Intermittent faults on Modbus Plus outstation

I wonder if anyone can help. We run a traffic tunnel and have recently updated the environmental control system within. The PLC network is based on Schneider/Modicon Quantum HSBY Master controllers, with Momentum outstations. Not all outstations have processors as it was felt that this would be done via Remote I/O.
The problem is thus, One of the outstations for air quality sensors is producing intermittent faults i.e switching between On and Off state all the time, now to me this would look as though it was a processor problem . However, the unit was replaced with a new (Good) unit but the problem remained. Power supplies and earthing were tested and seemed fine. I was just wondering if any body here has encountered anything similar, as it is probably just an oversight somewhere.
Thanks in advance,
Steve

Have any new module been added to the Modbus+ network? If so are their network addresses unique?

Is the actual node going on-off or just some of the I/O points?

What are the module's LEDs indicating?

Thanks for replying Jim
Further to the last post I would like to add that the data is still coming through, about once every few seconds, and outstations can be accessed either side of the 'dodgy' unit (Communications Adapter). Therefore network, addressing, cabling and connections seem fine. The problem is that this is all monitored via SCADA and the alarms database is being filled which will eventually cause it to 'fall over'. A temporary fix has been incorporated via an extendeed watchdog on the status of the unit but this is not the long term answer to this problem. The LED status on the unit does not seem to change.

If this is a MB+ network, are you writing commands over the
network? Too many MSTR function blocks will spoil the soup.
I have also seen intermittant problems with the connectors not being
properly made. It may check just fine then your problems return.
That problem was solved by making all of the MB+ connections again
with the proper tool. Or it could be that one segment of the network
is too short or too long. Do you use repeaters?
Or is this a modbus ethernet network?
I have been experiencing problems with an ethernet network that has a
Quantum and Wonderware on it. I have not had a chance to figure that one out yet. If you do solve your problem be sure to post the
solution. It may help me.

You replaced the outstation hardware, but did you replace the Modbus+ network connections?

The last time I had an excessive number of communication errors it was solved by replacing the network connection hardware. We had been messing with the Momentum hardware for a couple of months. Finally we took a lightning strike which "burned out" our MB+ tap. After installing a new tap (and cable) the errors disappeared.

Oh yeah, don't forget about those end-of-line thingys.
They go on the ends.

Jim,
The network connections were not replaced, however a laptop with the Concept 2.5 and Modbus Plus network diagnosis software (MBPSTAT) was plugged in at that point/connection and the entire network was accessible. I was expecting the lap-top to 'drop' in and out as the Momentum unit had been doing, Therefore indicating power supply errors, but it seemed fine.
On further investigation with MBPSTAT it was also discovered that the unit was generating Framing errors. All a little bit over my head really.
W.R.T. JDW's post, To the best of my knowledge the network has been correctly terminated.

My experience (limited as it is) has been that framing errors are caused by hardware. Often the message packet is corrupted by a bad cable connection (most commonly a broken cable) resulting in a communications retry. Many, but not all, Modbus Plus devices are fairly tolerant of this type of error wich may explain why your programming device worked correctly but your Quantum does not.
I still vote for re-terminating the Modbus Plus cables at the suspect node. Also review your grounding connections.

I second Jims emotion, RETERMINATE.
And get one of the tools from Schnieder to
do it with since it makes such a nice termination.

Have checked to see if the TOKEN has become worn?

Thanks for your help chaps, The connection has been re-made and the fault still exists. The next option is to try and relocate the unit to another node on the return leg of the network, the problem is that it is on a 'live' road and access is very limited. I will let you know how we get on, Cheers