The bending table and fixtures are close at hand to the steam box. With this table and fixture setup it is easy to move the bending forms around and securely fasten them to the table.  I made a guess on the amount of spring back to allow for when I laid out the radius for the bending form. After a trial bend, I decreased the radius to minimize the amount of tension when gluing the piece to the table edge.

After steaming for approximately an hour, a piece was bent and clamped around the form and allowed to cool for about 15 minutes. It was then removed from the form and held to the measured length with a bar clamp.  All the pieces  were bent and then set aside for a couple of weeks to dry thoroughly while we worked on other parts of the table. We made a few extra pieces as I had picked an interesting board of walnut for the edge band and we had some minor kinking problems in the bend.  Some of this can be dealt with in the fitting, which I will discuss in a further post.

From my experience on the outer edge, I laid out and cut the form for the inner edge band.  In the end though, these pieces continued to bend as they dried and were over bent. This turned out to not be a problem as it was very easy to clamp them in place.  The outer edge band was ¾” thick and the inner was ½” thick.

The next post will cover doing the inlays dividing the sections prior to cutting and fitting the edge banding.

When the veneer sections were edge glued as shown in the last post, they were fit tight to each other.  In the long plan there is an inlaid strip that goes over the joint area and connects with the inside and outside key blocks of the edge banding.  From the  board clamped on a section line we set 5 of the 6 sections in place and marked the outer table edge on the veneer.   We then spread Weldwood solvent based contact cement and when it was time to set them in place, we used strips of formica underneath so we could position the sections accurately.  We set one section at a time, not worrying about the joint between sections because of the future inlay. After 5 sections were laid, we final fit the last section and then glued in place.

Time to fit the last section

Tools for testing the bond.

The sections were rolled and pounded with a block and hammer. A light stick of wood was used to tap on the veneers to find any spots that needed more pressure. Using the wood as a drumstick you can hear the hollow spots.

Bottom Side Veneers

The bottom side veneers after they were trimmed.

. 

The top was then turned over to do the topside.

Veneer Plan

For this table we are going to do a pinwheel layout with a random match since we cannot get enough veneers out of any one board.  To make the veneer manageable for gluing to the ground (table top), the top was divided into 6 section to be divided by an inlay that ended at key blocks in the inner  and outer edge bands.  After looking over the stock of wood for the veneers, most of our boards were in the 7 ½” range so we could divide each section into 7 veneers with a cord length of approximately 7”. We then cut the 8/4” stock into rough lengths, followed by face jointing, planing, edge jointing and ripping to uniform width.

Resaw setup

A guide was set up on the band saw to cut pieces that were 3/16”, allowing us to get 7 veneers per piece of 8/4” stock

This is the pile as it came off the saw.  We always put a V mark with a felt pen on the edge of the board before sawing to be able to keep track of each board and its order.

The Grizzly Drum Sander

From there we went to the drum sander and sanded down to 1/8”.  The drum sander was set up with 60 grit and 100 grit on the 2 drums. They were then sorted into front side and backside piles, each board kept in its own pile so we could randomize easier when final layout was done.

A table saw jig was made to saw the angle on the veneers. It had to be done in 2 passes.  I made a template as accurately as I could of the finished veneer angle cuts. Setting the pattern on a veneer and drawing the lines, I was able to set the first cut.  All the veneers were cut for the backside veneers first.

Before the veneers were cut, we did a random layout to see how it looked.

From the Circumference numbers, we stretched a thin steel tape around the edge and made accurate marks at the 6 quadrant points. Setting a long straight edge across the top, we were able to clamp on an accurate boundary for a quadrant.  The fence was reset on the jig to cut the 2^nd angle and a  batch of 7 was cut to test the fit.  The jig was tweaked to get it as close as possible but the final fit was done by a tapering cut on the jointer.  Only 2 to 4 veneers were jointed in this manner per quadrant.

