PLC toggling bit by itself

[Fernando Castro]

I know that it sounds weird, but I don't have any other explanation.
first of all, let me tell you that I have one PLC (V700) on Wich on parameter seemed to be changing itself randomly. I don't think that its randomly because I was able to systemic repeat that problem (only on that PLC)
I also have the same version of software installed on other PLCs (so far at least 10 running with no issues). Today I decided to tackle that problematic PLC so, I downloaded the most recent version of my software as follows:
1. First I did a Blank project download, with a burn upload project.
2. I went into info mode and did a plc Init to leave it entirely "virgin".
3. I downloaded the program using Unitronics download manager as I usually do with all my other PLCs.
4. I entered manually all the parameters required to run the machine this bit must be "SET" in this particular machine.
5. I did the machine's normal operation and then could spot exactly when that tricky bit was changed (after I loaded a recipe from the SD card and the confirmation of successfully read appeared on my HMI).
6. Checked the setting and the bit had a "RESET" value

This is where that bit is being used, and can only be changed in 3 ways:

1. Is a setting, so naturally, it is linked to an HMI button.
2. When loading the default settings button (Buried 3 screens deep and needs to be manually triggered, also that defaults to "SET")
3. As you can see there is a RESET coil linked to that bit, and this is a feature Not a bug, let me explain:
That setting is to select which type of sensor to use. There is an external control loop that is driven by a valve and a sensor, we have brands A, and B. valve A works with either sensor B or sensor A, but B valve can only work with sensor B, given this scenario, proper settings need to be entered on HMI to scale the readings of the sensor depending on if its A or B.
so, if valve B is selected on HMI, sensor B is forced:

This is the usage of MB122:

Again, 
1. Linked to an HMI button.
2. When loading the default settings button.

So, I am confident that is not a problem with my logic. I also already look at all the MBs used on the function that loads files from the SD card (the trigger that happens to RESET MB129), thinking that maybe somehow, I am overriding that value with any vector operation and all MB'S used on that entire subroutine are in the 2000's range (I usually try to reserve my operands by functions to avoid those kinds of problems).

 

Finally, I gave up and tried to replicate the problem on another PLC in my bench, downloaded the same software, imported operands from the faulty PLC then exported the operands to the bench PLC.
The PLC works as intended. Doing the same steps MB129 is not being RESET after loading the recipe.

 

I really don't want to say it, but it seems that it is the PLC that is not doing what was programmed to do.

[kratmel]

8 hours ago, Fernando Castro said:

could spot exactly when that tricky bit was changed (after I loaded a recipe from the SD card and the confirmation of successfully read appeared on my HMI).

SD card must be changed for test.

 

[Flex727]

I'm going to say I had a similar issue many years ago. My memory is a bit fuzzy as to the specifics, but I also had a single bit that refused to cooperate. I never found the cause and finally decided to move the MB to another address and the problem was solved.

[John_R]

I've had odd things happen like this from a dirty touch screen, a funny spot in the touch screen right where the HMI button is...

Also, I use to leave the factory film on the screen, thinking it was good protection, but also found this can cause strange touch screen issues.

Just food for thought...🤔

JohnR

[Fernando Castro]

1 hour ago, John_R said:

I've had odd things happen like this from a dirty touch screen, a funny spot in the touch screen right where the HMI button is...

Also, I use to leave the factory film on the screen, thinking it was good protection, but also found this can cause strange touch screen issues.

Just food for thought...🤔

JohnR

Definitively not the case, but I have seen a considerably number of posts related to that specific problem with the touchscreen.
I was able to see on online mode the exact moment when MB 129 reseted, and when that happened the Screen that was being shown was other than the one that has that button
Thanks anyway.

5 hours ago, Flex727 said:

I'm going to say I had a similar issue many years ago. My memory is a bit fuzzy as to the specifics, but I also had a single bit that refused to cooperate. I never found the cause and finally decided to move the MB to another address and the problem was solved.

Thanks, I feel like I'm not crazy... Other colleagues that are PLC programmers did not believe me I must be doing something wrong in the logic they said 🤣.

I am not saying that I don't make mistakes, but being on just this one PLC and with a well-tested program I don't think that that is the case

 

 

10 hours ago, kratmel said:

SD card must be changed for test.

 

I did not think on testing that... I don't see how it is related but this Is oddly enough to make me stop trying to find a logical explanation.

