Tuesday, August 20, 2013

To the Extreme! Adventures in Neck Setting. Carson J. Robison

Sometimes repair brings surprises.  I recently did some further work on the Carson J. Robison guitar that I discussed a month or two ago. "Can of Worms" might be a better title for this blog.
Let's start at the beginning.  The action is the highest I've ever seen on a real guitar.
I, like most North American luthiers, tend to describe action in terms of 64ths of an inch. This, gentle reader, is a stiff action. 14/64". A full 7/32"!  The owner of this instrument plays slide, and I decided to get confirmation from him as to what we should be done. He asked if it was possible to lower it to within reasonable levels, and I agreed to do a neck re-set.  I spent some time doing the trigonometry and realized that changing the neck angle this much would remove a huge amount of material from the heel. I reasoned that it would be possible to insert a filler under the fingerboard extension to alleviate some of the problem. More about this later.
It's important to protect the fragile finish on the top from heat damage, so I made up a mask from heavy card. The tongue of the fretboard was heated using a regular household clothes iron.

Releasing the tongue was easy. You can see I added several layers of masking tape to insulate the area I didn't need released, and to protect the old plastic dot inlays.
While the 15th fret was still hot, I pulled it up. There are several ways to do this, but I like to use a thin chisel.  Having pressure against the board as the fret is pried up seems to keep chip-out to a minimum.
Pretty clean.
To reach the space at the end of the dovetail, one needs to approach it from a slight angle.
On Martin guitars, and most modern factory-made instruments the neck is finished separately from the body. This guitar was assembled before spraying. I would later find a lot of filler in the intersection between neck and sides. The dark stain hides the less-than-fastidious fit.  It's important to score along all the seams with a sharp blade to prevent as much damage to the finish as possible.
Here is a shop-built tool for applying steady pressure to the heel as steam is injected.  This is where where the fun begins.  I applied steam in the usual doses and cranked a little.  Nothing happened. I let it go a little longer, -still nothing.  I worried about letting the guitar get too hot, so I cooled it down and wiped off any moisture before re-trying.  Still nothing.   I tried using a mallet.  It didn't budge at all! (In retrospect it was lucky I tried this operation with the guitar all clamped up like this. Too much steam could have loosened the joints between the plates and side assembly.

I was stymied.  There was no reason for this to be happening!  I've taken the necks off a dozen other guitars, mostly Japanese dreadnaughts from the 70's and the odd Harmony.  This instrument was definitely put together with hide glue. What was the issue?

Something jogged my memory and I went searching through my archives.  I found a reference to a specific group of Gibsons in an old issue of the Guild of American Luthiers journal....
This guitar was shipped out of the Gibson workshops in 1936. What were the odds?!   I informed the owner and advised him that I would have to go oldschool and cut the tongue free.  Nobody does that anymore. It's a technique from an earlier time. It's not nearly as clean, but there seemed no way around it.

Suspicions confirmed -but it wasn't so easy. I cut the fingerboard extension off and revealed an unsullied soundboard surface. A little excavating revealed the dovetail pocket.  I had access to the cavity and steamed it for the fourth time. Still no luck. It refused to come free. If you look closely at the pocket above, you'll see that it was cut using multiple passes from a table-saw blade. (What a laborious way to do things - imagine being that guy who had to stand there dragging the same chunk of wood over the blade 15 times...)  This was not a fancy guitar. It was put together with haste. Its original price of $9 was roughly a quarter of the cost of a similarly appointed "official" Gibson. That included a case and a harmonica.  If you look closely at the top of the dovetail cavity you'll see it's really chewed up. The fibers have been squashed.
This thing was literally hammered together.  There was no neck "angle" originally. The neck and body were in one plane. No time was spent finessing the attachment. In engineering terms it's what might be called an "interference fit". The angles weren't quite identical, and with the addition of some glue to swell the wood - there was no way it was ever coming apart. I figured this out when a hairline crack started to appear on the heel surface. That did it.  Out came the finest Japanese flush-cutting saw, and I cut the neck free, leaving the dovetail in situ. I cleaned it out using a chisel.
(Sorry I didn't take a photo of the sawing. At that point I was a little exasperrated!)
I was seriously tempted to drill some holes and turn it into a bolt-on.  The purist in me knew that wasn't the way to go.  I pulled out a dovetail template and cleaned up the cavity to make it square and regular. I then found a suitable piece of mahogany and carefully laid out a tenon to mate with the now pristine mortise.
It took some time to get the fit just right. I then sanded the heel area, removing a wedge of material to effect the angle correction (as much as I dared). Getting everything lined up, I applied epoxy to the tenon and stood the neck upright on it, making certain that it aligned properly with the bridge.
 When the epoxy cured, I could then remove the neck from the guitar and drill for two large screws to really secure the tenon to the heel surface.  The joint is side-grain to end-grain - not ideal, so I also drilled for a large dowel, which was then cut flush.  I felt satisfied that this arrangement would remain secure.
Usually I like to see a straightedge lying on the surface of the frets land right on the top of the bridge, that way any action adjustment falls within the dimension of the saddle. In this case it's a little high -this is very low bridge, and I don't believe it's ever been planed down. In doing the math, it seems certain that the guitar left the factory with an exceptionally high action.  Marketed for strumming cowboy chords, I don't think it was designed to be played beyond first position.
The only way I could get things in reasonable shape was to use a shim of rosewood under the fingerboard extension. Without it, the board would drop off precipitously over the body, but it also allowed me a 1/16" "boost" to get up close to the string plane.  I still shaved more than 2mm of material from the distal end of the heel.  I would have made the fingerboard shim even higher, but I didn't want to pull the heel up past the edge of the back:

There's now a little step down to the heel surface where before it was flush.
There was also a lot of finish touch-up. The steam softened things a little in various areas. I used shellac and french-polished it along with alcohol dyes. The effect is okay. It's hard to know when to stop. This wasn't a refinishing project, after all.
I managed to reduce the action by almost 6/64".  Remember that any numerical change at the 12th fret is doubled at the bridge, so we're talking 3/16"(almost 6mm) theoretical change in string height at the saddle! That's *huge*.  You can see why the booster-shim was necessary.
This is the original saddle. It's glued in. Deep notches had been filed into it to lower the strings a bit, and the blank was very square and pointy looking.  I filed off the extra bone and gave it a more graceful profile, but didn't have to change the height.

What an experience.