Clarification of the purpose of the S7 MMC card?

[uptown47]

Hi all,

I was always under the impression that S7 300 series PLCs needed a MMC card to function. Therefore when you download a program it is going on to the MMC. If you then remove the MMC and put it into a different PLC it would transfer the software (and hardware config etc).

However, one of our guys reported to me a strange occurrence..

He wanted to do a memory reset of the PLC in order to try and clear a persistent fault.

He compared the blocks and found that he had identical software on his PC as the PLC.

He then performed a 'memory reset' (something which I have always thought deletes / formats the contents of the MMC).

He then said that when he completed the memory reset and did a comparison of the software, the blocks in the PLC were vastly different and seemed to be from an old program?

He managed to fix it by simply downloading all the offline blocks into the PLC but I just wondered how/if this is possible?

Is there a program on the MMC and a program stored somewhere in the PLC as well? I've seen the options in the memory to "copy from ROM to RAM" etc and wondered if there was a retentive area in the PLC that could have been harbouring an old copy of the software somehow?

Many thanks for clarifying how this set up works?

I always thought the copy ROM to RAM was a throwback to when the PLCs had EEPROMs in them?

Any ideas?

 

[JesperMP]

A memory reset with an inserted MMC card does not format the MMC card.
If the MMC card contains a valid HW configuration (which it does in the situation you describe), a memory reset sets the actual values of the DBs to the initial values. That is all.

In an S7-300, RAM to ROM copies the actual values in the working memory (RAM) to the actual values of the load memory (the MMC card).
Before leaving a plant I allways use RAM to ROM.

Were the differences reported by the compare function in code blocks or in DB blocks ?

 

[uptown47]

Thanks for the info Jesper. I wasn't there to see the comparison differences but apparently it was code (FCs and OB1 as well as DBs) - and it wasn't *every* bit of code. In other words, there was some code there which was presumed to be from an old project?

So does RAM to ROM copy only DB values as I'm presuming the FCs/FBs etc aren't in RAM are they? (or are they??)

 

[JesperMP]

The FCs and FBs gets loaded into RAM when the CPU encounters code that addresses them.

Also, there is a significant distinction as to if difference is merely the timestamp or in code. The comparison tool will tell that. You can turn to timestamp differences off to filter them out of the comparison so you only see real code differences.
Timestamps will change if you save a block, even if there are no code differences. Or if you do a block consistency check.

 

[uptown47]

The differences were actual code, not just timestamps apparently.

So if you download to the PLC does it copy to the MMC (ROM) and working memory (RAM) together? Or would you need to do a RAM-ROM (vica versa) everytime you make a program change?

 

[JesperMP]

That is how I have understood it.
The information is from what I have eeked out of Siemens manuals.
In order to know for sure, one would have to perform a number of experiments.
It has been a non-issue to me since I never leave a plant without having done a RAM-to-ROM.

 

[uptown47]

Yep, I'll maybe do a few experiements. I have an old test rig that I could get set up. One for a rainy afternoon I think. Thanks for all the info Jesper, much appreciated as always