Motion control is often seen as a difficult specialist subject,
typified by a mass of external wiring and complex programming that is
best left to the experts. To demystify the subject Mitsubishi Electric's
approach is to make motion control technology for PLC users, as David
Ramsay explains
Most automation engineers are familiar with PLCs and control networks,
but somehow motion systems seem several orders of magnitude more
complicated. Based on the universal truth that nobody likes a Smart Alec,
Mitsubishi has developed a range of technologies to simplify motion
engineering by making it similar to PLC integration.
One of the first things that the design team realised was that in order
to maintain accurate synchronisation of large numbers of axes or of axes
physically separated by some distance, high-speed data transmission was a
pre-requisite. Thus the network (or wiring) has to be an integral part of
the range, so Mitsubishi developed SSCNET, Servo System Controller
NETwork, specifically for the purpose.
Originally SSCNET had a standard cyclic speed of 3.5ms ,on a range of
four axes, and was sufficiently fast for the majority of applications.
For the most demanding applications, it has also developed SSCNET II,
which cycles at 0.8ms. These speeds are inherent within the SSCNET
network and do not depend upon the types of controllers or amplifiers
used in the system. This is because the SSCNET has a degree of inherent
intelligence, which causes the controllers and amplifiers to default
automatically to the highest speed applicable to all connected devices.
An interesting example of an early use of SSCNET motion engineering is
the opening and closing mechanism for the Toyota Stadium in Japan, which
figured so prominently in last year's World Cup. The roof is made up of
several sections, each weighing many tonnes and moving completely
independently of the others. It takes over two hours to fully open or
close the roof.
As SSCNET is a bus system, connection from controller to amplifier and
from amplifier to amplifier is simply a case of daisy-chaining units
together. This, at a stroke, removes the confusing mass of wires normally
associated with motion and eliminates noise problems, as all signals
(including position data) are sent as serial data not position pulses.
All data from the amplifiers is returned to the controller via SSCNET.
This is advantageous, as all data appears in pre-defined data register
areas, meaning that there is no need to set-up any specific handshaking
routines to acquire data.
The design philosophy - that motion control should be in a PLC format,
using dedicated CPUs that mount on the back plane - was seen as the most
user friendly and fastest to configure. It also allows seamless
integration with logic control. The Qn PLC platform formed the basis of
this design, and a series of Q-Motion controllers was developed. The
motion control units are referred to as the Q172 (8-axis) and the Q173
(32-axis).
Programming
One of the key features that aid the user friendliness of Q-Motion is the
use of Sequential Flow Chart (SFC) programming, identical to that used
for a PLC. Special motion requirements are generated through the
inclusion of specific processing blocks, which are cut and pasted into
the SFC program.
The basic blocks that make up a process include motion, arithmetic and
sequence commands. As well as these features, direct comparison and
setting of parameters gives the Q-Motion range all the features needed
for highly complex motion control. This is typified by a recent
application in the aeronautical industry where over 120 axes were needed
for a complex wing application that involved precision control of all
axes simultaneously to fractions of a millimetre accuracy.
SFC is held by most practising engineers to be the best method of
programming, and is