3 quadrants were set up at a time and the veneers were edge glued. Waxed paper was put under the veneers and then they were clamped in place. The edge clamping pressure for glueing  was provided by the wedge action from the shape of the veneers. They had to be clamped so that as you pushed on one the others wouldn’t pop back out.  After they were set up in this manner, we removed one at a time, added hot hide glue from a small opening squeeze bottle and quickly pushed the removed veneer back in place. Only a bead of glue was put on the stationary veneer.  The edges were checked for being flush vertically; end clamped and held down with a pad and brick.  The glue was cleaned off after a couple of hours with a scraper. Wiping off the excess glue with a wet rag would add too much water to the veneers and cause them to buckle.

We only had to repair one quadrant that was broken in handling and it did not break on the glue line.

After allowing the quadrants to remain clamped overnight, they were scraped on the back side and stacked with spacers.  A fan was used to drive all residual moisture from the panels. We moved on to steam bending the inner and outer edge bands for the table.  That process will follow in the next post.

Tracing the curve

We then used 80 grit 3M stickit paper  attached to the table edge to sand and true the block.  The next layer, 1/4″ MDF and the top veneers will be flush cut to this edge so we want it nice and smooth.  No mill marks or chips outs that the router bearing could find.

80 grit 3M Stickit paper

Tuning up the sanding block

Once the edge was sanded, it was time to add the Resin Impregnated Honeycomb.  It was very easy to rip the material on the table saw and use a straight edge and matt knife to cut the angle cuts. The sheets come in 2 ft x 4 ft pieces and we were able to use the angle cut offs to be more efficient in the use of material.

Filling out with Resin Impregnated Honeycomb

The last picture shows all the honeycomb in place.  We then took all the pieces out and vacuumed before gluing.  We used Unibond 800 as it has a long open time and rolls on easily. The compartment and the framing was coated and then the backside of the MDF was coated. We did 1/2 of each side at a time so as not to have any kind of glue up panic. We used a narrow crown staple  3″ on center.

After the glue dried overnight, we flush cut the excess MDF and flipped the table over to do the other side. We could tell the table was much stiffer after the first layer and very stiff  when the 2nd side was done. The next post will be on cutting the veneers to getting them edge glued in section.

Plan view with math

Table cross section

An important step is to build a platform without any twist (wind) that would translate into the top. Cross site the 2 levels and shim till they line up.

Making sure there is no twist

We built a flat framework to be roughly 6” inside the finished table OD.

2 sheets of 1/2”  multiply form the center layer of the core.  They were joined using Lamelloes and glue. The finished top will be 2 ¼” thick with 7 layers total.  With resin-impregnated honeycomb used for 2 of the inner layers, it becomes a torsion box structure.  Some of the tables I have built in the past have been so heavy that you have to get a crew of young men to move it.  The goal was to build a top that could sit on a minimally sized base and have its own structural integrity and be light enough not to have to get the young boys to move it around.

Eryx making sure we do it right

From our center point, the circles were drawn and the outer edge routed.

Router on a trammel.

All the poplar framing that gets glued and stapled to both sides could then be flushed trimmed with a bottom bearing router bit.

The inner circle was routed on the first side through the poplar and halfway through the plywood.  It was then cut out with a saber saw and flush cut after turning the table over to do the framing on the 2^nd side.

Adding the framing

We divided the circle into 12 segments for the framing and for compartments for the honeycomb material.  We used a long thin tape measure around the outside to easily do the layout.

Cut almost through

Ready to turn over

Ready for the honeycomb

This last picture shows the top with all the framing and ready to cut in the honeycomb material. We put in an extra batten over the plywood joint for a little extra stiffness.  The ¼” layer will go on crosswise so as not to create a hinge line.

If there are any questions on specifics I would be happy to answer.

