TOOLS AND GADGETS

1. Fret Saw Machine Design ..... by Frank Harris
2. Sanding or Scraping a Picture ..... by James Colter
3. Strip Cutting Jig ..... by Bruce Fairchild
4. An Analysis of Fret Saw Blade Angles ..... by Frank Harris
5. ... and Follow-up to Blade Angles (with table) ..... by Frank Harris

  Back to Canadian Marquetry Home Page

All the homemade machine mounted fret saws that I have seen and also the designs shown in marquetry manuals have a throat-piece inserted from the front. This requires you to cut into the throat-piece the first time you use it and perhaps to have two or three throat-pieces if you wish to change the angle or size of the blade. Finally the throat-piece will eventually wear to the point you start losing tiny pieces through the hole and you have to make a new one.

I believe there is a better way to design the throat-piece and it is very simple and could easily be done as a modification.

Instead of inserting it from the front, cut your dovetail groove from side to side right across the table and use two square-ended throat-pieces inserted one from each side. Where they meet is a slot no wider than the thickness of the blade and you didn't have to cut it. You don't need different throat-pieces for different angles or blades and when the inserts begin to wear just trim the ends. You will never again lose any small pieces. In order to have the slot no wider than the width of the blade when you are bevel cutting you must, of course, trim the end of one of the inserts to a slightly greater angle than you might use on the saw.

You might also consider changing the top surface of your table from wood to polyethylene. I did and am no longer bothered by the masking tape (used to hold the veneers together) sticking to the table.

You might also try the following suggestions:

1. Melamine coated particle board is probably as good as the polyethylene and considerably cheaper.
2. Round over the edges of the table surface. This helps to prevent your work from snagging on the edge as you rotate in the sawing process.
3. Circles of various diameters can be cut by mounting a small diameter pin in either of the throat-pieces for use as a pivot point.

by James Colter

For those of you who tuned in expecting a debate on the merits of sanding versus scraping, sorry Charlie. This is about something a little more mundane. How to hold the picture while sanding or scraping it. Well, sure, you could hold it, but what if you want to use both hands? That's the problem.

When using a good palm sander in the early stages of sanding, or using a scraper, or even a sanding block, the picture has a tendency of travelling in the direction of your force. How do you stop it, you ask? There are many ways of achieving this end. If you are not married, there is that plush carpet in the living room. Nothing will skid on that sucker. I find, however, if you are married you have to be a little more creative in problem solving.

There are still many ways to solve a particular problem. First, I tried nailing the picture to the work bench. Won't work you say? Yes it will. You just have to be sure to counter-sink the nails or the scrapper gets a little nicked. After getting tired of finding the right colour nail filler, and fixing the side of the picture (claw hammer), I decided to get smart, and glue the work to the table. Hey, you only do that three or four times (OK, four) before you realize those suckers don't come up. All was not lost, however. (See my upcoming article on Marquetry Designs for your Workbench).

It was then that I stumbled on the incredibly simple device who's design I lay before you here. It will allow you to apply pressure without the picture slipping.

by Bruce Fairchild

Assembly

Glue 18" x ½" x 1" back stop to 12" x 18" base board.

Glue 4 corner feet in place.

Glue 12" x 18" piece of arborite on top of base board, also glue arborite strip on sides and bottom to protect the particle board.

Drill 4 - 5/16 holes with ¾" centres from each side, ¾" from the top and 2" from the bottom.

Using a 5/16" straight router bit, set router guide to centre router bit in the pre- drilled holes and cut a 5/16" slot from the top hole to the bottom hole.

Repeat for the other side.

Drill 2 - 5/16" holes in the adjustable width stop to line up with the slot in the base.

Clear a 1/8" deep slot from the bottom to within 1¼" of each end.

This gives a tunnel about 15½" wide for veneer to slide freely in and out for the jig.

Drill 2 -5/16" holes in the adjustable cutting straight edge.

(prior to this I had to cut and epoxy glue a piece of wood into the hollow portion and sand to get a solid flat surface on the bottom.) A piece of sand paper is glued to the bottom of the cutting edge to hold the veneer steady while cutting.

Place a flat washer on each ¼" carriage bolt and above through the slots from the bottom of the base. Place each the width stop and the cutting edge onto the ¼" carriage bolts, place a flat washer and a wing nut on each ¼" bolt.

