Using a hammer-and-chisel approach should enable new kinds of lightweight, portable, but slow CNC machine tools, by way of eliminating toolhead side forces during positioning.
One of the greatest challenges in precision machining is toolhead side forces. A 3-D printer can get by with a low-rigidity frame, open-loop control with stepper motors, and triangular-thread (V-thread) leadscrews in large part because the forces at its toolhead are so small. But cutting metal with edged tools or abrasives, even very sharp edged tools, produces substantial side forces, and these forces change rapidly and even unpredictably as contact is established or contact angles change. Consequently, precision machining invariably uses closed-loop control and massive structural members to achieve the desired stiffness, and usually also uses ballscrews or acme-thread leadscrews.
I’ve speculated previously that it with sufficiently rapid control systems, it will be possible to provide such stiffness “virtually”, by countering a detected displacement with a larger opposing displacement in some actuator, effectively multiplying the rigidity of the frame by the ratio of displacements. But here I want to discuss a different approach, that of hammering.
A hammer or karate chop is a kind of simple machine: you apply a small force to the hammer over a large distance, and the hammer applies a large force to the work over a proportionally small distance. (Stevin and Galileo did not include the hammer in their list of simple machines, as they lacked, I think, the concept of kinetic energy.) Sufficiently rapid and precise feedback about the hammer’s path allows you to control the point of impact precisely with low energy, but an alternative mechanism is the hammer-and-chisel mechanism used by an automatic center punch; this allows you to position the chisel as slowly and precisely as you like, then transmitting a hammer strike anywhere on the chisel head precisely to the chisel tip.
The hammer-and-chisel mechanism has been used for this purpose since the Paleolithic, when it enabled soft, squishy human hands to precisely shape rocks harder than steel.
At one point I was pretty skeptical of hammering on things to precisely shape or fix them, since there’s inevitably a shock produced at the time of impact, and that shock can damage things as it propagates through the rest of the structure. But eventually I learned that hammering is, in many cases, safer and more precise than applying a large force more slowly through an elastic frame; applying a force F through a mechanism with stiffness k requires building up an energy ½F²/k. If the resistance diminishes once initial resistance is overcome, as in solid friction and many kinds of material fracture, this energy is released all at once in an uncontrolled fashion. By contrast, a hammer blow can have an arbitrarily small energy, and the energy at a given force is limited only by the stiffness of the chisel and hammer head themselves, which can be made orders of magnitude smaller than the stiffness of an entire frame.
This very feature is one of the reasons for the use of automatic center punches: their hammer energy is the same on every stroke, so every divot has fairly precisely the same depth.
You could imagine a CNC machine that precisely repositioned a chisel several times a second under very low toolhead forces, striking the chisel with a hammer separately moved with higher, but still fairly low, forces, in order to cut or form a piece of material by applying a precisely controlled amount of energy with a very large force at a precisely controlled location. Wood chisels have been used with this technique by carpenters to make precise cuts in wood since the paleolithic.
As the hammer approaches the chisel, it must of course be held in some kind of mechanism to apply force to it, which means that it needs some non-negligible stiffness to the chisel and workpiece. However, at the point that it actually strikes the chisel, it is not necessary for it to be held with any stiffness in the axis in which it strikes the chisel. It could, for example, be sliding down a tube under its own momentum, or held with a negative-stiffness cancelation mechanism like that used to isolate the LIGO from vibration. Such an expedient could prevent the shock of impact from being transmitted back into the struture holding the hammer, which would introduce a vibrational impulse into the machine frame.
For metal cutting, it probably is not possible to reach the optimal cutting speed with this approach; recall, as cutting speeds go up, up to a point, tool life goes up too, because the machine is cutting hot metal which applies less force to the cutting edge, wearing it less. (I think it also is more efective at producing a built-up edge, which worsens cutting precision but decreases wear.) So we should expect that cutting tools used in this way will not last as long as cutting tools used in the traditional cutting mode on a lathe with continuous movement. Milling-machine cutting points (whether inserts or just flute edges) are typically used in a similar discontinuous-cutting mode, but the metal of the workpiece does not have time to cool completely between strokes.