PLC monitoring means reading values from the programmable logic controllers that run your machines, such as temperatures, pressures, counters, run states and fault codes, and watching them from somewhere other than the machine's own panel. You don't need a full SCADA (supervisory control and data acquisition) project to do it. If the goal is to see a handful of values from several machines in one place and get an alert when one goes out of range, a gateway, a register map and monitoring software are often enough.
That distinction matters because SCADA projects are built for control. They take engineering time, they're often tied to one automation vendor, and changing them is a project of its own. Plenty of engineers simply want to know that the chiller on line 3 is drifting warm before the product does, and that the water treatment skid from a different supplier hasn't thrown a fault overnight. This article walks through how that works with Modbus, the protocol you'll meet most often, and what to look for when the PLCs come from several vendors.
In short
- Most PLCs and many field devices can share data over Modbus, either natively or through a gateway.
- Monitoring reads registers on a schedule, scales them into real values, and checks them against conditions.
- Alerts need a delay and a reset condition, or every flicker becomes a notification.
- For mixed vendors, an edge gateway that translates each protocol into one common format keeps the monitoring side simple.
How does Modbus monitoring work?
Modbus is a simple request-and-response protocol that Modicon published in 1979 for its PLCs. It became a de facto standard because it was open and easy to implement, and it's now found on PLCs, drives, energy meters, chillers, boilers and countless sensors. Modbus monitoring means a client (traditionally called the master) asks a device for specific values, and the device answers.
Data in a Modbus device sits in four tables:
| Table | Typical address notation | Contents | Access |
|---|---|---|---|
| Coils | 00001–09999 | On/off outputs | Read and write |
| Discrete inputs | 10001–19999 | On/off inputs | Read only |
| Input registers | 30001–39999 | 16-bit measured values | Read only |
| Holding registers | 40001–49999 | 16-bit values and settings | Read and write |
Modbus RTU runs over serial lines, usually RS-485, with up to 247 device addresses on one bus, though a single RS-485 segment typically supports 32 standard devices without repeaters. Modbus TCP carries the same messages over Ethernet, normally on TCP port 502. Neither version pushes data on its own initiative, so monitoring is always polling: the client asks every few seconds or minutes, and the answer is whatever the register holds at that moment.
Each register holds 16 bits, which is fine for a value like 1830 but not for a floating-point number or a large counter. Those span two consecutive registers, and here's the trap that catches most first-time setups: devices disagree about the order of the two halves. Read a 32-bit float with the words swapped and you get nonsense, such as a temperature of 1.2 × 10⁻³⁸ °C. When a value looks absurd, try the other word order before suspecting the device.
Serial buses add their own problems. An RS-485 line needs termination resistors at both ends, every device on the bus needs a unique address and the same baud rate and parity settings, and only one client should poll the bus at a time. When a PLC's serial port is already used by an existing HMI, adding a second poller usually means a gateway that shares the port or a spare port on the PLC, not a second cable spliced onto the first.
Scaling is the other common surprise. Many devices send integers that you have to convert. A temperature input configured as 0–4000 counts for 0–100 °C reports 1830 when the real value is 45.75 °C. The device manual's register map tells you the scale, and comparing one reading with the machine's own display confirms you've read it correctly.
How to set up Modbus monitoring alerts
Knowing how to set up Modbus monitoring alerts properly saves weeks of false alarms. A procedure that works:
- Get the register map for each device from its manual or the machine builder, and list only the values you need, since every extra register adds polling load and one more value someone has to name, scale and check.
- Test-read each value with a simple Modbus client utility on a laptop, and compare it with the machine's display to confirm address, data type, word order and scaling.
- Choose a polling interval per value. Temperatures and tank levels rarely need more than once a minute. Fault codes and run states benefit from faster polling, within what the device and network can handle.
- Give each value a clear name and unit, such as "Chiller 3 supply temperature, °C", rather than "40107".
- Define the alert condition with a delay. "Supply temperature above 12 °C for more than 5 minutes" ignores a brief spike when a door opens; "above 12 °C" alone does not.
- Define how the alert clears. A reset at 11 °C rather than 12 °C (hysteresis) stops a value hovering at the limit from creating a new alert every poll.
- Decide who receives it, how, and who's next if they don't respond. Then trigger a test alert on purpose before relying on it.
Fault codes deserve their own treatment. Many PLCs expose an active fault number in one register. Rather than alerting on "fault register is not zero", map the fault numbers that matter to readable text and severity, so the person on call knows whether it's a blocked filter or a tripped motor before they drive in.
What is the best PLC monitoring software?
