IMDHMIControls LibraryShop main store

How-to

Configure a PowerFlex 525 on EtherNet/IP from an Omron NJ301

Industrial networks Published 7 min read

Short answer

The tag data link fails after a good ping because the NJ, as the EtherNet/IP originator, must send the drive's chosen output assembly plus a 32-bit Run/Idle header: error 0x0127 means the output connection size is wrong. Import the revision-matched EDS, select the target assemblies it lists, create Input and Output network variables of exactly those sizes with the header included, and register them on the Tag Set tab.

On this page
  1. How do I set up the tag data link in Sysmac Studio?
  2. Which assembly instances do I pick for the PowerFlex 525?
  3. Why does ping work but the tag data link still fails?
  4. What does connection error 0x0127 mean?
  5. How do I add the Run/Idle header to the output variable?
  6. How big should the originator variables be?
  7. How do I start, stop and set speed once the link is up?
  8. Questions asked

Your NJ301-1200 pings the drive and the EDS imports, yet the tag data link will not open: the NJ, as the EtherNet/IP originator, must supply what the drive's connection definition asks for — originator variables sized exactly to the chosen assemblies, with the 32-bit Run/Idle header in front of the output data. When Sysmac Studio reports 0x0127, the output connection size from controller to drive is wrong, and on this pairing the missing header is the usual cause. Load the revision-matched EDS, take the target assemblies from that file, build the two network variables to those sizes, register them on the Tag Set tab, and the rest is drive-side wiring of words and bits.

Create two global variables with their Network Publish attribute set to Output and Input, register them on the Tag Set tab with Registration all, and point the connection's Target Variable fields at the assemblies your imported EDS lists — that is the entire mechanism.

The target variable names the drive-side assembly the connection uses; the originator variable is the controller-side variable that carries the bytes. Neither field is the name of your start bit or your speed tag. Work in this order:

  1. Import the EDS for your exact drive revision into Sysmac Studio and create the slave device from it.
  2. On the NJ's EtherNet/IP port, create a tag data link to the drive's IP address using the connection definition the EDS provides.
  3. Put the cursor in a Target Variable field and press Ctrl+Space. The list that appears is the set of assemblies that EDS defines; take one input and one output from it.
  4. Create two global originator variables, one to send and one to receive.
  5. Set Network Publish to Output on the variable you send and Input on the variable you receive, judged from the controller's side.
  6. Open the Tag Set tab and press Registration all, which turns those variables into assignable tags.
  7. Assign them to the connection, compare the compiled byte sizes with the EDS connection entry, then transfer and go online.

For the first attempt, two byte arrays are enough: the controller verifies that sizes match, not how you interpret the bytes, so layout can wait until the connection is proven.

Which assembly instances do I pick for the PowerFlex 525? #

Use the pair your imported EDS offers when you press Ctrl+Space in the Target Variable field — numbers like 1 and 2, or 100 and 150, circulate for this drive, but the first was a guess, the second a generic example, and the EDS in your project decides.

The connection entry in the EDS carries the input and output sizes along with the instance numbers, so a matched pair is chosen as a unit. An instance that connects but returns a different layout produces data that looks like a fault: feedback sitting in the status word, garbage where a reference should be. Read the drive's embedded EtherNet/IP adapter manual alongside the EDS and confirm the instances, their byte counts and their word maps before you trust a single value.

Ping proves the IP path; it says nothing about the cyclic connection, which is EtherNet/IP I/O running over UDP — so a drive that answers ping still rejects the link on EDS revision, assembly size or header grounds.

Work outward from the simplest arrangement:

  • Verify the drive's IP address and subnet against the controller's port settings and the address used in the tag data link.
  • Ping each device from your engineering station, then note that this exercises the station's path, not necessarily the controller-to-drive path.
  • If a managed switch sits between them, retest on an unmanaged switch or a direct cable; when the connection is multicast, IGMP snooping needs a querier or the cyclic traffic goes nowhere.
  • Compare the drive's revision with the EDS. One reported case had a revision 6 drive on an EDS written for revisions 1 through 5, and the link failed until the revision-matched file replaced it. Take EDS files from the manufacturer or its distributors, never modified versions.

Explicit messaging is the dividing test: explicit messages run over TCP port 44818 and the cyclic link over UDP port 2222, so explicit success with cyclic failure points at the connection definition, not the cable.

What does connection error 0x0127 mean? #

It is a connection-size error on the output path — the bytes the controller sends to the device — so fix the controller's output variable first; enlarging the input variable is the wrong response to this code.

In Sysmac terms, Output is the direction from the originator, your NJ, to the target, the drive. The drive's connection definition requires a 32-bit Run/Idle header ahead of the output assembly, so an originator variable sized to the bare assembly falls short by the header and the connection is refused. Add the header accommodation before touching any control-word logic. The originator-side discipline is the same whatever target you own — the same first questions come up in fixing a TwinCAT EtherNet/IP slave connection from an Omron PLC.

How do I add the Run/Idle header to the output variable? #

Declare a DINT member — holding 1, written 16#00000001 in Sysmac literal syntax — as the first member of the output structure, and keep it at 1 whenever your logic should command the drive; at 0, every output byte is treated as idle.

