Official Luthiers Forum!

Hi all,

I am a full-time archtop maker from Spain. I registered a few days ago mainly to describe a new type of side bender. I am currently building the fifth version of a design based on the bending of convex surfaces (a concave surface is convex if viewed from the other side) and detachable blocks. It has a guided wire rope mechanism for using tensioned slats which I am very proud of, actuated by pneumatic cylinders. This mechanism pulls the slats tangentially at every point of the contour. When you see all my attempts at this and the final machine, you may wonder why I have wasted so much effort, but those among you that bend curly maple "Venetian" cutaways will understand. I have bent my sides with a pipe/strap and with a Fox Bender, but these are dangerous methods when you have heavy figure. Look at this:



All those side sets are curly maple, but I wouldn't risk bending the one on the left for making a cutaway, at least if the thickness is around 2.2-2.5 mm. I know there are makers that go down to less than 2.0 mm, but I don't like that. Curls are wavy grain, so they are more fragile than straight-grain wood. After sanding and truing the sides, their thickness can easily go to 1/16" or less in some areas and, while that is enough for a violin, it is not enough for a deep guitar, that can break easily with a moderate blow.

I decided that I had to use my best sets, instead of being scared to touch them, but I should improve my bending methods. Initially, I decided to redesign the humble conventional iron. I arrived at what I called a “double bender”, and a machine shop made it from aluminum after my plans:





Both rollers have heating elements, 200W cartridge resistors similar to what you can find in common electric bending irons. Those heaters are inside the axles, and are controlled through two separate dimmers. The rollers can turn with a minimal gap between them and the axles, so that heat transfer is good. The gap between the rollers can be adjusted for different wood thicknesses. Their diameters are different, to improve the contact area along the guitar contour.

In use, this machine works well to avoid cupping, and that may be its main advantage. It also supports and heats the bend from both sides, but that doesn’t seem to be determinant to avoid fractures. And it has real problems. One of them is the degradation of lubricants between the rollers and their axles. Synthetic oil for Diesel engines works better than any other lubricant that I have tried (didn’t try many, though), but seizure is a constant danger. I thought of using some graphite/steel alloy bearings (self lubricating) or a high-temperature grease. Also, some springs would work better than a fixed gap between the rollers. However, after the tests (I bent the sides for a few guitars with cutaways, but had to use a bending strap as usual) I lost my faith in the idea, and didn’t think it was worth more development. You can see that I didn’t even install a well made handle, even less after thinking of a second, quite different alternative. I will describe that later.

A pasta maker

I admire you for trying to innovate

_________________
A higher purpose for wood.
Rich Smith
Issaquah, WA

I don't get how it works, sorry...is that just a roller you can pull the side through and bend as you go? I don't see the advantage over something like a Doolin-style bender (www.doolinguitars.com/articles, the one on the bender for description and pics) which will keep two bits of spring or stainless steel in contact with the side at all times, and adding one or two heating blankets with temp control means you get full support, over the entire side, with even heat throughout.

It looks interesting. Can you show or tell us exactly how it's supposed to function? As far as seizing is concerned, you might consider using shell bearings like those used between the crank journals and rod ends in automotive engines. They function under high heat and are lubricated with engine oil. Keep this thread going.
Craig S.

Works like a steel plate roller only heated the nore times through the roller the tighter the bend till you reach min bend radius which would be equal to approx. 1.25 time the larger roller dia.

Same way a fretwire bender works but with just 2 maderal rollers

Thanks all for your comments.

Mattia,

That machine was simply going to be an improvement over the conventional iron. I reasoned that supporting the convex side of the bend would be a big advantage... but it wasn't. Being little more than a conventional iron, it has the same advantages over more elaborate machines, as Fox or Doolin machines and derivatives; for example, you can use it for bending in repair situations, where making a form is usually not practical.

csullivan,

Working with it was a lot like using a conventional iron, but instead of a rocking motion I used mostly a side to side motion. That was easy to do, as the cylinders rotated with little friction. There's not a lot more to it and in fact, as I said before, after working some time with it instead of with my conventional iron, I lost my faith in it and I didn't use it anymore. I had thought that pressure from both sides was more important than it really was.

My second attempt was a machine based on the convex bending of all the curves. Here it is:



There you can see three blocks that define three different surfaces. The bending process doesn't begin as pictured. Instead, initially, you have only the block corresponding to the lower bout, while the others are not present. The sandwich (slat/blanket/wood/slat) is placed/clamped on it, heat is applied and then its perimeter clamped. Then, the piece marked "18-2" (which means "eighteen inches guitar, piece 2") is put in place with its three bolts, and the sandwich is clamped to it. Next, the piece marked "18-5" is put in place and clamped as before. For cutaways, there were two pieces instead of the "18-5". You can see an outline of one of them to the left of the "18-5". The rectangular windows are for placing some clamps in key places.

