IMDHMIControls LibraryShop main store

Selection guide

Selecting an EtherNet/IP to EtherCAT gateway for drives

Industrial networks Published 7 min read

Short answer

A ControlLogix scanner driving Lenze i700 EtherCAT drives needs a gateway that is an EtherNet/IP adapter toward the PLC and an EtherCAT MainDevice toward the drives. The documented part with that direction is the ABC3107-A; the reversed ABC4012-A cannot master the drive group. Confirm the ordering code and the PDO-mapping workflow, then prove the cyclic data at the tags before release.

On this page
  1. Match two roles, not one protocol
  2. The ordering code, and the code that is not on the page
  3. The check a green link cannot do: the cyclic mapping
  4. Worked example: sizing the cyclic exchange for a 1769-L33ER and three i700 axes
  5. Acceptance test before you release the design
  6. The decision
  7. Questions asked

Control a group of Lenze i700 drives from a ControlLogix scanner through one protocol gateway, and that box must hold two roles at once: EtherNet/IP adapter on the PLC side, EtherCAT MainDevice — the master — on the drive side. The documented product built for exactly that direction is the Anybus Communicator ABC3107-A, whose job is to put EtherCAT devices behind an EtherNet/IP control system. Its reversed-direction counterpart, the ABC4012-A, joins the EtherCAT network as a slave and cannot run the drive group. Picking the code is half the decision. The other half is cyclic data: a reported installation had this gateway talking to a 1769-L33ER and an Omron EtherCAT servo and still found no way to edit the PDO mapping from the Web UI. Link-up proves transport, not data. Here is the part to order and the tests that finish the job.

Match two roles, not one protocol #

A protocol converter is two network interfaces with a block of mapped memory between them, and each interface takes a role on its network. Both roles have to match your architecture; one right role paired with one wrong role still leaves you a part that cannot run the machine.

On the EtherNet/IP side your ControlLogix is the scanner — the originator that opens the connection and sets its timing — so the gateway must present the adapter role and accept that connection. This half is familiar if you have wired Logix to remote I/O before, and when the adapter shows connected but the mapped tags never move, the fault hunting is the same as for any stuck EtherNet/IP I/O link between an adapter and a Rockwell scanner.

On the EtherCAT side the drives are slaves, and an EtherCAT network has exactly one master. The gateway must be that MainDevice: it enumerates the axes, owns the bus timing and runs the cyclic exchange. A converter that joins as a slave talks into somebody else's EtherCAT network — in your architecture there is no other master, so the drives simply never run.

The selection rule is mechanical. Read the direction in the product name, match both halves, and only then compare codes:

Gateway part (documented direction) EtherCAT-side role EtherNet/IP-side role ControlLogix scanner plus i700 slaves
ABC3107-A, EtherCAT MainDevice to EtherNet/IP adapter MainDevice (master) Adapter Yes — the match
ABC4012-A, EtherNet/IP adapter to EtherCAT slave Slave Adapter No — nothing masters the drives

The ordering code, and the code that is not on the page #

The manufacturer's product page for this direction documents item number ABC3107-A, the Anybus Communicator EtherCAT MainDevice to EtherNet/IP adapter, described as connecting EtherCAT devices to EtherNet/IP control systems. That description is your architecture in one line, and it is the acceptance criterion for the order: both roles, named.

A second code, ABC3017-A, is quoted for this same product in some ordering and support replies. It differs from the documented item by one digit, and on a protocol converter one digit usually separates two different network pairs. The reported working installation ran ABC3107-A. Treat ABC3017-A as a different part until the manufacturer confirms otherwise in writing.

Decision: order the code whose printed description names both roles, EtherCAT MainDevice and EtherNet/IP adapter. If a quotation arrives with any other code, it carries written confirmation of that role pair or the design does not proceed on it.

