The Luthier: Part III- Steam Bending

Steam bending. Something that I am glad I have been able to do, but not something I ever want to do again. This process, nearly 9 weeks now since I did the first test bend, was immensely frustrating, stressful, and unsuccessful from the beginning to the end. While the sides to the mandola are not perfect, I am calling it done here before I do something irreparable to the wood. Beyond the damage that has already been done, that is...

Preparing the sides for this project is the first order of business and the first real glimpse of progress after so much planning and redesigning. I have a pair of book matched Bolivian Rosewood sides that are about 4mm thick and wider than I need. Since I expected to have some levelling and plane adjustments to make after getting the sides put together, I left them a little wide.  Because of the shape, each side cannot be one continuous piece of wood, but at the corners I want to preserve the grain alignment as closely as possible.

Out of wherever these were milled and resawn into the pair, the ends are not even close to square. For now, I left that alone but ripped them down to about 9cm. Since I expected to have some levelling and plane adjustments to make after getting the sides put together, I left them a little wide. In the end, they will be about 8cm tall.

And now we begin the steam bending. Here the first piece (longer part of the left side in the plan above) is in the steam box I built a while back. A rule of thumb is steam for an hour for every inch of thickness. This is very thin, so I put it in for a bit longer at 15min. Before entering the box, I wetted the surface slightly to help saturate the wood and encourage good bending.

Taking it out as quickly as possible, I set it in the press and clamped down as tightly as I could manage, using a length of pipe for extra torque on the screw.

A day later, I undid the press, and you can see that only about 80% of the bend took set in the wood. You can also see that the grain started lifting on the tight bend on the edges where it was not quite supported by the frame. This is, although annoying, not that bad as I will be trimming these down anyway. It did, however, serve as a warning to the limits of the wood which I push later on.

To get the last bit of bend out of the wood, I pulled the pipe from the box and turned my rusty old kettle into a steam gun. At this point, I had been bending this piece of wood for a few weeks and was ready to do literally anything to get the wood to stay bent.

Scratching my head and beating it against the wall, I let the project sit for a few days when I was out of town. On the trip, I talked with a friend who is a professional luthier and he suggested I try a hot pipe to bend the wood around. Having never heard of that, I was absolutely willing to give it a try. Apparently, this is a common method used in undeveloped countries who do not have access to a purpose made heated mandrel. All I had to do was heat the pipe (but not so hot that it burns the wood) and bend the wet wood around it. The pipe will boil the water and force steam into the wood, helping it both bend without breaking and take the set that is forced into it without it later relaxing.

For this, the pipe I had on hand was a bit tight of a diameter, but I later found a larger pipe that worked perfectly. Had I known about this from the beginning, I would not have ever bothered making the second press frame.

The pipe got the wood very close. One thing, however, was the ends were extremely difficult to bend without something to either clamp against the pipe or push onto it with sufficient pressure to bend the free ends. At the expense of burning myself a few times, I eventually stopped trying that and looked for another improvised solution. Had the wood been longer, it would not have been a problem as I would have cut the non bent length off, but that was not an option.

Instead, I pulled out my post vice from under my bed and clamped the
end in that, using weights and braces to force the wood into submission.

Craftsmanship is, as I see it, nothing more than persistence overcoming failure and the ability to creatively solve problems while starting with an educated guess at the solution.

Just as I finished the first bend, I realized I had to do something differently. There was no way I was willing to take five weeks to bend each of the six pieces. At that rate, I would never finish.

When I solicited help, I was told how thick sides of acoustic guitars are ordinarily. Having never made one or even played one before, I had virtually no idea and assumed that the wood (which came from a luthier's supply) would be close enough that I did not absolutely need to take it any thinner. That was wildly wrong.

I suppose that my initial reluctance to thin the wood was my lack of workspace. I did surface the sides before bending, but they were still about twice as thick as they needed to be. So, with a bit of annoyance at myself for not doing it earlier and at trying to do this without a workbench, I set to work. Between the joiner plane, files, and a card scraper, I managed to take the sides from 4mm to just over 2mm. This would not only make the wood easier to bend on its own, but would also allow the steam to do its magic more easily and leave a smaller discrepancy between the tension and compression surfaces of the bend for grain to lift.