There isn't one answer to "What is the best PLC monitoring software?", because the tools solve different problems:
- Vendor engineering tools from the PLC maker show live values inside the program. They're the best way to debug logic, but they're meant for engineers with the project file, not for 24/7 monitoring.
- SCADA and HMI software supervises and controls a process, with operator screens and alarm handling. It's the right tool when you need control, and heavy if you only need to watch.
- Historians store large volumes of time-series data for analysis, often alongside SCADA.
- IoT platforms collect data from many devices and sites, with dashboards, rules and alerts, and usually run in the cloud or on a server you host.
- IT network monitoring tools check whether a device answers on the network. Some read a few Modbus values, but they rarely understand industrial data in depth.
The same logic answers "What is the best Modbus monitoring tool?" For a one-off check of a single device, a free Modbus client utility on a laptop is often all you need. For continuous monitoring across sites, you need something that polls, stores, scales and alerts without a laptop plugged in.
A few best practices for PLC monitoring hold for every one of these tools. Read only: monitoring should never write to a PLC, and if the gateway supports it, block write function codes entirely. Keep polling modest, because a PLC also has to run its program and a flood of requests can affect its communication load. Put monitoring traffic on a separate network segment or through a gateway at the boundary of the control network. Classic Modbus has no authentication or encryption (the newer Modbus/TCP Security specification adds TLS on port 802), so it should never be reachable from the internet. And keep the register maps in version control or at least a shared document, because they're the knowledge everyone loses when the integrator who wrote them moves on.
What is the leading solution for multi-vendor PLC monitoring?
When plants ask about the leading solution for multi-vendor PLC monitoring, they usually mean: how do we watch Siemens, Rockwell, Schneider and a few smaller brands without a separate tool for each? The answer that has become common across the industry is architectural rather than one product: put protocol translation at the edge, and keep the monitoring software protocol-neutral.
OPC UA deserves a mention here because it solves part of the multi-vendor problem at the source. Unlike Modbus, which only carries numbered registers, OPC UA (standardized as IEC 62541) carries named variables with data types, units and structure, and industry groups publish companion specifications that describe common machine types the same way. Newer PLCs from most major vendors include an OPC UA server, which makes them much easier to read than an undocumented register map.
In practice, an edge gateway, a small industrial computer or software package near the machines, speaks each controller's native language: Modbus, OPC UA, Siemens S7, EtherNet/IP and others. It reads the values you've configured and publishes them in one common format, very often JSON messages over MQTT. MQTT is a lightweight publish/subscribe protocol in which devices publish to named topics and software subscribes to them. Above the gateway, nothing needs to know which brand of PLC produced a value.
Kilo follows the same edge pattern: you connect Modbus equipment, as well as BACnet and OPC UA devices, through MQTT edge gateways such as Modbus2MQTT, Moxa or Advantech hardware, or Node-RED flows, and Kilo's MQTT connector takes whatever they publish. Raw keys from different manufacturers are then mapped to one normalized metric, so "supply temperature" means the same thing on every machine and a single trigger can watch all of them.
PLC monitoring FAQ
Is PLC monitoring safe for a running machine?
Reading data is low-risk when it's done sensibly: read-only access, modest polling and a gateway that can't write. The risks come from writing to registers by mistake or overloading an older PLC's communication port. Agree the register list and polling rate with whoever is responsible for the machine before connecting anything.
What port does Modbus monitoring use?
Modbus TCP uses TCP port 502 by default, and the secure variant, Modbus/TCP Security, uses port 802. Modbus RTU doesn't use network ports at all; it runs over a serial line, so monitoring it over a network requires a serial-to-Ethernet or serial-to-MQTT gateway.
Can PLC monitoring read Siemens or Allen-Bradley controllers without Modbus?
Yes. Siemens S7 PLCs speak their own S7 protocol and newer models offer OPC UA, while Allen-Bradley controllers use EtherNet/IP. Many edge gateways and industrial software packages read these protocols directly, so you don't need to add Modbus to the PLC program.
How often should PLC monitoring poll values?
Match the interval to how fast the value can change in a way that matters. A tank temperature that moves over hours is fine at one minute. A machine state you want to count as downtime might need a few seconds. Polling everything every second mostly creates load and storage without adding information.
Point an edge gateway at three machines from different makers, read one temperature and one fault register from each, and bring them onto a single dashboard with one rule that alerts you when any temperature stays out of range for five minutes. The whole setup fits on Kilo's free plan, so the only purchase is the gateway. Kilo Electronics stocks Modbus devices for projects like this, and gateways from any supplier work as long as they publish JSON over MQTT.