Operation

The width of veneer to be cut is established by moving the cutting edge till the distance between the back stop is established. Tighten the wing nuts on the cutting edge bar to hold this setting, then move the adjustable width stop up to the back of the cutting edge bar and tighten the wing nuts on the width stop. Loosen the wing nuts on the cutting edge bar and slide veneer under the slot in the width stop and under the cutting edge bar. Tighten the wing nuts on the cutting edge bar, making sure it is tight against the width stop. You can cut your veneer using a craft knife. Remove the cut piece and repeat for as many cuts as you require of the set width.

Making a Chessboard

Using the jig set the width of the cut to 1½". Take a piece of maple or birch about 15" wide and cut with the grain nine pieces that will be 1½" x 15". Next to a piece of walnut about 15" wide and cut with the grain eight pieces that will be 1½" x 15". Now join the alternately maple or birch and walnut strips together edge to edge and tape along the joints with gummed tape. Rotate the assembly by 90 deg. and slide into the jig. It is important to position the assembly at 90 deg to the back stop (this can be established prior to the setting the 1½" width by sliding a try square under the width stop and cutting edge to the back stop and marking the 90 deg edge).

The first cut will not be 1½" but enough to establish a square cut across the assembly. Now proceed to the 8 strips across the grain from the assembly that are 1½" wide. With the 8 strips move alternate strips forward by one square, to bring the light and dark squares together and tape them together. Trim off the unwanted ninth square which protrude at the end of each now.

by Frank Harris

Since joining the Marquetry Society of Canada a couple of years ago I have been puzzled by many things, and one of those "puzzlements" has been the proper angle to set the fret saw to obtain a good fit between mating parts. This puzzlement came about because I was unable to find any consensus among other members, or in books, as to the proper fret saw angle for certain conditions (different saw blade and veneer thicknesses). The few members I have asked about this say they set their fret saws to angles between 10 and 15 degrees. The Modern Marquetry Handbook by the Marquetry Society of America (page 73) says, "12 degrees will form a perfectly tight fit that requires little or no filler", without mention of either blade size or veneer thickness. The Marquetry Manual by W. A. Lincoln makes two or three references to angles, and sets out one table of angles which vary from 8 to 20 degrees and are related only to veneer thickness, not to blade thickness, (although he says in a previous paragraph that you have to experiment with practice cuts to establish angles).

It is obvious from the high quality of work done by many marquetarians that the saw settings being used are working very well, even though they may not be calculated mathematically or even well understood. It may also be, of course, that those skilled at the craft know that approximations are all that are needed, and that perfect geometrical accuracy is a waste of time. Nonetheless I have set out to do a little trigonometry to determine the theoretically-correct angles for certain conditions, if only to satisfy my own curiosity.

In order to understand why only one fret saw blade angle is correct for a perfect fit for any one combination of veneer thickness and blade size, please refer to Fig 1. In this diagram, a "plug" (shown in sawing position) is to be sawn with a fret saw blade of a certain thickness (kerf) set at an angle a, such that the resultant fit of the plug to the "body" of veneer will be exactly flush at the top face of the body. In other words, the dimension of the top face of the plug must exactly equal the dimension of the top face of the "window" cut out of the body in order for the resultant top face to be flush. If this is to be the case, it can be seen that the points shown as A on the top face of the body veneer must be directly over the matching points B on the top face of the plug after sawing. It is obvious from the diagram that, for any given thickness of "body" veneer, and a specific blade thickness, this condition can be met at only one blade angle. (If the blade angle is increased using the same blade thickness, point B on the left side of the diagram will end up to the left of its present position and will no longer be directly under point A. Similarly, too small a blade angle will move point B to the right, away from its position directly under point A.

So for any given thickness of body veneer and of saw kerf, (I am using "blade thickness" to represent the resultant kerf after sawing) how do we find the correct blade angle? From the diagram, note that the blade angle a is also one internal angle of a right triangle (ABC), of which the hypotenuse (AB) is the body veneer thickness, and the side (BC) opposite the angle a is the blade thickness. From trigonometry we know that the opposite side of a right triangle divided by the hypotenuse is the sine of the angle, so the sine of angle a is the blade thickness divided by the thickness of the body veneer. We can find the blade angle a from a set of trigonometric tables.

