Ethernet vs Wireless for Industrial Automation: A Scenario-by-Scenario Guide

By Dustin Guttadauro, Product Line Manager - Telecom & Fiber, Infinite Electronics  

Key Takeaways 

• Short answer: Use wired Ethernet (PROFINET IRT or EtherCAT) where you need sub-millisecond determinism and functional safety. Use wireless (Wi-Fi 6 or private 5G) where cables are physically impractical — AGVs, rotating fixtures, remote assets. 

• Wired protocols like EtherCAT achieve cycle times below 100 µs with jitter under 1 µs — a performance floor no current wireless technology can match for motion control. 

• Wi-Fi 6 and private 5G have closed the gap significantly for non-time-critical applications: condition monitoring, remote diagnostics, and mobile robotics are all viable wireless use cases today. 

• Most real plants use both: a fiber or copper backbone carrying PROFINET or EtherCAT, with wireless only at the 'last meter' to mobile or rotating equipment. 

• Picking the wrong medium is expensive — rewiring a cell or retrofitting wireless APs after installation costs far more than making the right call in the design phase. 

Walk any modern production floor and you'll see both approaches running side by side. CNC machining centers hardwired into PROFINET IRT rings. AGVs weaving between them on Wi-Fi 6. Neither is universally right. The question is which one fits your specific application — and where the wrong choice causes real pain. 

This guide skips the abstract debate and works through five concrete automation scenarios, recommending wired or wireless for each with the reasoning behind it. There's also a latency table comparing the major protocols, and a section on hybrid architectures — which is, honestly, what most well-designed plants end up with anyway. 

Why Does Wired vs Wireless Even Matter? The Core Tradeoff 

The fundamental issue is determinism — the guarantee that data arrives within a predictable time window, every cycle, without exception. Wired industrial Ethernet protocols are built around this guarantee. Wireless protocols are not, at least not to the same standard. 

PROFINET IRT (Isochronous Real-Time) and EtherCAT deliver cycle times measured in microseconds with jitter so tight it's essentially invisible to the control system. That's what motion control and synchronized robotics require: the servo drive needs to know exactly when the next setpoint is coming. 

Wi-Fi introduces variable latency caused by radio contention, retransmission, and channel switching. Even Wi-Fi 6 with its OFDMA scheduling sits at 2–10 ms under good conditions — and conditions in a factory aren't always good. Private 5G with standalone core improves on this, but still doesn't reach the sub-millisecond window that synchronized axes demand. 

The full picture of how these protocols sit in the broader connectivity stack is worth understanding — the  covers the architectural differences in depth. For now, here's where it matters in practice. 

Latency and Jitter: The Numbers That Drive the Decision 

These are realistic operating ranges from published vendor and standards documentation. Your actual numbers will depend on network load, cable quality, and AP placement. 

Table: Latency and jitter comparison for major industrial protocols. 

5 Automation Scenarios: Wired or Wireless? 

The right answer changes depending on what's actually moving, what's at stake if a packet is late, and whether cabling is physically practical. Here's how it plays out across the scenarios engineers encounter most often. 

Scenario 1: Multi-Axis Motion Control — Use Wired (EtherCAT or PROFINET IRT) 

Synchronized motion control — coordinating multiple servo axes for a CNC machine, a pick-and-place system, or a linear transfer line — is where wireless simply doesn't belong yet. The axes need to receive setpoints in a deterministic, isochronous window. Miss that window and you get positioning error, following error faults, or worse, a collision. 

EtherCAT is the dominant choice for high-axis-count motion because its distributed clocks mechanism synchronizes all nodes to under 1 µs. PROFINET IRT achieves similar timing through its scheduled communication channel. Both require the right cabling: Cat 5e minimum, Cat 6 preferred, with proper shielding in high-EMI environments. 

For cabling into servo drives and motion controllers,  rated for continuous flex cycles are worth specifying from the start — standard patch cables fail early in moving machine axes. 

Scenario 2: Robotic Work Cells — Use Wired, With One Exception 

Robot controllers communicate with the plant PLC over PROFINET or EtherNet/IP, typically at cycle times of 4–8 ms for standard robot motion. Wired is the right call here for the same reasons as motion control: low latency, high reliability, no RF interference from welding equipment or other robots in the cell. 

The exception: tool changers and end-of-arm tooling that rotate past the cable management limit. Here, a short-range wireless link — sometimes WLAN, sometimes a dedicated industrial wireless solution operating in the 2.4 or 5 GHz band — handles the 'last meter' from the robot wrist to the end effector. This is a deliberate design choice, not a workaround. 

