Rename a smart camera network name without conflicts

Two identical smart cameras come out of the box with the same default network name, and on a DHCP line their IP addresses are not stable either. Sooner or later a discovery list shows two entries that differ only by address, and an application that opens devices by name connects to the wrong camera. The fix takes minutes per unit: give every camera its own network name, set once at commissioning, sized to the discovery field, and proved by rediscovery before the cabinet closes. This page is for integrators and maintenance engineers standing up or standardising a multi-camera line. You will plan a naming convention that fits the field, rename each camera through its configuration client, and run three checks that prove the rename took hold.

What the network name actually is

A networked camera carries three identifiers at once. The MAC address is fixed at the factory and never changes; it is unique but unreadable on a floor list. The IP address is whatever the DHCP server hands out today, which makes it a poor identity: renew a lease or move a cable and it changes. The user-defined network name is the one you control. The camera stores it in non-volatile memory and announces it during discovery, so it survives reboots, re-addressing and re-plugging.

The name is also bounded. The GigE Vision discovery response carries the user-defined name in a 16-character field, and GenICam calls the parameter DeviceUserID. Sixteen characters is the budget for every camera discovered over GigE Vision, which covers the Fast Ethernet and GigE smart cameras sold here.

Discovery problems — cameras missing from the list, duplicates, ghost entries — are a network fault more often than a naming one; the network camera discovery troubleshooting for multiple cameras guide covers that diagnosis. And when no DHCP server answers, the camera's IP stack falls back to a link-local 169.254.x.x address, so the name becomes the only friendly handle left in the list.

Plan a convention that fits the 16 characters

Encode place, line and unit: a site code, a line number, a camera role and an index. Keep it to uppercase letters and digits — several camera families accept nothing but alphanumerics in the discovery name, and uppercase avoids case-sensitive string comparisons in scripts.

Worked example: two-character site A1, two-digit line 02, three-character role CAM, two-digit index 03. Without separators, A1L02CAM03 uses 2 + 2 + 3 + 2 = 10 characters, leaving 6 spare. With separators, A1-L02-CAM-03 uses 13, leaving 3 spare. Both fit. If you later add a station segment of two characters plus one separator, the separated form reaches 16 exactly, so the unseparated form is the one with growth room.

Record the plan as a table and fill in the MAC suffix from each camera's label. The MAC never changes, so this column is what settles an argument between two identical units on a bench.

Rename the camera, step by step

You need the manufacturer's configuration client — the same tool you use to set IP addresses — a laptop on the camera's subnet, the name table, and a window when the line is idle. Take the client from the camera maker or its distributor, matched to your camera family and firmware generation.

1. Discover and confirm. Find the camera in the client and identify it by serial number or MAC label, never by IP alone. If the tool shows nothing at all, fix addressing first; see change a camera IP address and update tools stuck on the default.
2. Open the device settings. Locate the user-defined name field — DeviceUserID in a GenICam client, the device or network page in a smart camera's configuration utility.
3. Type the planned name. Alphanumeric, uppercase, 16 characters or fewer, exactly as the table spells it.
4. Write and reboot. Apply the setting, then power-cycle or reboot the camera as the client instructs, so the name loads from non-volatile memory instead of living only in the session.
5. Rediscover and match. The list must show the new name on the entry whose MAC matches the table. If you saved the rest of the camera's set-up as a user set, the name rides along on every boot; see save camera parameters as a user set that loads on startup.
6. Repeat one camera at a time. Never batch-rename a subnet while production runs; ambiguity during the window is how wrong-camera swaps happen.
7. Re-bind the host side. Search the vision application and PLC code for the old name and bind to the new one. Where a controller addresses cameras through job or recipe codes, keep one unique code per unit, as in RS232 vision recipe selection: one command code per recipe.

Older camera families sometimes expose a raw console over Telnet, with a parameter-dump command that prints the current network name and a name-update command that takes the new string. If the manual shows that console, run its exact commands: dump first, read back the name, update, then reboot as the manual says. Do not improvise around them, and leave the console disabled on networks with no other use for it.

Every networked smart camera and vision sensor in the catalogue follows the same client-driven procedure, over Fast Ethernet or GigE:

Verify the rename took hold

Three checks, in order:

1. Persistence. Power-cycle the camera and rediscover. The name must come back exactly as typed; if it reverts, the write never reached non-volatile memory and the client was showing you a session label.
2. Unplug test. Disconnect the camera you believe is A1L02CAM02 and confirm the discovery list loses that entry, and only that entry. A different entry vanishing means your table and the floor disagree.
3. Application test. Start the host program and confirm it opens the camera by the new name. Then disconnect that camera and confirm the application faults with the new name in the message — proof the binding moved with the rename.

Common mistakes that undo a rename

• Renaming while a client holds the device open. Applications can cache the old identity until the next re-enumeration, so a live rename leaves the host answering to a string the camera no longer carries; the camera discovery after connection in a running application guide covers that stale-entry problem.
• Leaving two units on the default name. Enumeration order is not stable across power cycles, so a pair of identical names can behave differently every morning.
• Treating the name as a fix for connectivity. A name is a label. If the camera does not appear in discovery at all, rename nothing; diagnose the link, the subnet or the addressing first.
• Mixing case. Name the camera A1L02CAM02, write a1l02cam02 in the script, and some string comparisons fail. One case, everywhere.
• Renaming the host computer by the same habit. The PC that runs server-based vision software has its own name, and changing it can break services that recorded the old computer name at install time — see what happens to server software when the computer name changes.

When to bind by more than the name

At stations where a swap costs a shift, pin the device by serial number or MAC in the application where the client supports it, and keep the name for people. The multiple depth cameras on one host: select each by serial number guide shows the same select-by-serial principle; it transfers unchanged to smart cameras.

When several cameras share one host, a dedicated vision controller keeps naming and addressing tidy: one subnet per camera group, and built-in light channels that save a separate amplifier. For example:

Before you close the panel

• Every camera answers to a unique, alphanumeric name of 16 characters or fewer — duplicates are the failure this whole procedure exists to prevent.
• The name table with serial and MAC suffix is filed where the next engineer will find it.
• The application and PLC code open devices by the new names, and no reference to an old default name survives a text search.
• The unplug test and the power-cycle persistence check have both passed on every unit.

The decision in one line: name every networked camera at commissioning — unique, uppercase, alphanumeric, 16 characters or fewer, recorded against its MAC — bind the software to those names plus serials, and re-run the unplug test after any config restore. To pick the cameras themselves, start from the smart cameras range or the camera finder; for quoting a multi-camera line, the contact page is the direct route.