Panelview Plus DeviceNet problems

We have a Panelview Plus 1000 communicating with a warm backup CompactLogix (L32Es with a 1769-SDN card) pair over DeviceNet. At this stage, the CompactLogix are configured to nodes 0 and 1 respectively, the HMI is configured to node 3 and there are 3 Flex IO racks set to nodes 10-12 respectively. The DeviceNet bus is currently operating at 125k.
The HMI project runs without any issues until we have to restart the program (either the panel loses power, or we go back to the configuration screen and then run the project again). At this point (probably 90% of the time) the comms don't seem to come back properly. The HMI can't write to any tags (error messages appear in the diagnostics popup) and seems limited in reading tags (most indicators show 0 although I did get an alarm which came through). However, it doesn't get any tag errors etc - it's like it's sort of communicating, but just not properly. The DeviceNet bus doesn't show any errors, and the indicator lights on the PV+ DNet module (2711P-RN10H) are all solid green, indicating that it is all working properly.
We've tried reloading the program and then running it again (replacing the comms). On startup, we have tried both replacing the comms and not replacing the comms. The only thing that seems to work is downloading the HMI program again.
Any ideas about what is causing this, and better yet, how to fix it? It's obviously not acceptable for the client to have to redownload the program everytime they want to cycle power to the panel.
Thanks

I suspect that the errors you are seeing are essentially "nuisance" errors, where the Polled connection hasn't yet been established, so FTView can't read or write the alias tags attached to it.

Is the PV+ configured as a Slave or as a Master ?

How many bytes Produced and Consumed data are being exchanged ? Is it a Polled connection ?

And the usual questions: which model of PV+, which firmware revision in the PV+ terminal, which firmware revision in the 2711P DeviceNet adapter.

An often-overlooked feature for diagnosing PV+ errors is the ability of the terminal to write its Diagnostic Log to a remote PC. You need a PC connected over Ethernet, with the "Accept UDP messages" feature turned on in FactoryTalk View, then you have to configure the Remote logging of errors, information, and audit messages. They'll then be viewable in the PC's FactoryTalk Diagnostic Log, rather than just on the PV+ local diagnostic log.

I happen to be working on a PV+ 1000 with ControlNet this week so I'm particularly aware of the ups and downs of these network interfaces.

The PV+ is configured as a slave, and has the full 240bytes produced/consumed data mapped using a polled connection.
The 2711P DNet adapter is v3.016 and the PV+ is v5.1 (model 2711P-RDT10C).
If it is just that the polled connection hasn't been established, is there any way to trigger it? It doesn't seem to matter how long we wait, it just doesn't work.

If there was a problem with the Polled I/O connection, the 1769-SDN scanner would show you on the 2-digit display and in the Status tags inside the CompactLogix database.

So we have to figure out what exact sort of errors are being logged by FTView ME when it's trying to communicate with the RSLinx Enterprise I/O object.

Can you open up the diagnostic log and post some of the error messages ? Look into that remote logging feature; it's really an excellent troubleshooting tool.

If you can create a *.APA archive and post it (or send me a PM and we can trade it by e-mail) I'd be happy to have a look.

For anyone with similar issues, we suspect that the issues were caused by macros that we had written in the PV+ project (for multiplexing data). We have since removed the macros and handled multiplexing different, and the problem seems to be resolved.