The hardest part about designing a control system is to figure out which components to use in the design. As someone who is the main designer and programmer for a packaging company, I know that I can come up with a set of schematics and a matching program and get all the way to the finished PLC and HMI applications without giving more than a cursory thought to the actual components to be used on the machine. The challenge comes when the design gets down to the detailing needed to order components.
For example, a normally open photo sensor is just a symbol on a schematic, and its behavior can be very binary in a PLC application, but it becomes a living thing when it is mounted on a machine. This requires examining some of the considerations for deploying a design to a machine. My industrial electrician might think that the selection of component is easier, but the right choice of components can make his job significantly easier.
I have to admit that my journey through control design has been blessed in many ways. I always had other team members who were mainly focused on the actual on-machine components while I was concerned with the drawings and then the programs. By the time I got to start commissioning the machine, the magic of getting the components on the machine and the wires back to the control panel were already taken care of.
In my current role, I am best described as an OEM inside an end user. My former team of 15 to 20 skilled engineers and technicians is now a small group of five or six people who came to this place in this time with a set of skills that allows us to do great things with limited resources.
With the ongoing restrictions due to the pandemic, we are challenged even more to ramp up automation projects with that same small group of people. For me, this means paying closer attention to what my electrician/fabricator is doing to make my designs come to life so that I can make his life easier and improve on the delivery time.
Previously, I have discussed the actual components on a machine in some detail, but how do we bring it all together? My own journey through the design process mirrors the evolution of on-machine connectivity.
I have always tended to stay with wired solutions because the cost of the machine cable is relatively low and distribution of signals from the main control cabinet out to the end devices is a well-established tradition of junction boxes connected by wire way and conduit.
Later on, I started using some distribution blocks for I/O to take care of connecting to upstream and downstream equipment because of the ease of using an M12 connector to quickly hook up to other pieces of equipment and quickly disconnect when we wanted to move that equipment around.
It’s a fairly easy progression to move from a machine-mounted distribution block connected to a junction box and eliminate the junction box altogether by using a longer home-run cable to bring that field-mounted I/O directly back to the main control panel.
It’s easier yet to move from landing those wires on terminal blocks in the main panel to installing a bulkhead on the side of the cabinet and pre-wiring from the bulkhead to the terminals in the panel. Now the machine cable can be quickly connected during the machine-building process.
For the team at my place of employment, these early steps in the design journey came about because we regularly put new control systems on older packaging equipment. Our approach, up to my arrival, was to duplicate the existing control system using the original control panel.
We were literally putting new components into a 50-year-old design. The features that we added, early on, were to facilitate the need to set up and tear down production lines to suit the needs of our clients. The need to go from a horizontal packager making pouches for a vertical cartoner to a vertical packager making pouches to be inserted into the flights of a horizontal cartoner, with all the upstream and downstream equipment, would take long periods of time.
Any time a line isn’t producing product, our company isn’t making money. Our early, on-machine, efforts were intended to reduce the time needed to separate all the various pieces of equipment (electrically) so that new equipment could be brought into the room and connected back together.
The more significant change in our approach to control design happened when we decided to move away from replicating the existing, decades-old designs and come up with new control systems.
As we quickly learned, the amount of cabinet space needed for the old control systems was huge compared to what modern technology allows us to do. What we ended up with was a huge enclosure that was only partially filled with the new control design.
The idea quickly developed that we could reclaim some of the physical space occupied by these huge enclosures. The nature of our business leads us to put a lot of packaging equipment in a small footprint, and these large enclosures often hinder the ability to look across or down a line to other equipment.
We also wanted to pay close attention to the needs of arc-flash protection, so we came up with a concept to split the single enclosure into two smaller enclosures.
The immediate improvement on the sight lines of our equipment was dramatic. We used a windowed enclosure for the control side and everything in that enclosure is 24 Vdc or less.
Once focused on reducing the time to execute a build or a rebuild, we really dug into areas where we could leverage the pre-build time to reduce the actual time on the machine once it hit our shop floor.
One key way to do this is the use of field I/O. We had not previously used much of this technology but started by putting a point I/O station inside our operator station enclosure. All of the operator devices could be wired directly to the point I/O terminals, and the net result was a significant reduction on the amount of on-machine wiring time.
We simply ran a small machine cable with switched and un-switched power signals, as well as an Ethernet connection for the point I/O and human-machine interface terminal. We even found a DIN-rail-mounted breakout board that would allow us to use the machine cable conductors to provide our Ethernet connection between the main PLC and the remotely located I/O station, further reducing the number of cables to be pulled.
If one point I/O station could work, certainly more would be better. Most of our machines now utilize three or more of these stations, and we have already reaped the benefits of having field wiring that only has to be run a few feet to the I/O station, rather that 10 or 20 feet back to the main control panel.
The added benefit of the local I/O stations is we have reduced the number of modules in the main PLC chassis, reducing even more space in the main control enclosure. We enclose these remotely located I/O stations in a windowed enclosure, again so that we can check on the status of the communications, as well as the I/O points without having to open the enclosure.
Most machine-mounted I/O technology now employs in/out ports for communications. This means that, in most cases, a single communication cable is all that is needed to pick up field I/O, human-machine interfaces and even items such as encoders that would normally need a separate network connection for each, and the requirement for a remotely located communications switch.
The cost reductions are significant when compared to the cost of a skilled person wiring up all this technology in a more traditional dedicated wiring system.
It would be remiss to not talk about the emergence of IO-Link as a game changer for on-machine components. By putting an I/O master in place of the usual input or output module in the point I/O station, a wealth of additional information and diagnostics are available through the very same components on the machine as most manufacturers have IO-Link embedded in the common machine components we have used for years.
There are no wire changes, but that master module in the point I/O block can make a serious improvement in the capabilities of a control system.
Finally, as a programmer, I like to get as much of the control design tested out at my desk, long before the on-machine wiring takes place. The relatively small footprint of remote I/O stations means that I can set up the entire machine, control-wise, in an area that is only a little bigger than the area my keyboard occupies on my desk. Everything can be pre-configured and tested for communications, and most of the PLC algorithm proven out before handing the components over to the electricians and fabricators.
In our situation, we have come up with a way to use the same controls architecture for a horizontal or vertical cartoner, for example, that we would use for a horizontal packager. The individual on-machine sensors and controlled components might change, but the architecture and major program elements are the same, regardless of the machine that we are building. Our development times are significantly reduced by using on-machine distribution systems to make it all come together.
