Hello,

I have a very strange problem using GET.

There are three S7-1200 in PROFINET network. One (CPU 1212C DC/DC/DC, Firmware V3.0) is collecting some data from the other two (CPU 1212C AC/DC/Rly, FirmwareV4.1), namely several sensors and statuses.

I use two GET instructions in a sequence (used manual "PUT / GET sequencer, in attachment).

The problem is that read values (sensors) are changing value every 5-10 second to something completely other.

The problem is not with the sensors, because I read DB (for GET) and data on remote controlers and they are OK.

What is really strange is that I transfer some binary statuses and counter (number of seconds from controller reset) and they are steady!?

I suspect that there is some problem in GET, but I don't know what.

ADDED: In CPU which collects DB data with GET I have MODBUS functions (CPU is a slave) and SMS communication with remote (the fourth) CPU. Is it possible that this interfere with GET?

Does anybody know where is the problem?

Did you allow the FW4.1 CPUs to use PUT/GET commands?

Under the hardware configuration - protection.
Move the page to to see below the passwords you fill in to protect the PLC code.

William_B

Did you allow the FW4.1 CPUs to use PUT/GET commands?

Yes, I did.

This is a video what is happening.

Values at offset 4, 8 and 12 are unstable values.

The other are stable, including value at offset 20 which is seconds counter.

The left window is controller DB1 from which GET data is collected, the right is a "unstable" DB36 which is GET.

I tried to fill source DB1 with a constant value (81.12345), but DB36 oscillated again.

Dear all,

after much of code digging I have found a bug and a solution.

I 'll post detailed description later.

THE SOLUTION

As you can see in the video from previous post, there are unstable values at master side (the side which calls GET) and stable values at slave side (the side which is called).

I cropped software on the master side to the point that there is no doubt that there is no mistake. Having no other solution, I inspected the slave side and found errors in the code.

Namely, I was using memory zones (which master read via GET) for temporary calculations of final values! Because asynchronous communication can occur at any time, it read temporary calculations occasionally, i.e. senseless values are displayed.

The frequency of the "wrong values" approximately matches the ratio of the temporary calculations time and total execution time of OB1 on the slave.

BEWARE: never use memory zones which are part of the asynchronous communication for temporary storage or calculations! Otherwise, this leads to the errors which are very hard to discover!

That applies e.g. to the GET/PUT, MODbus slave,...

The reason why these wrong values are not displayed in TIA Portal monitoring is presumably because it is synchronized with the end of OB1 cycle.

I appologize for this late answer.