AC500 CS31 bus spontaneus failures
Each group has an AC500 PM573-ETH CPU and 7 RIOs (CI592) on CS31 bus.
Each CS31 network is about 600m so a repeater from phoenix contact (PSM-ME-RS485/RS485-P) was used in the middle of the bus. The cable used is a profibus cable and it is properly terminated at each end (R1->B1 & R2->B2). There are no VFDs around nor any other equipment that would produce electrical noise.
The system was commisioned (without the repeaters) and it seemed that all is ok but after a few days the client copmlained that the system is down. I visited the site and I read the errors from the PLC. The problem was that the PLC could not see the RIOs through the CS31 bus. On the RIO modules(CI592) the CS31 LED flashed indicating that the module is in diagnosis mode. The problem only went away after a power cycle of each RIO. I left and the system failed again after a week or so. At that point I thought that the problem was the bus length exceeding the mark of 500m. So today I went on site and installed the repeaters thinking that this was the problem. The repeaters worked straight away without any issues. I left and then after a couple of hours I get a call from the client that the system cannot see two of the RIOs.
I asked him to reboot them and the problem went away.
I dont know what the problem may be. So any help is highly appreciated.
Voted best answer
Sometimes you learn the solution to some problems the hard way and you would not wish this to happen to your worst enemy, so I post the solution to the problem.
For some reason I had set the PLC task to be freewheeling and not cyclic. Because of that the PLC had longer cycle times and CS31 did not like that. The task was changed to cyclic and all is good now. The bad thing is that IMO this was not well documented in the manuals and we were searching in the wrong direction (EMC, repeaters etc).
1. I have been told that adressing modules in CS31 you need to have some space between them. It might give some issues. FX 1,5,10,15,20 could solve this issue.
2. Update you software. There are a new module for the cs31 bus. You can finde the updates on www.abb.com/plc right hand corner, software updates.
3. Which firmware are installed in your equipment? You can get a new one at www.abb.com/plc in the right corner there you will find the firmware updates, sort by publish date. The version is 2.3.3 at the moment.
Then I have some questions.
1. Which error codes are there in the CPU??? If log in using the control builder which error number do you get?
2. Where are you based? I think you need to contact a real support line and I would like to guide you there.
Regarding your notes:
1. The addressing for the first network is 1,2,4,5,8,9,11. For the second is 3,6,7,10,12,13,14,17. From control builder I right click the bus node and then I use the check function and it does not complain.
2. Which software do needs updating? I failed to find an update for CI592 module in the link you provided.
3. Regarding the firmware please see the attached files KKM4.txt and KKM7.txt.
It is the two CPUs that have the problem.
Regarding your questions
1. Please see the attached screenshot.
2. I am located in Greece and the local ABB office is already dealing with the issue.
My question is the following:
On the CI592 Module, the CS31 LED flashes, indicating as per the manual that the module is in diagnosis mode. However, nothing more is mentioned in the manual regarding that state. Nor what is diagnosis mode, neither what caused it and nothing about how to fix it.
Please also note that the CI592 modules have some Analog inputs confiqured where the instrument is not connected yet, so this produces a measurment underflow error.
Could it be that this makes the module drop to diagnosis mode?
Actual addressing should be based on the IO modules you are using in the network.
But as a thumb rule you can simply use address with 5 digit difference.
Thanx for your help. Today I went to site and I have the following notes:
One of the two networks was down. I did change the addresses like suggested (5,10,15...) and I did a firmware update to the CPUs. I also disabled some analog inputs without instruments connected that produced underflow errors. I also rechecked the cabling, the earthing ,and the terminations. After all that and a reboot of all stations, all looks in order. It has been running since 2 oclock in the afternoon without problems.
I also tapped a digital ocilloscope on the CS31 bus and took some screenshots of the bit steams. As it is just a DSO it cannot trigger from a specific slave (as or Profibus analyzer from Procentec does) and so I cannot tell for which sleve or from which slave the bit streams were. I judge that the waveforms look like standard RS485. Do you see anything unusual in these waveforms? I will also spend a whole day at site tomorow monitoring the behaviour of the system.
I still have not got a solid feeling if the problem is fixed or not but I hope for the best.
I came at site this morning and found out that the las RIO of both CS31 buses was lost again.
My mind is drained out from ideas on what to fix. I checked with the oscilloscope the powersupply and it looks like a proper sinwave, so I guess there are no harmonics. Because we have a multicabinet configurations I am now adding capacitors between the shield point of connection on the CI592 Module and the cabinet plate as indicated in the manual topic "CS31 system bus".
What other reason can cause the module to drop out of the network.
The strange thing is that in the missing modules I can see the DI/DO of the module but not the analog values, so I guess that when it says "Module lost" it does not mean that the PLC cannot communicate with the CI592 at all, but it cannot see the analog module. By analog module I guess it is meant both the analog signals on CI592 and on AI561.