IMDTouchKnowledge BaseShop main store

Troubleshooting

Why a PC Fails to Reach an HMI When Wi-Fi Owns Its Address

Industrial networks Published 6 min read

Short answer

The wired programming connection fails because the laptop's Wi-Fi adapter already owns 192.168.0.20, the exact HMI address, so Windows treats the target as local and never sends packets out the Ethernet port. Disable Wi-Fi or move the HMI to an address no local interface owns, such as 192.168.0.25, and the connection returns.

On this page
  1. Why does one laptop reach the HMI while an identical laptop cannot?
  2. How does Windows route a packet addressed to your own adapter?
  3. Is this the duplicate IP problem seen on a shared network?
  4. Which commands prove which adapter holds the HMI address?
  5. Should I disable Wi-Fi or re-address the HMI?
  6. Can both maintenance laptops keep the same wired address?
  7. What if the address match is gone and the connection still fails?
  8. Questions asked

The wired programming connection failed on one laptop and worked on an identical one because the failing machine's Wi-Fi adapter held 192.168.0.20, the exact address programmed into the HMI, so Windows delivered the connection attempt to the laptop itself and never put a frame on the wire. The working laptop's Wi-Fi held 192.168.0.19. Disable the wireless adapter, or re-address the HMI to 192.168.0.25, and the connection returns with everything else untouched. What follows shows how to see the conflict with two built-in commands, why no duplicate-address warning ever appears, and what to check once the exact match is gone.

Why does one laptop reach the HMI while an identical laptop cannot? #

Because the failing laptop's wireless adapter was assigned 192.168.0.20, the exact address set in the HMI, so the destination was one of the machine's own addresses; the working laptop's wireless adapter held 192.168.0.19 and had no such collision.

The two laptops carried identical wired settings, which is what made this look like a bad network controller. Both were used one at a time, never together, so the shared wired address was never itself in conflict. The difference that decided success sat on an adapter nobody was comparing.

Setting Working laptop Failing laptop
Wired IPv4 address, static 192.168.0.105 192.168.0.105
Wi-Fi IPv4 address 192.168.0.19 192.168.0.20
HMI target address 192.168.0.20 192.168.0.20
Programming connection Opens Fails

The comparison that closes the case is the application target against every active adapter on the failing machine, not the wired dialog of one laptop against the wired dialog of the other.

How does Windows route a packet addressed to your own adapter? #

The IP stack compares the destination with every address on the machine before it consults the routing table, and an exact match means local delivery: RFC 1122 requires a host to accept a datagram addressed to one of its own addresses as its own, never to forward it.

That local delivery happens ahead of any forwarding decision. No route lookup runs, no ARP request goes out, no frame is queued to the Ethernet port. The programming software waits on a connection the operating system quietly answered itself, times out, and reports a failure while both wired ports sit idle.

Worked example, using the 255.255.255.0 mask this address range normally carries: 192.168.0.20 AND 255.255.255.0 = 192.168.0.0, and 192.168.0.105 AND 255.255.255.0 = 192.168.0.0. Both interfaces therefore sat on the same connected network, which is legal but ambiguous; with two equally specific on-link routes, the lower interface metric picks the sending adapter. The exact-match fault is settled one step earlier than all of that, and it is mask-independent: the destination equals a local address, so no route is ever selected. Once the duplicate was removed, the overlap remained, the wired route still won, and the connection worked.

Is this the duplicate IP problem seen on a shared network? #

No, a classic duplicate is two devices holding one address on the same link, while here the PC itself owns the HMI's address on a separate wireless link, so nothing on the wired segment is duplicated.

Two devices on one shared segment fight where everyone can see it. Here the wireless network and the wired point-to-point link are physically separate, and each is internally clean: the wired link held the HMI at one address and the laptop at another. The fault is ownership, not collision, and address-conflict detection only hears the link its own adapter sits on, so nothing flags it. Replacing the cable, the switch or the Ethernet controller cannot touch a destination the host believes is itself.

Which commands prove which adapter holds the HMI address? #

Run ipconfig /all and compare the HMI target against every adapter's IPv4 address, wireless, VPN and virtual adapters included, then route print to see which interface the stack selects; an exact match in that list is the whole diagnosis.

  1. Read the target out of the connection settings of the programming software and write it down.
  2. Run ipconfig /all and record each adapter's IPv4 address, subnet mask, default gateway, and whether the address is static or DHCP.
  3. Compare the target with every address in that list, including Wi-Fi, VPN and virtual adapters; an exact match anywhere is the diagnosis.
  4. Run route print and find the route covering the target, noting which interface index it uses.
  5. Run netsh interface ipv4 show addresses for a second, scriptable view of the same inventory.
  6. Run arp -a after a ping attempt; no ARP entry for the target on any interface means no frame was ever prepared for it.

Capture this output before changing anything, so you can prove afterwards exactly what moved.

Should I disable Wi-Fi or re-address the HMI? #

Disable Wi-Fi first, because it proves the diagnosis in under a minute; re-address the HMI second, because that is the durable fix, and 192.168.0.25 let both laptops connect with Wi-Fi active.

