ENBT-1756 Connection Time OUT

I am able to Ping from any switch with 1ms respond time, Go online with PLC, download, upload. But receive communication timeout from ENBT-1756 when the PLCs are trying to communicate with each other through produce/consume

I have increased the RPI to 100ms in the consume and i/o configuration from 35ms
I have 7 PLCs using , 7 ENBT-1756, 2 HMI and they are communicating through a fiber optic redundant ring with 1 Managed N-tron 608MFX and 6 N-Tron 508 FX2-a with IGMP enabled.

The N-Tron support team requested me to add static address filter to all Ntron 508FX2 series, however it did not solve the problem.

under my I/O configuration, ENBT-1756 receive random frequency of Timeout error Code 16#0203 every 2 to 3 seconds

I also used their software N-View to see the diagnostic of each ports in each switch and they do not have any alignment, FCS, collision, symbol, fragment, undersize, oversize, dropped packet error.

Is this problem related to multicast?

From

To

Communication Status

PLC1

PLC2

Time Out

PLC1

PLC11

Ok

PLC1

PLC12

Ok

PLC1

PLC13

Time Out

PLC1

PLC14

Time Out

PLC1

PLC15

Time Out

PLC2

PLC1

Time Out

PLC2

PLC11

Time Out

PLC2

PLC12

Time Out

PLC2

PLC13

Ok

PLC2

PLC14

Ok

PLC2

PLC15

Ok

PLC11

PLC1

Ok

PLC12

PLC1

Time Out

PLC13

PLC1

Time Out

PLC14

PLC1

Time Out

PLC15

PLC1

Ok

PLC11

PLC2

Time Out

PLC12

PLC2

Time Out

PLC13

PLC2

Ok

PLC14

PLC2

Time Out

PLC15

PLC2

Time Out



The communication time out change from PLC to PLC when I cycle the power to the switch.

Wondering if any one out there had a similar problem with N-Tron switches.

Using this search term in google, error Code "16#0203" , I found this and a couple of others that may be useful.

http://forum.*******.com/index.php?t=msg&th=1179&rid=&S=b964074fffc6bdee1e27681e8836c100&pl_view=&start=0#msg_2906

The * is forum.com??trol.com remove the ??

That has to do with I believe DeviceMasters

If I do not have any alignment, FCS, collision, does that mean my fiber is ok, no leakage in the fiber?

Try disconnecting one side of your redundant ring and see if you still have the problem. I have seen N-Tron switches with the redundant ring enabled (this is also true with trunked connections also) have issues like you are describing.

Darren

If a device is communicating normally and gets disconnected, cyclic UDT multicast data continues for a while (in your case, 4x100 ms) and then an error code 16#0204 gets declared.

The error code 16#0203 is different; it indicates a circumstance where a connection was negotiated between two devices, and one device (the originating 1756-ENBT, for example) began sending cyclic UDP Multicast data, but never received any corresponding UDP multicast data in response.

I have seen this mostly in SoftLogix systems that had the firewall turned on; the TCP packets that make up PING and RSLinx browsing and connection establishment went through fine, but when it was time to exchange cyclic UDP Multicast data, the "replies" from the remote devices were blocked by the firewall.

In the NTron management software you should be able to figure out which of the ring ports is the blocked one... that might be instructive.

I'm going to have to write out your list of connection status to view and puzzle over it. I was confused by the numbering of the controllers.

You're probably in for some Wireshark network sniffing and port mirroring work on this.

It would be helpful as well to have a connection diagram and some additional configuration information about the fiber ring setup.

Be careful with the word "redundant". Do you really have two fiber rings, or do you only have one ?

You are right. I should have said disconnect one side of his ring so that it is no longer a ring, but a linear backbone.

Darren

If he has the 508F2 switches he has a single ring and not a redundant ring.

You are correct, it is not a redundant ring.



The problem is fixed. It was because there was too much traffic across the switches at the same time, either the ENBT can’t handle it or the switch is not sending the data fast enough.





In the previous msg, i stated that I changed all consume message RPI to 100ms, however i have up 10 sets of consume array that request data every 100ms from every PLC at the same time and this create sudden surge in traffic every 100ms



I set each consume RPI with different time such as 100ms, 150ms, 200ms….. sort of scheduling it.



The CPU usage went from 70% down to 38% after the changes





Note: after scheduling RPI changes, I still had the problem but the problem was solved when I reset the switch. I also reset the switch many time when I had the problem but it didn’t fix it.

I got that issue also, and they way i solved the problem is by changing "Rack Connection" under module definition to NONE instead of Optimization.

Hope this helps