Carving on a hollowed turned ball

The biggest challenge for me so far, other than doing the layout drawing was to figure out how to hold the ball while carving and not break it since the glue joints are so small.    This jig allowed me to support the ball from the part of the ball taking the pressure of the carving, ( the 2 metal dowels) and to give me just enough  pressure from the padded clam shell jig to keep it from moving roo much.  One side of the clam shell is fixed to the plywood and the other has one screw to pivot on.  The quick clamp can thus adjust the squeeze pressure,  Hope I don’t break the ball after all this work.

Model for 8' dining table

This is a concept model for an 8′ round dining table with an inset lazy susan.  The top will be veneered with shop sawn radially laid walnut veneers in 6 separate sections. The edge banding inside and out will be steam bent black walnut. The table top is 2 1/4″ thick with a multilayer resin impregnated honeycomb core for strength and lightness. The process with pictures and text will post soon.

I just finished a run of 8 balls made out of woods of different thicknesses. Using the proportion formula that I came up with (see last post), I made a list of the finished box size for each one. I sanded the interior face of each board to 220 grit and gave it 2 sanded coats of Seal Coat shellac followed by paste wax. I also gave the exterior face a seal coat to help keep the board as flat as possible. Using a glue joint rip blade, I sawed a 45-degree bevel along one edge, leaving the width slightly wider than the finished box. I like to keep track of the pieces so that I can align the grain when I glue up the box. Each board was numbered and then square rough cut to length on the miter saw.

From there I go to the table saw with a dedicated sled for 45 degrees. It is very important to be on the money with the bevel angle as there is no fitting done after the sawing. Each piece gets an adjacent bevel cut as shown in the right hand pile above. All the different boxes get this first cut on the sled. After this is done, a stop is set up and the other 2 cut are made for each box size. Accuracy is key so make sure all cuts are true 45 and square to each other.

Assemble the pieces in order, buff out the waxed face and sign one of the interior faces at this time.

I then layout 4 of the 6 pieces in order face up and tape their edges together, keeping them lined up against a straight edge. Flipping over this grouping, I use a toothbrush to apply glue (Titebond III) to all the joining edges. Since I align the pieces with the grain, these joints are all end grain so make sure you don’t end up with a dry joint. I coat them once and then go back and apply a light coat again. Fold the box closed and press and tape the last joining edges together, making sure that the top & bottom edges align. Glue the top and bottom in using the same procedure and taping in place. Do not press either piece in to hard, as they will wedge the box apart. I orient the grain of the top and bottom in the same direction. I start with 3 quick action clamps, tightening each one a little in turns so as not to spread the other joints. Look over your joints carefully and add opposing clamps as necessary. Remember that only a small amount of the gluejoint will remain in the end, the most important part being at the short point of the miter. I don’t touch any of the glue squeeze out on the interior of the box as it pops out later after turning.

Some of the balls from this run in the finish room.

After making quite a few balls, I decided I wanted to understand the relationship between the material thickness, the ball diameter and the size of the resulting hole. It would be nice to have a simple formula to figure out how thick the stock should be for a certain size ball or if I found a nice piece of wood, how big could I make a ball with it. Looking at the balls that I had already made, I looked closely at the ones that had a nice size hole for the size of the ball and took measurements. So it was trial and error to find what I liked and then to develop the math behind it.

For this example I will use a material thickness of .750”. Easier to work in decimals at this point. The ball I made with this material was 4 5/16” or 4.335

Knowing those 2 pieces of information gives me the proportion of the material thickness to the ball diameter.

EF/ 2(CF) = X

.750”/4.335” = .173 This number was generated by trial and error.

So if you want a 6” ball , The material thickness = .173 (6”) = 1.038” or 1 1/32”

If your material thickness is 1.750” , the box size would be 1.750”/.173 = 10.116” or 10 1/8”.

I figure I lose about 1/8” in diameter in the making of the ball so if you want to get close to a certain size, factor that into the initial box size and size your material accordingly.

The hatched area in the drawing is the plan view of the opening and how it was generated. I would need a mathematician to prove if my drawing is correct but it seems to work in the real world.