Half way to the needed thickness. The top one is the thinned, bottom the original.

Final thickness. Making it consistently thick across the 9cm width by hand was difficult, especially the thinner it became. In the end, it was an average of ,05mm variation across the entire width and length, which was close enough for me.

Eventually, after what felt like an eternity battling this thing, I moved onto the second piece of the side. This one is for the top left that has the recurve. I figured it would be easier to bend it half way and then move to the other shape of the press instead of forcing the opposing curves straight off.

The new piece is on the bottom of the above picture, and took almost no set. For retrospectively obvious reasons, the first curve unbent in the steam as I bent the second, leaving one curve instead of two. At this point, the extreme radius of the curve started splintering the outside grain, so I stopped there to reconsider the overall design. In the end I decided that I would use the neck brace block as the outer wood for the top rather than trying to force the rosewood into place and likely breaking it beyond use and repair.

By now, I more or less totally abandoned the press as it was originally designed. Trying to wrestle opposing curves in one pressing operation was not working, so I used the one corner of the press from the original piece to bend virtually everything. Here, I also added two pieces of leather between the press halves and the side. These both helped distribute clamping pressure, retained heat, and kept the moisture from dissipating too quickly. That combination was enough to prevent most of the damage I had experienced up to this point.

Now that the first half of the sides is done, time to move to the second. This one is more complicated and has two more pieces and one additional corner to deal with.

For the first of four pieces, I measured and marked square. This one, as with the others, leaves about an inch of extra material to accommodate the bend and potential splintering.

Back to my trusty saw horse. Cutting wood only 2mm thick was difficult in this configuration to say the least. After this first cut, I did the rest with a knife and straight edge, which was faster and cleaner.

Due to using only part of the curve of the frame for this piece, I had to reconfigure the clamps to have even pressure both into and across the frame. For whatever reason, all the positions of clamps which I had not initially thought I would need (like the one on the top side of the frame above) were only about 2cm too far apart for the clamps to reach. So I had to cut some channels to accept the screw...

Next up is the small piece that serves as the double corner. I was worried that the piece, being so short, would break along the grain while bending. Fortunately, the leather served its purpose and it bent without incident.

Third is the three curve piece. This one was particularly tricky. Had the wood been inherently more flexible, I would have been able to use the press as shown above, but it was not so I could not. S it was back to that one corner to set the outer bends.

Again, as with the pipe, it was troublesome to have the bend carry all the way through the end of the side, so I had to start improvising again. It's a wonder how many configurations you can manage out of a few simple clamps and vices.

Adding the recurve...

Every bit of extra weight and spacing makes a significant difference in the end result.

If you take it slow and use all the steam at your disposal, the wood will eventually bend however far you need it to go. Outside the press, it takes a bit more overbending to keep the set you need because there is no pressure to help submit the grain.

I thought I had learned my lesson with the two top pieces, but with the success of the pipe and the other stuff, I found my confidence again. Or rather, over confidence. This bend did not work and nearly ruined me a second time.

So it was back to the pipes. This time I had the larger pipe to serve as the primary bending curve, and the smaller one for tightening up the fine radii.

Using a scrap bit of chrome tanned leather (nasty! Glad I held onto it though...) as a buffer from the hot pipe and steam, I slowly worked the wood to shape. These are the top pieces that meet the neck, and I stuck with the decision to not have the recurve this time for good.

Unlike the press and steam box approach, the wood takes its set and remains there almost immediately. There is no need to keep it clamped or tied. It is what it is, once it is dry.

Progress!

And finally, here are the six pieces of the sides. There is a slight amount of bending that will be tightened up in assembly, but not more than a few millimetres. Next up, trimming the final lengths and fixing them together.

The Luthier: Part II- Framing II

When I built the first frame for this project, it was intended to be used as the one stop forming solution, but that of course never works. Trying to clamp down with the needed pressure in that configuration was not working, so I needed to think of another solution. And that happened to be building another frame, this time in the form of a press rather than whatever that other one is. The first will still be invaluable during the building process, so it was not entirely wasted time.