Bit 0 of the header's first byte is the run flag, and the drive consumes the header itself before your logic command word ever arrives. With the flag clear, the drive ignores your command and reference and follows its configured idle action, which makes the flag a clean per-connection enable. Note the direction: this header exists only on data you send. The reverse arrangement, an Omron controller as the target of someone else's scanner, is the case where you oversize the input instead. Toggle the run flag for the first time with the motor isolated or the machine locked out, because a correct header plus a set start bit starts the motor.

How big should the originator variables be? #

Output variable: the 32-bit header plus the output assembly's documented size; input variable: the input assembly's size exactly — as a worked example, a two-word output assembly needs 4 + 2 + 2 = 8 bytes and a two-word input assembly needs 2 + 2 = 4 bytes.

Say the connection entry you selected defines an output assembly of two words, a logic command word and a reference word, and an input assembly of two words, a status word and a feedback word. The sizes then work out as:

  • Output structure: header DINT 4 + command WORD 2 + reference WORD 2 = 8 bytes total.
  • Input structure: status WORD 2 + feedback WORD 2 = 4 total.
  • The drive sends no header back, so the input variable is nothing but the assembly.

Read the sizes from the connection entry in your own EDS rather than from this example — the arithmetic is the method, and the assembly sizes are whatever your EDS documents. Two byte arrays of the example sizes open the same connection for a first test. If 0x0127 persists after the header is added, the sizes still disagree with the EDS: recheck the compiled variable sizes and confirm the registered variables are the ones actually assigned to the connection.

Point the drive's command source and reference source at its EtherNet/IP port, set the run flag, write the start bit into the logic command word, and put the speed command in the reference word — the drive answers in its status and feedback words.

No predefined PowerFlex structure ships with Sysmac Studio, so build one in the data type editor from the adapter manual's word map; bit members inside the structure replace word-to-bool conversion functions. The same manual fixes the reference word's scaling against the drive's maximum frequency, so set that before judging speed tracking. The word-and-bit handling on the drive side is the discipline used to start and stop a PowerFlex 700 over EtherNet/IP from a CompactLogix; only the tool changes.

One rule does change: the tag data link makes your NJ the single owner of the drive's network commands. If two controllers must command one drive, that is the architecture of controlling a PowerFlex 525 from two PLC controllers, not a second tag data link. Prove the header works by clearing the run flag on a connected, commanding link: the drive must drop to its idle action immediately. If it keeps running, the flag is not in the bytes the drive reads — go back to the structure layout.

Questions asked

Does the input variable from the drive need extra bytes too?

No, not in this configuration. The header travels on the data the originator sends, so your output variable carries it and the drive's input assembly comes back exactly as documented. The advice to oversize an input by the header size belongs to the reverse arrangement, where an Omron controller is the target of a scanner that sends the header and must therefore receive it.

Can I test the drive with an explicit message before configuring the cyclic link?

Yes, and it separates cable problems from configuration problems. An explicit read of the drive's identity data over TCP confirms the IP path, the open port and the device revision without opening a tag data link. If explicit reads succeed and the cyclic connection still fails, keep working on EDS revision, assembly instances and the header rather than replacing hardware.

The link is healthy but the status word reads nonsense — what do I check?

Work through the received bytes against the adapter manual's word map one member at a time: word order first, then signedness, then scaling. A connection that stays up with wrong-looking data almost always means the layout or the assembly instance is wrong, because the controller only verifies size. Compare raw bytes in a byte-array variable before trusting any converted value.

What happens to the drive if the cable is pulled or the NJ stops?

Two different protections apply. With the run flag at 0, the drive is idle and follows its configured idle action. With the connection lost entirely, the adapter times out and executes its communication-loss action. Configure and test both actions on the bench — coast, ramp or fault — before the machine runs, because a drive that holds its last reference through a fault is a hazard.

Does an NX controller need the same Run/Idle header treatment as the NJ?

As the scanner, yes: whichever Omron controller originates the connection must supply the header the drive's connection definition calls for, and the output-side fix is identical. The firmware-specific cases reported — an NJ at any firmware, or an NX below firmware 1.21 — concern the controller acting as a target to a Rockwell scanner, where the header arrives instead of being sent.

Keep reading

  • Troubleshooting

    Fixing a TwinCAT EtherNet/IP slave connection from an Omron PLC

    The CIP1 fault is a Forward Open mismatch: each Omron direction must carry the full serialized TwinCAT assembly, not the process variables alone. With 2 bytes in and 2 bytes out, the 4-byte ConnCtrl on outputs and ConnState on inputs make each direction 6 bytes, and the instances must follow producer and consumer roles. If a capture shows the request and a rejection, read the sizes and instances the originator actually serialized.

    Industrial networks

  • How-to

    Control a PowerFlex 525 drive from two PLC controllers

    Yes — add the PowerFlex 525 to the Logix controller's Ethernet tree and its command tags appear, but a cyclic drive connection has exactly one owner. Give the Logix controller that ownership, read the MicroLogix machine's request into it, and let one rung select the command the drive follows. Never let two controllers write the speed reference.

    Drives and motion

  • How-to

    Set up Micro850E implicit messaging to a Kinetix 5100 servo

    The implicit connection to a Kinetix 5100 is created in the Micro850E controller's EtherNet/IP module configuration under Ethernet > Modules, not by a UDFB. The drive is added as a natively supported motion slave, the UDFBs then run above the active connection producing commands and decoding status, and the same controllers accept up to eight generic EtherNet/IP slave devices.

    Industrial networks