BENDING THE SIDES.

I finally feel I'm getting the to core of the build. Last week it was the soundboard thickness and this week it's the side bending.

It was a process I was looking forward to with both enthusiasm and anxiety. It was completely new territory to me, as exciting as it was scary. To prepare for this I read up on as many books I could, most notably Cumpiano and Courtnall. I also posted a thread at Luthierforum and got many responses with helpful hints.

It all lead to the fact that, the best way to start bending was to practice on some scrap and so I did. For this I used some flat sawn walnut I had and a cut-off from the bubinga I'm building with. This gave me a chance to gain some experience with different bending techniques and setups.

I started with the walnut, sanded to 2mm thickness, and went at it haphazardly on purpose to see what mischief I would get myself into and then tried to get myself out of them again. All in all the walnut felt very leathery in it's elasticity and I had many problems controlling it. It was very difficult to manipulate one area while retaining the bend in another. I also ended up with some kinking, making it impossible to lay the bent side down on a flat surface without one of the bouts twisting upwards. I tried bending both dry and sprayed with distilled water and the wet was definitely the more responsive of the two.

Eventually, after lots of unruly bending and unbending, I kind of got it into a shape somewhat close to the template I was measuring it against, and left it as such. However, the next evening I noticed that it had curled up quite a bit, meaning an overbend had occurred in both bouts. I should mention that I only sprayed the walnut on the face that contacted the pipe, and my guess is that it could have been a result of moisture still being in wood that had evaporated overnight; Maybe a good argument for soaking or spraying both faces regardless of which way the side is being bent. After some dry touch-up I managed to get it back in shape again.

I also had a quick go at the bubinga. It felt considerably tougher than the walnut. Whether this was because of it's density or the quatersawn grain orientation I don't know but, it took considerable more force to manipulate and once bent, it seemed to hold it's shape much better. I tried bending it wet, both after 10 minutes of soaking and from spraying it on both sides but I didn't notice the wood behaving much differently between the two methods. However, like the walnut, it was considerably less manageable when bent dry. I really tried to push it and did manage to fracture it a little when bending it on the tightest spot on the iron, but only after using considerable force.

I also experimented with the bending iron in different positions - horizontally, vertically, with the flat side facing me, with the round end facing me, etc. I ended up with it clamped vertically to the bench with the tip of the egg shaped pipe to my right. I came to the conclusion that the great advantage with the vertical setup was that my eyes were in line with the pipe as I was looking down upon it, making it easier to check whether I was pressing the wood against the pipe with even force across the side or not. On the down side it is more difficult to see whether you are holding the side perpendicular to the pipe or nor. It also took more muscle to bend the wood compared to a horizontal setup where gravity would be working in your favor.

After these practice sessions I felt a little more confident and eventually took the plunge and started on the 'real' sides. I started by laying both sides on my bench in a bookmatch. After measuring my template with a sewing tape measure, I marked both sides with a line across the position of the the heel, waist and butt and named these accordingly, leaving about 5cm extra stock at each end. I also added arrows to each line indicating which edge was going to be glued the top. This way I was able to avoid bending both sides the 'same way' and thus ruining the bookmatch. However, when I was done bending both sides I realized that I had bent the sides so that the true bookmatch on the inside and that I should have bent them both the other way. It's not a big deal as the sides are perfectly quartersawn thus having a good match on both faces, but nevertheless a shame I didn't get the better of the two on the outside. We live and learn!

After the marking was done, I wetted one of the sides by spraying it with distilled water on each face, put it to the iron and started bending the waist. I used the flatter side of the iron to heat the wood and slowly fed it across with my right hand while pulling it towards me with my left. Soon the water started to sizzle and evaporate into steam and I felt the wood giving in under the pressure. Slowly the waist emerged and after a few goes it fit the template. After cooling it a little I went on to bend the upper bout. This was done in a similar way, though because the bend was more gentle, less force and a faster feed was needed. The upper bout was a little more tricky to match as it had to be done in a more free form fashion, whereas the waist fitted the bottom of the pipes oval quite well. But after a few goes and some areas of unbending and rebending, it eventually settled quite nicely. The lower bout was done likewise, though with even less pressure and faster feed.

After the whole shape was done a few problem areas became apparent - some where the bend had gone bumpy, a few overbends and some areas where the bend had gone off the template and back again. I had to respray the side to correct the more severe ones, while others could be dealt with by bending dry, either locally or by sliding the bend across the pipe applying pressure by wrapping my hand around the wood.

Eventually it was pretty much within the tolerance of ±0.5mm and I left it overnight to cool and dry ready for a final touched up the following evening to get it as close to dead on as I could. More touch-up might be needed once the sides are ready to be glued on.

MARKING THE SIDES

WETTING THE SIDES BY SPRAYING

BENDING THE WAIST

CHECKING WAIST BEND

BENDING THE UPPER BOUT

CORRECTING AN OVERBEND

CHECKING UPPER BOUT BEND

BENDING THE LOWER BOUT

CHECKING THE BEND AGAINST THE TEMPLATE

FINISHED SIDES

FINISHED SIDES

FINISHED SIDES

FINISHED SIDES

SIDES IN SOLERA

SOUNDBOARD THICKNESSING AND DEFLECTION TEST - PART 2.

After much debate and recommendation from fellow builders, I decided to thin my soundboard even further. Same procedure as last year Miss Sofie.........

After careful planing, I ended up with a final thickness in the bridge area of 2.2mm and 1.8mm around the periphery. The 2.2mm thickness extends all the way from the bridge, up past the rosette, to the top of the soundboard, slowly graduating out towards the sided in a pattern parallel to the center seam. Below the bridge area the thinning follows the contour of the lower bout. In other words the thinning forms a kind of a U-shape, rather than the shape of the guitar itself.

