Troubleshooting
Fix EtherNet/IP I/O between a ctrlX adapter and a Rockwell scanner
Short answer
Start at the Studio 5000 module status: a connection that never establishes points to the IP path or the assembly fields, while a connection with wrong or frozen data points to swapped input/output instances or the task interval. Match instance numbers and byte counts at both ends, instance 100 for consumed data with configuration instance 102 at length 0 in the documented example, and set the adapter I/O task interval equal to the scanner's RPI, 4 ms at priority 20 to 39.
On this page
- Read the scanner status before you reset anything
- Confirm the IP path and the selected interface
- Match assembly instances, byte lengths and direction
- Working the numbers for 10 consumed and 15 produced bytes at a 4 ms RPI
- Keep the I/O mapping refreshing
- Bring the module into Studio 5000 with or without the EDS
- When the link runs and then stops
- Questions asked
When a Rockwell scanner and a ctrlX CORE acting as the EtherNet/IP adapter stop exchanging I/O, work the fault in this order: scanner module status, IP path, assembly instances and byte counts, then the adapter task interval. The status reading tells you which fault you have. In the documented setup the adapter consumes 10 bytes and produces 15 bytes, the consumed assembly is instance 100 (16#64) with a configuration instance of 102 at length 0, and the scanner runs a 4 ms RPI that the adapter task must match exactly. Power-cycling the CORE can bring the link back, but it erases the evidence and repairs nothing.
Read the scanner status before you reset anything #
Open the Studio 5000 module properties and the controller diagnostics before touching either controller, and record:
- Whether the connection shows as established, faulted or idle, plus any module fault code shown with it.
- Whether input data on the scanner changes when the ctrlX application changes its produced variables, and the reverse.
- The ctrlX PLC application state and the adapter diagnostics at the same timestamp.
That record routes the rest of the work:
- Never establishes: go to the IP path, then the assembly fields.
- Established but the data is reversed, wrong or frozen: go to the instances, the direction and the task settings.
- Runs normally, then stops until a power cycle: capture both sides first, then work the stall checks.
The diagnostics variables generated under the adapter and module are not self-describing, and the reset bit is a control, not a diagnosis. Treat any bit you cannot tie to a documented meaning as unknown rather than assuming it healthy.
Confirm the IP path and the selected interface #
A correct I/O definition cannot connect across an unreachable network, so settle the path before touching assemblies:
- On the ctrlX Ethernet Device, check which network interface is selected. The example puts the adapter on the XF51 port of a CORE X3; selecting the engineering port instead sends the connection out the wrong jack.
- Compare the adapter IP address and subnet against the scanner module's configured address and the network plan.
- Check for a duplicate address on that subnet, then the cable and any switch between the two ports.
Change addresses only deliberately: pick the assigned, non-conflicting values from the plan and mirror the change in the controller configuration. Probing with a borrowed address to see what happens creates the duplicate-address fault you are hunting.
Match assembly instances, byte lengths and direction #
EtherNet/IP, ODVA's CIP network, locates cyclic I/O through the Assembly object (class 16#04, decimal 4): the scanner must be told the same instance numbers and byte counts the adapter declares, for each direction. The direction labels are where most of these faults live, because the two controllers name the same bytes from opposite ends:
| ctrlX adapter side | Studio 5000 scanner side | Documented example |
|---|---|---|
| Consumed data, received by the adapter | Output assembly | Instance 100 (16#64), 10 bytes |
| Produced data, sent by the adapter | Input assembly | Instance from the exported module definition, 15 bytes |
| Configuration | Config assembly | Instance 102, length 0 |
Check each row at both ends:
- The instance number for each direction is identical on both controllers.
- The size is a byte count, identical on both controllers; a tag count is not a byte count.
- The example documents instance 100 only for the consumed direction; read the produced-direction instance from the exported EDS or module definition instead of assuming the reverse is also 100.
With a healthy IP path, a connection that refuses to establish in this setup is an instance-number or byte-count mismatch until the module definition proves otherwise.
Working the numbers for 10 consumed and 15 produced bytes at a 4 ms RPI #
Inputs, each from the documented example:
- Consumed data: 10 bytes, received by the adapter, sent by the scanner.
- Produced data: 15 bytes, sent by the adapter, received by the scanner.
- Scanner RPI: 4 ms, set in the Studio 5000 module definition.
- Adapter task: ENIPAdapterIOTask, priority 20 to 39, interval equal to the RPI.
The rule in words: each direction's instance number and byte count must match at both ends, and the adapter's I/O task interval must equal the scanner's RPI.
