When I last left you, we were right at the point of beginning construction on many of the jigs and forms that will make this endeavor much easier. Conveniently, the timing works very well since I’m still awaiting the opportunity to make it down to Austin to pick up that Performax 22/44 that I got off eBay (thanks to my cousin, Brady, for getting it for me). This, of course, will allow me to get all that beautiful wood down to thickness easily and more precisely.
In the meantime, this blog entry will detail some of the progress being made on the main mold, which will hold the sides during construction, and the side bending “machine.” Yes, I decided to build something that would make this guitar, and future guitars, without all of the headaches traditionally involved with bending over a heated pipe. More on that later.
In Blog #1, I mentioned that part of my inspiration for building a guitar at this time came from seeing a You Tube series of video from the “Hand-tooled Guitars” guy. I gained the confidence required for tackling something I always thought would be one of the more ultimate expressions in woodworking. But, more than that, I really liked the design of the guitar, which he classified as an 12-fret OM thin body guitar. I’m not too interested in building a thin body design, yet, but I did like the basic shape, so I ordered an acrylic template sent my way for $12.
In viewing the template and comparing it with my “grand auditorium” (GA) style Taylor guitar, I realized that I wanted the lower bout to be a little wider. So, in tracing out the shape onto the sheets of high-grade plywood that I would be using, I used the template as a “french curve” to connect the markings of the new dimensions I would employ on this guitar. A typical GA guitar is similar to a Dreadnaught, being almost identical in its dimensions except for its width at the waist, where the GA curves are more reminiscent of Marilyn Monroe as opposed to Charlize Thearon. Therefore, this guitar body, and its mold, will have the following dimensions…
Body Height – 20.5”
Upper Bout Width – 11.75”
Waist Width – 9.875”
Lower Bout Width – 15.5”
Depth of Body – 4.75” (tapering to the neck)
Creating the Main Mold
After tracing the shape, a jigsaw is used to rough cut half of the guitar out of the plywood. The new Ridgid spindle sander is then used to sand to the line. Five more “halves” are roughed out with the jigsaw and a straight pattern bit is used on the router table to duplicate the clean lines of the first half, which is used as the pattern for the rest.
The rough cut outs from the mold are placed stacked together as well to be used as a form for the bending machine. In the meantime, I clean these up simply with the spindle sander. It does not need to be perfect, since minor errors in the form will be averaged out a bit by bending through spring metal slats (more on that later).
Three halves are then stacked per side, showing the complete shape of the final mold…
And those are then clamped up and screwed together…
And finally, a board is screwed into the bottom to hold the halves together and a large clasp is employed at the top to keep the halves closed and registered with each other. The bottom board flexs to allow an opening at top top of perhaps a 1/2” or so, which is enough to allow the guitar sides to slide out.
That is put aside for a while and I begin work on the “machine.”
Bending Machine Construction
Now the fun part…why do something minimal when you can have tons of fun building something really complicated?
I did my studying and I realized that the highest rate of success when it comes to bending guitar sides happens by bending entire sides (or ribs) over a large form, using uniform heating across the entire board. That leads luthiers (and pretenders like me) towards one of three different designs.
1.) An “unheated” machine where the board is boiled and simply pressed onto a form.
2.) A machine that uses powerful, dimmer-controlled light bulbs under a hollow form to provide heat (Fox Bender)
3.) A machine that uses a silicon rubber heating “blanket” and a temperature controller (or dimmer) to provide heat atop a solid form.
My machine of choice is #3. Why? Three reasons.
1.) I already had the solid form from having constructed the mold.
2.) The success rate for this machine is very high, being versatile and precise with its heating controls
3.) I wanted to fool around with a cool temperature controller. :)
To start, I build towers for the device from high grade 3/4” plywood. These towers would be slotted and would hold a shaped board that would be screwed down by some sort of screw/clamp/vise thingie. This board, shaped to match the curve of the guitar’s waist, presses and secures the sides at that point. Truthfully, I had no idea what I would use for the vise “thingie” when I started this step…
Slots for the towers are added via hand-held router, straight bit, and guides. A 3/4” plywood bottom is added.
The boards for the form are connected with lots of double stick carpet tape and a bottom board is screwed onto the bottom, which will serve as a guide to properly position the form into the machine. Boards are added to the front and back of the base to receive the form.
Shown is the curved press board that will be laminated to another sheet of ply. The board is shaped on the spindle sander. Did I say I love that machine?
So, what to do for the vise-clamp-pressing-down thingie?
After work one day, I thought I’d get inspiration while walking around my local Harbor Freight store. I knew they wouldn’t have large screws, but I thought about maybe getting one of their cheap bench vises and pilfering just the parts I needed. No, there’s gotta be something else. Think!