If all veneer were the same thickness, or very near the same thickness, only one calculation would be necessary and all marquetarians would be using the same saw angle for any given thickness of blade. Unfortunately such is not the case. In a measurement survey of 52 pieces of veneer (from the two or three hundred that I have accumulated) I found thicknesses from .019" up to .041", and a size for every .001 increment in between. This means a lot of thickness measurements and calculations to get the correct blade angle for each cut in a piece of marquetry if you want to have perfect fits. To make this easier, I have provided a table (Table 1) of blade angles for 26 veneer thicknesses (in increments of .001") and 9 different blade sizes. To use this table it is necessary only to look down the column for the blade number or blade size that you are using, to the row corresponding to the veneer thickness, and pick out the correct blade angle in degrees and tenths. For example, for a 4/0 saw blade and a veneer thickness of .032", the correct blade angle for a geometrically perfect fit would be 15.6 degrees (or 15 degrees, 36 minutes).

For the type of fit shown in Fig. 1, flush at the top face, the computation of blade angle depends on the thickness of the body veneer, not on the thickness of the plug. Any variation in the thickness of the plug will show up as irregularities on the bottom face of the assembly. This is not always desirable. For many if not most types of marquetry work, it would be preferable to have the bottom face of the assembly flush, so as to form a solid base, and the irregularities in the upper or "picture face" can be removed by sanding or scraping prior to the finishing process.

This is the situation shown in Fig. 2. In this diagram, it can be seen that, in order for the bottom face of the assembly to be flush, it is necessary for the points A on the bottom face of the body veneer to be directly over the points B on the bottom face of the plug. Once again it can be seen that for any given veneer thickness and blade size, this condition will exist for only one blade angle. However, now the computation of blade angle depends not on the thickness of "body" veneer (as in Fig. 1), but on the thickness of the veneer in the plug. Once again, the blade angle forms an internal angle of a right triangle (ABC), and the side BC opposite the angle divided by the hypotenuse (AB, or the thickness of the plug veneer) gives us the sine of the blade angle. Again, the angle can be found from a set of trigonometric tables, or you can use Table 1 . This calculation is completely independent of the body veneer thickness, and all variations in veneer thickness will appear as irregularities in the upper face of the assembly.

Finally we come to Figure 3, where I have included the cellotape that is used in the bevel-cutting technique. Note now that points A and B are separated by the plug thickness, plus a double thickness of cellotape. For a fit similar to the case of Fig. 2, where the bottom face of the assembly is flush, all that needs to be done to compensate for the cellotape thickness is to add one double thickness of tape (.004") to the thickness of the plug, and use this sum in the calculation of the sine of the blade angle (or as the veneer thickness in Table 1).

All that remains is to decide whether all this has any real practical value, and if so, how to use it. I believe it has, if only to give some marquetarians a clearer understanding of what is required to get those tight "invisible" joints that are so desirable. The mathematics and geometry involved inexorably lead us to a "theoretically-perfect method", and this is what every perfectionist wants, isn't it? Of course everyone knows perfection is impossible, so it is your choice as to how far you wish to go along this road. Perhaps you will find among these numbers only a slightly better angle than the one you have been using all along.

1. Measure with a micrometer each piece to be inset (i.e., each "plug");
2. If you are using cellotape, add .004" to this thickness;
3. Look up the correct angle in the Table for the saw size you are using;
4. Set the fret saw to this angle for that particular inset piece.

This method will give a theoretically-perfect fit, with a flat back face for easy and solid gluing to a substrate, and all variations in veneer thickness will show up as irregularities on the top face which makes for easy removal by scraping and sanding. When edge gluing by this method, it is important to lay the piece to be inset (the plug) face up on a flat surface, and place the window (the body) over it. If you try to place the plug into the "window" when the body is lying with its upper or picture face against a flat surface, and if the plug is thicker than the body, it will not fit properly.

The equipment required would be a micrometer or caliper capable of measuring to .001", and a means of setting a blade angle accurately on your saw. The latter can be as simple as a protractor fastened to a block of wood with double-sided tape, and stood on your saw table behind the blade while you adjust it. It might also be as complex as a built-in protractor complete with vernier graduated in increments of five minutes of angle.

