Resolving S7-400 PROFIBUS DP Bus Faults at 19.2 kbps

Problem Overview

An S7 414-2DP master reports periodic PROFIBUS DP bus faults (1-2 events per 5-minute window) when running at 19.2 kbps against a CP 342-5 slave attached to a third-party S7-300 furnace controller. The same network is stable at 1.5 Mbps. The peripheral footprint is small: 24 bytes I/O for each of three distributed stations and 80 bytes of consistent SFC14/SFC15 data on the CP 342-5 link. Because the furnace PLC mandates 19.2 kbps, the speed cannot simply be raised.

This document walks through the root cause tree, the correct configuration of Siemens OLM (Optical Link Module) G11/G12 in STEP 7 HW-Config, diagnostic buffer analysis, and bus parameter verification required to clear the faults without changing the requested baud rate.

System Architecture and Topology

The reported master station is a SIMATIC S7 414-2DP (6ES7414-2XG04-0AB0 or comparable order number). The DP master port serves four slaves, with the following physical arrangement:

Two OLM variants are deployed: OLM G11-1300 (6GK1503-2CC00) for the first fiber link, and OLM G12-1300 (6GK1503-3CC00) for the second. The G11 has one optical port; the G12 has two. Both operate at 1300 nm and are designed for distances up to several kilometers on multimode fiber, well above the 300 m used here.

PROFIBUS DP Baud Rate, Cable Length, and Slot Time

PROFIBUS DP transmission speed vs. segment length per IEC 61158 / EN 50170 is the first thing to verify when low baud rate problems appear. The copper segment length is not the limiting factor at 19.2 kbps (1.2 km is supported), but the slot time T_sl and the min Tsdr values change significantly compared to 1.5 Mbps.

The slot time at 19.2 kbps is the maximum possible (1023 bit times). Each bit time at 19.2 kbps is 52.08 us, so T_sl is approximately 53.3 ms end-to-end. This huge timing window is precisely the source of the issue when the OLM is not declared to the master.

Root Cause Analysis

The dominant cause of "runs at 1.5 Mbps but not at 19.2 kbps" symptoms on a multi-segment PROFIBUS network with OLMs is that the OLM is not entered in HW-Config as a passive station. Without that entry, the master calculates bus timing (slot time, min/max Tsdr, quiet time, retry value) as if the entire network is one copper segment. When the OLM is then inserted in the physical path, the additional propagation delay and signal regeneration alter the bit-level timing window, and at 19.2 kbps the slot time calculation is large enough that the master's watchdog for slave response trips sporadically.

Secondary contributors worth ruling out:

1. Terminating resistors on the CP 342-5 subsegment. With a 50 m copper run and a CP 342-5 at the end, both ends must have termination ON, with all intermediate nodes OFF. The OLM copper ports have built-in selectable terminations.
2. Shielding/grounding of the PROFIBUS cable. At 19.2 kbps the signal rise time is long enough that 50 Hz/60 Hz hum from poor shield bonding appears as bit errors.
3. CP 342-5 operating mode mismatch. The CP must be configured as DP Slave in the S7-300 project, and the S7-300 side must be in RUN or RUN-P for the slave to be reachable. A stopped CPU on the furnace side still toggles the bus, which is consistent with the report that disconnecting the furnace PLC reduced but did not eliminate faults.
4. Diagnostic repeater presence. The diagnostic repeater (6GK1500-0AA10) actively monitors segment health and can be inserted between master and OLM to capture which segment drops the token.

Solution: Configure OLM in HW-Config

STEP 7 / SIMATIC Manager must know about the OLM so that the master's bus parameters include the additional repeater-related timing overhead. Procedure:

1. Open the STEP 7 project containing the S7 414-2DP station.
2. Launch HW-Config and select the PROFIBUS subnet on the DP master port.
3. Open Properties > Network Settings on the PROFIBUS subnet object.
4. Switch to the Options tab.
5. In the repeater / OLM section, add the OLM as a passive station. Siemens provides GSD-style catalog entries for the G11 (6GK1503-2CC00) and G12 (6GK1503-3CC00) under PROFIBUS DP > Other field devices > Repeater > OLM.
6. Assign the OLM the same PROFIBUS address it occupies physically (typically 0 is reserved for masters; OLMs commonly use 1-3 in a small network, but the OLM address is purely a placeholder for timing calculation, not for token passing).
7. Save and recompile (Station > Save and Compile).
8. Download the HW configuration to the S7 414-2DP CPU while it is in STOP.
9. Restart the CPU and observe the diagnostic buffer.

