If you make puzzles this is a must read...
The Goal by Steve Strickland
The Goal in puzzle making is precision in the size and shape plus repeatability in the dimensions of the puzzle pieces and the subunits used to build the pieces. This is not an easy goal to achieve because of the variables in wood and the environment.
Wood is its own substance, quite unlike anything else. Wood compresses, expands, twists and warps seemingly at whimsy. These "movements" in wood are not random, they are effects that have causes. One task in achieving the Goal is identifying and controlling, wherever possible, the causes of wood movement.
Sanding is also a big problem in puzzle making. I just don't know of any method to accurately finish sand pieces. The rule here is "sanding degrades geometry" so I minimize sanding effects as much as I am able. You'll hear me repeating this mantra often.
Wood movement and sanding are not the only set of variables that must be dealt with. The behavior of our tooling and machines becomes important, too. I'll show how I deal with the variations of my tools with a technique called Kinematics that I'll explain in detail later in the book. The main thing here is to be extremely vigilant, always on the lookout for possible sources of variation.
Personal behavior also effects the outcome of puzzle making projects. These kinds of things are quite difficult to identify and it's natural to blame problems on the environment, the materials or the tooling. The puzzle maker is strongly advised to include himself in the error analysis and eliminate personal ego along with any ideas of right and wrong. An open mind is a most useful tool.
Never Assume A Single Damned Thing!
I cannot stress the importance of this statement enough. This is the most critical technique the puzzle maker can develop. It means the puzzle maker must constantly be taking measurements of everything and must constantly examine his tools and techniques with a skeptical eye. Once the puzzle maker starts making assumptions it is an absolute certainty that a project will eventually be ruined.
I speak from hard earned experience. I made an assumption and turned $75 worth of top quality Pecan and 60 hours labor into firewood. I learned that top grade Pecan makes a superb and very expensive cookfire.
It is valid to know what kind of tolerances and precisions we are setting as the Goal. First, I'd like to mention that perfection is simply not possible, so I'm going to ignore it.
When I say "tolerance", I really mean "tolerance envelope". Tolerance is the entire range of errors in a batch of pieces. When I say "precision", I really mean "deviation from design specs". These two terms describe completely different things. Precision tells us how close to design specs we are and tolerance tells us how much variation we're getting in a batch of pieces. As a general rule of thumb, precision is a static variable and tolerance is a dynamic variable. Now, this rule doen't always hold perfectly true because the 2 can interact with each other. It's really just a way to visualize these two concepts. It is important to keep these two types of variation separate.
The tolerance envelope of a batch of pieces can be graphed. We decide to measure each of a thousand cubes. We then count the number of pieces that're smallest, and count each size up to the largest. We could then graph the distribution of sizes. This would illustrate our distribution of error in the batch. We'd have a picture of our tolerance envelope.
Having detailed data on the tolerance doesn't tell us if we are making the pieces the correct size. We might have an excellent tolerance while all of the pieces are too big or too small. Getting the pieces the right size is the precision.
Here's another way to think of this. Precision is how accurately we have set the sideguide on the jig (a static variable because it doesn't change) and tolerance is how accurately we are able to repeat that dimension cut after cut (a dynamic variable because it does change).
Based solely on my personal experiences, I have learned that an amateurish puzzle can be constructed with a tolerance of 0.020" (0.5mm). This means that all pieces are within +0.010" and -0.010" (0.25mm) of design specs. A "good" puzzle requires a tighter tolerance of 0.012" (0.3mm), all dimensions within +/- 0.006" (0.15mm) of design specs.
An "excellent" puzzle requires an even tighter tolerance of 0.006" (0.15mm), all dimensions within +/- 0.003" (0.076mm) of design specs. On occasion everything will come together just right and puzzles of even closer tolerances can be built. However, excellent puzzles are a reasonable, achievable goal day after day, month after month. Puzzles made within 0.003" of spec usually fit pretty well.
Some of the more challenging designs will require a superb tolerance of 0.004" (0.1mm) and a precision of 0.001" (0.025mm). It takes experience and to achieve superb tolerance. For the tyro puzzle maker, a Goal of 0.006" (0.15mm) tolerance is a good one to work towards.
I also want to point out that just because I am currently unable to do much better over long time spans does not mean that this is the limit on what can be done. On short time spans a much better tolerance can indeed be achieved. I also believe that better results can be achieved over long time spans and I simply haven't figured it all out yet. This is certainly something that you should be constantly working on. Reducing the tolerance envelope has direct benefits and improves the quality of the puzzles.
Long term tolerance can easily be improved by cutting the pieces one at a time. This eliminates the variables of 1 entire dimension on the jigs. However, I must cut pieces in batches to make money so this technique is largely left to the amateur. If you are building puzzles only for yourself then this is the recommended approach as it will consistantly yield better results.
The Goal is therefore a tolerance of 0.006" (0.15mm) and a precision of 0.001" (0.025mm). To achieve the Goal the puzzle maker must take a highly systematic approach. Everything must be measured, tested and proven. Nothing can be assumed or left to chance.
The most difficult part of achieving the Goal is to develop a systematic approach and an analytical way of thinking. Mental discipline is important and it can be employed to a high degree of success.
I don't intend for you to read this book and follow the construction details slavishly. I much prefer that you grasp the underlying principles and then apply them in your own way. I'd really like it if you could discover some new principles that makes things easier, faster and more reliable. You undoubtedly enjoy using your brain and puzzle making is a wonderful way to give it a good workout!
The puzzle maker only needs to know a few things to be able to construct a wide variety of puzzles that achieve the Goal. First, we'll learn the Kinematic Principles. This information forms the foundation for a systematic method of jig design. Then we'll develop 3 fabrication techniques that exploit the Kinematic principles in jig construction. We'll therefore be able to construct a bewildering variety of precision jigs from just a few simple ideas and construction techniques.
The idea of deriving a lot of variety from a few simple ideas doesn't stop with the puzzle maker's jigs. Just being able to make accurate cubies allows the puzzle maker to construct hundreds of different puzzles. Add the ability to accurately notch sticks and thousands more designs become possible. A jig for cutting 45 degree angles opens up a whole new realm as does a jig for cutting 60 degree angles. Using the exact same jig construction techniques, the puzzle maker can also explore the world of compound angles. Using the same type of jigs on a router table opens up even wider vistas of puzzles that employ exotic jointery.
In conclusion, the puzzle maker need only grasp the principles of Kinematics and develop a few jig construction techniques that exploit these principles. He need only to keep the Goal in mind and to develop a systematic way of thinking and working. With these few ideas in hand the puzzle maker can successfully construct many thousands of different types of 3D mechanical puzzles.
Author