The arithmetic:
- Cyclic payload per connection cycle: 10 B + 15 B = 25 B.
- Connection cycles: 1000 ms ÷ 4 ms = 250 per second.
- Cyclic payload rate: 250 × 25 B = 6250 B/s, before protocol overhead.
Decision: set the Studio 5000 RPI to 4 ms, set ENIPAdapterIOTask to a 4 ms interval at a priority between 20 and 39, and declare 10 and 15 bytes at both ends with instance 100 on the consumed side and configuration instance 102 at length 0. If your scanner uses a different RPI, the rule is interval = RPI, not 4 ms; on the Rockwell side the RPI lives in the module definition, the same field behind configuring RPI for consumed tags on a ControlLogix controller.
Keep the I/O mapping refreshing #
A configured connection can still leave you reading stale variables:
- In the Ethernet/IP Module I/O mapping, assign variables to the consumed and produced data.
- Set the mapping's update option to Enable 1, which refreshes mapped data even when no other task uses it; the example notes the bus cycle task for exactly that case.
- If the connection is up and the data is frozen, this update option and the task interval are the two suspects, in that order, before any restart.
Bring the module into Studio 5000 with or without the EDS #
The documented workflow on the ctrlX side, with ctrlX PLC App V2.6:
- Install the Ethernet/IP Adapter V4.5.1.0 package from Tools > Add-On Installer, closing ctrlX PLC Engineering first.
- Create the project on the matching template: ARM64 for a CORE X3, x64 for a CORE X5 or X7.
- Add the Ethernet Device with its interface and IP, then the Ethernet/IP Adapter beneath it, then the Ethernet/IP Module beneath that; configure the assemblies and the mapping.
- Export the EDS file from the Ethernet/IP Adapter.
On the Studio 5000 side:
- Import the EDS with the device description file installation tool.
- Add the module (Vendor: CODESYS), set its IP address and the RPI, and open the module definition to confirm the instances and sizes.
If you skip the EDS import, add the module as a generic Ethernet module and type the same fields by hand: the assembly instances, the byte counts per direction, and the configuration instance with its length. The checks are identical to the ones behind sizing a Generic Ethernet Module connection for WAGO 750 I/O. From ctrlX OS V3.6 a dedicated EtherNet/IP app configured in the ctrlX OS web interface takes over this role; the add-on workflow above is the documented path for the V2.6 generation.
When the link runs and then stops #
A link that dies until the CORE is power-cycled is a real fault even though the recovery is easy:
- Before restarting anything, capture the module status, the fault code, the connection state and the adapter diagnostics from both sides, with timestamps.
- A power cycle that restores the link is a recovery, not a fix; without the pre-restart capture, the outage taught you nothing.
- Then work the physical path: cable, switch ports, and a duplicate IP address on the adapter's subnet.
- For stalls tied to the adapter port, resolving EtherNet/IP stalls on a ctrlX CORE XF51 port works the port-specific case.
If any part of this link carries stop or interlock commands, move those functions to safety-rated I/O: a cyclic connection that drops and recovers on its own is not a safety function.
Questions asked
Can I use a generic Ethernet module in Studio 5000 instead of importing the EDS?
Yes. Add the module as a generic Ethernet module and type what the EDS would have carried: output assembly instance 100 for the adapter's consumed data, the input assembly instance from the exported module definition, configuration instance 102 with length 0, and the exact byte counts. You can disable matching or enter the vendor ID and product code manually. The connection matches instances and sizes, not the file they came from.
The connection is up but my data lands in the wrong tags. What did I get backwards?
Almost certainly the direction. Data the ctrlX consumes is the scanner's output assembly, and data the ctrlX produces is the scanner's input assembly; swapping the two instance numbers moves everything the wrong way. Write a free-running counter into the produced variable on one side and watch which scanner tag increments. That identifies the real direction in one test, then swap the instance entries.
The link runs for hours, then stops until I power-cycle the CORE. What should I record?
Before the restart, capture the module status, any fault code, the connection state and the adapter diagnostics from both controllers, with timestamps, and note how long it ran. A repeating run time points you toward a timer or watchdog, a random one toward the cabling; either way the captured status, not the recovery, identifies the fault. The reset bit will not restart a dead connection.
What interval and priority should the adapter I/O task use?
Set ENIPAdapterIOTask to the same interval as the scanner's RPI, 4 ms in the documented example, with a priority between 20 and 39, and change it whenever the RPI changes. Running the task faster than the RPI refreshes nothing that reaches the scanner sooner; running it slower leaves the mapped variables stale between wire updates. Set the mapping update option to Enable 1 as well.