M12-coded connections are standard at the robot and tooling interface.  in D-coded (100 Mb/s Ethernet) or X-coded (Gigabit) versions are designed for the vibration and repeated mating cycles robotic work cells produce. 

Scenario 3: AGVs and AMRs — Use Wireless (Wi-Fi 6 or Private 5G) 

This is wireless's strongest industrial use case. Automated guided vehicles and autonomous mobile robots move continuously through the facility — cables aren't an option. The communication requirement is also more forgiving than motion control: position updates, load assignments, and traffic management don't need sub-millisecond timing. 

Wi-Fi 6 (802.11ax) handles most AGV/AMR fleets well. The OFDMA scheduling reduces contention in dense AP environments, and target wake time (TWT) extends device battery life on battery-powered platforms. For larger facilities or environments where RF coverage is complicated by metal racking and high interference, private 5G is worth evaluating — its deterministic scheduling and better wall/obstacle penetration can simplify the RF design. 

The critical design factor for AGV wireless isn't the protocol — it's AP placement and roaming behavior. An AGV that stalls mid-floor while its Wi-Fi client searches for a new AP is a bigger operational problem than most teams anticipate. Design for seamless roaming from day one. 

AGV and AMR connectivity is also where PoE infrastructure matters — powered APs and network switches that don't require separate power runs simplify installation.  rated for DIN-rail mounting and wide temperature ranges hold up in environments where standard IT switches fail. 

Scenario 4: Condition Monitoring — Either Works, Wireless Often Wins on Cost 

Vibration sensors, temperature transmitters, and acoustic emissions monitors on rotating equipment are a natural fit for wireless. The data rate is low, latency requirements are loose (a few seconds of lag won't hurt a bearing health trend), and cabling to motors and gearboxes in confined machine bases is genuinely difficult. 

WirelessHART and ISA100.11a are purpose-built for this use case — mesh topologies that handle the RF environment in process plants, with security built to IEC 62443 standards. For OT-adjacent condition monitoring that reports into a standard Ethernet network, Wi-Fi 6 endpoints work fine. 

Where wired condition monitoring makes sense: safety-critical measurements where data loss is unacceptable (pressure relief valve status, for example), or where the sensor is physically close to an existing Ethernet drop and the extra cable run costs less than a wireless node. 

Understanding which  your condition monitoring platform supports will narrow the hardware choice significantly. 

Scenario 5: Remote Pump Stations and Distributed Assets — Wireless, Sometimes Cellular 

A pump station three miles from the nearest control room isn't getting a fiber run without a serious business case. This is where wireless — and sometimes public LTE or private 5G — is the only realistic option. 

Private 5G standalone networks are increasingly viable for distributed industrial assets: a refinery with tank farms, a water utility with lift stations spread across a municipality, a mining site with equipment at multiple pit faces. The coverage radius of a single 5G base station, combined with deterministic scheduling and network slicing, handles asset monitoring and SCADA connectivity for geographically spread infrastructure. 

For smaller deployments — a single remote skid, a pipeline monitoring point — cellular (LTE or 5G public network) with a secure VPN tunnel to the plant historian is often the most cost-effective answer. The bandwidth requirements for SCADA polling are low enough that public cellular handles it fine. 

What Does a Hybrid Architecture Actually Look Like? 

Most real plants don't choose one or the other — they use both, deliberately, with the boundary placed where it makes engineering sense. 

The common pattern: a fiber backbone running between electrical rooms, connecting industrial Ethernet switches that distribute PROFINET or EtherNet/IP to fixed machinery. Wireless APs connect to those same switches and provide coverage for mobile equipment, handheld terminals, and any assets where cable management is impractical. 

The 'fiber backbone + wireless last meter' architecture keeps deterministic traffic on wire and reserves wireless for applications that can tolerate variable latency. The separation is both physical and logical — OT traffic on a dedicated VLAN, wireless devices authenticated through 802.1X, with firewall segmentation between the OT and IT networks. 

A few things this architecture gets right that all-wireless designs don't: if the Wi-Fi goes down, the production line keeps running because motion control never touched the wireless network. Troubleshooting is also simpler — you're not chasing a latency problem across a mixed wired/wireless control loop. 

Which Factors Should Override the Default Choice? 

There are situations where the usual recommendation flips. Knowing them saves time in the design phase. 

