How to Resolve Conflicts in the IP Address of CNC
Two devices on the same subnet claim one address, and the machine drops off the network mid-cut. This guide is for maintenance engineers and shop IT who need to clear the fault and stop it coming back. You will get the diagnostic order, the parameter ranges that matter, and the architecture that keeps duplicate addresses from forming.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
Key takeaways
Why the IP address of CNC machines collides
A CNC controller, a robot cell, and a vision camera all need a unique address on the same subnet. When two of them answer to the same number, ARP replies race each other. The switch keeps updating its table, and packets land on the wrong device. From the operator panel it looks like a random network failure, not an addressing problem.
The usual cause is a copied setup. A technician commissions a new machine by cloning the parameters from an older one, including the network page. Or a router hands out a DHCP lease from a pool that overlaps the static range. Both faults are invisible until the second device powers on.
The third common cause is a stale lease. A laptop was patched into the machine VLAN months ago and still holds an address the controller now uses. It may sit in a drawer, but it still answers when powered up.
- 1Cloned parametersNew machine commissioned from an old backup file.
- 2Overlapping DHCP poolRouter leases addresses from inside the static range.
- 3Stale leaseA forgotten laptop or gateway still holds the address.
Symptoms that point to a duplicate address
An IP conflict rarely announces itself with a clean error. The controller usually reports a communication alarm, a timeout, or a dropped remote connection. On some controls the alarm names the interface but not the cause. That is why you check the switch log before you touch any parameters.
A second symptom is intermittent behaviour. The machine runs fine for hours, then drops when a second device is switched on. Pay attention to the timing. If the fault appears at shift start, someone is powering up a device that shares the address. If it appears during a program transfer, the conflict may involve the office network or a CAM workstation.
A third symptom is that several machines drop at once. One duplicate can flood the subnet with ARP noise and slow every device on it. If two or three cells alarm within the same minute, look for one bad address rather than three separate faults.
Static or DHCP for the IP address of CNC equipment
Machine controllers should hold static addresses. A DHCP lease can move, and a controller that changes address breaks every CAM post, remote monitoring link, and DNC path that points at it. Static also makes the fault easier to trace, because the address on the panel matches the register.
Reserve the low end of the subnet for infrastructure, the middle for machine controllers, and a narrow band for DHCP. On a typical /24, keep 192.168.10.1 to 192.168.10.49 for switches and gateways, 192.168.10.50 to 192.168.10.199 for machines, and 192.168.10.200 to 192.168.10.250 for leases. That gap is what stops a lease from landing on a machine address.
Do not rely on link-local or automatic addressing on the shop floor. When a controller cannot reach a DHCP server it may fall back to an address in the 169.254.0.0/16 range. Two controllers doing this at once can collide, and the fault looks like a hardware failure.
Designing a subnet that does not repeat addresses
Put machine control traffic on its own VLAN. Office printers, guest Wi-Fi, and CAM workstations should not share a broadcast domain with a spindle controller. A VLAN boundary also limits the damage when a duplicate does appear, because the noise stays inside one cell.
Assign addresses by rule, not by habit. Tie the last octet to the cell number or the asset tag, so a new machine has an obvious slot. Keep the register in one place and make it part of commissioning. If the address is not in the register, the machine is not released to production.
Monitor the subnet rather than waiting for an alarm. A simple script that logs ARP replies and flags one IP answering from two MAC addresses will catch most conflicts before an operator sees a fault. On larger sites, enable DHCP snooping and dynamic ARP inspection on the switch.
- 1One VLAN per cellKeeps broadcast noise inside the cell that caused it.
- 2Rule-based addressingLast octet tied to cell number or asset tag.
- 3ARP monitoringFlags one IP answering from two MAC addresses.
- 4DHCP snoopingBlocks rogue servers from handing out leases.