First off, I wanted to solve the problem of height. The original frame was about an inch too short for the height of the sides, which is good for laying the supports and back onto the sides, but not so much for the actual bending. To solve this, I went with stacked plywood. Four layers thick brings me to within a few millimetres of the side height, which is close enough for me (famous last words...)

The problem was breaking it down by hand. Without a workbench.

Eventually I got it cut to size, four rectangles that I strategically sized so I was able to use it in three different positions relating to the three shapes I need to bend with it.

The layout for the shape was a challenge in design, having to think about the reverse cuts without being able to align them first. The double line in the centre is for the main curve of the body and the first corner, which will also be used for the other side that does not have it but is shorter in length. The other side is broken down into two parts. First (bottom left) is for the other half of the second side, which takes two compound curves. Second (bottom right) is for the part that meets the neck. That will be two pieces, mirrored. All together, I will be getting five different shapes out of the two piece frame. (the fifth being a small bit for one of the corners only an inch or so long)

Instead of trying to draw the lines four times, I just cut the first and used it as the master for the other three. Unfortunately, I had to make all the cuts with a coping saw that is just a bit too short to reach the full depth of the curves.

Also, without a work bench, I am forced to use my computer desk with a block of steel to weigh it down. Far from a good solution.

To reach the inner curve, I needed to take off the outsides first, that small distance being critical to reaching all the way across.

It was tedious and difficult with the space constraints, but eventually I had the first sheet cut. Three more to go.

In the end, it was finally done. Several painstaking hours with the coping saw and they were close enough in shape that I was satisfied to glue them together.

After a considerable amount of shaping and sanding, the sides were flush and square.

Next up I had to drill some holes to accept the clamps, as I do not have any deep enough to span the entire press.

With a hand brace, it was slow going. Despite only having to drill three holes, it took nearly two hours. It did not help that the teeth on the hole saw were worn out and chipped from drilling through sheet metal a while back.

As expected, the saw was too shallow to drill all the way through the form in one pass, so I had to extend the pilot hole to the other side and start again.

Periodically, I would chisel the pieces out to help relieve the trapped sawdust, which was not expended by the saw.

At long last, the holes are drilled and the form is ready to press.

For the first configuration, here are the clamps. Now finally, I can get to work on the actual project.

The Luthier: Part I- Framing

For the last 14 years I have been a percussionist, playing mostly symphonic and swing with four years of the marching variety, but since moving down south I have been unable to find a musical home. Music has, for such a long portion of my life, been one of the constant creative outlooks, and now that it has a glaring absence, I decided it was time to learn an instrument that was a little less oppressive to neighbours than drums and bagpipes (haven't forgotten about those just yet!). 

Despite my love for the violin, and although I may someday explore that as well, I decided to pursue a more celtic design. Mandolins, although in the right setting very enjoyable to listen to, I prefer the deeper acoustic tones. Guitars are fine and well, and there is such a wealth of resources available to learning them, but again, I am after a different sound. Mandolas, the larger cousin of the mandolin, are perfect, but harder to find. And then it came to me. To me, the best way to understand an instrument is to first understand how it is made. So, without any knowledge of the luthier's beautiful trade, I decided to build an instrument on which to learn. Because I am doing it myself, I am free of the commercial constraints of availability and reasonability.

The design will be a sort of hybrid between a mandola and guitar- the body of the former and the fretboard of the latter. In theory, it will yield similar sounds to the mandola while holding to the technique of the guitar. A 'Mandolar' if you will.

First, I needed a frame. For hollow body instruments, the sides must be bent into the requisite shape. Following the pattern of an A body mandola, I glued together a bunch of wood hiding out in the laundry closet in a rough trapezoid. Because it will be under a bit of pressure during the bending process, it had to be rigid and roughly the thickness of the side height. Due to my lack of a steam bending mandrel, I decided to instead turn to the route of a steam box. When the wood is sufficiently pliable, it is clamped tightly into the shape of the frame where it will dry and conform to the inset bends and curves.