[kratmel]

18 minutes ago, Fernando Castro said:

I did not think on testing that... I don't see how it is related but this Is oddly enough to make me stop trying to find a logical explanation.

The SD card is the most unreliable device I've ever dealt with. As a matter of fact, the old small size cards were Flash... I have some even 16MB in size.

Now they are large nand flash inside with a built-in table of corrupted cells. The actual rewriting to the card from the card in the computer makes it possible to check everything based on the CRC code. The PLC always writes to the card in the same way... and in the same memory.

So I actually thought you were reproducing my old experiments with the ABB robot - which I treated every time after reinstalling a new system on a damaged flash disk - it always wrote files to the same damaged location. I renamed the file in the damaged place and wrote a new copy to the disk - everything started.

[Ausman]

I'm with Kratmel's observations on SDs, but I'm curious to see whether b/n your steps 2 & 3 you did an online look at parameters, and also ran an Operands Not Reference query.

Some years back I had a similar thing happen that did not run an SD that sent me a bit nuts, and the only thing that cured it was a number of Init and Resets....not just one.....on a blank program.  I found some elements hanging around with numbers other than 0, even though this shouldn't happen.  Also make sure your data tables are completely cleared during doing all of this, too.

And finally, totally left field, did you try things with a brand new (checked ok) battery?

cheers, Aus

[Fernando Castro]

15 hours ago, Ausman said:

but I'm curious to see whether b/n your steps 2 & 3 you did an online look at parameters, and also ran an Operands Not Reference query.

I did not understood what you mean, could yo please clarify ?

 

15 hours ago, Ausman said:

Some years back I had a similar thing happen that did not run an SD that sent me a bit nuts, and the only thing that cured it was a number of Init and Resets....not just one.....on a blank program.  I found some elements hanging around with numbers other than 0, even though this shouldn't happen.  Also make sure your data tables are completely cleared during doing all of this, too.

Now that you mentiomed that, I tought that the blank project cleared the operands, then on Info mode saw some of them still not "Blank"  So i did the PLC init, is there any other way to clear entirely the PLC. Regarding the the data tables in my experience if not being re-written manually usually keep random values, is there a way to clesr all of them at once?

[Ausman]

On 10/8/2022 at 12:45 PM, Fernando Castro said:

2. I went into info mode and did a plc Init to leave it entirely "virgin".
3. I downloaded the program using Unitronics download manager as I usually do with all my other PLCs.

After no.2  "virgin" many elements may not have been "null".  Numbers may still exist and some operand references might still be hanging around etc. which you have found out yourself.  I had to do a few init and resets in my case.

7 hours ago, Fernando Castro said:

data tables .... is there a way to clesr all of them at once?

I don't know for sure how to clear all data tables areas in one go.  I do manual clearing on the few times needed.  I would think that getting the plc to do a Clear Table on each one in use would be the trick, but as to how to effectively do an Init and Reset on all the memory used for all tables I don't know.  I think it would be a bit tricky given the dynamic memory aspects.   (Incidentally, the dynamic memory bit was my thinking of why you were having trouble in one of the other questions on the forum, with memory not being used exactly the same way and thus old bits perhaps upsetting things) 

@Joe Tauser   , @kratmel ,  @Saragani or others.  Is there a way to fully clear out the entire PLC for sure in one hit?  Without battery removal?

cheers, Aus

[kratmel]

47 minutes ago, Ausman said:

Is there a way to fully clear out the entire PLC for sure in one hit?  Without battery removal?

We have in our hands a Vision PLC that has two separate programmable sections. One of which is data tables.

Regarding the operands and their quick reset with one command, I only see the SB 24 bit.

As for the data tables - here I am not sure about the method of erasing them (perhaps removing the battery for a short time will not help, because I am not sure that after that the controller will initialize this part of the memory).

Therefore, the preparation of an empty project (with guaranteed zero operands) and with a Data Table (with zero values) that occupies the entire size of the RAM memory allocated for this appears to be the most effective and fastest (together with data tables).

I did not set myself the task of experimenting, but I think that the operation of cloning such a project should bring the PLC to a standard appearance.

P.S. With data tables built into the project (placed in FLASH) is another unexplored problem.

[John_R]

We all know that any predefined Data Tables are not downloaded with the program, but rather dumped into the plc/program afterwards.

and so far as Data Tables that are filled in from the program, I am not aware of where they reside in the plc memory, or how to actually clear them.

But always assumed that the Data Table F/B "Clear Data Table" did just that?