Hello, I'm in the beginning of the usage MB_CLIENT FB , wrote a single test to read 1 word , everything worked fine, I left office for a while and on my return I noticed that data read by block are frozen, i.e. not being read. Modbus server I read from was Ok. I provide a link to video (when all Ok) with the code. 

http://screencast.com/t/LIDYYnXIX

the stat word was #local for OB so I was not able to analize cause , which would have recorded in other case with global type of tag. 

I wondered of possible cause , so far have supposed 2 reasons -

1. On my notebook I use V13  and once I 've connected to the PLC from other place with V13 SP1, so MB_CLIENT in my program was 3.1, though must have been of 3.0 version for TIA V13 - fixed this
2. REG input changes to 0 after calling FB (it is 1-scan long tick) - will fix this. 

Also I supposed broken link between 1200 and modbus server, and tried to duplicate by unplugging network cable from 1200, but then it restored the connection  successfully.

What else could lead to this ? ( afraid of " strange things" happening in the field )

Thanks, 

HI, thanks for answer, something with sreen capture tool...i checked screencast before uploading it was Ok.

I changed code according to to my guesswork - now I'm testing this chunk. (see attachment ) Hope it is Ok and won't stuck. 

Unfortuntely, the MB_CLIENT still gets frozen, it was not finished because DONE output did not reset request, and status = 0, so I can not analize what's happening. (see attached sceenshot of the moment )

Pls advise way to solve this 

I tried both 3.0 and 3.1 versions of MB_Client (and CPU formware is 2.2. )

those 3 networks are all that exits in project, I did not tried more.  

Timer once  per second sets MB_request (only if no current request ).  And request is reset with Done output via local tag #mb_done. No other mb_client FB  in the program. 

mb_err copies status word to global tag in case of error (to analize later, but it is 0) this netwok is like in example in the system manual 

I'll try advise from huggy_d1 - remove REQ after BUSY goes OFF, not with DONE

Thanks for support.

Evgeniy 

Hi, I put off for a while MB_CLIENT issue, and now working further with it (still have errors ). 

I've looked thru the examples provided here,  There a different method of timing is proposed .  

But those examples are for V12 of TIA and MB_CLIENT is of 2.0. version there.

And mine - V13 has V3.0 . or 3.1 for V13 SP1 

Please advise wether I can follow this sequence , 

start

        - wait for MB_CLIENT.CONNECTED & NOT BUSY then set ADDR data_1 and  set REQ input 

        - wati for DONE and read data_1 , set next ADDR for Data_2

       - wait for DONE and read Data_2 , set next ADDR and so on 

end  

// so , they don't reset REQ for each ADDR they need to read, only upon reset or swtiching  to another node.

for errors handling - they offer timeout which is enough for polling all slaves in chain, if all slaves fit into this timeslot, this timeout just resets polling again, if there is some problem the same timeout lead to error handling networks, so they don't even analize errors codes

Please lead to example of MB_CLIENT you mention so I could take a look at. 

Take a look at how Siemens uses their own comm blocks in FAQs with project downloads.  That is where I got the above steps as the way they are meant to be used.

Hi, I tried method of just keeping REQ input always set and changing actual parameters (addr and so on) for MB_CLIENT and this did not work.

When I turn off REQ after DONE , wait a bit 1 second, put new data for MB_CLIENT inputs and again set REQ it works , no errors in status word. 

But sometimes I read wrong data from server. I.e. server exposes room temperature which is nearly same s 26C, and MB_CLIENT sometimes reads  0  (no any error in status word ) !!! another utiity (modbus poll) read always 26C , which is correct.

 what could cause such MB_CLIENT behaivour ?