Being extremely rough cut, the frame needed a tremendous amount of work to reach its final shape. However, my workspace is limited to the size of a bath towel folded in two. Without a workshop to use, or even a space that I can be unconcerned with producing dust and shavings, everything must be carefully done by hand.

Smoothing out the large rounds was the easiest part. My trusty No 52 spokeshave left a clean, even surface that required minimal correction later.

Periodically, I checked the frame for square, which I assume is extremely important when bending the sides. Torsion in the thin sides would be difficult to correct later, and this frame will also be used in the assembly and gluing of the sides to the back.

For the tight inner radii, I had little other choice than to use chisels and gouges. The process was even slower than before, and by this point roughly a week elapsed since first gluing the frame together.

After all the sides were finished, smoothed and sanded, I drilled a few strategic holes in the frame to accept clamps. Combined with some outer blocks, this is what forces the steamed sides to shape.

6 holes later (one for each pair of inner curves, one for the bottom, and one for each of the middle of the large outer curves), it's onto the joint braces. I have no idea if there is a technical word for any of this, as again, I have no experience being a luthier.

These strips will be in the corners of each joint where the sides cannot be bent to be one continuous piece of wood. It will both reinforce the joint as a structural anchor for the tension that will inevitably still be in the wood after bending (based on the test bends I did, the final curvature was never perfectly tight enough, requiring the wood to be bent into place).

Because the frame is not quite deep enough, the ends of the braces hang over a bit, although they are still longer than the mandola sides and will be trimmed after assembly.

First, the width of the brace is cut, sized to the frame.

Then it's back to work with the plane. Each shape is made individually to account for the direction of the sweep. Note the two finished pieces whose ends you can see above the plane.

Gradually, the waste basked filled with layers of shavings, changing from tan to brown to purple and back again.

Those four little pieces were the easiest, but the inner corner was a bit less so. Compound curves met at a skew angle, which made the wood too shallow to accept the needed dimensions. So I split the wood and glued it together to doubled its thickness.

First is trimming the width. I could have just worked from one side, but the two sides were not perfectly aligned from gluing, so I split the difference.

Then with a number of chisels, I cut out the inner corner and checked for square.

Transcribing the shape squarely onto the frame was tricky, but it does not matter all that much. The frame will only be lightly holding the corner block for bending, and during assembly, I do not want it to be tight anyway. The tighter the fit, the more difficult it will be to remove after the sides are glued together.

Scribing the lines using a handy knife, cutting the grain on the faces and across the height of the sides. Unfortunately, as I struggled with throughout the entirety of building this frame, the wood grain changed directions with every piece that were glued together. It made a sort of plywood, which is great for the strength but awful for shaping.

Slow work. Should have cut the long lines with a saw, but still would not have been able to chip out the rest because of how it is glued...

And there we are. The fit is a bit sloppy on the corners, but as I mentioned, not necessarily a bad thing.

Now we turn to the first major piece to be used in the actual mandolar. This block is what reinforces the butt of the instrument and serves as an anchor for a strap. Because it is not a piece of scrap, I did not want to make nay errors here.

Same as with the others, the block gets inset into the frame, and the necessary curvature traced. When seen from above, the shape of the frame should be uninterrupted for all these extra blocks and braces.

Close, but the corners need to go.

To the block plane to ease the geometry.

This, two weeks later, is where I thought I was finally done with the frame. Wrong! I still needed to make all the reverse curvature clamping blocks.

Fortunately, most of them were easy outside curves.

Trimmed out of pieces cut from the original roughing of the frame, the seven pieces were conveniently the perfect height.

Rasps, saws, planes and chisels made decently fast work (still monotonously slow, mind you) of the first five. Those remaining two, however, were a bit more complex.

Due to the grain directions changing in the block, I had to make several cuts with the saw and chip out the ribs, finishing with files.

Aside from one errant cut, it was not bad. Doesn't need to be perfect, as this will merely be a pressure distributor so the clamp does not destroy the side wood.

Finally, the last piece is carved and fitted to the interior corner block from before.

And that's it! About a month of work with simple hand tools over the foot of my bed as a work bench, and the frame is done. This will (hopefully) be by far the most tedious part that has no real place to be showcased in the final instrument. Next up- bending the sides to shape.