thumbnail group

Connect With Us:

ME Channels / Machines & Automation

Don't Ignore PKM Machines


Parallel kinematic architecture continues to be studied and implemented, and may soon find a place on a shop floor near you

By Yon San Martín
Senior Researcher

By Marta Giménez
Senior Researcher
San Sebastian, Spain

By Vincent Nabat
Senior Researcher
Montpellier, France

When the first parallel kinematic mechanism (PKM) machine tool appeared at IMTS 1994 in Chicago, this new concept was considered the machine tool of the future, and some predicted the end of conventional machine tools. As happens with all new technologies, some problems appeared, and PKM machine concepts were soon dismissed as completely unsuitable for milling, and probably useless for manufacturing purposes. Today the situation has changed, and the future coexistence of conventional and parallel kinematic machines is a sure thing.On one end of the struts (the spindle end) spherical joints are employed. On the fixed frame universal joints are used. With this parallel kinematic configuration, it's not possible to transmit torsion or flexure stresses through the arms, while the six struts contribute to support cutting forces.

Unlike serial (Cartesian or SKM) machines in which one actuator has to carry the weight of the following ones, in PKM architecture none of the axes must carry another moving axis; all of them connect the fixed frame with the cutting head or end effector. This arrangement reduces the moving masses of the machine, and, as a consequence, machine dynamics can be improved.

The main objective of building PKM machine tools is to achieve higher acceleration and increase machine stiffness. Increased acceleration can be obtained mainly due to two factors: the reduction of mass and the multiplication factor. The multiplication factor appears because all the actuators contribute to the motion of the end effector. In some PKM configurations, with the same moving mass on the end effector as a Cartesian machine, the required torque on the actuators in the SKM machine is four times higher than in a PKM structure.

Usually, the stiffness of PKM structures is considered to be greater than that of conventional machine structures. The main reason for this increased stiffness is that the elements connecting the spindle head and the frame support only tensile or compressive stresses, compared to the flexural stresses that act on Cartesian machine tools. Another important factor is that all of the connecting elements contribute to support the forces. When compared to Cartesian structures, the PKM can dramatically improve the mass/stiffness ratio. It is possible to obtain the same stiffness using less moving mass, and in some cases the stiffness improves a great deal.

It's said that one of the main drawbacks of the Stewart-platform-type PKM is the ratio of working volume to machine size. Typically, it's assumed that a PKM architecture produces big machines that have a relative small working volume. The size of the machine especially increases if high tilting angles on the tool are desired. Also, it seems useless to have six controlled axes if only three degrees of freedom (DOF) are desired on the tool.

To improve these two drawbacks, a new generation of PKM machines has been created. The common strategy is to reduce the DOF by employing passive elements. In some cases, the resulting machine is fully parallel (passive elements only constrain degrees of freedom), and in some cases the machine is combined with serial elements (very interesting when one of the axes is very long).

Another hot topic of research in the PKM community is calibration. A calibration procedure is necessary to compensate for the insufficient precision of the assembly and some of the components (especially joints). Calibration can improve machine precision by more than four times.

A benchmark for PKM and SKM was established by the Research Institute of Communication and Cybernetics of Nantes (IRCCyN), University of Nantes (Nantes, France). This benchmark includes SKM and PKM machines.

Several tests were performed to compare the performance of the different architectures. These include:

  • Productivity with optimum cutting parameters
  • CAD parameters
    • Linear interpolation (G1)
    • B-spline interpolation
  • NC parameters
    • Feed forward
    • Acceleration profiles
    • Grouping of interpolation block (for example Compcad in Siemens)
  • Contouring and rounding parameters (G64, G642 in Siemens)
  • Cutting conditions
    • Cutting direction
    • Orientation of part in table

The analysis has been realized based on machining tests of a complex part with four profiles (see illustration of the benchmark test part). After the part has been machined, it's measured on a CMM. To improve the analysis, the IRCCyN has developed a high-speed machining simulator that includes the real behavior of the machine and tool, based on information extracted from the machine axes' encoders.

The main conclusion of this study is that each machine, whether it employs SKM or PKM architecture, requires the optimization of all parameters and cutting conditions to obtain the best possible result. It's clear, however, that PKM machines are able to perform high-speed cutting tasks with a performance comparable to that of SKM equipment.

During the design of an SKM, it's important to know the objective of the application, as well as the machining processes the equipment will perform. When designing PKM machines, these factors become even more important, to permit optimizing the necessary working volume and machine stiffness.

According to recent studies, there are now more than 100 different PKM architectures, although very few of them are practical architectures. And few of these practical structures result in machines capable of industrial use.Robotics is an important application for parallel structures. Shown above, the Quattro robot, which is sold in the US by Adept Technology Inc., can provide an acceleration of 15 g and handle a payload of 2 kg.

Some of the commercial machines are fully parallel: three-axis designs include the Quickstep from Krause & Mauser (Macomb, MI), SKM400 from Heckert (Chemnitz, Germany), and Ulyses from Fatronik; a five-axis machine is the P800 from Metrom (Chemnitz, Germany); six-axis configurations are the HexaM from Toyoda (Schaumburg, IL) and the PM600 from Okuma (Charlotte, NC). There are hybrid structures, like the Ecospeed from DS Technologie (Cincinnati), and the HERA and VERNE from Fatronik, all of them combining a three-axis parallel head (two rotations, one translation) with two serial axes. Another interesting hybrid structure is the Tricept concept from SMT Tricept A.B. (Vasteras, Sweden), which employs a three-axis parallel structure and a two-axis serial rotation head, resulting in a five-axis hybrid machine. The Pegasus machine from Reichenbacher (Dorfles-Esbach, Germany) uses linear drives, and the same secondary is used by both primary drivers.

