Troubleshooting
Why an Industrial PC Pings an S7-1500 but Runtime Stays Disconnected
Short answer
Ping proves the IP path only. When Runtime shows the S7-1500 disconnected with an unsafe-connection error, make the IE General interface, the Windows adapter and the Runtime connection resolve to the same NIC, make the station name and PROFINET device name match, then trust the certificate in both directions.
On this page
Ping success and a busy packet capture prove the IP path and nothing above it. Runtime stays disconnected with the unsafe-connection message because three PC-side settings still disagree: the PC station's IE General interface is not bound to the Windows adapter that carries the PLC subnet, the station's name and PROFINET device name do not match the project, and the security negotiation has no trust path. Fix them in that order: rebind the interface, make the identity match, then trust the certificate in both directions with correct clocks. While another client already reads data from that PLC, the switch and the controller configuration are not the fault.
What ping and the capture prove #
A Runtime link is built in layers: Ethernet transport, IP addressing, an application session, an authorized session, then tag data. Ping tests the second layer only. A packet capture shows whatever travels the wire, which includes broadcast and PROFINET traffic plus the sessions of every other client on that subnet, so a busy capture proves the cable, the switch and the capture adapter, never that your IPC holds a session. Establishing and authorizing the session is the Runtime's job, and that is where this fault sits. The layer split is runtime-independent: the case traced in Fix CoDeSys communication error 0 when ping succeeds fails the same way.
Check the adapter binding first #
An industrial PC with several network adapters can ping a PLC through one interface while its communication software is assigned to another. Setting a static address inside the PLC subnet on one card does not assign the PC station to that card.
- Record every adapter with its address and subnet mask, and mark the one with a link into the PLC subnet.
- Open the PC station's local configuration, select the IE General component and read which adapter it is assigned to. That assignment is a separate setting from the address.
- Rebind IE General to the adapter that owns the PLC-subnet address, or move the static address onto the adapter IE General already uses. One card, one role.
- Retest with a capture filtered to the IPC's own address and the PLC address while you retry the connection. Frames present with the connection still failing move the fault to authorization; no frames at all confirms the binding.
If the ping itself fails with media disconnected, the fault is the Windows adapter or its driver rather than the station configuration, and the adapter-side checks in troubleshooting media disconnected on an industrial PC NIC are the ones that apply.
Match the station name, device name and slots #
Installing SIMATIC NET supplies the communication software; it does not create the Runtime connection. The PC station is the configured set of interfaces and application components the IPC hosts, and the project and the IPC must agree on every identity in it.
- The station name lives in the IPC's local station configuration. The Windows computer name is not the station name, so renaming Windows changes nothing here. Set the local station name to the PC station name used in the project.
- The interface's PROFINET device name must equal the configured name of the IE General interface. PROFINET identifies a device by that name, so a mismatch leaves the station outside the configured network even while every PLC on the subnet is visible to the scanning tool.
- The OPC, Application and IE General components must occupy the same index positions in the local station as in the project. One slot off and the connection never opens.
- An XDB file is compiled station configuration for the workflows whose documentation calls for it. Importing one is not a remedy for an unsafe-connection message; use it only where the installed version's documented method says so.
Clear the authorization path #
The message text, connection unsafe with manual authorization impossible and error code 50, marks a security negotiation that failed before any tag data flowed. The number belongs to the component that raised it, so its documented meaning is in that Runtime version's own help; in this configuration the causes behind such a failure are an untrusted certificate, a certificate outside its validity window, or a security policy the two ends do not share.
OPC UA is standardized as IEC 62541, and its rule is short: each application identifies itself with an X.509 certificate, and a secure channel opens only when each side holds the peer's certificate in its trust list and that certificate is inside its validity period. That gives three concrete acts:
- Trust the S7-1500 server certificate in the Runtime client's trust store.
- Have the PLC's server settings accept the Runtime client's certificate. With that trust in place, the server can accept the client without user credentials.
- Correct the clocks. The documented responses to a time error during certificate checking are the status codes BadSecurityChecksFailed and BadCertificateTimeInvalid, and a PC that lost its backup battery fails validity after every power loss until the clock is set.
Set the same security policy on both ends, the strongest one they share, rather than widening what the PLC accepts. Do not clear the message by disabling security on a live controller or accepting a certificate you have not verified: that removes the controller's protection instead of fixing trust.
Worked example: two adapters, one subnet #
Illustrative inputs for an IPC with two adapters: adapter 1 = 192.168.0.10 with mask 255.255.255.0 and default gateway 192.168.0.1; adapter 2 = 192.168.10.20 with mask 255.255.255.0; S7-1500 at 192.168.10.50 with mask 255.255.255.0.
- Adapter 2's network: 192.168.10.20 AND 255.255.255.0 = 192.168.10.0.
- The PLC's network: 192.168.10.50 AND 255.255.255.0 = 192.168.10.0. Same /24, so Windows sends the ping out adapter 2 and the ping succeeds.
- IE General bound to adapter 1: 192.168.0.10 AND 255.255.255.0 = 192.168.0.0, which is not 192.168.10.0. The station presents an address from the wrong network, the partner the PLC was configured with never answers, and Runtime reports the PLC disconnected.
Add the identity fault: the project names the interface ipc-edge-01 while the interface on the IPC carries no matching name, so the scanning tool lists the station as outside the network.
The fix and the test: rebind IE General to adapter 2, set the station name and the device name, reload the station, then trust the certificates in both directions. The proof is a capture filtered to 192.168.10.20 and 192.168.10.50 showing the session frames while Runtime reports connected.
The decision #
- Ping works, and a capture filtered to both addresses shows no frames during a connection retry: adapter binding. Rebind IE General; you are done when the frames appear.
- Frames are present and the unsafe-connection message with error code 50 follows: authorization. Trust both certificates, set the clocks, match the policy.
- The scanning tool shows every PLC but lists the station outside the network: device name or station name mismatch. Set both names and reload the station.
- Another client already reads that PLC: the controller and the switch are fine, and every remaining fault is on the IPC.
- A test OPC UA client connects where Runtime does not: the difference is the client stack and its trust store, the split traced in Troubleshooting Node-RED OPC UA When UaExpert Connects.
Questions asked
Does renaming the Windows PC to match the station name help?
No. The station name is read by the SIMATIC communication software from the local station configuration, not from Windows. Matching the two names buys nothing, and changing the Windows name can break remote access or domain logons that rely on it. Set the station name where the station itself is configured and leave the Windows name alone.
What does the PC clock have to do with the certificate?
Certificate validity is checked against the local clock, and a wrong date fails that check during the handshake. Set the correct date on the PC and point both the PC and the PLC at one NTP time source, so the question never returns after an outage or a battery loss.
Should I import an XDB file to fix the disconnected runtime?
Not for this fault. An XDB file carries compiled station configuration into the workflows whose documentation calls for it, and importing one replaces the local station settings, so a wrong file can undo a correct slot layout. It does nothing for an adapter binding or a trust problem. Use it only where the installed SIMATIC NET version documents that step.
The connection works after a reboot but drops days later. Related?
Usually the same fault returning. A PC with a failing backup battery loses time on every power loss, and each invalid clock fails certificate validity again, which reads as a fresh disconnection. Check the date as soon as it drops; if it has reset, replace the backup battery and source time from one server for both devices.