Siemens MMC's & those damn actual values

[guest]

Its stated very clear in the manuals- The processor will not run without the MMC card.

Heres another thing you might not be aware of-

Try a memory reset with the switch (the keys are gone on the new ones)- you know, the one where you hold down the mres button and watch the lights- then release and hold it down again for the faster blink and viola- your blocks should be gone.

Well on the new processors using the MMC card-
NOPE- your blocks are still there.

The only way to delete them (except with the prommer) is to go through accessible nodes, select the blocks and hit the delete button.

I wouldnt get in to the habit of removing the MMC with the power on.

 

[DesertDog]

Here's examples of the save & restore I have to do for the MP-370 WinAC, it really $ucks not having any retentive memory.

PS. Guest: There is a sequence to reset the 300 series with the run/stop switch. I've used it before but I don't remember what it was, something odd you wouldn't do by accident.

 

[JesperMP]

Roy,
when reading your posts I am not 100% sure that I have understood you right.

If you remover the MMC while the CPU is powered up - then it goes into fault.
What happens if you remove the power to the CPU - re-insert the MMC - and then power up the CPU again.
(No reset is performed !).
Does it go into RUN ?
Or does it stay faulted ?
What happens to the actual values ?

 

[RMA]

I wouldnt get in to the habit of removing the MMC with the power on.

While I agree it's probably not a good idea to get into this habit, in fact, there is no problem for the MMC card as such in removing it with power on (apart from the original problems with retentive memory), however there is a VERY big problem removing the card with the CPU running. If the CPU happens to be writing to the card when it is removed, you can finish up damaging the format of the MMC. If that happens you have no other choice than to send the MMC back to Siemens for reforming.

By the way, a word of warning to anybody meeting the MMC cards for the first time - the MMC is formatted with a proprietary Siemens format - DO NOT TRY AND FORMAT THE CARD BY PUTTING IT IN A NORMAL CARD SLOT IN A PC - if you do you'll find yourself in the same situation as described above and have to send the card back to Siemens for repair.

BTW2, the newest generation of MMCs appear to have an "improved" encryption system for the format which results in older CPUs not being able to work with them without a Firmware upgrade.

They also appear to be a bit unreliable at present - I recently wasted a couple of days, when I used a new 4MB MMC to upgrade the firmware of my 317 CPU (which had a few bugs, which together cost be nearly a week of wasted time bonkhead ) - the upgrade went through with no problem, but I then couldn't get the program to download to the MMC. I spent the best part of two days trying everything possible out - download from PG, download over the network, write the program directly with a prommer - all to no avail. Then I phoned the hotline, who listened to my description of the problem for all of ten seconds, before interrupting to say "No problem the MMC's faulty, send it back and we'll send you a new one."! banghead

So don't forget, just because you can download the firmware upgrade and write it to the MMC with the Prommer, then put the MMC in the CPU and update the Firmwaare with no problems - IT DOES NOT MEAN THE MMC WILL RUN CORRECTLY WITH S7!

Edit: Just saw your bit Jesper,

If you remove the MMC with the CPU powered up, as soon as you remove the MMC (even with the program not running) the STOP lamp goes into slow blink and you cannot restart the CPU without doing a memory reset.

Because I assumed this memory reset only reset the CPU memory (I'm pretty sure I remember reading somewhere in the manuals that memory reset does NOT clear (anything in - my interpretation!) the MMC), I thought that if I left the CPU in stop after the memory reset and then switched the power off, then set the CPU switch to RUN and then switched the power back on, the CPU would do a normal power recovery and fetch the rententive memory values from the MMC, which I assumed would not have been changed, since the CPU HAD NOT RUN since the memory reset.

However when I tried it the DM- and M- values were again zero.

All my initial values were non-zero and in no case did I ever get the initial value. Either I got the value stored at run time. I'm pretty sure the only time you'll get the initial value back is if you download the original program again from the PG/PC.

If anybody has any specific ideas they'd like to try out, let me know the 314's still sitting around doing nothing, just like me!

 

[RMA]

Jesper,

I added some substantial edits, possibly after you started reading the last posting, might be worth refreshing it just in case.

