S7 - Rubbish in X-Reference

I've two versions of my project which were backed up a few days apart. The second version has quite a lot of minor mods (more than I would normally do at one go without testing, but I couldn't get on the PLC because acceptance tests were running in parallel) and needless to say it's not working properly.

It's proving quite tricky to find the fault but I've managed to narrow it down to a fairly small area and looking for a bit more info I opened up X-Reference. Here are the relevant Networks of the two Versions of the program, as you can see:

1) They are identical
2) The IDB's are being called correctly





When I look in X-Ref this is what I find





As you can see:

1) They're different!
2) According to X-Ref in Ver C, Instance 1 of FB31 (I_Normal) is being called correctly with it's Instance DB DB31, however, for Instances 2 and 3, the IDB and the CPU_Out DB have been swapped.
3) For Ver J, the Instance 2 call appears to be OK but in Instance 1 the IDB has been swapped with the CPU_Out DB and in Instance 3 the IDB has been swapped with the CPU_In DB

Questions:

1) Has anybody ever seen anything similar?
2) How do I get things sorted out and
3) Is it likely to have anything to do with my problem (given that Ver. C works and Ver. J doesn't)?

Further info

It took me a little while to find the place under "Extras" in Simatic Manager where I could force X-Reference to re-generate the data, rather than just updating it. Result - no change, identical to the pics posted above.

When I have (rarely) have problems with the X-ref, I select the "Delete" option under the "Reference Data" menu.
It usually solves the problem.

No, this time that didn't help (I've just tried it - previously I had just regenerated the data without deleting them first).

In the meantime I've deleted the remaining 20-odd Networks in OB1 (in a new copy of the project) and the X-Ref is now correct. Now I'm going to start copying the deleted Networks back in one by one in the hope that I'll either end up with a correct X-Ref or at least I'll find out which Network is causing the problem.

Culprit found (or at least one of them)

Well it didn't take long to find the first troublemaker, it turned out to be Network 1 - the check to see which Modules are present and to disable those that are missing.

The Network is too long for a screen-dump, so I'll post it as a source code text file. If anybody has any ideas what's causing the problem I'll be grateful to hear them.

In the form posted, all three calls to I_Normal (FB31) are screwed up (funnily enough the calls to I_Debug have never been affected). Simply deleting the first network and regenerating the X-Reference file clears everything up.

I think this one's going to land on the Siemens hotline!

Commenting out all the command lines in NW1 resulted in correct X-Ref entries. Simply enabling the first 8 command lines (down to the SET command) was enough to cause the X-Ref entries to be screwed up.

Guess I'm going to have to look for my program problems some other way for the next few days!

How about replacing the very first OR with an AND - just a long shot.

We've had some other problems this morning, so I've only just got round to trying this out.

Just changing the first instruction to AND didn't help, but everthing was OK again if I just left the AND and the SET in (commenting out the rest of the ORs). I thought I might have found the suspect when, after removing the comments one by one everything was OK until I removed the the comments from the O "GLOBAL_DIAG_DB".BUSY command. However, after commenting out all references to the "GLOBAL_DIAG_DB" and the FC125 call, the X-Ref was still screwed up.

I have a feeling the FC125 / DB100 combination is part of the problem, but it doesn't seem to be the whole story, since even deleting FC125 and DB100 didn't solve the problem.

.. even deleting FC125 and DB100 didn't solve the problem.

Click to expand...

To remove the DB100 completely, you have to also clear any symbol that is defined for it.
(just in case you've forgotten it).

Oops, hadn't thought of that! However, a quick check in the Symbol Table shows no Symbols associated with DB100.