This was my first trial at bending convex forms, and I see it so naive now. You needed to use thick gloves unless you wanted to get deep burns, and the effort to adapt the sandwich to the forms was very high. There was a toughness problem with the machine having a single side support too (I did that for removing the sides more easily). In short, I soon realized that this thing was not going to work without profound changes. The Fox design supports wood from both sides, and this design doesn't do that. Trying to bend with it was a sure way to break some nice sides. Then I went to correct some of its problems -- I will leave that for later.

I tried to fix the important problems with my following design. The result was the most Rube Goldberg-ish machine in the line:



This version had two lateral supports (the "chassis"). Removing the wood would be more difficult, but its toughness was improved. I redesigned everything, including the blocks, so that I could use the same chassis for my 16, 17 and 18-inches guitars.

The mode of operation was similar to my previous machine. A difference was that the lower bout curve was formed using a shoe (not shown), similar to the Fox bender sliding presses (although reversed: it began near the tailblock and ended near the waist). The rest of the operation was similar to my previous design (the "All Convex Bends" concept), using detachable blocks, but the clamping was done with the screws. They were tightened against the sandwich with the intermediation of a series of metal pieces similar to the bar at waist in Fox design, but obviously different as they pressed from the outside for all the perimeter.

Due to the many clamps, those that pressed the waist area were located inside the machine:



This design could not bend the non-cutaway side. I left that for later, as I knew that there were going to be things to improve. For example, clamping the sandwich was very slow. Too many screws to tighten, even after some improvements that I made later (among other, springs for each screw clamp, not shown). Also, the slat edges became damaged, mainly because it was impossible to tighten the screws to both sides of the bar at the same time. Another problem was due to the blocks being fastened to the chassis using bolts. They were not stiff enough (8 mm diameter) to support all the pressure exerted by the screw clamps unless they were tightened a lot, but that deformed the machine. Here you can see one of those blocks:



The holes are for the 8mm bolts, that screw into the inserts that can be seen here:



After the clamps were tightened, the machine was so deformed that it was difficult to hit the insert with the bolt.

In spite of these problems, I managed to bend some curly maple with it. Although the bent side was satisfactory, the process to achieve it was not. There was a lot to improve, so I started to think of new improvements.

Ok, now you're scarin' me!
Seriously Fernando, welcome to the forum...we need all the proven archtop input/experience we can get! Interested to see where this is headed.

_________________
Dave
Milton, ON

It's been a long time since I posted here. The objective of the thread was to explain the "guided rope" mechanism that I mentioned at the beginning. Before doing that, I decided to explain my previous approaches, but I didn't explain the final machine in the line. Here it is:

http://www.guitarrasjaen.com/English%20 ... Bender.htm

It's a somewhat tricky concept, I hope you'll understand my explanations...

Fernando, it is so good to see you back here again!
This statement of yours, "you may wonder why I have wasted so much effort, but those among you that bend curly maple "Venetian" cutaways will understand.", really hits the nail on the head. Those of us using highly figured maple sides can really appreciate your efforts to take the variables out of the process.
Keep up the good work and, again, it is so kind of you to share.
Best regards
Nelson

Thanks, Nelson.

The machine works very well now, but I have to make some improvements. For example, guides for the rope (and everything else) with less friction, screwless attachment of the external blocks, possibility of adjusting the alignment, etc. I will have to make almost a new machine for all that, and it will take some time. When I have that new version, I'll try to post the construction details and even make a video. I wouldn't be surprised if you have troubles trying to understand the operation from what I said yesterday. I have had people here watching me demonstrating the machine for them and a few minutes later they didn't have a very clear idea of what they had seen.

I use the machine at 10 bar with 40 mm cylinders. This makes a force of around 1055 N at each rope (235 lb). At first I was scared, but now I see that the machine keeps up very well and deforms minimally. I don't go much above 100ºC (212ºF). I have even tried once to bend curly maple cold (but wet) and, although it obviously didn't keep the bend, it didn't suffer any damage.

http://www.luthiersforum.com/forum/viewtopic.php?f=10106&t=22422&hilit=fernando
http://www.guitarrasjaen.com/English%20Version/DesignForFeder1.htm
Fernando--I'm hoping that you will someday have time to finish the above translation. Great stuff!!

