I’ve been talking pneumatic systems (valves, cylinders, actuators, etc.) recently with my customers and I’m finding among these engineers some common pains coming out of the system design. It seems that many people are researching networked valve islands with I/O built-in. These seem to be a great way to consolidate lots of I/O into one IP address, but there are some new issues cropping up similar to the above photo:
- When assembling these at a machine builder the routing of cables with piping is more cumbersome with cables hanging off the valves, larger cable tray installations and large amounts of piping all running to the same spot.
- For machine builders, with all of the valves centralized in one place, the pneumatic lines have to be longer. This causes many issues such as slower responsiveness due to air volume, air inertia, and lower air quality.
- When trying to perform maintenance at an end-user, it becomes a nightmare to troubleshoot with a cluster of cables and pipes. The zip-tied and clean runs installed by the machine builder are cut, tangled and re-routed as the machine ages and becomes more difficult to troubleshoot.
- Also at end users, if the manifold needs to be expanded, updated, retrofitted with new valves or I/O, there are big hurdles to jump when doing this: re-piping the valve due to mounting position shifting or even having to edit and repair code in the PLC to adapt to new bitmaps generated by the new valve manifold configuration.
- When closing the loop with magnetic field sensors mounted on the cylinders, typically reed switches are used which are prone to failure. In addition, these switches typically have two sensors & cables per actuator to give extend or retract position, these cables cause larger cable trays and long cable runs back to the centralized manifold and I/O.
What these problems really boil down to is that we need to be able to distribute our valves and I/O across our system design without compromising the value we have gained from a machine mounted I/O device and networked manifold. Improving upon this boils down to three simple to implement steps.
Step 1: Distribute Valve Architecture
Are your actuators centralized on your machine or decentralized? I am going to assume they are all over. By distributing the valves across the machine close to a cluster of actuators, you gain responsiveness (less air volume and inertia) with better air quality.
Step 2: Spec Distributed Modular I/O Architecture
Remember all those cable runs, like in the photo above? What if we could distribute the I/O inputs across the machine close to a cluster of sensors? Our cable runs could be shorter, easier to troubleshoot and reduce our cable trays and cable management. If you are unfamiliar with the distributed modular I/O architecture, check out these blog entries or this webpage.
Step 3: Implement Solid State Cylinder Sensors with Integrated Connectors
Reed switches are mechanical devices that are prone to failure in repeated use automation. There are technologies available like GMR that are solid state, much more reliable and are competitively priced to Reed technologies. In addition, these products are now available with two sensors with one connector, this is a big deal because the number of cables required for feedback is cut in half!
Final Result: Distributed I/O & Pneumatic Systems
So by implementing these three steps into your pneumatic system design you can improve the performance of the machine, clean up the look, make it easier to troubleshoot and simplify future expansion or modification. Please check out this webpage that talks about potential distributed valve architectures by valve manufacturer brand.