We have some PowerFlex 40 VFDs with 22-COMM-E adapters communicating with an L61 ControlLogix over Ethernet through Cisco Network Switches that keep losing the connection. About a minute and a half after the hour, every hour, the VFDs lose the connection with the ControlLogix. It reestablishes a connection and then is fine until the next hour. The same thing is happening at the same time with a CompactLogix and iFIX. The ControlLogix and iFIX were set up on our office network (192.168.xxx.xxx) years ago. When the PowerFlex 40s and the CompactLogix were added recently, it was decided to set up a new VLAN (10.120.xxx.xxx) and put all the new devices on it and move the rest over when time permits. So this connection loss seems to be between the two networks. Ideally we would just move everything over to the new VLAN but that would require some down time and may not happen for a few months. My question is, what could be causing this loss of communications at such a predictable time? It is within a 15 second period of time every hour.
Devices on Ethernet losing connection every hour
- Thread starter Rich21
- Start date
bulletin blues
Lifetime Supporting Member
If it is a managed cisco switch, see if you can log into the switch and look under all of the set-up parameters. Look for any one that has an update rate of around 3690 seconds. I would start looking at an IGMP query or update. Ideally, if it is a managed switch, I would use port mirroring with wireshark to monitor what is happening on the network at that exact moment. This looks like your first post, welcome to the club!
Timbert
Member
I would definitely second the recommendation for wireshark.
Are there any big once an hour data transfers on the network going on? Such as production database update etc. What happens every hour on the enterprise network?
Are there any big once an hour data transfers on the network going on? Such as production database update etc. What happens every hour on the enterprise network?
Ken Roach
Lifetime Supporting Member + Moderator
Welcome to the Forum !
I'm a little confused about which devices are affected. A network diagram might help.
Are the PowerFlex 40 VFD's being controlled by the 1756 ControlLogix, or by the 1769 CompactLogix ? Are the VFDs on the old 192.168.xxx.xxx network, or on the new 10.120.xxx.xxx VLAN network ?
How many physical switches are involved, and what models are they ?
If you were just losing the I/O connections, an IGMP configuration would be my first suspect. The fact that you're also losing messaging connections doesn't eliminate IGMP Snooping and filtering as a possible source of the problem, but it bumps it down the list a bit.
Fortunately you're in a good position with managed switches to go ahead and mirror the ControlLogix Ethernet module port to an unused port, and connect up Wireshark to capture all the network data from that port for a minute or so before and after the connection failure.
www.wireshark.org is the first place to get Wireshark downloads. Excellent tutorials are available at www.lovemytool.com (yes, really).
I'm a little confused about which devices are affected. A network diagram might help.
Are the PowerFlex 40 VFD's being controlled by the 1756 ControlLogix, or by the 1769 CompactLogix ? Are the VFDs on the old 192.168.xxx.xxx network, or on the new 10.120.xxx.xxx VLAN network ?
How many physical switches are involved, and what models are they ?
If you were just losing the I/O connections, an IGMP configuration would be my first suspect. The fact that you're also losing messaging connections doesn't eliminate IGMP Snooping and filtering as a possible source of the problem, but it bumps it down the list a bit.
Fortunately you're in a good position with managed switches to go ahead and mirror the ControlLogix Ethernet module port to an unused port, and connect up Wireshark to capture all the network data from that port for a minute or so before and after the connection failure.
www.wireshark.org is the first place to get Wireshark downloads. Excellent tutorials are available at www.lovemytool.com (yes, really).
Hi and welcome...
Another thing that catches a lot of folks out is ARP(Address Resolution Protocol), especially when using industrial Ethernet equipment with limited receive buffers. The 20-COMM-E and 22-COMM-E only have 3 receive buffers each, which can easily get hammered with ARP requests.
ARP in the switch has to resolves IP addresses to the physical MAC addresses of each hardware device before it can send any data. If using dynamic ARP, the switch has to resolve IP/MAC addresses routinely using broadcast messages at a very fast rate which can easily overflow small receive buffers. If an ARP request was being sent out to all the drives at once, their receive buffers could be maxing out and they'll drop some of the ARP packets. The switch will wait for the ARP responses and hold I/O connections from communicating, potentially timing them out.
Is the CISCO managed switch using dynamic or static ARP?
Whether this is the issue or not, I'd advise using a static ARP table in the switch, which maps out the IP/MAC addresses, so these ARP requests are then not needed.
G.
Another thing that catches a lot of folks out is ARP(Address Resolution Protocol), especially when using industrial Ethernet equipment with limited receive buffers. The 20-COMM-E and 22-COMM-E only have 3 receive buffers each, which can easily get hammered with ARP requests.
ARP in the switch has to resolves IP addresses to the physical MAC addresses of each hardware device before it can send any data. If using dynamic ARP, the switch has to resolve IP/MAC addresses routinely using broadcast messages at a very fast rate which can easily overflow small receive buffers. If an ARP request was being sent out to all the drives at once, their receive buffers could be maxing out and they'll drop some of the ARP packets. The switch will wait for the ARP responses and hold I/O connections from communicating, potentially timing them out.
Is the CISCO managed switch using dynamic or static ARP?
Whether this is the issue or not, I'd advise using a static ARP table in the switch, which maps out the IP/MAC addresses, so these ARP requests are then not needed.
G.
The Plc Kid
Member
The best way to rule out an ARP issue would just be to do a show ARP command in the CLI and look for any incomplete listings but IOS devices have a really good cache and refresh rate.
robw53
Member
I have a similar issue going off but my drives only talk to a PLC locally to them even though its connected to a plant network, in fact no communications are taking place over the network from either the VFDs or PLC but once a day I get a F071 fault.
Please Help!
Please Help!
Similar Topics
I am using a VM for my Rockwell software, and have been having a weird issue. Essentially, I configure my devices in RSLinx using the Ethernet...
I have numerous 24VDC/50mA beacon/horn devices in the field. How do I convert these devices to IP based devices where each device gets an unique...
I can't seem to find an answer in Rockwell's site or here. I am curious if CCW can connect to 3rd party devices via E-I/P like other brands and...
Hi All
Is it possible to connect non AB Ethernet devices to a Compactlogix 1769-L35E, without the use of RSnetwork for Ethernet?
Thank you
Is there a way to ping an Ethernet/IP device using a 1756-L81E? I want to set an alarm in the event that my Ethernet device disconnects from my...