How to determine which FB is affecting my PLC Scan Time.

Hello Everyone,

At our place PLC scan time generally is between 1-2ms, But sometimes it shoots upto 20ms,

Can you please tell me how can i identify what is causing such sudden changes in the scan time.

Thanking you for your help.

you can check time at start and end of each OB, FB and FC. Time difference is runtime for each part. Then check if runtime is greater than 2 ms. If so, log the time.

By that you should soon find the problematic code.

regards

i KNOW I HAVE TO CHECK EACH SCAN TIMES TO KNOW THE EXACT CODE ERROR,

BUT I DONT KNOW HOW AND HENCE THE QUERY.

Prayag

i KNOW I HAVE TO CHECK EACH SCAN TIMES TO KNOW THE EXACT CODE ERROR,

BUT I DONT KNOW HOW AND HENCE THE QUERY.

Hi.

I do not know the style of coding, but I can suggest what I would have done if it was my program. I hope that one our friends or experts can fill in the blanks. The following assumes you can make program changes that can be used for diagnostics.


1. Create two global variables. As example MW20 as #INDEX and MW22 as #STORE.

2. Create a time interrupt OB.. it should be executed "every" 10ms. Inside this OB, have the #STORE := #INDEX.

3. At the start of OB1 - re-trigger the time interrupt OB. In other words, the 10ms time should start from there.

4. At the start of OB1 - make a variable #INDEX := 0.

5. Inside OB1, before each suspect FB / FC call - have #INDEX get a new value. Alternatively have #INDEX take a value inside the FB / FC.

6. Inside all special OBs have #INDEX get a unique value.

7. Download and execute the code. Note if #STORE gets a new value.


I do not know from memory how to do point (3) - hopefully someone (please Mr. Hosseini) will be kind enough to teach us. The above is not a sure thing that you can sort out the problem.. it very depending on the program structure. It is my hope however that it can help you.

Does it help you?

Good luck.
w

eisenpferdreiter

you can check time at start and end of each OB, FB and FC. Time difference is runtime for each part. Then check if runtime is greater than 2 ms. If so, log the time.

By that you should soon find the problematic code.

regards

 

in each OB, FB and FC calculate runtime. Store minimum runtime, maximum runtime and average runtime for each in a DB.

After running your program for a while, the DB should give you a good overview which OB, FB or FC changes a lot.

Then you should look inside that part of code and do the same game for each network.

Finally you will know the problematic networks then.

regards

eisenpferdreiter

 

eisenpferdreiter

you can check time at start and end of each OB, FB and FC. Time difference is runtime for each part. Then check if runtime is greater than 2 ms. If so, log the time.

By that you should soon find the problematic code.

regards

 

in each OB, FB and FC calculate runtime. Store minimum runtime, maximum runtime and average runtime for each in a DB.

After running your program for a while, the DB should give you a good overview which OB, FB or FC changes a lot.

Then you should look inside that part of code and do the same game for each network.

Finally you will know the problematic networks then.

regards

 

Maybe combine the above suggestion with the attached information.

Thank you all for all your expert opinions,

At our place no program is called in OB1, they are called through cyclic interrupts . JFI

After going through the suggestions we found that a Task-scan will help analyse our problem.

Thank you all once again.

 HI Guys,
Please, I have this same issue.

I am doing some calculation by calling some FCs in different OBs and NOT OB1. This process increases my scan cacle to 75ms and sometimes shoots above 150ms which stops my PLC.

May I know how to calculate the amount of time each FC uses or OB?

Would it be ok to put a binary at the ENO of the FC and check in trace?

Could someone please tell me what the Task-Scan is and hos it can be done?

Thank you.

Prayag

Thank you all for all your expert opinions,

At our place no program is called in OB1, they are called through cyclic interrupts . JFI

After going through the suggestions we found that a Task-scan will help analyse our problem.

Thank you all once again.