Go wireless even for time-sensitive applications when: 

• The machine involves continuous rotation past cable limits (robot wrist, carousel, reel) and slip rings are ruled out 

• Retooling frequency means cables would be disconnected and reconnected weekly — connector wear becomes a reliability problem faster than most teams expect 

• The facility layout changes often enough that fixed cable runs would need to be redone within the equipment's useful life 

Go wired even for non-critical applications when: 

• The RF environment is genuinely hostile — induction heating, high-power VFDs, and large metallic enclosures can make wireless coverage unreliable in ways that are hard to predict and harder to fix after installation 

• Security policy requires air-gapped OT networks where wireless introduces an unacceptable ingress point 

• The sensor or device is stationary and within 100 meters of a switch — the incremental cost of a cable run is almost always less than the ongoing management overhead of a wireless node 

How L-com Helps 

L-com supports Industrial Internet of Things (IIoT) infrastructure with rugged connectivity solutions designed for demanding industrial environments. From industrial Ethernet components and ruggedized cabling to fiber connectivity and edge networking support, L-com helps organizations build reliable physical-layer networks that support long-term industrial system performance and uptime. 

Frequently Asked Questions (FAQs) 

Is wireless industrial automation reliable enough for safety functions? 

Not yet for most SIL-rated safety functions. IEC 61784-3 (functional safety communication profiles) covers wired PROFISAFE and similar protocols, but wireless channels introduce failure modes — packet loss, interference, handoff gaps — that complicate safety analysis. PROFIsafe over PROFINET RT does exist and handles some wireless safety applications, but the engineering burden is high and most safety-rated systems still run on dedicated wired paths. If your application requires SIL 2 or SIL 3 coverage, design the safety loop on wire. 

What's the difference between industrial Wi-Fi and standard Wi-Fi? 

The hardware, mostly. The 802.11 protocol is the same, but industrial Wi-Fi access points and clients are built for temperature ranges, vibration resistance, and ingress protection that standard IT equipment doesn't meet. Industrial APs also typically support faster roaming protocols (802.11r, 802.11k) out of the box — important for AGVs that need to hand off between APs without dropping the connection. Managed industrial Wi-Fi infrastructure also tends to offer deterministic QoS settings that help prioritize OT traffic over IT traffic sharing the same network. 

Can PROFINET run over wireless? 

Yes, with caveats. PROFINET RT (standard, non-isochronous) can run over Wi-Fi using PROFINET over WLAN (specified in PROFIBUS International guidelines). This works for device-class communication that doesn't need sub-millisecond timing — remote I/O, HMI connections, some robot interfaces. PROFINET IRT, the isochronous variant used for synchronized motion, requires a scheduled Ethernet medium and cannot run over standard Wi-Fi. Private 5G with time-sensitive networking (TSN) extensions may eventually enable IRT-class wireless, but that's not a standard production deployment as of 2026. 

How much latency can an AGV network tolerate? 

For typical fleet management traffic — position updates, route assignments, obstacle detection data — AGV systems generally tolerate 50–200 ms without operational impact. The more sensitive path is emergency stop propagation: most AGV safety systems require the e-stop signal to actuate within a defined time window (often under 100 ms from trigger to motor cutoff). This is typically handled by a dedicated safety-rated wireless channel or an onboard safety PLC that doesn't depend on the fleet management network for its stop function. 

What should I specify for cabling in a new automation cell? 

For fixed machine wiring: Cat 6 shielded (S/FTP) at minimum for any runs over 30 meters or in high-EMI zones. Cat 5e works for short, low-interference runs but limits your upgrade path. Specify M12 D-coded connectors at the field device end — they're designed for the vibration and IP67/IP68 requirements of machine tool environments, where RJ45 connections fail prematurely. For robot cables and any moving axis, specify continuous-flex rated cable and connectors with cycle ratings appropriate to your machine's duty cycle. 

Making the Call 

The wired vs wireless decision isn't really about preference — it's about matching the medium to the application's timing requirements, the physical environment, and the operational lifecycle of the installation. 

Motion control and synchronized robotics: wired, full stop. AGVs, remote assets, and condition monitoring on rotating equipment: wireless makes sense and often wins on installation cost. Everything else lands somewhere in between, and a hybrid architecture usually handles that middle ground better than forcing a single medium across the whole facility. 

Getting the cabling and connectivity specified correctly in the design phase costs far less than fixing it after commissioning. If you're working through the infrastructure decisions, reviewing the specific protocol and hardware options early saves time in procurement and avoids redesign after the equipment arrives on site.