Because every process application and site has its own particular characteristics—and obstacles—system integrators and their clients must still perform site surveys and assessment to identify and resolve those difficulties.
“When technologies reach maturity and ‘just work,’ they become expected commodities. Wireless is reaching this point, but site surveys and assessments, and individualized designs and implementations are still required for successful wireless applications,” says Corey Schoff, senior network security engineer and team lead at Malisko Engineering Inc., a system integrator in St. Louis, and a certified member of the Control System Integrators Association (CSIA). “This is why we do wireless surveys for clients with existing operations, so we can determine what they’ve got, and develop recommendations and designs to improve it. This is also important for greenfield projects, so we can predict what they’ll need to quote and assemble. Surveys also enable troubleshooting, remediation and homing in on performance targets.”
Get your subscription to Control's tri-weekly newsletter.
Just like most chores, wireless applications are never completely settled because their networks always need to be reworked or redesigned as equipment moves and/or processes change, which impacts their spaces and environments, and shifts wireless signals and requirements.
“When aisles of pallet racks or other metal equipment is added, they can impact a wireless network’s design requirements, which must be redone,” explains Schoff. “We recommend making these revisions whenever a large facility is altered, and looping in the wireless team, so they can help build wireless into their facilities, and analyze and plan for later changes.”
Schoff reports that Malisko often adds more wireless access points to production, distribution and warehousing facilities operated by its food and beverage and pharmaceutical clients. Most wireless access points can efficiently and safely handle a maximum of about 20 client devices, such as sensors, I/O and laptops. Higher-bandwidth applications, such as video from cameras or guest users like contractors, will require more network capacity.
“Basic building characteristics don’t change that much, but if clients add new tanks and supporting equipment, they can change what wireless is needed where. Stainless-steel, fluids and even water are all hard for wireless signals to get through,” adds Schoff. “Users need more robust network designs because the days are gone of just putting a few, wireless access points in the middle of a facility. Newer wireless technologies use higher frequency signals. These typically range from 2.4 GHz to 5 or 6 GHz, but the tradeoff is that higher frequency means reduced signal propagation, at least half or shorter available distances, and reduced ability to navigate around machines, walls and other obstacles, which requires more access points and better network design.”
On-the-ground design
To build a robust and effective wireless network, Schoff reports that onsite surveys and assessments are still essential, and that at least one full day should be spent at the facility, even if floorplans and predictive materials are already available. “Onsite time is crucial for identifying challenges that floorplans don’t show,” says Schoff. “These include large tanks, new piping and ceiling heights. “
Schoff adds that useful physical attributes of wireless networks that should be considered include:
- Positioning access points at the same height, and also attempting to have I/O, laptops, automatic guided vehicles (AGV) and other devices sending data at about the same heights, which can enable better transmissions;
- Placing antennas and arrange the distances between them in optimal ways for each specific coverage pattern, as antennas focus signals like a lens to further aid communications;
- Deploy more wireless devices than initial estimates required, which will allow redundancy, and avoid relying on covering any area with just one access point.
- Because the power levels of sensors, I/O, other transmitting devices, and access points can vary depending on how often or intensely they’re used, maintain medium power levels for access points, so they can reach end-point devices and be found by them, and coordinate their efforts for maximum efficiency.
- Beyond obvious device characteristics, evaluate the contents of larger settings, buildings and environments where the wireless network is located. This includes what their walls are made of, and how often their tanks, racks and other structures are full or empty.
Because the radii of their coverage areas should overlap to ensure effective communications, Schoff reports that wireless networks need to establish and maintain a higher density of access points in case one fails due to running out of power or other issues.
“You don’t want transmitting components, access points, or other devices running at maximum power all the time,” explains Schoff. “Access points can operate at a maximum, average power output of about 23 decibel milliwatts (dBm), which is a logarithmic scale. For comparison, a laptop PC with wireless capability usually runs at 10-17 dBm, while an FM radio runs at 80 dBm. You also want to run access points at lower power, so they’ll continue to outdistance their client/sending devices. This is because a 10 dBM sensor or I/O may scan successfully for a 23 dBm access point, while the access point may not be able to hear those 10 dBm components.