I found it very interesting that you could bend cold curly maple with your current machine. That speaks well for the concept.
I had to study your illustrations for a while before grasping the concept of the tensioned cables. It seems somewhat akin to the two tensioned cables used in some duplicators and also drafting machines where tensioned cables are used for alignment.

I too was very fascinated by the Taylor bending machine but I do think that your design puts it within reach of us small builders. One thing that I did note about the Taylor machine was the wood not being heated until it contacted the inside mandrel which allowed the the rest of wood to remain cool until bent. The heated rollers you show in the earlier post to this thread look like they would have some potential in that type of design. I could see, however, a challenge in how to place the inside blocks and connect to the power source. I don't want to redesign your machine for you, just thinking outside the box a little.
Nelson

Fernando
What is your technique for removing the sandwich after bending? Seems like that can be an issue when dealing with the springy steel slats?
Thanks
Nelson

Great topic! I am currently bending figured maple cutaway sides on a fox bender with a cutaway attachment. It works ok, but only ok... I am using a blanket and SS slats but my technique is under constant refinement. It definitely seems to be more art than science, and bending cutaway sides is always a high stress event.

I have spent countless hours trying to design or at least concieve an improved bending system, but usually end up at the same place: The shopmade testbed doesn't get it done, or is too cumbersome to get the side bent in the time frame required, or the machining costs to build a workable model would be prohibitive. The most frustrating aspect of all this is that SOMETIMES a Fox-bent cutaway side will come out perfectly (or nearly) and the next side bent under the same time/temp regime will crack. All I am looking for is some predictability, and so far it has been elusive.

I envy the classical builders bending modest simple curves in EIR... and I try not to resent them, but it's hard.

I'll be following this thread,
Gordo

http://www.guitarrasjaen.com/English%20Version/JazzGuitars.htm
For those of you just joining this discussion this is the link to Fernando's website.

http://www.guitarrasjaen.com/English%20Version/KitsEn.html
This is a link to his exciting offering of an archtop kit.

http://www.guitarrasjaen.com/MK16_V100.pdf
And here is a link to the instructions draft for building the kit.

http://www.guitarrasjaen.com/English%20Version/Bender.htm
His latest side bending machine:

I'm trusting that Fernando doesn't mind having these links posted.
Great stuff!!!

Gordo--You are not alone in your "bending frustration".

Nelson

Thanks for your comments, especially Nelson!

Removing the wood is not elegant, but it is safer (for the wood I mean) than it looks. I remove the pressure to the cylinders and then remove the inner blocks corresponding to the convex curves. For the cutaway side, the inside of the cutaway and the waist blocks. For the non-cutaway side, just the waist block. The slats are spring steel, and are much thinner than what LMI sells, but anyway everything jumps out of the machine. There must be a more elegant way of doing that, but I didn't find it.

This machine is not the same as others, and everything needs more research. For example, I told you that I tried to bend cold. It was an experiment, because at first I was having problems with the temperature. I was bending at 300F (150C), as I read everywhere. When I arrived at the cutaway, the wood was very dry there and it scorched. Then I tested if I could bend cold. I used a piece of curly maple, but it was narrow (some leftover from a guitar side). Narrow pieces have less surface, and the slats put more pressure on them, which is great. It was succesful, so I tried to bend at a lower temperature, and I saw that I didn't have more drying/burning problems.

In two or three months I will have another machine and then I'll make a video. I want to improve a few things, as I said. It is not easy to decide modifications for this machine. Some time ago I wrote a list of drawbacks:

- The machine is not adequate for a high production environment. As with the Fox Bender, you must wait a few hours to remove the wood.
- It is not easy to design and make. There are lots of small details that have costed me many hours of work and, in some cases, frustration. I have designed and made every single part more than once and, for some, more times
than I want to remember.
- The machine has a learning curve. A side can be bent in a few minutes, but you will have to go through a list of operations in a strict order.
-This machine can be set to bend different guitar models, but changing from one to another is not as easy as one may desire. If you work with small batches of sides for different models and you cannot change that, this machine may not be for you.

Also future improvements, that I will try to implement in the next version:

- The machine will be easier to use if there are no bolts for the outer blocks, that can go into recesses in the chassis.

- Individually heated inner blocks (or smaller independent heating blankets) would add more control. At tempco.com they have many small blankets that could be great for this.