Disabling the wireless adapter removes both the duplicate address and its routes in one action. In the case at hand the connection also needed a laptop restart, so clear the software's connection attempt or reboot rather than trusting the first retry.

One caution belongs with the address change: re-addressing the HMI drops every controller or SCADA client currently using that address, so make the change when the machine is not depending on the link. Afterward, verify with all adapters up. Inventory the addresses again and open the programming connection with Wi-Fi active, which is exactly how both laptops were confirmed working.

Can both maintenance laptops keep the same wired address? #

Yes, but only while they are never on that network at the same time, because the moment both plugs go in, 192.168.0.105 becomes a true on-segment duplicate and both connections degrade.

Their one-at-a-time success proves nothing about simultaneous use; the collision only exists when both cables are connected. Two hosts sharing one address on one segment produce ARP entries that flap back and forth and intermittent loss on both sides. Give each laptop its own fixed address in the range and the risk is retired. The HMI does not care which maintenance address dials it.

What if the address match is gone and the connection still fails? #

Then the fault has moved off the PC's address list: check the subnet masks for overlapping connected networks, the firewall rules on the wired profile, the service port (789 in this configuration stack), and the cable itself, in that order.

The service port, 789, is the TCP port the programming client dials once a route exists. It lives two layers above the address decision that failed here, and no port value can steer a datagram that never left the machine. If the target answers ping and the download still refuses, you are in application territory, the same class of fault as why an HMI is detected but the project download fails. If ping itself fails with the duplicate gone, repeat the AND comparison on the masks actually in use; overlapping connected networks want either a separate subnet for the HMI link or a deliberate interface metric that forces the wired route.

The decisions, case by case:

  • Target equals one of your own addresses: take Wi-Fi out of service now to prove it, then move the HMI to a fixed address no local interface owns.
  • Two laptops sharing one maintenance address: separate the addresses before they ever meet on the same switch.
  • Target answers ping but the software refuses: application service, its port, or the firewall profile, no longer routing.
  • Target fails ping with no duplicate in the inventory: masks and metrics first, cable last.

Questions asked

Why did Windows never warn about the duplicate address?

The warning comes from ARP exchanges on the link where an adapter takes its address. The clashing address was taken on the wireless link while the HMI sits on the wired link, so the probe and the panel never meet and the system stays silent. That silence is exactly why this fault hides from a firewall-first troubleshooting pass.

Will ping still answer if I unplug the Ethernet cable?

Yes, and that is the giveaway. A destination equal to one of the machine's own addresses is answered by the local stack in well under a millisecond, cable or no cable. If replies keep coming with the cable out, the address is local; if they stop once Wi-Fi is disabled, traffic is finally leaving toward the HMI.

Can I change the laptop's Wi-Fi address instead of the HMI's?

You can, and the connection will recover the same way, but it rarely holds. Maintenance laptops usually draw their wireless address from DHCP, so the next lease can hand the HMI's address straight back. Re-addressing the HMI to a fixed value no local interface owns, as the move to 192.168.0.25 did here, survives lease renewals and both laptops.

Does changing the subnet mask clear the conflict?

No. The mask decides which destinations are on-link; it has no vote on whether a destination equals one of the host's own addresses. Tighten or loosen the mask as you like, a datagram addressed to the machine itself still never reaches the Ethernet port. Masks only matter afterwards, when two connected networks overlap and a route must be chosen.

Does the HMI's programming port number play any part in this?

None at all. The port in this stack, 789, is the TCP service the configuration software dials once a route exists, and it sits two layers above the address decision that fails here. Port settings are worth checking only after the target answers ping and the exact-address match is gone from the laptop.

Keep reading

  • Troubleshooting

    Diagnosing intermittent HMI to PLC Ethernet comms hangs

    Stop changing the topology. Put one packet capture where it can see the PLC-to-HMI unicast exchange — a mirror port or a tap, not a spare switch port — run it through at least three hangs with device logs on the same clock, set every panel update cycle to 1000 ms, and confirm the three devices hold unique IP addresses. Whether panel requests stop or PLC responses disappear decides which side you repair.

    HMI and touchscreens

  • Troubleshooting

    Why Two Bridged VMs Cannot Ping Each Other in a PLC Simulation

    Both bridged guests carried the same MAC address, so the bridge could not forward frames between them: VM-to-VM and VM-to-simulated-PLC pings failed while every host-side test passed. Generating a new MAC address on one guest restored communication; distinct IPs, a shared subnet mask and ICMP-friendly firewall settings are the follow-up checks.

    Industrial networks

  • Troubleshooting

    Why One PLC to Two HMIs Fails Over RS-422 Multilink

    A 1:1 serial test proves one cable and one framing set; it cannot prove a shared n:1 multilink. Fix two HMIs on one RS-422 trunk by running the multilink driver entry on both panels, giving each a unique link identity, matching every serial parameter, and terminating only the trunk ends. The documented limit is 16 displays, so two panels are a configuration problem, never a capacity one.

    Industrial networks