How-to
Bridging a PLC-5/15 DH+ network to a Kinetix 300 servo drive
Short answer
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.
On this page
- What the bridge replaces, and what it cannot
- Choose the gateway mode by what is documented
- Define the eight-word map before you buy
- Split the 32-bit position into two 16-bit words
- Handshake the pair so a torn read cannot pass
- Commission the bridge in stages
- Bridge it or migrate: the decision
- Questions asked
A Kinetix 300 and a PLC-5/15 can talk, but only through a gateway, and only as data. A unit such as ProSoft's EtherNet/IP to Remote I/O or DH+ gateway moves command and feedback words between the DH+ side and the drive's EtherNet/IP assemblies; it cannot generate pulses, count encoder edges, or plan a trajectory. The Kinetix 300 is an indexing servo drive that runs its own move profiles, so give it everything time-critical and let the PLC-5 sequence it. You are retiring a stepper controller card, a pulse-expander card, and an encoder-counter card, so ladder changes are unavoidable and the no-changes-to-PLC-code claim does not apply. Define the word map before any hardware is ordered.
What the bridge replaces, and what it cannot #
The retrofit deletes three functions: step generation, pulse expansion, and encoder counting. Assign each one before touching communications. The drive's stored index profiles take over step generation and the rate schedule; the position feedback in the drive's input assembly replaces the counter card; the pulse expander's job disappears outright, because there is no pulse train left to expand. The PLC keeps what it is good at: enable, index selection, move trigger, and fault reset, in machine sequence.
This division of labor works because of the drive you chose. A Kinetix 5500 on integrated motion is the opposite case, and why a Kinetix 5500 needs a motion-capable Logix controller covers how that family leans on the controller to plan moves. The Kinetix 300 does the opposite: its user manual, publication 2097-UM001, documents the communication settings that set the drive's IP address and the object parameters carried in the Input and Output assemblies.
Choose the gateway mode by what is documented #
| Gateway mode | PLC-5 side | Drive side | When to use it |
|---|---|---|---|
| DH+ bridge | Joins DH+ at up to 230 kbps; PLC-5 MSG reads and writes a gateway data table | Gateway originates the EtherNet/IP connection to the drive's Input and Output assemblies | The committed choice for a Kinetix 300: your own map against documented assemblies |
| Remote I/O drive replacement | Emulates a rack at quarter, half, or full addressing; auto-detects the Remote I/O baud rate; data moves as block transfers of up to 64 words | Documented for up to 4 EtherNet/IP drives and all PowerFlex drives | Only with the vendor's written confirmation that the Kinetix 300 is supported |
Plan the project on the DH+ bridge. The drive-replacement mode converts applications that already had a Remote I/O drive on the cable; your stepper cards lived in the PLC chassis, so there is no Remote I/O drive to replace, and a PowerFlex data map is not a servo map. If you still want that mode, the gating constraint is configured rack and group addresses on the scanner channel, not the empty chassis slots, so read the scanner configuration and confirm a free address before ordering anything.
Define the eight-word map before you buy #
| Word | Direction | Contents |
|---|---|---|
| O1 | PLC to drive | Command bits: enable, move trigger, stop, fault reset, command-sequence toggle |
| O2 | PLC to drive | Index or rate selection |
| O3 | PLC to drive | Target position, low word, counts |
| O4 | PLC to drive | Target position, high word, counts |
| I1 | drive to PLC | Status bits: ready, moving, move complete, faulted, acknowledge toggle |
| I2 | drive to PLC | Drive fault code |
| I3 | drive to PLC | Actual position, low word, counts |
| I4 | drive to PLC | Actual position, high word, counts |
The exact instances, word positions, and scaling inside the drive's assemblies come from the drive manual for the firmware you receive; map each of the eight words to its documented place and record the mapping in the project. Eight words ride in one 64-word block transfer or one short DH+ message with eight times the headroom, so the map can grow without a second transfer or a different gateway.
Split the 32-bit position into two 16-bit words #
Position counts outrun a 16-bit word at 65,536 counts, which is why the position crosses as a 32-bit value in two words. The split and the rebuild are plain integer arithmetic:
- High word = count ÷ 65,536, integer result.
- Low word = count − (high word × 65,536).
- Rebuild: count = (high word × 65,536) + low word.
Worked example, commanded position 1,234,567 counts: 1,234,567 ÷ 65,536 gives an integer result of 18, so the high word is 18. Then 18 × 65,536 = 1,179,648, and 1,234,567 − 1,179,648 = 54,919, so the low word is 54,919. Rebuild check: 18 × 65,536 + 54,919 = 1,234,567 counts.
Treat the pair as unsigned everywhere. PLC-5 integer words are signed, so a low word of 54,919 displays as −10,617, because 54,919 − 65,536 = −10,617; sign-extending it or comparing it as signed corrupts the position. Prove the path with fixed test values: 0, 1, 65,535, 65,536, 65,538, and the machine's largest commanded count. The value 65,538 must read high word 1, low word 2; if the halves arrive swapped, the drive sees 2 × 65,536 + 1 = 131,073, and the swap is caught before any move is made.