The field evidence behind this article is one reported installation: an ABC3107-A, a 1769-L33ER CompactLogix as EtherNet/IP scanner, and an Omron R88D-1SN30F-ECT EtherCAT servo established communication. Read it for what it is — a reported symptom that supports the protocol direction, not a verified compatibility result for the Lenze i700 or for your PDO set.

The same report carries the finding that matters more: its Web UI offered no option to edit the PDO mapping. Take that as a design constraint until the manufacturer documents the supported workflow, because the mapping is where an EtherCAT master earns its keep. EtherCAT slaves advertise the cyclic objects they support; the master's configuration decides which of them are exchanged, in what order and at what offsets.

An i700 on EtherCAT runs the CiA 402 drive profile, so the cyclic channel carries the control word, the status word and the target and actual values, and the whole shutdown, switch-on, enable-operation sequence — control word 0x06, then 0x07, then 0x0F — travels inside those bytes. A mapping that is off by an offset leaves the network green while your enable command lands in a tag that does nothing. The state-machine mechanics are common to every CiA 402 axis; running CiA 402 CANopen drives from an S7-1500 through a gateway walks the same profile through a comparable two-network architecture.

Decision: if your axes can run on the drive's default cyclic set, order the part and prove the data at the tags. If you need a custom cyclic set, the documented mapping workflow is a purchase prerequisite — know which tool or import method assigns the set, in writing, before the purchase order goes out.

Worked example: sizing the cyclic exchange for a 1769-L33ER and three i700 axes #

The same arithmetic sizes any candidate part before money goes on an order.

Inputs, and where each comes from:

  • Scanner: one 1769-L33ER, EtherNet/IP, acting as scanner — the PLC model from the reported working installation; the role is what any ControlLogix takes toward an adapter.
  • Drives: three Lenze i700 axes, EtherCAT slaves — the drive family is the application's; three is an illustrative small group, so substitute your own count.
  • Cyclic set per axis, velocity control under CiA 402: control word (object 0x6040) 2 bytes out, target velocity (0x60FF) 4 bytes out, status word (0x6041) 2 bytes in, actual velocity (0x606C) 4 bytes in — the profile's standard object widths.

Rule in words: add the widths per axis in each direction, multiply by the axis count, and compare each direction's total against the process-data size documented for the ordering code you buy.

Arithmetic: per axis, outputs are 2 + 4 = 6 bytes and inputs are 2 + 4 = 6 bytes, so each direction is 6 × 3 = 18 bytes per cycle. Through the gateway that is 18 + 18 = 36 bytes per cycle.

Result: 18 bytes out and 18 bytes in, 36 bytes per cycle.

Decision: at 18 bytes each way, capacity does not disqualify this design — eighteen words of fixed drive data in each direction is the everyday workload of a drive-class converter, so the binding check is the mapping, not the byte count. Turn the numbers into acceptance criteria: the adapter assembly must offer at least 18 bytes in each direction, and the control word and target velocity must sit at fixed, documented offsets so the ControlLogix tags hold what you wrote. Scale for your real group — bytes per axis times axes — and apply the threshold: if either direction exceeds the process-data size documented for the part you priced, the direction is right and that part is too small, so step up a size, never down a role.

Acceptance test before you release the design #

  1. Both links live. EtherNet/IP connection established from the scanner, drives enumerated by the gateway on the EtherCAT side. Note what this proves — transport, nothing about content.
  2. First cyclic checks with the drives safe. Run them with the drives disabled or at a limited safe speed, because a wrong mapping can put a command value in a slot that starts motion you did not order.
  3. Handshake on the tags. Write the enable sequence — 0x06, 0x07, 0x0F — from a ControlLogix tag and watch the mapped status word follow; then command a small speed reference and read actual velocity back. A value that stays frozen while the network shows green is a mapping fault, not a link fault.
  4. Every word accounted for. Each offset must land in the tag you intended, checked word by word — the same discipline as verifying a scanner map, as in mapping a 1747-SDN scanner for DeviceNet drive control.
  5. Loss behavior defined. Decide what the PLC should do when the drive network drops, and confirm before unattended running that the gateway signals the loss in a bit your logic watches.

