DeviceNet strange issue

Hi all,

We have a small devicenet network running for more then 4 years without a single problem.
CompactLogix+1769SDN scanner
nodes 1,2,3 = PowerFlex 700
nodes 4,5 = PowerFlex 755
node 6 = D0-Devnets.
All connected in same order. Thin cable, good resistors, length < 70 feet, power supply only for network.

Out of blue, node 3 decided to cycle between online/offline therefore drive associated stopped.( is commanded over port 5 - 20-COMM-D ) creating havoc on production people

In 10 minutes I put drive on manual , cancelled port 5 command, started local from pushbuttons with freq change possible also from pushbuttons.

Now I'm waiting clearance to stop all to be able to play with Interscan Delay. Now it's minimum 10 ms, baudrate 125k.

Until then I'm trying to understand why node 3 got that behaviour.

We forgot yesterday AC off, drive 3 having at stop moment 86degC on IGBT. Inside room about 42degC.

I started back AC, today having 64degC on drive and 30DegC inside room, and 20-COMM-D all solid green.. hmm

Can temperature influence behaviour of 20-COMM-D ?

Morevoer started to monitor the network with DeviceNet Traffic Analyzer connected over 1784-PCD card.

I do not know all dnet under the hood frames, but at least can look at statistics.

And : I saw lots of group 3 frames going to nodes 4,5 less to 1,2 and slightly more to node 3. Node 6 no group 3 frames.
: on interpreted data showing lots of UCMM going to node 1,2,3 but also to nodes 38-61 .... hmmm again

In plc I have not explicit messages, I am using only polled I/O.

So, where these UCMM and explicit messages comes from?? And more why they are going to nonexistent nodes?

I have attached some printscreens and interpreted data.

Any thoughts on this? and what to do to get to the bottom of cycling on/off?

First thing I will do when possible is play with Interscan Delay, baudrate and also heat the 20-COMM-D to over 50DegC to see what happens ( book says operating 10-50degC, but in past there were times with environment 45degC and IGBT 100 and nothing happened )

sebastianp,

The first thing I would do is ask what has changed?
Drive replaced, comms card to drive replaced, and so on.
Communications settings just don't up and change, programs don't up and change. Something must have changed somewhere, either malfunction or human intervention.

I would reseat the comms card and network cable to the node and the associated device, their may be trash or a loose connection.

If that doesn't work, look at the cable & connectors.
What you are suggesting may create more problems elsewhere.

these are just my thoughts.
james

Hi James,
thanks for heads up

As hardware/parameters/programming nothing has changed, 100%
Human intervention again not in discussion.

When I'll be allowed to touch again the network I will check as well comm card seating as well as connection.

I do not see what problems I can create, maybe only kill the comm card during heating process, but I have a spare.

Paremeter wise I do not care much, I built the network, I know it inside out so I can recreate any time.

However need some expert advice on why there are explicit messages on traffic analyzer and where they come from? they are not initiated by plc, .... i hope

Can temperature influence behaviour of 20-COMM-D ?

Click to expand...

The room was very hot, the drive was extremely hot, and heat is a natural enemy of electronics. The 20-COMM-D's rated operational maximum temperature is a reasonable 50 C (122 F), so the fact that the interface board malfunctioned when the system overheated is not a surprise.

Are there any PCs that might be browsing across the network to the DeviceNet with RSLinx ? That would be my first guess for traffic that was attempting to make UCMM connections to devices that aren't in the scanlist.

Some of that traffic you showed in the PDF documents shows the scanner opening up a Polled I/O connection with electronic keying, for example to Node 3. When precisely was the capture taken; during a malfunction, or after the system was working correctly ?

Thanks Ken for jumping in.

Your guess about other PC's was spot on.
So I took them out to get a clear reading.

That capture was from a moment when I believed node 3 is OK ( no more code 72 on scanner - all 00 and full solid green on 20-COMM-D ) .

Today I have closed all RSlinx's to stop interfering with my Dnet and got a new reading. By far we can see node 3 is the star here. At random intervals ( seconds : 692-368-23-75-548-341-570-611-600-256-28 etc ) scanner interrogates node 3 about vendor/device type/product code/revision, he is setting the EPR for a value of 19200, reading in/out bytes and closing the connection.As you were saying, opening up a polled connection.

Meaning that my card is not quite OK, having loose connection inside or outside, or some other electronic problem.After replacement I will check again the network with the analyzer.

At this moment IGBT temp is 62degC ,30 inside room, 31 on top of 20-COMM-D and 35 on connection between 20-COMM-D and drive.

This being said at the moment of malfunction most probably temperature on connection was around 50 given or taken 3 degrees so poor card have an excuse and start to give signs wants to retire.

By the way, I am not at all a Siemens fan, not by far, but we have other installation where are 2 Micromaster 440. Last summer was so hot inside , that glue on serial plate melted and serial plate slide down on bottom of panel. I would guess 45 deg inside room, 55 inside panel and 90 on IGBT. No AC whatsoever, never cleaned. Drives up and running OK for more then 4 years

Here I have AC, room cleaned twice a year, drives/panels/bus bars vacuumed by dust each year and had 2 blown drives out of 5, 3 mainboard replaced and now that Dnet card. I would say that RA should work a little more on their PowerFlex MTBF.

Thanks all again for you support