MB_Client TIA Issue

Lare said:

You probably meaned to write
"But it doesn't for MB_server anymore so I guess it is the same with MB_Client."





As MB_Master is same than MB_client and MB_Server is same than slave.

Click to expand...

No. MB master is the same as MB client, like you say. So I think it works the same on same instructions for TCP and RTU..

m_turk said:

No. MB master is the same as MB client, like you say. So I think it works the same on same instructions for TCP and RTU..

Click to expand...

OK




As a Side note some manufactures don't use slave and master anymore for RTU as it can hurt someone's feelings.

I allways confuse which one server and client is meaning on communication thought.

Lare said:

As a Side note some manufactures don't use slave and master anymore for RTU as it can hurt someone's feelings.

I allways confuse which one server and client is meaning on communication thought.

Click to expand...

So no robot slaves in the future...

The Client is always right, so he is the Master, and the Server brings you what you need, but you'll tip him so it is not slavery..


For the topics sake, let us just agree that he should go with clock_1s -|P|- on the REQ input, and it should work.

I had already tried messing with a timer, though never used r trig. Tried that this morning but still the same issue. So I've used varying times up to 5 sec, a constant 1, r trig etc, all give the same problem.

If you go into the MB Client db and monitor you can see in the temp data area that the values I expect to see are being read as they change. The values (int) I expect to see however in the temp data area are split into two bytes. The expected values are 0-100 so one byte is zero and the other shows the value. The first value shows correctly for maybe half a second, then the other byte shows an unexpected value and the 'correct' byte shows zero. Whenever the correct value shows, the data is transferred to my db, otherwise all zeroes are written to my db even though the correct values are showing for the remaining 23 registers in the temp data area. Does that make sense?

I don't really understand you 100%..
I think the simulator might not be working as intended.. But I just noticed you created some #mb_data_pointer, not sure what is there.. You should just put your Array instead..
The byte swap, if it exists because of the Endianness you can solve with SWAP, READ_BIG, READ_LITTLE instructions. But, you have to get some stable data first..
Also, create some simple debugging..

Try this the thing in attachments..

Now see if you are getting the same stuff in buffer and data, how many errors vs good data are you getting and what are the errors..

if you have located #data_pointer to FC or FB temp area then it is only valid for one program scan and only when your MB_client done bit is "true".


Locate your MB_data_area to independent DB-block instead of temp.

Lare said:

if you have located #data_pointer to FC or FB temp area then it is only valid for one program scan and only when your MB_client done bit is "true".


Locate your MB_data_area to independent DB-block instead of temp.

Click to expand...

Thanks Lare! This was the issue.

bb76 said:

Thanks Lare! This was the issue.

Click to expand...

Yeah, everyone have had same mistake with siemens sometimes.


Use temp area only if you first write to temp and later on same program scan read same temp variable. On another program scan data is carbage and overwritten from other FC and FB blocks.