Manufacturing companies often feel they need to choose between using vision systems or radio frequency identification (RFID) to trace products throughout their supply chains. In fact, in most cases, vision and RFID are complementary technologies, and this is particularly true when one looks at vision sensors.
An RFID system provides identity information, so that a process can keep track of what is being worked on at some point in the manufacturing process. The RFID system does not identify product flaws, but rather can store a flaw detected by different means (for instance, the vision system or vision sensor), so that a carrier or pallet can be pulled out later for repair.
RFID systems are used as a complementary technology with a large number of other technologies. If we are talking about “error proofing,” these other technologies can be as simple as an inductive sensor (to detect the presence of a metal part before a weld process is activated), a photoelectric sensor (to make sure a plastic cap has been installed), or a vision system.
A vision system can be viewed more like a high-power, highly configurable sensor. The vision system or sensor checks the part by evaluating features. Because most manufacturing lines produce a large number of virtually identical copies of a part, the vision system is typically not used to determine the identity of the part; this is where RFID comes in.
One exception to this rule needs to be mentioned. In some cases, a part contains a barcode or some other mark that can be visually identified. In this case, a vision system (less so, a vision sensor) can read the mark and provide identity information, in addition to checking the required feature.
Vision systems used to be complex, with cameras going to frame grabber boards in high-powered (at least for the time) computers. This was expensive and complex. Now, vision solutions are getting smaller and self-contained.
The vision sensor today acts more like a standard sensor but uses vision technology for the parts evaluation. This is a big step. Such vision sensors are available for around $1,000, whereas older-style, multi-component solutions could run well over $10,000.
In RFID, the most important development of the last few years was based on the realization that “openly available chips” have to be used — in particular, RFID tags based on 13.56-MHz tag technology. This allows RFID manufacturers to offer low-cost tag solutions that are very stable (i.e., can be used thousands of times) with chips that are also inexpensive.
It is important to note that “inexpensive” does not mean “for pennies.” The reason is, while significant costs have been moved out of the chip, costs are still present in high-quality housings able to withstand factory environments. While there are tags available for less than $1 (mostly in the shape of ISO cards), they are typically not robust enough for manufacturing applications.
A second important development in RFID is connectivity to industrial networks. This allows nearly every PLC (programmable logic controller) to use a simple and standardized method of controlling the RFID system and accessing tag data.
In the past, the biggest issues when it came to installing and using this technology were related to incorrect wiring. This challenge could be solved by using only quick-disconnect connections on all RFID products.
Today, the biggest challenge is picking the right hardware and mounting it correctly. Many customers think that putting the tag as close as possible to the read head is a good idea. However, this is not the case.
Also, users sometimes do not consider the fact that reading data takes time, and the size of a tag and read head has an influence on the maximum allowable (safe) passing speed.
Helge Hornis, Ph. D., is manager of intelligent systems at Pepperl+Fuchs, a provider of automation technologies.