Ruling out hardware before you re-address a machine
Not every dropped connection is a duplicate address. A failing patch cable, a dirty fiber connector, or a switch port running at the wrong speed will produce the same operator complaint. Before you change any parameter, swap the patch lead and check the port error counters.
Look at the link LED and the port statistics. A port with rising CRC errors or a speed mismatch at 100 Mbit on a gigabit link points to cabling, not addressing. If the controller drops only during a spindle ramp, suspect electrical noise on the cable run rather than the network layer.
If the machine is a recent build from GreatLight, the controller network settings ship documented, so you can compare the delivered address against what is on the panel now. That comparison tells you quickly whether someone changed it on site.
5 steps to clear a conflict in the IP address of CNC
Work in this order. Each step narrows the search before you change anything.
- 11. Read the controller alarm and event logOpen the alarm history on the machine panel and note the exact time. Cross-check it against the switch log. Most managed switches keep a MAC-to-IP table, so a look at the same minute often shows two MAC addresses reporting the same IP. Write both down before you clear anything.
- 22. Map every device on the machine subnetRun an ARP scan across the subnet, for example 192.168.10.0/24, and export the list. Include controllers, robots, vision systems, gateways, and any laptop patched into the cell. Compare the result against your address register. Anything unlisted is a candidate for the conflict.
- 33. Confirm the duplicate before you change itPing the suspect address from two different hosts and watch the MAC reply. If the reply MAC changes between pings, two devices hold it. Do not reassign anything yet. Changing an address while a second device is live just moves the fault.
- 44. Take the second device off the subnetPower down or disconnect the device that should not hold that address. If it is a stale laptop, release the lease and remove it from the VLAN. If it is a new machine with cloned parameters, change its address on the controller network page while it is isolated from the subnet.
- 55. Reassign and verify end to endGive the controller a static address outside the DHCP pool, then reload the program and confirm the remote connection. Run a full program transfer and a tool change to prove the link survives load. Log the final address in the register with the device serial number and date.
Choosing an addressing method for CNC equipment
Match the method to the device role, not to convenience.
| Device role | Addressing method | Why | Risk if wrong |
|---|---|---|---|
| Machine controller | Static, documented | CAM and DNC paths stay valid | Broken post processors |
| Robot cell gateway | Static, documented | Fixed handshake with controller | Dropped cell handshake |
| Vision system | Static in machine band | Stable image transfer | Lost inspection data |
| Operator laptop | DHCP lease only | Temporary, easy to release | Stale lease conflict |
| Office workstation | DHCP on office VLAN | No machine traffic | Subnet flooding |
| Guest device | Isolated VLAN | No route to controllers | Broadcast noise |
Common questions
Can an IP conflict damage the machine?
No. The conflict is a network-layer fault, so it cannot harm the spindle, drives, or control hardware.
The real cost is lost production and a scrapped part if the machine drops mid-cut. Clear the fault before you restart the program.
How do I know which device holds the duplicate address?
Ping the address from two hosts and compare the MAC replies. If the MAC changes, two devices hold it.
Then check the switch MAC table and your address register to identify both devices by name and port.
Should I just enable DHCP on every machine?
No. A controller that changes address breaks CAM posts, remote monitoring, and DNC paths.
Use DHCP for laptops and office gear, and static addresses for anything that a program or a monitoring tool points at.
Why did the conflict appear only after a new machine arrived?
New machines are often commissioned from a parameter backup that already contains a live address.
The fault stays hidden until the new controller powers on and answers ARP for an address that is already in use.
How often should the address register be checked?
Review it at every commissioning and at every network change, such as a new switch or a new cell.
Once a quarter is enough for a stable shop, provided every new device is added before it goes into production.
Need parts that arrive documented and network-ready?
Send your drawings and we will return a quotation with a free DFM analysis within 12 hours.
12-hour quote100% inspectionNo minimum order
More shop-floor notes
We publish setup notes, tooling trials and inspection data from the factory floor.
