is brighter, louder, warmer, everything.  Needless to say, most instruments I build these days have a side sound port unless it is declined by a commission.

I bind the edges of the port, which is after all, a hole cut in the side.  I bind the edges with three layers of dyed veneer, usually black-white-black but I have done black-red-black to match black-red-black purfling.  I used to pre-bend stripes of veneer (thin for the inner two layers and a thicker black outside layer) on a hot pipe, like bending other wood.  This worked OK, but the bend was only an approximation of the oval shape of the sound port, and gluing them in required a certain amount of muscling things in, with the occasional cracked veneer.

I have gone to a new system.  There are a number of veneer softeners on the market, with Super-Soft being popular.  This stuff softens wood fibers, making the wood bendable.  When things dry out it leaves no residue, no color change, it is as if it was never there and the wood holds the shape it was bent into.

So –  I made some little ovals of plywood which match the oval of the sound hole.  Bigger for baritones, smaller for tenors and concerts.  I soak the veneer strips in Super-Soft for 5 minutes or so, and then the veneer is pliable enough to bend around the oval forms without cracking, even the thicker black outer layer.  I clamp the ends, let things dry, and the oval shape is held very well.  They go into the sound port when binding it with very little need for muscling things to fit.  Easy, cheap, and works great.

Here is a little trick I learned from Beau Hannam (Luthier extraordinaire).  When shaping a neck one carves it to the final shape but there are still little inconsistencies.  To straighten these out one makes a sanding block that is just shorter than the straight part of the neck.  I use PSA sticky-back sandpaper to stick onto the block.  You mark pencil lines all over the neck and then sand with the block cross-wise to the neck.

The pencil lines gets taken off where they are the high spots, and the low spots are revealed as remaining pencil lines.  Keep sanding, maybe giving the high spots some extra attention with a file.

Almost done, only one little low spot.

All done.  An easy process that makes the necks feel better under the hand because they are absolutely level and consistent.

I thought I would go into a bit of detail as to how I cut a spiral rosette. To do this I use three different jigs that I developed and cut out of ¼ inch acrylic on a laser cutter (which I use at out local community college, a great resource.)

The first jig is the spiral template itself. This is used to cut the outside of the spiral. It is indexed to a center hole which fits over a 1/8” pin on a routing board which comes through a corresponding hole in the top. This pin will also be the center of the inner circle of the rosette, and the smaller eventual sound hole. The spiral jig is positioned so the the point of the spiral is out from under the area that will be covered by the tail of the fretboard. The spiral jig is held in place by clamping a straightedge to the routing board along one side of the jig. Since the jig also is fitted over the centering pin this holds it steady.

The second jig is a base for a StewMac dremmel router base. This is a simple base, with a bit of 3/16” brass tubing glued to a hole in the center. This tubing extends a bit below the base allowing the base to ride on top of the spiral jig and follow the outside profile. A small dremmel bit is fitted to the dremmel and extends through the tube just far enough to cut the desired rosette channel, the depth of which is set by the thickness of the pearl which will make the rosette.

Using this setup the dremmel router is run along the outside edge of the spiral jig to cut the outside of the spiral.

When this is done one shifts to the third jig, which is the dremmel router base with which I cut my other more circular rosettes. This base has a series of holes which fit over the indexing pin and when the dremmel router rotates around the pin a circle is cut. The holes are spaced and numbered such that moving from one hole to the next larger increases the diameter of the circle cut by 1/16”, or only 1/32” increment in the radius.

Using this base, and a small router bit, one cuts circles aiming to just meet the tip of the spiral path already cut. One can start with a slightly larger diameter and run the cutter up to the spiral path. One then decreases the size of the circle by using successively lower numbered holes till the circle just meets the spiral.

The very pointy tip of the spiral is cut out using a scalpel. When this is done you use successively larger numbered holes in the router base to clean out the inside of the spiral channel.

Now that the spiral channel is cut, all that remains is to line it with purfling and cut pearl to fit. Cutting the pearl is actually easier than doing a completely round, even with rosette. Since the spiral is tapered, one cuts a piece of pearl and can then slide it around from a larger part to the smaller part, compressing the purfling tight against the side of the channel. The very end of the spiral will be hidden under the tail of the fingerboard.

Result:

The fret spacing on an instrument determines the scale length. Since the 12’th fret is an octave above the open string, the distance between the bridge and the 12’th fret is 1/2 of the scale length. Mostly …

Since when you press a string you stretch it a bit, if the 12’th fret to bridge distance were exactly the scale length, the note would be a bit sharp, from the added stretching. This stretching effect increases as you move up the fretboard since the string is getting ‘shorter’ as one uses the higher frets. The amount of stretching, and the amount that this stretching raises the note is dependent on a number of factors including the scale length, the type of string, the diameter of the string, and the height of the action.
To avoid having the note go sharp as one moves up the fretboard the saddle must be moved out a little bit beyond the actual scale length. This little bit of extra length is called the ‘compensation’. There are a number of ways of approximating the compensation, but since it is dependent on a number of factors specific to an instrument I developed a way to empirically set the bridge/saddle in the right position.

To do this, when preparing to glue the saddle down, I put the instrument on the bench, with a block with string holes in it at the bottom end, and the neck held down by a block in the middle of the fretboard. With this I can stretch the 1’st and 4’th strings over the instrument, across the not-yet-glued-bridge, in playing position. I set the action close to the final setting by adjusting the height of the saddle (as one would do on the finished instrument). Since the bridge is not glued down I can move it around and using a very accurate tuner I can get the bridge position so the open string and the string fretted at the 12’th fret play exactly an octave apart. I use the actual strings that I’m going to put on the instrument too. Once I get things right (1’st and 4’th string is an average for the other 2 strings in the middle) I mark the bridge position and proceed to glue it down. A straight saddle is always going to be a bit of an approximation for the compensation because of the differences in the strings, but I can get very close.

The example shown here is a Kasha braced baritone which is why there is an unusual bridge shape and an off-center sound hole.  The top is protected with a low-tack tape since the instrument is fully finished at this point and I draw the bridge position on the tape when I get it right.

I got this idea from one of Beau Hannam’s videos.  It gives me much better results, much more easily than what I did before when trying to get necks straight and smooth during the carving process.

Make a block with a flat side (length determined by the instrument neck size) and glue sandpaper to the flat side.  (I use PSA sticky paper so this is easy.)

This is used to sand the neck cross-wise which gets things really flat, and find all those little uneven places.  I get a really good neck with little effort.  The one thing I do which Beau does not is that I mark pencil lines all over the neck so I can see where the high spots are, and when sanding can give those high spots a little bit of extra sanding.  Also the pencil lines make sure that I have hit everything.

just starting out

half done

only a couple of low spots left.

This is what the bench looks like while putting together an inlay design (a woodpecker in this case).  One needs to get out a number of pieces of pearl, in a number of colors, and select the right colors and patterns, particularly when an overall design is composed of a number of pieces.

I print out the design in multiple copies, and then each piece is glued to a piece of pearl, cut out with a jewelers saw and very fine blades, and then assembled with a bit of filing to get pieces to fit together well.  The pieces are glued together with CA glue, face down, on a Teflon mat from which CA releases easily.

Cut this from an old big Haden mango tree taken down by a local mango grower.  Looks like pretty good stuff.  Nice quarter-sawn billets and spalted without being at all punky and soft.  Plenty big enough for baritones.