Troubleshooting
How a SCADA client reads a small PLC over EtherNet/IP
Short answer
An Ethernet link to a MicroLogix is not a working SCADA read: the controller serves data through PCCC commands carried over EtherNet/IP, and a client without that application service gets every read rejected. Check the driver's documented controller list for PCCC over EtherNet/IP or use the controller's documented Modbus TCP server, then prove one read-only value end to end.
On this page
A live Ethernet link to a MicroLogix proves almost nothing about a SCADA read. The controller serves its data through PCCC application commands carried over EtherNet/IP, and a client whose EtherNet/IP driver only speaks the tag services used by newer Logix controllers can open the TCP connection and still have every read rejected. Check the driver's documented controller list for PCCC support before you touch the PLC, or plan on the controller's documented Modbus TCP server, and prove one read-only meter value end to end.
Match the application service, not the protocol name #
EtherNet/IP is a transport envelope: it defines how a message rides on TCP and UDP, not what a device is willing to tell you. The application service inside the envelope decides that. Logix5000 controllers publish named tags over CIP services; the MicroLogix family predates that scheme and answers PCCC commands addressed to data files instead. A driver can honestly list EtherNet/IP support because it speaks the tag service, and still have nothing to say to a data-file controller.
The practical consequence: ping can succeed, and even a TCP connection to the EtherNet/IP port 44818 can open, while every data request comes back rejected. That is not a broken network and not a bad meter value; it is a missing application service. The fix lives in the SCADA driver selection, not in PLC settings.
Check the driver before the PLC #
Work down this list in order and stop at step 6:
- Record the SCADA client version and the exact name and version of the communication driver or server you intend to use.
- Open that component's supported-device documentation and look for the controller family by name together with PCCC over EtherNet/IP. The bare phrase EtherNet/IP is not enough.
- Confirm the addressing model it documents: data-file addresses for PCCC controllers, symbolic tag names for Logix. If every example shows tags, it is the wrong driver for this controller.
- Check that the driver is licensed and actually installed in the runtime doing the polling. A protocol printed on a datasheet proves nothing about the component in service.
- Read the change log. Driver capability moves with versions: one SCADA platform's EtherNet/IP driver only gained PCCC support in 2021, aimed at MicroLogix-class controllers, so an old supported-device PDF is not evidence about your installed build.
- Decide. PCCC documented means configure and test. Absent or unclear means leave the PLC untouched and take one of the paths below.
What each symptom tells you #
| Observed symptom | What it points to | Next check |
|---|---|---|
| Client cannot reach the controller at all | IP settings, routing, cabling or a filtering firewall | Fix the network path first; an alarm on link loss keeps cable-down separate from request-rejected afterwards |
| Reachable, but every read fails or shows bad quality | Missing PCCC service or wrong driver selection | Run the driver check above; this is also why HMI values show question marks when reads time out |
| Some items read, others fail on the same connection | Address or data type outside the documented range | Compare each failing item against the controller project |
| Reads report good quality but the values are wrong | Word order, scaling or a stale mapping | Walk the worked example below against the project |
A reachability test cannot verify PCCC support. An ICMP reply or an open TCP session says the transport works, and the transport was never the question.
Two documented paths when the driver has no PCCC #
Path one is a communication server that documents PCCC over EtherNet/IP. Install it, point the SCADA client at it through the interface both products document, and match the client's interface type to what the server actually exposes. A working Modbus TCP OPC connection proves that one path; it does not configure a second server's items or permissions.
Path two is the controller's own Modbus TCP server. The MicroLogix 1100 user manual documents Modbus TCP connectivity on the Ethernet port, so controllers running firmware that supports it can answer the Modbus TCP driver you already trust. The register map still has to be defined in the controller project: a PLC that only collects Modbus RTU meter data for its own program exposes no registers until you map them. Never translate a data-file address into a guessed register. The two address spaces are unrelated, and the mapping, the register numbering convention and the word order come from the project and the manual, not from a pattern half remembered from another controller.
Worked example: one mapped energy value #
Say the project maps the meter's energy total, a 32-bit integer in watt-hours, into two consecutive holding registers at register address 100, high word first. The meter reads 1,000,000 Wh, which is 1,000.0 kWh.
Split the value across the registers: 1,000,000 divided by 65,536 is 15 with a remainder of 16,960, so the high word is 15 and the low word is 16,960. Check it: 15 × 65,536 = 983,040, and 983,040 + 16,960 = 1,000,000.
A client configured to read one register picks up the low word alone: 16,960 Wh = 16.96 kWh, with a good quality flag and a plausible-looking number. It is 98.3 percent low, because (1,000,000 - 16,960) / 1,000,000 = 98.3 percent. Read the two registers as one 32-bit request in the project's word order and the client shows 15 × 65,536 + 16,960 = 1,000,000 Wh = 1,000.0 kWh. The register range, word order and scale in this example are illustrative; your project's values govern.
Prove one read-only value end to end #
- Pick one meter value whose PLC location, data type and units you have verified in the project.
- Run the test on the machine that will host the client in production. If that station lives in the cabinet door or on the console, that means an industrial panel PC, not a laptop that leaves after commissioning.
- Configure the connection in the server and confirm the raw read in the server's own diagnostics before opening the SCADA client at all.
- Create one client item using the server's documented address syntax, with its data type matched to the PLC value.
- Watch the PLC value, the server value, the item quality and the client display together.
- Change the field value and confirm the new number arrives within the configured poll interval.
- Only then check scaling and screen formatting, and keep every write out of the test. This is a monitoring application, and a read-only connection cannot stop anything.
The decision for each case #
- The driver documents PCCC over EtherNet/IP for MicroLogix: configure it, address the data files as its manual shows, and the job is done with no PLC change.
- The driver has no PCCC, but the controller's Modbus TCP server is documented for your firmware: keep the Modbus TCP driver, define the register map in the project, and read every multi-register value as one request in the project's word order.
- Neither applies: stop editing addresses, because a missing application service is not fixable from either end's settings. Add a documented converter, either a communication server with a documented PCCC driver or a protocol gateway whose mapping you can read on both sides.
- The rule that started this: a link that answers ping and rejects reads is a service mismatch, so the next move is always the driver's supported-device list, never the PLC.
Questions asked
Can one SCADA driver read both Logix tag controllers and MicroLogix data files?
Only if it documents both services, and they are normally separate selections in the driver setup. The CIP tag service reads named tags on Logix5000 controllers; PCCC reads data files on MicroLogix and SLC-class controllers. Select the entry that matches the controller family in front of you, because the addressing syntax differs: data-file addresses on one side, symbolic tag names on the other.
Which Modbus function codes should the commissioning read use?
Use function code 03 for holding registers or 04 for input registers, whichever the project's mapping defines. Keep the write codes, 06 and 16, out of the commissioning test entirely. A monitoring connection that only ever issues reads cannot stop a machine, which is the safest state to prove a new path in.
How often should the SCADA poll an energy total like this?
Poll a cumulative energy counter every 2 to 5 seconds. The value changes slowly, so faster polling adds controller and network load without buying information, and slower than roughly 5 seconds makes the trend look broken. Keep the poll interval comfortably longer than the controller scan time so the number you read is a settled value.
Why does the client show a healthy connection status and still no values?
Because connection status tracks the transport, not the application service. The client can hold an open session to the EtherNet/IP port while every PCCC request inside it is rejected. Judge the path by item quality and by the server's own read diagnostics, and treat a healthy status with bad quality as a driver or service mismatch, not a network fault.