To conclude, this article could not have been written (and some may wish it hadn't been!) without the able assistance of an engineer friend and former work associate, Mr. Roy Bourke, who with his knowledge of mathematics set me straight on basic trigonometry (which I had forgotten), and who provided valuable expertise and computer time in preparing the table.

Well, there you have it, for better or for worse!!

by Frank Harris

Since writing the above theoretical article, I have had some time to experiment and try the theory. As you might expect, some of it did not work out exactly as theorized, but it did work.

To begin with, I tried cutting-in nine randomly shaped pieces using the "theoretically-perfect method". The body veneer was .037" thick, and the plugs were .021", .025", .030", .032", .034", .036", .038", and .040". Each piece was overcut such that all pieces ran into at least two others and sometimes three or four. The back surface was reasonably flat but not flush, and the irregularities did show up on the top face. Initially I was quite disappointed that the back was not flush, but on reflection I realized that I had not allowed any tolerance for working inaccuracies, e.g. inability to set the saw perfectly, slightly rough kerf preventing intimate contact, wood swelling due to the white glue, and perhaps other reasons.

I then experimented to find out how far undersize I needed to cut the angle to get a flush fit. I tried ½ degree, 1 degree, 1½ degrees, and finally settled on 2 degrees. I cut-in five more pieces, which were much better than the original nine.

The next experiment was to cut into a piece of thick veneer (which had itself been cut into the body veneer), three pieces of dyed white veneer to illustrate how a thin veneer can be cut-in so it is flush on the top or bottom surface, as you choose. A third white piece was first laminated to double thickness before cutting-in, leaving it flush on the bottom and slightly proud on the top.

Another demonstration experiment I performed was to cut-in three pieces approximately 1¼" in diameter of a veneer which had the same plug and body thickness. A perfect fit in this case should be flush on both top and bottom. This was the case with an angle two degrees smaller than theoretical, while the theoretical angle cut would not fit flush. It was slightly proud on the back, and depressed on the top. The third piece, cut two degrees larger than theoretical, would not fit into the window. This third piece illustrates why so much pressure is sometimes required to get the plug into the window and roll the edges flush. What happens is the fibres are crushed, and the main body is forced out of shape.

While doing these experiments, it occurred to me that there was another way to obtain good fits without changing the angle. I reasoned that if a double thickness of cellotape effectively increased the thickness of the plug, then I could effectively make the plug any thickness I wanted by simply putting shims between the body and the plug in preparation for sawing; e.g. for a plug .045" thick and a 4/0 saw, I would set the angle at 9 degrees (2 degrees smaller than theoretical) and no shim is required. I can now cut-in every other size of veneer without changing the angle by shimming the plugs to .045" thickness; e.g. to inset a .020" plug, use .025" shims. This effectively increases the plug thickness to .045", for a theoretical angle of 11 degrees and the saw is set to 9 degrees. When working from a minimum thickness to a maximum as in my example, the amount of shimming becomes quite large so one might set the angle to 13.1 degrees (2 degrees smaller than theoretical 15.1 degrees), and shim .013" instead of .025". Remember that with this method, whatever angle you set on the saw, this angle is 2 degrees smaller than the theoretical and only those veneers at and above the theoretical angle can be cut-in by using shims.

Shims could be anything that can be cut by a fret saw blade. I used heavy waxed paper which worked very well. It was easy to cut and also lubricated the blade. Make sure you pack and tape the assembly (body, plug and shims) tightly. Remember that whenever you apply tape to the bottom side of the body or to the top side of the plug, you are using shims and if you overlap the tape edges you are doubling the thickness; e.g. with overlapped tape on both body and plug (.020" thick plug), the plug is effectively increased to .028" which is a 40% increase, and the theoretical angle changes from 25.5 degrees to 17.9 degrees, a change of 7.6 degrees.

I considered concluding this article with some recommendations or suggestions but I have decided that, due to my lack of practical experience, this would be presumptuous. I am confident that, mathematically and theoretically, everything I have written is true and I have experimented enough to prove it to my own satisfaction.

I hope that the material in these articles is not dismissed arbitrarily as being too complicated or difficult to incorporate, but will be considered carefully to see what might be of practical value.