Troubleshooting
Why IO-Link hub outputs stay off despite a valid process image
Short answer
Setting output bits in the master's I/O mapping commands a hub point, but it cannot turn an input-configured point into an output. Configurable hubs ship with points defaulting to inputs, documented for the Balluff BNI004L and BNI007Z families, and the direction is assigned through the IO-Link parameter path at ISDU index 41h on those hubs. A vendor function block or the master's parameter tool writes it; after that write the same cyclic bits energize the points.
On this page
- Two paths decide whether the point energizes
- The cyclic image and the direction parameter are two channels
- Checklist before writing the direction parameter
- Energizing two points on a 16-point hub with a 4-byte output image
- Writing ISDU index 41h on the installed hub
- Limits that travel with this fault
- Questions asked
Your mapping is not the fault, and that is the point: setting output bits in the master's cyclic image commands a hub point, but it cannot convert an input-configured point into an output. Configurable IO-Link hubs ship with their points defaulting to inputs — documented for the Balluff BNI004L and BNI007Z families — and that direction is assigned through the IO-Link parameter path, ISDU index 41h on those hubs, not through the cyclic image. In the case traced here, two points on a 16-point hub showed no 24 VDC with 4 bytes IN / 4 bytes OUT mapped and the commanded bits toggling. A vendor-supplied Sysmac Studio function block wrote the direction parameter, and the same cyclic bits then energized both loads.
Two paths decide whether the point energizes #
Trace the installation as two connected paths. The cyclic command path runs from the controller through the fieldbus and the IO-Link master to the hub point. The actuator power path runs from the output supply through the hub's output circuitry to the load, with its own return. A valid command image cannot energize a point still configured as an input, and a correctly configured output cannot supply a load whose power feed is absent.
A hub that communicates proves only the communication supply, never the actuator supply. The same split between true logic and a dead point appears when outputs fail to energize despite true rung logic on directly networked I/O: the command is present and the terminal still carries nothing. Different network, identical discipline — prove both paths.
The cyclic image and the direction parameter are two channels #
| Channel | What it carries | What the controller sets | What it cannot do |
|---|---|---|---|
| Cyclic process data | Point states and commanded values, mapped as bytes in the I/O image | Which bits are on and how many bytes exchange | Change what a point is |
| ISDU parameter access | Device parameters, including point direction | A parameter write routed through the master to the port | Toggle a point in real time |
The two never substitute for each other. Enlarging the output image in the mapping dialog grew the command space; it said nothing about the hub's points. Process-image layout follows the connected device, not a master-side default: on the Beckhoff EL6224-0090 IO-Link master terminal, the process data depend on the connected IO-Link device, and the default is a 16-bit unsigned integer mapped at byte offset 0 from IO-Link channel 1.
Cyclic data is a request the device must be configured to act on, the same discipline as cyclic telegram data for a PROFINET servo drive. Which bits in the image command which point is fixed by the hub's documented process data; the master mapping only positions that image in the controller's address space.
Checklist before writing the direction parameter #
- Read the full order code off the master and hub labels before selecting documentation; the installed hub's manual governs its parameters, ratings and point numbering.
- Put the machine in a controlled commissioning state so that energizing an output cannot create unintended motion.
- Measure the hub's actuator supply at its terminals against the return for that circuit, separately from the communication supply, and check connector seating, cable continuity and the load return.
- Confirm the uplink is healthy and the IO-Link port reports the hub connected.
- Compare the mapped sizes, 4 bytes IN and 4 bytes OUT, with the hub's documented process data, and identify the bits that serve your two points.
- Toggle the commanded bit and confirm it changes in the mapped output image, with no other write or retained force overwriting it.
- Only then write the direction parameter and retest the point.
Energizing two points on a 16-point hub with a 4-byte output image #
Inputs, each with its source:
- Hub: 16-point configurable IO-Link hub; exact model not recorded in the fault report, so its own manual governs.
- Master: EtherCAT IO-Link master, installed hardware.