Handshake the pair so a torn read cannot pass #
The gateway forwards the map on its own schedule, not on the PLC scan. If it forwards the map between the write of the low word and the write of the high word, the far side accepts a mismatched pair, and the value is wrong by a multiple of 65,536 counts. Close that hole with one bit in each direction:
- The PLC writes both position words, then toggles the command-sequence bit in O1.
- The drive side copies the pair only when that bit changes, then sets the acknowledge bit in I1.
- The PLC holds its command until the acknowledge arrives, drops it, and waits for the acknowledge to clear before issuing the next one.
Hold every command as a state, never a one-scan pulse. The PLC scan runs far faster than the DH+ hop, the gateway, and the EtherNet/IP connection, and a pulse can be over before the far side ever sees it.
Commission the bridge in stages #
- Set the drive's IP address from its drive tools, configure the gateway on both networks, and confirm both sides report healthy, with no motion command issued.
- Map one command word and one status word. Toggle single bits and confirm no adjacent bit moves.
- Push the fixed test values through the position path and verify word order, signedness, and scaling at the drive, then read them back into the PLC data table.
- Run the full handshake: change one half of the position pair first and confirm the far side rejects the incomplete value.
- Prove stop, inhibit, and fault reset before the first enabled move, with the axis unloaded or guarded, because this is the first motion under the new command path.
- Command one low-risk index and compare the commanded count, the drive's own feedback, and the value returned over the bridge; all three must match to the count.
- Exercise every rate change the machine sequence uses, and record the command-to-acknowledge time on a logging station: an industrial mini PC in the cabinet runs the drive tools and those timing logs through the whole acceptance run.
Read the timing against the machine, not a rule of thumb. A rate change that must land within one PLC scan will not survive this path; a rate change per product or per batch, landing in tens of milliseconds end to end, will. Measure the command-to-acknowledge time once and let that number decide.
Bridge it or migrate: the decision #
Bridge when the machine's motion fits drive-local profiles: fixed indexes, selectable rates, position feedback for the sequence, and no per-scan trajectory. Migrate when the PLC-5 must compute the profile, stream setpoints every scan, or coordinate this axis with another to millisecond order; no gateway closes that gap, and the honest fix is a controller that speaks EtherNet/IP natively. On a current platform the same start, stop, and status job is one EtherNet/IP connection with no bridge at all, as start-and-stop a PowerFlex 700 over EtherNet/IP from CompactLogix shows. If the decision lands on migration, the planning steps in migrating a MicroLogix 1500 machine to a current PLC apply line for line: inventory the I/O, freeze the sequence, cut over once.
Questions asked
Does the gateway's no-changes-to-PLC-code claim apply to this retrofit?
No. That claim belongs to the drive-replacement mode, which is documented for up to four EtherNet/IP drives and all PowerFlex drives, and its premise is a drive that already hung on Remote I/O as an adapter. Your stepper controller and encoder counter sat in the PLC chassis, so there is no Remote I/O drive to replace, and the pulse and counter logic is rewritten as word commands either way. Buy the gateway for the data path, not the code savings.
How do I split and rebuild a 32-bit position in 16-bit words?
Divide the count by 65,536 and keep the integer result as the high word; the remainder is the low word. For 1,234,567 counts the high word is 18 and the low word is 54,919, because 18 times 65,536 is 1,179,648 and the remainder is 54,919. Rebuild with high times 65,536 plus low. Keep both halves unsigned: a low word of 54,919 shows as minus 10,617 in a signed 16-bit integer, and any sign extension corrupts the count.
Why would the reported position jump by exactly 65,536 counts?
That signature is a torn read: one half of the pair updated while the other half was being copied. If only the high word advances, the rebuilt value moves by 65,536; if only the low word rolls over, the error equals the difference between the old and new low words. Write both words, toggle a data-valid bit, and let the receiving side copy the pair only on a change of that bit, acknowledging after the copy so stale pairs can never pass.
Which gateway mode should I plan on for the Kinetix 300?
Plan on the DH+ bridge mode: the gateway joins DH+ as a node, so confirm a free node address, the PLC-5 reaches its data table with MSG instructions, and the gateway originates the EtherNet/IP connection to the drive's documented assemblies. Use the Remote I/O drive mode only if the vendor confirms Kinetix 300 support in writing, and only after the scanner configuration shows a free rack and group address. Chassis space proves nothing about logical addresses.
When is a bridge the wrong answer altogether?
When the PLC must feed the drive a new setpoint every scan, or coordinate this axis with another in milliseconds. The path is PLC scan, DH+ transfer, gateway processing, EtherNet/IP connection, drive: expect tens of milliseconds, not one scan, and measure it during acceptance. If the machine's rate changes can live as selectable index profiles or a velocity override inside the drive, bridge it; if the PLC-5 computes the profile live, migrate the controller instead.