The key to minimising the impact of production errors in manufacturing processes is to spot problems as soon as they occur, making inline quality checking a very attractive proposition. Unfortunately, many businesses believe that inline checking is too expensive or too complex for them to implement. Tim Dodd of ifm electronic explains why this is no longer the case.
Which manager or engineer in charge of a production line wouldn’t want the processes and products monitored and checked at every stage of production? The potential benefits would be enormous – wear and tear on moving parts detected before they start to cause serious problems, out-of-tolerance products detected before time and money are spent taking them through the whole production process, missing or incorrectly oriented components spotted before the end product is packed: the list is almost endless. But surely, providing such comprehensive monitoring and inspection would be both costly and complicated?
Once that might have been true, but recent developments in sensing and detection technology have changed everything. Adding inline checking to an existing plant – or designing it into a new plant – need no longer be a dream. It can be made into reality much more easily than you might imagine and at a surprisingly modest cost. Let’s take a look at how this can be achieved.
The first step is to realise that inline checking doesn’t necessarily have to be synonymous with complex vision systems. Of course, these have their place, but in the majority of applications, much simpler solutions will do all that’s needed. Consider, for example, ordinary inductive proximity sensors. Not long ago, these had just one function: to indicate whether or not some piece of metal was present in a particular place. But modern proximity sensors that incorporate IO-Link technology can do much more.
In addition to sensing the presence or absence of a metallic target, they can be configured to operate at a particular and very well defined distance or range of distances, which opens up a whole range of new applications. They can, for example, be used to spot wear and tear in guides and clamps and, in some applications, to check component orientation, always giving a clear go/no-go indication. These new abilities mean that proximity sensors have an important role to play in inline checking solutions, especially as they have the additional benefits of being inexpensive, robust and extremely reliable.
Proximity sensors are not the most appropriate solution in every application, especially when non-metallic targets are involved. In many of these instances, however, optical sensors will provide the required functionality. The latest types, which again benefit from integrated IO-Link interfaces, use time-of-flight laser technology to provide a long-range measurement of distances, as well as other useful functionality.
They can, for example, check the height of a product on a conveyor belt with an accuracy of 1mm or better, even though they are mounted a metre or more from the conveyor and are thus protected against accidental collisions. As well as measuring distance, these intelligent optical sensors can simultaneously check the target’s colour and reflectivity. If it suits the user’s needs, the outcome of all the measurements – height, colour and reflectivity – can be logically combined in the sensor to produce a single pass/fail output, which saves on I/O capacity for the control system PLC.
Where even more detailed inspection is needed than can be achieved with a normal optical sensor, a laser profiler will often provide a straightforward solution. Effectively line-scan devices, laser profilers use time-of-flight and triangulation technologies to produce an accurate and detailed 2D profile of a reference object. After they have been configured – which is a simple teaching task supported by a software wizard – they can then be used to scan target objects to determine how similar they are to the reference. The profiler provides a pass/fail output based on the percentage match between the reference object and the target.
Profilers are well suited to applications where ordinary optical sensors do not provide sufficient functionality, but a vision system would be overkill. As well as being less costly than vision systems, profilers are also much easier to configure and, as they use their own internal laser light source to scan the target, they have no special lighting requirements. Finally, advanced versions are now available that can store up to ten reference object profiles, which can subsequently be selected via the IO-Link interface. This makes them eminently suited for use in applications where a production line is required to handle multiple product types.
As we have seen, relatively simple sensors can satisfy many of the inline checking requirements associated with typical production processes, but there are some applications, such as complex object recognition, where only a vision system can provide the functionality needed. Even in these cases, users now have a choice between traditional (and costly!) full-function vision systems and much simpler and less expensive vision sensors.
Vision sensors are available in 2D and 3D versions, with the principal difference being that 2D sensors look at only the width and length of an object, whereas 3D sensors also take into account the height. Apart from this difference, 2D and 3D sensors offer similar functionality, allowing them to be used to assess the size, shape, position and orientation of objects as well as their colour and reflectivity. They can also count the number of objects in their field of view.
As with full-function vision systems, vision sensors required good lighting if the best results are to be obtained, but they are much easier to set up and use in every other respect. Typically, they are configured using a software wizard to store multiple models against which they can compare the images they capture, working with user-defined percentage tolerances on parameters like object size and position. In inspection tasks, they can readily be configured to provide a go/no-go output which saves a lot of PLC coding compared with evaluating the more complex results delivered by conventional vision systems.
Users of vision sensors report that the time required for setting them up is measured in minutes compared with hours for full-function vision systems, the prices are very much lower, and the onboard logic provided by vision sensors significantly simplifies overall control system design and implementation.
Inline quality checking – or error-proofing of production – was once difficult and costly to implement if you prefer. As this article has explained, this is no longer the case.
Today’s smart sensors – even the most basic types like the humble proximity sensor – can perform invaluable quality checking roles and, by using combinations of proximity sensors, optical sensors, laser profilers and vision sensors, it is possible to develop versatile and highly effective quality checking systems that will quickly repay their cost in terms of enhanced product quality, increased efficiency and reduced production costs.