Every design I had seen for this machine used a centrally-located screw, but in the back of my mind I was curious about what keeps the curve uniformly flat during the bending process when bottoming out the curved block? What’s to keep this block totally square to the forms? Would it lift up at one side of the waist? Well, if I built it right, the answer is probably not, but I didn’t want to take that chance if given another choice.
Then, walking down the aisle with the clamps, I saw a 12” wooden screw clamp for $10. Illuminated, I didn’t even use a coupon on it. Yep, I paid full price at Harbor Freight.
I took my prized clamp home and I did this…
After further trimming and rearranging of the barrel bolts and screws, I did this…
The holes in the top of the curved block are recesses for the screws as they are loosened (lifted) away from the form. Because the screws are double-threaded, with left and right threaded halves, I knew that I’d have a total travel of TWICE the distance from the bottom of the holes to the barrel bolts in the blocks. And, yes, those blocks are re-purposed from the clamp. In total, it gives over 4” of travel. Plenty enough to slide in a straight side and slide out a freshly bent side (with its form).
And when fitted to the towers, you get this…
Of course, the screws would have to be cut, but that’s a piece of cake.
The blocks are attached with pocket screws to the curved press. Here is the bender with the “waist vise” complete and in the down position…
With a nice fit at the waist! The only real complication is that by using this woodscrew and reversing one of the handles, tightening the clamps means that you can throw “lefty-loosy, righty-tighty” right out the window! But really, it’s no big deal and it feels right when you do it. Plus, it assures that I get equal pressure across the waist.
Playing with Electronics
About the time I finished the machine’s construction, the toys came. Namely, a 6” x 36” silicon rubber heating blanket ($88 from Omega.com), a “k” thermocouple ($9 on Amazon), and a 1/16 DIN PID temperature controller ($26 from Amazon).
I had my buddy, Ross, who is smarter than me, come over to assist with the fun. Once we realized that the controller is merely a relay and that both the controller AND the blanket need their own power sources, we got the thing correctly wired and working properly…
PID stands for Proportion/Integration/Differentiation, which uses calculus to control how the controller reacts when the temperature gets to its setup point. Though I’m a math teacher and teach calculus, I had no interest in delving into manually setting up the calculus, so we took it through its “auto-tune” setup. We had some issues at first. Compared to my Ideal voltmeter (with its temperature sensor), the thermocouple just didn’t seem very responsive…taking a long time to register. At least, that is until I realized that my placement of the thermocouple at the heating element was leaving the tip of it exposed. Doh!
Once corrected, the PID controller just zoomed. Heat of the blanket began to match the voltmeter reading very closely and once it reached its set-points the PID auto-tuning really got locked in.
Lets play! I had a remnant of a thin walnut board that would be perfect for a test piece. So, I soaked it in my swimming pool for 10 minutes, put it underneath the blanket, and began putting pressure on the waist. Because the spring clamps aren’t installed yet (spring loaded clamping cauls work the board ends down and hold it in place), we awkwardly held everything down with parallel clamps. It produced this…
Woot! Ross and I did a giggle dance. I felt like I was Tom Hanks on a tropical island having just started a fire! Too cool. I took it to school this week for “show and tell,” whereas my high school kids were mildly amused.
PID Temperature Controller Enclosure
If you purchase a bending machine from LMII (Luthier’s Merchantile), you pay over $500 for their setup. With it, you get a controller in an actual box.
So, I needed a box.
I have a couple of “boxes” around the shop that were experiments while learning the Incra TSLS system. One box, just the right size for this application, was 1/2” milled marblewood sides joined with corner post dovetails made of walnut. The box had some chip outs, but it was salvageable enough to sand it up a bit, fill some gaps, fashion a walnut lid, and cut out some holes for the controller face and the cables. This is the result…
And, still without the spring clamps for the ends, and a slight modification to the end pieces to raise it off the table a bit (more clearance for longer sides), here is a larger beauty shot…
It’s a long build, but I think it will make bending my sides a breeze. The test piece I used was pretty much bent by the time the controller reached the set-point…and I only left it cooking for about 10 minutes afterwards, before I let it cool off just enough to pull it out of the machine. A little spring back, but not much considering that you would typically leave them in the form overnight (without heat). The test piece was a little thin on one end, but a little thick on the other, so it was a good test.
I will add the spring clamps later…unless I can device something a little more creative. I will also be receiving some spring steel slats from McMaster-Carr that are sized to match my heating blanket. The idea is to create a sandwich of the layers, whereas the steel uniformly distributes the heat, provides smoother action during the bend, protects the blanket from the clamps, and provides a great visual to any places where the board isn’t properly against the form.
I’m excited to add this tool. Whereas you certainly do not need this to build a guitar, it will definitely encourage me to build MORE guitars. And it’s given me thoughts to even other woodworking projects that might incorporate bent boards and laminations.
NEXT BLOG: Rosette design and Performax 22/44 thicknessing
-- jay, www.allaboutastro.com