“This situation is just like buying a residential access point with an integrated antenna for a home network. Users typically have 15-17 dBm laptops or other devices running at about 5 GHz over distances of 100-150 feet, so their access point creates a small-scale, enterprise-like, wireless network. This is why it’s important for access points to maintain medium power levels, so they can match the power and distances of client devices and serve them properly.”
Wireless on the move
More recently, wireless is also affected by the amount of mobile equipment that’s ballooned in industrial settings. Where wireless networks in most facilities previously mainly dealt with trucks driving in and out, now they have to face multiplying AGVs, robots and support devices rolling around, and potentially impacting wireless signals.
“Increasingly mobile equipment can be a big challenge for wireless. If your business is looking to add AGVs, don’t keep it a secret from your wireless team,” says Schoff. “AGV’s use wireless because they need continuous telemetry about where they are, where to go, what to pick up and drop off, and how to avoid conflicts and crashes. Many AGVs are also increasingly autonomous, and without sufficient telemetry, their efficiency will be reduced, and they can’t navigate or coordinate with their fellow AGVs.”
Schoff points out that AGVs are similar to laptop PCs because they operate at about 15 dBm, and their suppliers each have different wireless requirements that end-users and their facilities need to support.
“In the past, AGV vendors would come in and set up their own, independent wireless network. Now, more users have existing wireless infrastructures, so we recommend that they test and certify that their existing network can support the new AGVs or other equipment,” explains Schoff. “They need to ask whether the AGV could link to existing access points and types, and can the present coverage area work with the expected AGVs? This is similar to adding a laptop, I/O or barcode scanner because users want to certify that all of them can link with their network and function as intended, even though they’re overall numbers are multiplying.”
Revising a CPG network
For example, Malisko recently worked with a consumer packaged goods (CPG) client to redesign the wireless network for its mostly discrete production, assembly, packaging and shipping operations, along with some metal-plating and other process applications. This client runs about 2,000 connected devices, including I/O, PLCs, switches, machines, torque screwdriver and other tools, and went from 37 access points to 103 in the same space, or an almost threefold increase.
“This increase in access points wasn’t because the user needed to reach more client devices. It was because they needed greater density to provide better coverage and more consistent signals,” explains Schoff. “Thirty-seven access points weren’t enough to give them the consistency they needed, plus they had some dead spots in their plant. Even though the physical space was the same, using only 37 access points meant that each client device and access point were effectively further away from each other, which reduced speed and bandwidth. Using more access points over shorter distances also allowed the client to go from 2.4 GHz to 5 GHz communications, which further improved signal penetration and consistency.”
To upgrade the CPG client’s wireless network, Malisko conducted its usual onsite survey, analyzed the results to determine that 103 access points were needed, cut over to 5 GHz after they replaced the previous 37 access points, and even added a secondary, 5GHz network. The new components included Cisco’s Catalyst 9130 access points running on 802.11ax Wi-Fi, two high-availability Catalyst 9800 wireless local area network (WLAN) controllers, and several antenna models, including internal, external omni-directional, external directional, and ceiling-mounted directional that can cover more circular planes.
“The new network improved wireless coverage, eliminated dead zones, and provided more consistent signals,” adds Schoff. “However, some spots still had issues despite the new design. These spots didn’t match the simulation for the new network, so we resurveyed to identify and remediate them. We readjusted some antennas, moved some access point and returned and adjusted some software settings.”
Overlapping with the edge
Schoff adds that Malisko is also implementing some edge-computing functions, lately with help from Cisco’s Secure Equipment Access software that runs on cloud-brokered gateways, as well as Cisco’s Cyber Vision software that also runs on switches, such as Rockwell Automation’s Stratix 5800 with remote, edge-computing module. The software lets users employ gateways or switches to broker links with machines and other production assets that are as close as possible to operations and their data.
“Instead of establishing a virtual, private network (VPN) for a whole plant, and then drilling to an individual PLC, it’s simpler to use a brokered link on a switch that provides users with the same granular access,” says Schoff. “These switches can also use a container-based extension to run Cisco’s Cyber Vision software for collecting data that can help with visibility, analytics and threat detection. We also run Docker software containers on Cisco’s IOx firmware modules on Stratix, but they could also run on any Docker-compatible application.”
