In machining, vibrations, also called chatter, are the relative movements between the workpiece and the cutting tool. The vibrations result in waves on the machined surface. This affects typical machining processes, such as turning, milling and drilling, and atypical machining processes, such as grinding.
A chatter mark is an irregular surface flaw left by a wheel that is out of true (off-center) in grinding, or regular marks left when turning a long piece on a lathe, due to machining vibrations.
As early as 1907, Frederick W. Taylor described machining vibrations as the most obscure and delicate of all the problems facing the machinist, an observation still true today, as shown in many publications on machining.
The explanation of the machine tool regenerative chatter was made by Tobias. S. A. and W. Fishwick in 1958, by modeling the feedback loop between the metal cutting process and the machine tool structure, and came with the stability lobes diagram. The structure stiffness, damping ratio and the machining process damping factor, are the main parameters that defines the limit where the machining process vibration is prone to enlarge with time.
Mathematical models make it possible to simulate machining vibration quite accurately, but in practice it is always difficult to avoid vibrations.
Avoidance techniques
Basic rules for the machinist for avoiding vibrations: Make the workpiece, the tool and the machine as rigid as possible Choose the tool that will least excite vibrations (modifying angles, dimensions, surface treatment, etc.) Choose exciting frequencies that best limit the vibrations of the machining system (spindle speed, number of teeth and relative positions, etc.) Choose tools that incorporate vibration-damping technology (with structure damping using high damping material in the joint areas and with mass dampers using a counteracting force to stabilize the motion).
Industrial context
The use of high speed machining (HSM) has enabled an increase in productivity and the realization of workpieces that were impossible before, such as thin walled parts. Unfortunately, machine centers are less rigid because of the very high dynamic movements. In many applications, i.e. long tools, thin workpieces, the appearance of vibrations is the most limiting factor and compels the machinist to reduce cutting speeds and feeds well below the capacities of machines or tools.
Vibration problems generally result in noise, bad surface quality and sometimes tool breakage. The main sources are of two types: forced vibrations and self-generated vibrations. Forced vibrations are mainly generated by interrupted cutting (inherent to milling), runout, or vibrations from outside the machine. Self generated vibrations are related to the fact that the actual chip thickness depends also on the relative position between tool and workpiece during the previous tooth passage. Thus increasing vibrations may appear up to levels which can seriously degrade the machined surface quality.
Laboratory research
Industrial and academic researchers have widely studied machining vibration. Specific strategies have been developed, especially for thin-walled work pieces, by alternating small machining passes in order to avoid static and dynamic flexion of the walls. The length of the cutting edge in contact with the workpiece is also often reduced in order to limit self-generated vibrations.
The modeling of the cutting forces and vibrations, although not totally accurate, makes it possible to simulate problematic machining and reduce unwanted effects of vibration. Multiplication of the models based on stability lobe theory, which makes it possible to find the best spindle speed for machining, gives robust models for any kind of machining.
Time domain simulations compute workpiece and tool position on very small time scales without great sacrifice in accuracy of the instability process and of the surface modeled. These models need more computing resources than stability lobe models, but give greater freedom (cutting laws, runout, ploughing, finite element models). Time domain simulations are quite difficult to robustify, but a lot of work is being done in this direction in the research laboratories.