CP 342-5 Configuration on the Furnace Side

The CP 342-5 is configured as a DP slave in the S7-300 project that controls the furnace. Confirm:

The S7-300 user program on the furnace side must call SFC14 'DPRD_DAT' and SFC15 'DPWR_DAT' for the 80-byte consistent CP slots. The S7 414-2DP master program must do the same on the master side. Verify that the SFC calls are placed in OB1 (or a time-driven OB if non-cyclic transfer is required) and that the RECORD parameter points to a DB of sufficient length:

// S7-400 master side, reading 80 bytes from CP 342-5
CALL  "DPRD_DAT"
  LADDR  := W#16#0100      // I/O base address of CP 342-5 in master view
  RET_VAL:=#status         // Any non-zero value indicates DP error
  RECORD  :=P#DB100.DBX0.0 BYTE 80

// S7-400 master side, writing 80 bytes to CP 342-5
CALL  "DPWR_DAT"
  LADDR  := W#16#0100
  RECORD  :=P#DB110.DBX0.0 BYTE 80
  RET_VAL:=#status

Wrap both SFC calls in an error-handling block that reads RET_VAL; for low baud rates a small percentage of retries is normal, but the master should not drop the slave within 10 s. The SIMATIC S7-300 CP 342-5 manual specifies that SFC14 returns W#16#8xxx on slave failure and W#16#0xxx on success.

Diagnostic Buffer Analysis

Open the S7 414-2DP online with STEP 7 and navigate to CPU > Diagnostic Buffer. The PROFIBUS-related events arrive in this format:

Sort the buffer chronologically. If the failing slave ID varies between CP 342-5 (slave 4) and ET 200M (slave 2/3) and the OLM address appears in the failure window, the OLM is the suspect. If only the CP 342-5 fails, concentrate on the 50 m copper segment, termination, and the furnace-side CPU state.

For slave-level diagnostics from the master, use:

CALL  "DPNRM_DG"   // SFC13, read complete diagnostic from DP slave
  LADDR  :=W#16#0100   // Slave diagnostic address
  RET_VAL:=#ret        // Error code
  RECORD :=P#DB200.DBX0.0 BYTE 32  // Diagnostic buffer in master DB

Bus Parameter Verification

After the OLM is added in HW-Config, export the calculated bus parameters and confirm:

Fiber Segment and OLM Checks

For the OLM G11-1300 and G12-1300 fiber links, verify:

1. Optical power budget. On 1300 nm multimode fiber, 300 m should yield 1-3 dB attenuation. The OLM receiver threshold is typically -30 dBm; if a dirty connector pushes the link above -25 dBm, intermittent loss-of-light events will occur that are invisible to the user but visible to the DP master as a slave dropout.
2. Connector cleanliness. Clean both SC connectors with a proper fiber optic cleaning pen (e.g., Cletop S-250). One OLM fault often masquerades as ten.
3. Termination on the OLM copper side. Switch the OLM's termination ON only at the bus ends, OFF for any OLM in the middle of a copper chain. Each OLM has a slide switch for the terminating resistor and bus termination pull-up/pull-down.
4. OLM operating mode DIP switches. The G11/G12 ship with all DIP switches OFF (transparent mode). Do not enable segment monitoring mode unless a diagnostic repeater is part of the design.

Verification

After the HW-Config change, follow this verification protocol:

1. Download HW Config to the S7 414-2DP while in STOP.
2. Switch to RUN, clear the diagnostic buffer.
3. Run the bus at 19.2 kbps for a minimum of 4 hours continuously while exercising all 80 bytes of SFC14/SFC15 data at 100 ms cycle.
4. Confirm zero new 0xE0C0, 0xE0C1, or 0xE0C4 events in the diagnostic buffer.
5. Force a slave failure (unplug the CP 342-5 PROFIBUS connector) and confirm the master logs 0xE0C0 within 2 s and recovers within 10 s when the connector is reinserted.
6. Use a PROFIBUS cable tester or diagnostic repeater (6GK1500-0AA10) to capture a 24-hour bus trace at 19.2 kbps and verify no spur errors, retries, or illegal responses appear.
7. Compare the failure rate per 24 h to the original 1.5 Mbps baseline. They should be statistically indistinguishable (near zero).

Frequently Asked Questions