 

[JesperMP]

If you remove the MMC with the CPU powered up, as soon as you remove the MMC (even with the program not running) the STOP lamp goes into slow blink and you cannot restart the CPU without doing a memory reset.

Even if you power down, re-insert the MMC and power up again ?

I am completely baffled.
How does the CPU know that there was a fault, when its memory (the MMC) is removed ?

By the way.
You can delete that post in the other thread that you posted by error.
Click the edit button in the lower RIGHT corner - NOT the QuickEdit button. You will get the option to delete the post.
It seems to be a little known fact that posts can be deleted. I think that Phil added this at some time - it wasnt there in the beginning.

 

[JesperMP]

OK now...

I also have experimented. This is with an S7-313C.

If you remove the MMC while the CPU is in RUN, then it will fault - the STOP lamp will flash.
Then, if you power down - re-insert the MMC - and power up, then the CPU will go in STOP.
With diagnostics you can see that it reports that a "memory reset was executed". This probably means that the RAM was cleared when it sensed that the MMC was missing.
To get the CPU to RUN mode you have to switch the little lever to STOP and back to RUN - OR - cycle power once more.
In other words, it is NOT required to do a memory reset on the CPU !
The DBs keeps the actual values as they were from the last powerdown (not the actual values when the MMC was removed).

 

[RMA]

I hadn't actually tried pulling the MMC while the CPU was running, as that struck me as being potentially too dangerous.

However, in the interests of science - and knowing that I had set an M-Bit so that the DBs were not being accessed , I tried it.

Result, as you say you can now restart the CPU simply by going into STOP and back to RUN, or by cycling the power - BUT my M-locations and DBS are still zeroed.

I tried it again, stopping the CPU first, so that the STOP LED goes into slow blink to demand a memory reset, then after trying to switch to RUN, (which as expected didn't work) powered down and back on and sure enough switching to RUN started the CPU as you said, but also as you said, the Diagnostic buffer shows that a memory reset has been performed so functionally it's no different from performing the memory reset manually.

The bottom line is that in both cases, my M- and DB-memory was zeroed after the reset.

If it was OK for you then maybe you've got a newer CPU/Firmware Version.

It looks like we're getting into pretty fine details of differences between different CPUs and Firmware versions here.

For the record, I'm using a 314C-2 DP from late last year,

Order No.: 6ES7 314-6CF00-0AB0
Firmware: V1.0.3

All in all, I'm coming to the conclusion that the safest way is simply to power down, remove or replace the MMC and power back up - WITHOUT stopping the CPU, i.e. creating a genuine Power Fail condition.

It goes without saying, that before doing this you must check that the plant or machine is in a safe condition!

You can delete that post in the other thread that you posted by error.
Click the edit button in the lower RIGHT corner - NOT the QuickEdit button. You will get the option to delete the post.

I knew it ought to be possible to delete the posting - but I hadn't even noticed the EDIT Button in the RIGHT corner! -Blind as usual!

 

[JesperMP]

Its an old 313C.
Forgot to check the Merkers, but DBs vere definitely retained with the actual values.

Roy, when you check, do you do like this:
1. Set the Merkers and DBs that you want to test.
2. Cycle power.
3. Remove MMC (while CPU is in RUN).
4. Power down.
5. Re-insert MMC.
6. Power up.
7. Cycle power.
8. Examine the Merkers and DBs.
Maybe you forget step 2. It is required that one "normal" power cycle is performed for the values to be stored on the MMC.

the Diagnostic buffer shows that a memory reset has been performed so functionally it's no different from performing the memory reset manually.

I was thinking that maybe a memory reset with a MMC plugged would force the DB values to be initialised. The reset that the CPU performs by itself just clears the RAM.

 

[JesperMP]

more experimenting .....

Aha, its not as simple as I thought.

It turns out that the actual values get stored on the MMC upon power down if it is the first time after the DB was created !
Subsequent changes does NOT get written to the MMC.
Only if you use the COPY RAM TO ROM function do they get saved again.

The Merker words are not retained when you remove the MMC card.

