AC800FR is stopped due to error
Voted best answer
hi
During plausibility check (Compilation), user gets error messages regarding limit violations,
- number of objects which is bigger than defined in CPU properties;
Yes, in plausibility error check list you will get error message.
- amount of synchronized data between primary and standby CPUs.
For redundancy memory, Interface objects & Network buffer you will get error message in plausibility error check list.
- number of objects which can’t be transferred to CPU because remaining CPUs memory is too small;
Practically details of CPU memory usage can be retrieved when CBF is running in Commissioning mode hence while compilation this can’t be detected because during commissioning mode compilation is not possible.
Find more details regarding memory in attached file.
Answers
HI,
Might be possible, I have faced similar scenario while testing, but make whole station download will do…. If it is running process take required precautions. Good thing is controller redundancy is there.
And UDP Send/Receive you will be using with EI 813F or EI 803F card ???
Whole station download rectifies your problem or error led is continuously on in your case ?
For UDP we are using FI813F card.
If I start primary CPU without standby program with all communications runs OK. Just starting standby primary goes in error. So probably, problem not in communication itself.
Ok it is EI 813F, anyways card is not a concern with issue here I feel,
It may be because of 'synch' of primary and standby not processing properly ?
does primary goes into error inbetween both controller get 'synch' or after that ?
Sumit,
thanks for answer.
It seems to be problem in synchronization. How it could be avoided? Some CPU settings?
Don't know as synchronization is internal processing, but if redundant controllers are not in plant or if there is no harm in whole station download, take required precautionary steps & you can try to make whole station download to the resource in Project tree allocated to redundant controller hardware and then both primary and secondary controller will get downloaded in sequence and will get synchronized.
Hello Eugene,
I work in tech support in the USA, I can tell you that there can be many reasons for a failure. Anything from faulty harware components to wrong configuration, even a network overload can cause the controller to failover.
I suggest you save the controller log files of both the primary and secondary controllers and send them to your local supportline for analysis. They have the necessary tools and experiences to help you solve this issue.
To access the controller log files, type the IP address of the controller in a web browser. Click on the Diagnosis button and then Write to File button.
Hello!
Thanks to all for answers. Support investigated the logs and their conclusion - too much variables has to be synchronized between CPUs. Their proposal - increase number of CPUs, decrise number of tasks and so on. There is no chances to play with new CPUs so will change program code.
Question to ABB - when CBF at compiling stage will inform project engineer that there are problems with at least:
- number of objects which is bigger than defined in CPU properties;
- number of objects which can not be transfered to CPU because remaining CPUs memory is too small;
- amount of synchronized data between primary and standby CPUs.
What are restrictions in CBF that you will know this only when trying to download????
Dear Sumit,
do you have screenshot of error message window with errors regarding limit violations you wrote?
Find attached here with one of the case...
Hello Sumit,
thanks for reply. "...at least 1369...". Does it mean that current number of objects in project is 1369?
In continuation:
if in redundant PM803F with redundant Profibus-DP communication to S800 one task is defined as redundant (TASK/RED, PID loops) and one task defined as non-redundant (TASK, AI monitoring):
1) will be data of non-redundant task synchronized between CPUs?
2) if primary CPU goes offline (for example, power supply is switched off) non-redundant task will be executed in standby CPU?
3) if Profibus-DP communication from primary CPU to S800 goes offline will standby CPU synchronize I/O with primary CPU so primary CPU will continue process this non-redundant task based on I/O data from standby CPU?
Thanks.
Following to your question...
1.Logically NO,I am not sure otherwise why task and task/red both are exist? Check this statement from online help “Not only the redundant task (TASK/RED), but also tasks of the previous type (TASK) can be configured under the redundant resources. Thus only those functions of the user program which also have actually to be redundant need to be configured redundantly within a redundant resource.”
2.In this case primary power OFF triggers redundancy toggle and previous standby take over as primary. Hence non-redundant task is running/active in primary(previous standby) controller and anyway now standby(previous primary) is dead due to power OFF.
3.There is lots of dependencies on type of profibus configuration (profibus line redundancy/ slave is redundant or nonredundant/ type of RLM used 2-1 or 1-2) and case of S800 offline( Profibus cable cut/ slave not responding/ FI830 is having problem), to keep it simple if profibus cable cut/removed from FI830 triggers redundancy toggle and red. Switchover happens and non-redundant task of acting primary processes I/O data.
Nevertheless as in your case I/O’s are in non-redundant task & non redundant task don’t get ‘synch’ then the if S800 I/O’s are not configured for safety value then there could be bump/data loss for redundancy switch over period.
I hope these answers your question If I understood correctly……
Thanks
Hello Sumit,
thanks for your answers. Your understood me correctly.
All of task, which are working with outputs (mostly PID loops) I will do as redundant.
If I will do analog monitoring (read and processing AI only) as non-redundant task it
will not interfere with operator in case of redundant switchover. The only question - what will be with settings of analog monitoring FBs in non-redunant task: standby will process with settings from project?
Dear Sumit,
question is still open - how one can know if number of objects can’t be transferred to CPU because remaining CPUs memory is too small?
Thanks.
Hello, answer lies into your question that by knowing CPU memory is too small one should not try to load or overload the controller, you can get the available/free memory status via zzzz.PRAM_FREE or zzzz.RAM_FREE system variables. Where zzzz is your resource id.
Also if you overlook all this and load your configuration than while performing change load object or whole station download you will get some error message that memory is too small. And I suggest to take local L3 support help incase you have memory configuration issues.
Hello Sumit,
thanks for reply.
This is main question - how one can know the size of project? Before downloading I know PRAM_FREE and RAM_FREE but I don't know size of changed objects I'm trying to download and don't know size of whole project. So, am I right - noone can predict absence of memory in CPU?
Add new comment