- Mapping: 4 bytes IN / 4 bytes OUT in the engineering tool.
- Command: 2 points set on from the controller; loads are magnets.
- Measured: no 24 VDC at either point, reported reading.
- Documented default: hub families BNI004L and BNI007Z default to inputs, with direction assigned at ISDU index 41h.
The rule in words: a point delivers voltage only when three things hold together — its direction is output, its cyclic bit is set, and the output supply is present. The master's mapping sets none of the first condition.
The arithmetic:
- Output image: 4 bytes × 8 bits per byte = 32 bits commanded by the controller.
- Points configured as outputs before any parameter write: 0 of 16.
- Image bits able to energize a load before the write: 0 of 32. Every commanded bit arrived at the hub with no output transistor behind it.
- Writing index 41h for the two points: 2 of 16 points become outputs; 14 remain inputs.
- The two commanded bits then stand behind live output circuits: 24 VDC measured at both points.
Decision: when a mapped output bit toggles and the point stays at 0 VDC, write the direction parameter first, before touching the mapping or the wiring. After direction is set, a point still at 0 VDC is a supply or wiring fault, not a data fault.
Writing ISDU index 41h on the installed hub #
Confirm the exact index, any subindex, data length, encoding and allowed values in the installed hub's documentation. Index 41h is documented for the named families; the hub in your cabinet answers to its own manual, not to a forum lead.
Three documented routes perform the write. The master's parameter interface: an EtherNet/IP master series documents a device ISDU interface that asks for the ISDU block index. The engineering environment's device catalog: a hub listed there has its ISDU parameters edited and downloaded with the configuration. Controller code: a function block requests the parameter service through the master, which performs the device access on the port — the route that worked in this case as a vendor-supplied block for Sysmac Studio. On another platform, a PLC-side ISDU write to an analog hub through an IO-Link master follows the same request path.
Limits that travel with this fault #
Output ratings belong to the specific device: the BNI006A EtherNet/IP master documentation lists IO-Link 1.1.3, eight Type A IO-Link ports, 16 PNP inputs and 16 PNP outputs with 2 A maximum output current. Those numbers describe that product, not the hardware in this fault, so the installed hub's label remains the authority for what its points may switch.
Installation conditions ride along: the BNI IOL-302-002-K006 user's guide sets a 20 m maximum standardized sensor cable length and requires blind caps on unused sockets to maintain IP67 — documented constraints for that hub, and none of them a direction setting. The order of attack stays fixed: identify both devices, prove the actuator supply, set direction through the parameter path, then command the bit.
Questions asked
Does adding output bytes to the master's I/O mapping turn hub points into outputs?
No. The mapping only sizes and positions the cyclic image the controller exchanges with the master; point direction is a device parameter reached through indexed service data. The distinction matters because IO-Link checks the content of a parameter write for consistency before accepting it, so a direction change either succeeds or returns an error rather than silently doing nothing.
Is the ISDU write different on an EtherCAT master than on an EtherNet/IP master?
No. The master carries out the indexed service on the selected IO-Link port regardless of its uplink; the controller or configuration tool only requests it. An EtherNet/IP master series documents a device ISDU interface that asks for the ISDU block index, and in the EtherCAT case traced here a vendor function block issued the equivalent write from the controller. Same service, different door.
What does the pin 2 setting in the master actually control?
The master port's connector, not the hub. On a port class A M12 connection, pin 2 is an optional second switching line on that port, and enabling it in the master changes what the connector pin does. No setting on the master connector reassigns the direction of a point inside the downstream hub, which is why the pin 2 bit had no effect on the dead outputs.
How do I confirm which cyclic bit commands which physical point?
Read the process-data definition in the hub's documentation or device description, which fixes the bit-to-point order; the master mapping only places that image in the controller's address space. Prove it by toggling a known input on the hub and watching which input-image bit changes, then command the corresponding output bit and measure at the point. If the bit moves and the point stays at 0 VDC, the point is still input-configured or its output supply is missing.