The decision #

  • Architecture is an EtherNet/IP scanner plus EtherCAT drive slaves: order the ABC3107-A, the documented EtherCAT MainDevice to EtherNet/IP adapter, and never the ABC4012-A for this architecture.
  • Quotation shows a different code: written confirmation of the MainDevice-plus-adapter pair, or no order on it.
  • Axes run on the default cyclic set: order, then pass the five checks above on your own drives.
  • Custom cyclic set required: the documented mapping workflow comes before the purchase order, then the same checks.
  • Required bytes exceed the documented process-data size: same direction, larger part.

Questions asked

Is ABC3017-A the same gateway as ABC3107-A?

Treat them as different parts until the manufacturer says otherwise. The product page documenting this direction, EtherCAT MainDevice to EtherNet/IP adapter, lists item number ABC3107-A, and that page's printed functional description is the acceptance criterion: both roles named. ABC3017-A differs by one digit and is quoted in some ordering and support replies, but it is not the code on the documented page. If a quotation arrives carrying ABC3017-A, require written confirmation that it delivers the EtherCAT-master plus EtherNet/IP-adapter pair before you release the design.

Why can't an EtherNet/IP adapter to EtherCAT slave gateway run the i700 drives?

On that product the gateway enters the EtherCAT network as a slave, and a network of slave drives has exactly one master. With no MainDevice in the gateway, nothing enumerates the axes, assigns the cyclic exchange or owns the bus timing, so the i700 group never runs. The reversed part, documented as the ABC4012-A, serves the opposite architecture: an existing EtherCAT control system masters the network and an EtherNet/IP system exchanges data with it through the gateway as an adapter.

Can the ControlLogix master EtherCAT without a gateway?

Not on its built-in ports. A Logix CPU's embedded and standard communication ports speak EtherNet/IP, and EtherCAT master function is not part of that firmware, so the scanner you configure toward the gateway cannot also enumerate an EtherCAT bus. The alternatives are an in-chassis EtherCAT master module for the Logix chassis or a controller family with EtherCAT built in. Both change the architecture; the gateway keeps the existing PLC, chassis and control program in place and only adds the conversion.

How quickly does drive status reach the PLC through the gateway?

The end-to-end delay is the sum of the two update periods, one per network. The EtherNet/IP side updates at the RPI you configure in the Logix scanner connection; the EtherCAT side updates at the cycle time the gateway runs the drives at. The gateway swaps its mapped memory between the two every cycle. For drive control, both periods normally land somewhere between a millisecond and a few tens of milliseconds, but the real numbers are your RPI and cycle time settings, not a property of the part number.

Keep reading

  • How-to

    Bridging a PLC-5/15 DH+ network to a Kinetix 300 servo drive

    Yes, a DH+-to-EtherNet/IP gateway exchanges command and feedback words between a PLC-5/15 and a Kinetix 300, but it only moves data: the drive must execute the motion profile locally. Replace the stepper controller, pulse expander, and encoder-counter functions with an eight-word map that carries a 32-bit position as two 16-bit words behind a data-valid handshake. If the machine needs trajectory points streamed per scan, migrate the controller instead of bridging it.

    Drives and motion

  • How-to

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

    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.

    Industrial networks

  • How-to

    Connect an Omron PLC to Smart Function Kit Pressing

    Use EtherNet/IP between the Omron NX/NJ/NY PLC and the Smart Function Kit Pressing: Omron's EtherCAT runs CoE, the kit's runs SoE, and the two profiles cannot exchange data. Load the Sysmac Studio example project and drive the kit with the SFK_Command, SFK_Fast_Variable_Access, SFK_SerialNumberConverter and SFK_GetNotification blocks.

    Industrial networks