I also repeated the deflection test and it came out showing 7.5mm of deflection; Significantly more than before but still a little stiffer than the 8-9mm target. However, at this point the soundboard felt very flexible indeed and flexing it cross grain with my hands, I was afraid of snapping it in two. It was also hard for me to detect any change in pitch, or rather character of the sound, when I tapped it.

BRIDGE AREA THICKNESS

EDGE THICKNESS

UNWEIGHTED DEFLECTION

WEIGHTED DEFLECTION

DEFLECTED SOUNDBOARD

DEFLECTION TESTING.

Over the weekend some further development came about on the soundboard front.

Suspecting that I might have left my soundboard to thick/stiff, a luthier friend of mine suggested trying out measuring the soundboard deflection as a more scientific approach to thinning it to it's optimum stiffness. He advised that on a classical guitar, a weight of 380 grams should be able to deflect a suspended soundboard 9-10mm, measured cross grain at the bridge location.

Doing the math, I should have aimed for a thickness of 2.2mm around the bridge area and about 1.8mm around the periphery of the lower bout with this smaller instrument. But, though these dimensions are thinner than what I ended up with, I feared it was just too thin, mainly because the top started to appear a little floppy, but also because I started to have difficulties hearing the pitch dropping any further at this point - It was as if the pitch had become a low rumbling indistinct sound. Mind you, I made these assumption with no practical experience to base my judgments on.

Loosely following his instructions, I quickly devised this crude and simple jig for this purposed. To each of the sides of the lower bout I secured a 5/8" think block of MDF with 1" plastic spring clamps, each block protruding 1/4" into the plantilla or outline. I centered the MDF blocks at the widest point at the lower bout, a little more than 1" below the bridge.

I propped up the soundboard on the jig, resting the MDF blocks in line with the bridge. A small MDF block supporting the top of the soundboard was also added to the upper part of the jig to keep things stable and level. I used big washers for weights and they proved to have the added benefit of allowing a ruler for measuring to rest inside them.

As this guitar is about 11% smaller than the standard, I scaled down the weight and deflection targets accordingly to 340 grams and 8-9mm.

The deflection test came out showing 6mm difference between the weighted and unweighted soundboard, indicating that it indeed is too stiff. However, I'm still not sure how accurate my measuring was. I don't know how much factors like friction between the MDF blocks and the jig played, how much the placement and size of these have entered the equation, measuring with the soundhole uncut versus being cut, etc. etc.

Nevertheless, I might have to revisit the thicknessing procedure again.

DEFLECTION JIG

SETUP

WEIGHTED SOUNDBOARD

UNWEIGHTED SOUNDBOARD

THICKNESSING THE SOUNDBOARD

After some serious procrastination, I finally dived in and did something about the marks around the rosette created by rosewood dust from cutting groves for the rosette segments as well as some careless leveling of ditto. Somewhere in the back of my head some little persistent devil was sitting, insisting I should do this with a smoothing plane. Last night I couldn't take it any longer, and had to act against convention and advice from books and fellow builders, suggesting the use of sandpaper for this operation, first 150 grit then 220 grit. I figured I could always resort to sanding if the planing didn't work out.

OK, I'll be the first to admit it; I'm a stubborn bugger!

Off I went. I started by sharpening my 38º blade to the best of my ability and slit it into the Bevel-up smoother, adjusted it to the finest setting possible and secured the soundboard to the bench with a long MDF caul. With even pressure I slowly drove the plane through the spruce, up over the rosette, to the full length of the soundboard. My biggest concern was that the blade would bite into the rosette and tear it up and cause further damage, but it didn't. Instead I got whisper thin slivers of spruce and rosette. I slowly and methodically planed my way twice from one side soundboard to the other, with full strokes along the grain, until the marks and dust had gone.

Next came the thicknessing. I flipped the soundboard over clamped it to the bench with the MDF caul and started planing the back in a similar fashion: With full slightly overlapping strokes along the grain, slowly working my way across the soundboard. Once the first half was planed I unclamped and rotated it and continued on the other half. At the starting point the soundboard had an even thickness of 3.2mm all across.

After each complete cycle I measured the thickness and marked any area that would be a little thicker and leveled these first, before starting the next cycle. I also tapped the soundboard by holding it by the upper bout with one hand and gently tapping the bridge area with the end of my middle finger of the other. In the beginning the tapping produced a clear ring but as the thicknessing progressed the pitch slowly started to drop and became lower and lower. Supposedly you are striving to thin the top until you reach a pitch where the ring has 'dropped out' and been replaced by a low and indistinct tone, but as this was the first time for me it was hard to determine when that point was reached. Instead I resorted to measuring, tapping and listening along the way, ending up with a final thickness of 2.4mm in the middle, from the bridge area and up, and just under 2.0mm along the periphery of the lower bout.

It was interesting to follow the pitch drop as the soundboard got thinner. In the beginning the change was quite noticeable, but as I got closer to the final thickness it became increasingly difficult for me to distinguish the change, which on one hand probably was due to the fact that I started to proceed in smaller and smaller increments, but also lack of experience. One thing I was quite aware of is the it is supposedly typical for the novice builder to overbuild. I just hope I didn't got too far in the other direction.

ROSETTE AND SPRUCE SHAVINGS

THICKNESS PLANING

TAPPING THE TOP

FINAL THICKNESS OF BRIDGE AREA

FINAL THICKNESS AT THE EDGE