Networking issues

741C · Oct 30, 2019

We just setup controls for a new job, and we are having issues with a few devices dropping out of the network with what is suspected to be a duplicate IP address.

We aren't on-site anymore, so troubleshooting makes it difficult. What are your recommendations to troubleshoot? We think something might be doing DHCP on the network.

Red flag 1: there is a stratix managed switch. It was configured for VLAN. I didn't configure it.

Red flag 2: We setup some powerflex 525 drives with 25 Comm E2P adapters on them. The previous integrator had them configured so the drive was on DHCP, whereas the ethernet adaptor was a static IP. So that's how we set them up. I'm can't find much useful informaiton on AB's site to determine if this is how they are supposed to be setup, or if perhaps the DHCP on the drive is getting assigned to be a duplicate IP somewhere on the network. Here is all I could find:

https://imgur.com/VrlSZeh

Thx.

dmroeder · Oct 30, 2019

The most useful information will probably come from the web page of the PLC that the drive is connected to. Look at the overview page and the application connections page. Look for missed/rejected packets, FSC alignment errors, etc. You should have exactly 0 of these types of errors. The first place I go to in these situations missed packets on the application connection page.

Are you using a DLR network, or did you use the E3P to simply daisy chain the drives?

741C · Oct 30, 2019

Just as a daisy chain. There is a ring there, but we keep on getting a ring fault, because the ring isn't completley in-tact somewhere.

Do you know if we configured the drive correctly? DHCP for the drive, and static for the E2P?

Dravik · Oct 30, 2019

DHCP is probably a bad idea, I would configure everything static if it were up to me.

Phil Buchanan · Oct 30, 2019

Using DHCP you have to set a reservation in the DHCP server using the devices MAC Address so that the same IP address will be given to the same device each time or you will have problems.

As Dravik said above DHCP on an automation device/network, in general, is a bad idea and all items that can be should be set to a static address.

The only place I use DHCP on an automation network is to give an IP address to say technician laptops that also may go home with that employee or off-site for other business.

Phil Buchanan · Oct 30, 2019

741C said:
We just setup controls for a new job, and we are having issues with a few devices dropping out of the network with what is suspected to be a duplicate IP address.

We aren't on-site anymore, so troubleshooting makes it difficult. What are your recommendations to troubleshoot? We think something might be doing DHCP on the network.

Red flag 1: there is a stratix managed switch. It was configured for VLAN. I didn't configure it.

Red flag 2: We setup some powerflex 525 drives with 25 Comm E2P adapters on them. The previous integrator had them configured so the drive was on DHCP, whereas the ethernet adaptor was a static IP. So that's how we set them up. I'm can't find much useful informaiton on AB's site to determine if this is how they are supposed to be setup, or if perhaps the DHCP on the drive is getting assigned to be a duplicate IP somewhere on the network. Here is all I could find:

https://imgur.com/VrlSZeh

Thx.

Most if not all Stratix switches come out of the box setup for VLAN 1 so that may possibly be what you are seeing.

thingstodo · Oct 30, 2019

741C said:
Just as a daisy chain. There is a ring there, but we keep on getting a ring fault, because the ring isn't completley in-tact somewhere.

Do you know if we configured the drive correctly? DHCP for the drive, and static for the E2P?

I have not done the DHCP thing myself. I have read that using DHCP, per port with exactly 1 address in the DHCP pool. This supposedly allows you to pull out a failed drive, plug in 'any' new drive to the same ethernet port, and have it take the correct IP address. From there it operates normally.

I have not gotten around to trying it.

As for ethernet devices dropping out and coming back, I have some experience with Cisco 2955 switches. If port 1 and port 2 are duplicate addresses, the switch turns ON port 1 turns off port 2 for 60 seconds, then turns off port 1 and turns on port 2 for the next 60 seconds. It does not appear to be consistent exactly what is done across various models of Cisco.

dmroeder · Oct 30, 2019

741C said:
Just as a daisy chain. There is a ring there, but we keep on getting a ring fault, because the ring isn't completley in-tact somewhere.

Do you know if we configured the drive correctly? DHCP for the drive, and static for the E2P?

You are probably not using the drive's Ethernet port since you are using the E2P, so its configuration is irrelevant. It's probably just at default.

I assume that you are intending on using DLR since you are seeing that fault, it's just that the ring is not connected currently? Maybe because someone couldn't get it fully operational?

How is the baud rate set for the E2P modules? In almost every single case, the baud rate needs to be left at auto negotiate. The exception is when you are using DLR, they need to be set to fixed 100MB full duplex. Every device in the ring. Another key to the ring working is making absolutely sure that the devices connect consistently from port 1 to port 2 throughout the entire ring.

Ken Moore · Oct 30, 2019

Do you have devices being powered up and down? I have run across Stratix switches that someone just plugged in and no management setup. By default anytime it sees a new device, it stops all the ports until it confirms no duplicate IP's. Caused drives to loose connections etc...

einnh · Oct 31, 2019

- If you have multiple switches, make sure either only one is assigning DHCP addresses, or if multiple are you aren't overlapping in dhcp pool addresses.
- If the problem is immediate, check by disconnecting devices and replugging in one at a time until the problem appears.
- What the system log on the stratix switches for dhcp assignments or port information. See if there are duplicate addresses.
- Some devices don't like multiple controller connections. Test the device with only one controller accessing it. (I have some network encoders that drop off the network if more than one compactlogix tries to access them).

JosephMont · Oct 31, 2019

Phil Buchanan said:
Using DHCP you have to set a reservation in the DHCP server using the devices MAC Address so that the same IP address will be given to the same device each time or you will have problems.

As Dravik said above DHCP on an automation device/network, in general, is a bad idea and all items that can be should be set to a static address.

The only place I use DHCP on an automation network is to give an IP address to say technician laptops that also may go home with that employee or off-site for other business.

Hmm. You like setting the IP of 100's of devices? I can not imagine the agony having to set up a 100 powerflex 525's and having them all static. I DHCP all devices with each port having an assigned DHCP address.

thingstodo said:
I have not done the DHCP thing myself. I have read that using DHCP, per port with exactly 1 address in the DHCP pool. This supposedly allows you to pull out a failed drive, plug in 'any' new drive to the same ethernet port, and have it take the correct IP address. From there it operates normally.

And this is amazing when you DHCP drives. It makes changing a drive really easy for unskilled workers.

JosephMont · Oct 31, 2019

I have had a similar issue with what I think was EMF. The DLR ring ethernet cords were run in the ducting with the 3 phase power for the drives inside the panel.

I moved the ethernet cords outside of the ducting and my comms loss went away.

The panel isn't pretty but I don't have any issues.

Networking issues

741C

Member

dmroeder

Lifetime Supporting Member

741C

Member

Dravik

Member

Phil Buchanan

Lifetime Supporting Member

Phil Buchanan

Lifetime Supporting Member

thingstodo

Member

dmroeder

Lifetime Supporting Member

Ken Moore

Lifetime Supporting Member

einnh

Lifetime Supporting Member

JosephMont

Member

JosephMont

Member

Similar Topics