2. The teak box was glued together with epoxy but for subsequent balls I used Titebond III. If the triangular holes end up fairly large, you can see inside the ball easily so pre-finishing the inside and waxing will make it easier to remove any glue squeeze out in the final part of the process. I also sign the ball on the inside before gluing together.

3. In the picture below, the Hickory ball had greater thickness for its diameter and shows the smaller openings. In a post coming up I will show the geometry and math for sizing the material for a given box/ball size or vice versa.

4. I picked the axis that I wanted to turn on and located the centers, drawing the circle and the material thickness. . See figure 1 for the way I orientate the grain of the wood. With the axis as shown, the first turning will be standard spindle turning. It is important to locate your turning centers because if you are off center, the diameter of the cylinder will be less than its length, making the holes bigger and glue joints smaller.

5. I rough cut the box to a cylinder and then turned and glued on mounting points for attachment to the lathe using CA glue.

6. Mounted on the lathe and ready to turn.

7. I turned to a true cylinder, the goal being to make the diameter the same as the length. I am usually within 1/16”. The more you lose in diameter at this point, the larger your openings become and less final glue joint size. I figure that the final ball will be about 1/8” smaller in diameter then the original box, assuming that everything goes according to plan during the turning.

8. I located the centerline with a solid line and shaded on either side, an area to stay away from when rough shaping into ball form. Thank you Richard Raffan for this and much of the following method.

9. I made hardboard templates for checking the progress while turning. It is easy to see where the material needs to come off. If you orient your pieces as per figure 1, the turning techniques will go between parallel and face grain as you round the corner during this second stage of turning. I go between a spindle gouge and a bowl gouge for doing this part of the turning.

10. The first end is rounded and the openings are showing.

11. Close enough at this stage although the closer to true round at this point, without overturning, the easier it is to remount in its next orientation and keeping it there for the next turning stage.

12. I turned the ends at the attachment points before parting off. In subsequent balls I shear scraped more of the nub left under the mounting point to get it truer to round. This makes it easier to get the ball to round in the next stage. The ball is ready to be mounted in its new orientation.

13. I turned cup centers, which are padded, to hold the ball for turning. It is important to make sure that the old centerline is now parallel to the new axis. After much trial and error, I found that gluing 80 grit sandpaper into the cup for the next stage, gripped the ball well for turning to truer round. Because the ball is hollow and the glues joint are fairly small, care has to be taken to not break it with too much pressure from the tailstock. I glued a leather pad into the tailstock cup.

14. After initial remounting, I use a dial indicator to make sure that the ball is also centered front to back. I try for + or – .005” to keep the ball as close to round as possible.

15. My goal is to turn away the ghost, getting down to round in its new orientation. This shows that I am close, a little overturned at the center, but close enough for this stage. The ball can be rotated 90 degrees on the pencil line axis, re-centered and turned once more to get it ready for sanding.

16. I make a custom hard felted block to do the rough sanding to bring it into truer round. I started with 80 grit, being very careful not to get the ball too hot, which would soften the glue joint. After a couple of passes. I stop the lathe and rotate the ball in a different direction, doing this as many times as needed to get the to smooth round. It is not as important to get it to true center when rotating as some off center wobbling can be tolerated. After this initial rough sanding, I change to a drive center using a cupped drive with a piece of rubber shelf or toolbox liner glued in the cup. As I tighten up the tailstock, I grab and try to rotate the ball, allowing me to feel when there is just enough pressure to hold the ball for sanding. You can skip the felted block for the rest of the grits unless you start to get a wavy surface from the difference in grain hardness in early and late wood. Douglas fir would be an example of where you would continue using the felted block.

17. After sanding I cleaned and trued up the openings and carefully removed the excess glue on the inside. Now to make a stand.

In a following post I will show the geometry and math for sizing the box /ball for a given thickness of material.