 

[RMA]

I'm fairly sure I've tried all possible variations, but since I'm forgetting what I did when and in what order, I repeated the tests (using my lap-top as PG) with your post on the Desktop beside me to be sure I did things in EXACTLY the same order.

First I checked after point 2) that everything was still OK, which of course it was.

Then I went into the Diagnostic Buffer, after everything was up and running and there was no mention of a memory reset. However, that notwithstanding, my M- and DB-memory has still been reset!

Just realised I should have checked after point 6) - will now do so.

Ah-Hah - there it is.

Point 6) - request memory reset following memory change
Point 3) - memory reset carried out


The only thing I can think of is that under certain circumstances (specifically, following a memory reset) the background task responsible for saving the retentive memory to MMC starts its cycle by writing the current (reset) value in memory to the MMC, instead of first recovering the saved version from the MMC.

Given the problems I had with the 317 because of what turned out to be Firmware faults, I think I'll go do a search of the Siemens Support site and see if there have been any upgrades for the 314 Firmware. (with the 317 I went from V2.0 to V2.1 and there were 1 1/2 pages of fault corrections for that! Since then there is a completely new Version of the 317 out with new Firmware V 2.3 - not upgradeable for the earlier model and as far as the 314 is concerned, on principle, I don't like V1.0 of ANYTHING!)

Siemens seems to pushing some of it's newer equipment out to market a bit too quickly in the last year or so, I remember reading in another thread that somebody had been warned off the latest Siemens' product (might have been an OP) by the salesman himself!

However getting back to the original problem, the summary still seems to be - do a normal power fail before you remove and replace the card and you shouldn't have any problems!

Sorry about the wide picture, but at 800x600 it was barely readable!

Been an interesting exercise though!

 

[Mylo]

Bloody hell,

its not clearcut at all really, I think I'll need to verify this the next time I'm on site and manually take down all the settings (Variables) just in case, I thought there may have been a better way around this but i'll just have to be careful,,

JesperMP the copy ram to rom would have to be carried out with a PG online I presume, or can this be set up in blocks ??

Cheers

Mylo

 

[RMA]

JesperMP the copy ram to rom would have to be carried out with a PG online I presume, or can this be set up in blocks ??

Desert Dog's post shows how to do it, you'll find the SFCs 83 and 84 in the System Function Blocks folder of the Standard Library.


As I summarised in my last post, if you just do a normal power fail before removing the MMC , then there's no problem - it's trying to be too careful that causes the problems this time!

Cheers

Roy

 

[RMA]

I've just been and had a look round the Siemens Support site and the 314-6CF00 has had one further upgrade to V1.0.5.

There is now a 314-6CF01 which started at Firmaware V2.0.5 and has now been superceded by V2.0.8

Since I can't see any difference in the Spec of the two 314s, I assume the newer one is an upgraded version - the Firmware version 2.0.5 as opposed to 1.0.5 might seem to support this.

All in all that seems like quite a few changes for only one year of product life and although I couldn't find any references to our funnies with the MMC, that doesn't mean they aren't a Firmware fault.

 

[JesperMP]

further experimenting .........

Have found the following:

If you download a fresh new DB, and cycle power, then the CPU retains its actual values.
And the MMC stores these actual values.

If you change the same DB, and cycle power, then the CPU retains its actual values.
But the MMC keeps the previous actual values.

If you remove the MMC (while the CPU is running), power down, re-insert MMC, power up, flip RUN-STOP-RUN, then the CPU fetches the actual values from the MMC card.
The actual values that were active when the MMC was pulled are discarded .

From this follows that the actual values do NOT get stored on the MMC every time that CPU is powered down.
Also, the actual values must be stored somewhere else than on the MMC card when a DB isnt new. In capacitor backed RAM ?

The above I have found and can replicate every time on a 313C.
Maybe it varies with different models or firmware versions.
Siemens manuals simply says that the DBs are allways retentive on the new CPUs with MMC cards.
It appears that it is a combination of MMC card and capacitor backed RAM (possibly).

Also, it is important that the first time you power down a CPU with a new MMC card, the actual values will be stored on the card. So you better have all the settings set to something reasonably acceptable at that moment !

And the Merkers ? They are probably also retained in capacitor backed RAM and not on the MMC card.