- The wire rope must pass through some guides for pulling from a different direction, depending on the side being bent (cutaway/non cutaway). After those guides, it must pass through the external blocks. It is not surprising that friction is a problem. When the rope follows that complicated path, it loses tension quite noticeably. I'll replace the wire rope guides with bearings. The friction at the external blocks can be diminished by carving them from PTFE, Delrin or some other low-friction material. Or, even better, they can be substituted by circular pieces running on a central bearing or by pieces with several bearings. While the shape of the outer and inner blocks is not too different, everything would be much the same. A good design of the external blocks (determining their size and shape) should also allow the bridge to stay in equilibrium (or almost) while the blocks are placed/removed. Currently, I use metal pieces that are not elegant as they must support a good part of the weight of the bridge (you can see them in the photos at my web site).

- I must have two alignment adjustments for the sandwich. Currently I have one at the inner block for the lower bout, but it is not enough (and it is not well designed).

- A lighter bridge is possible.

I forgot to say that, just before removing the inner blocks to extract the sides, I "separate" the slats (disconnect the two wire ropes that join them, visible in the photos).

Thanks for sharing all this with us, Fernando.

I've been considering building a more modest version of what Taylor uses (convex bends only, tensioned slats) but your design is definitely beyond anything I've come up with. Please do keep us up to date when you finish the next round of changes.

_________________
-Clint Beacock

I like the idea of individual blankets on the inner blocks as that would allow only heating the current bend during the process.
The bends could be sequenced without being in a rush before all the moisture boils off.
For small builders, the time spent actually bending sides is a small percentage of the total build time.
Is there usually a short unbent area between bends on most guitar shapes and could that area be left relatively cold?
Nelson

This diagram explains the machine a little more:



The guides are the blue pieces. There are two, because the rope must pull from different places depending on the side being cutaway or not.

The green block is exclusively for bending the non-cutaway sides. The purple block is fixed: when the sides are put inside the machine, they assume the curvature of this block while they are cold. I've never had any problem damaging a side because of this, so I found that was a good decision: a removable block for that area would make the rope and the bridge much longer.

You can see the shape of the side as a black line; it is quite close to the blocks that bend the concave surfaces (waist and cutaway inside), but farther from the others. This is because the heating blanket goes inside the guitar, to avoid burning the outside. When I make independent heating blocks, I will have to change them because of their sizes.

There are strong metal tubes inside the inner blocks (the circles in them). I learned that from a previous design, where I used screws and they bent due to the pressure. By the way, it is the outer block that is represented in the drawing for inside the cutaway; the inner block has two tubes, but the outside block is smaller, as shown. There must be a reason why I did this...

The machine can bend 16, 17 and 18-inch guitars, so I have a lot of blocks:

I'm going to go out on a limb here and say that I'm surprised that there isn't a tremendous amount of interest shown here in Fernando's invention.
I could be wrong but it looks like this is a tremendous breakthru in side bending technology and within reach of small quantity builders.
It's very difficult to completely grasp (at least if was for me) what's going on with the machine in spite of Fernando's great efforts in diagramming and explaining.
That may partially explain the lack of interest.
I realize that not all builders are concerned with tight cutaway bends in curly maple as it's used primarily on archtops.
Anyone watching the Taylor bending videos on Factory Fridays will see that Fernando's machine takes the technology "over the top".
My two cents worth.
Nelson

I agree, Nelson.

There's quite a bit going on with this design, but once you get your head around it, it really is a wonderful design. Brilliant even in some aspects.

I suspect your correct that for most people, the fox-style will always do them just fine, but since I've not yet built a bender, I plan on going this route.

Again, PLEASE keep us informed when you finish the next revision, Fernando.

_________________
-Clint Beacock

I made this video to show how the machine works:

http://www.youtube.com/watch?v=FPREblh4ZiA

Beautiful, just beautiful.

As Nelson said, it's a Taylor-like bending solution that can be pulled off by a small shop builder!

_________________
Bob Garrish
Former Canonized Purveyor of Fine CNC Luthier Services

Very interesting design! However in my opinion it fall short on a couple counts. There are far too many adjustments to make in 4-6 min time frame. A lot of the adjustments are unnecessary because the tension of the bend when using single point of contact radial pull cauls and a press screw caul will produces everything that all those individual screw jacks are trying to control. The design is innovative but unneeded complication makes it fall sort in my eyes

In my mind's eye; the requirements for a great bending machine are;

1. Easily interchangeable form shapes.
2. Proper support for the wood during and at completion of the bend.
3. Simplicity of operation, speed of operation, accuracy of operation.

Machines like that last one are great exercises in over engineering for requirements of the task at hand. (just my opinion). Butt I have to admit there was great bit of thought and detail attention put in to that design.