One important application of parallel structures involves robotics, especially pick-and-place applications. The first PKM robot dedicated to pick-and-place operations, the Delta robot, was invented in Switzerland in 1986 by R. Clavel. This mechanism moves a platform in three dimensions and has very lightweight moving masses (a few kilograms; for similar performance, serial robots have moving masses of several tens of kilograms). An external kinematic chain is added to this robot to produce a rotation around the vertical axis, which is useful in pick-and-place operations.

This kind of robot has rotational actuators fixed on the frame, and has a set of inner and outer arms made of carbon fiber that link the moving platform to the motor. The mechanism can achieve high dynamic performance. Indeed, the industrial robots that use this type of mechanism (which are sold by ABB and Bosch) can reach accelerations of 10 g, while serial robots of the same size are limited to 2 g. However, this maximum acceleration can hardly be exceeded by the Delta robot, because of the external kinematic chain that provides the rotational degree of freedom.

To develop a robot that can go faster, a new parallel mechanism dedicated to pick-and-place operations has been developed by the LIRMM (Le Laboratoire d'Infor matique, de Robotique et de Microélectronique de Montpellier) of the University of Montpellier (Montpellier, France), the French Centre National de la Recherche Scientifique (CNRS), and Fatronik. Designated the Quattro, this robot was patented in 2005, and an exclusive license has been purchased by Adept Technology Inc. (Livermore, CA).

Test part made to benchmark SKM and PKM architecture.

The Quattro avoids the use of the external kinematic chain. It has four identical chains that link the actuators, and an articulated mobile platform. To rotate the end effector, Quattro uses the articulated platform, obtaining a rotation of ±180°. The four kinematic chains increase the system's overall stiffness, and produce an acceleration of 15 g; a payload of 2 kg is possible. With these values, a cycle time of 0.25 sec is achieved, which represents a production rate of 240 parts/min. Other common applications of PKM robots are for ultra-high-precision manipulation of small components, mainly with six-dof platforms.

Nowadays PKM machine tools are a reality, and can be used in typical manufacturing operations in five-axis or three-axis milling and drilling. They can be used without any limitation in medium to low-precision applications, starting from a precision of 20 µm. Because of the increased productivity that they can achieve, they deliver lower costs per manufactured part, and machine acquisition costs are equivalent to the costs associated with conventional machines. The next step for PKM machines is to reach a higher precision level, with values tighter than 20 µm and close to 5 µm.

To achieve this objective, several approaches are being taken. These differ, depending on the viewpoints of the persons doing the work.

From the user's point of view, the main concerns are:

  • Calibration: The calibration procedure must be integrated as a standard feature of the machine, and must be easily performed by the machine user in a short time—a maximum of a couple of hours. The current situation demands—in most cases—the presence of the machine manufacturer, and a complex measuring process that requires several hours.
  • Thermal stability: Because PKM machine tools use long struts to connect the system's fixed structure to the tool head, temperature changes can have a significant influence on machine precision. Strategies to compensate for these effects must be integrated on the machine.

From the viewpoint of the machine manufacturer, the main concerns are:

  • Mechanical design: Optimization of the machine performance in the design stage and workspace analysis, including singular-point analysis.
  • Integration of the controls: Integration of basic transformation algorithms (direct/inverse kinematics); calibration and thermal effects control; development of integration of velocity and acceleration transformations; dynamic models of the machine.
  • Trajectory planning: Improvement of machining precision and/or productivity due to nonconventional trajectories.
  • Dynamic prediction: Creation of dynamic models of the machine, and use of these models to maximize productivity and precision.
  • Calibration: Development of industrial calibration strategies and procedures.
  • Thermal stability: Thermal modeling of the machine. Definition of sensors for measuring and compensation strategies.

These areas are the current hot research topics at companies and development centers. During the next five years, the first new-generation PKM machine tools will appear.

The sector where PKMs will have the greatest impact is robotics. In some applications like pick-and-place, where the precision required is low (0.1 mm or more), PKM-architecture robots increase productivity to levels that are impossible with conventional solutions. Some projects that seek to create two-DOF PKM designs are underway, and in two years these solutions will be available in the market.


Machine Structure Benchmark

Two main parameters have been used to compare the performance of the machines: contouring precision and productivity. Results show that the contouring precision is comparable for SKM and PKM machines and independent of the machining direction.

Productivity of these machines is defined as the ratio of the machining time versus the precision. The graph shows the productivity of some of the SKM and PKM machines that were tested.


Fundación Fatronik

Aprivate Research Center that employs 140 persons, Fatronik specializes in applied research in mechatronics. The group is organized in two divisions: Industrial Systems—oriented to provide solutions for the means of design, manufacturing, maintenance, and end-of-life of products and services; and Health—oriented to the development of new technologies in the socio-sanitary field at the service of individuals, particularly elderly and disabled people. Fatronik is a member of EURON (European Robotics Research Network), the High Level Group of European Technological Platform for Researching in Manufacturing (ManuFuture), coordinates the ManuFuture-EU Spanish Technological platform, and is a member of the European Technological Platform for Researching in Robotics (EUROP).


This article was first published in the March 2008 edition of Manufacturing Engineering magazine.  

Published Date : 3/1/2008

Editor's Picks

Advanced Manufacturing Media - SME
U.S. Office  |  One SME Drive, Dearborn, MI 48128  |  Customer Care: 800.733.4763  |  313.425.3000
Canadian Office  |  7100 Woodbine Avenue, Suite 312, Markham, ON, L3R 5J2  888.322.7333
Tooling U  |   3615 Superior Avenue East, Building 44, 6th Floor, Cleveland, OH 44114  |  866.706.8665