Have I told you guys how much I hate programming using Step7 today?

[surferb]

This thread's outta control! Can't believe I missed the action until today. And it seems to have calmed down...

To find a bullet to fire at Peter...Vista's great. Microsoft rules. Step7's awesome. Touche! Seems you're a bandwagon Mac user - guess I am too.

 

[Peter Nachtwey]

You can transfer a maximum of 128 bytes of consistent data. So this confirms what you say.

If you had doubts you should have challenged this back in 2006.

http://www.plctalk.net/qanda/showthread.php?t=23942&highlight=profibus+consistent

The PTC ( Profibus Technology Center ) has one of our controllers in the lab to test Profibus DP masters to make sure they can transfer the 128 bytes with data consistency. I think we have one of the few slaves that transfer so much data.

To find a bullet to fire at Peter...Vista's great. Microsoft rules. Step7's awesome. Touche! Seems you're a bandwagon Mac user - guess I am too.

The bullet just bounced off. I have had Mac Mini for 3+ years now so I know what I am getting into.

I have little tolerance for waiting for the computer. When I enter a command the computer should get to work. Most of the time Vista seems to take my commands as a suggestion that it will get around to doing when it feels like it. I use the performance monitor to see what Vista is doing. When I give it a command I expect the CPU to go to 100% unless it is waiting for the hard drive. In that case the hard drive better be transferring data at 100% until it is done. My Vista machine has Raid 0 which means I have double the access speed because I am reading or writing two drives in parallel and still the loading of files was slow and the transfer rates never seem to get to the maximums. I have what should be a hot rod computer but it runs on only 3 cylinders. I have 4GB of memory too so I know it wasn't paging in and paging out all the time. My Mac Mini only has 1 GB of memory. After booting up it thinks there is about 770 MB free. Rarely do I see page outs on my Mac Mini unless I am using Handbrake. You know you are running out of memory when the number of page out is a high percentage of the number of page ins. Vista takes 1 GB just to run. Something is wrong with this picture.

surferb, have you seen?
http://www.islayer.com/index.php?op=item&id=28

 

[LadderLogic]

"You can transfer a maximum of 128 bytes of consistent data".

Was ist "consistent data", Donnerwetter? Persistent data? Contiguous data?

 

[504bloke]

The bullet just bounced off. I have had Mac Mini for 3+ years now so I know what I am getting into.

I have just got round to using Ubuntu at work on a seperate machine, had little time to play yet as been too busy but so far it has run 24/7 since turned on, not fallen over, not had any hiccups and connected straight to the network and internet with no help from me...

XP2 on my work machine has decided the 2 dvd burners i have in it are cd-roms and i cant/havenot time to see what wrong with this, along with the odd occasional lock up its very irritating and it just seems to get slower every day!

Sorry for digressing

Ubuntu btw running openoffice, firefox and thunderbird all at present work at light speed.

 

[Peter Nachtwey]

I don't like the term either.

Was ist "consistent data", Donnerwetter? Persistent data? Contiguous data?

I don't think we have a good substitute for 'consistent data' in the English language.

I would call it 'all at once' or 'all together' data transfers. If someone knows of a English word that means all at once or 'all together' I would really like to know. I looked in the thesaurus and didn't find anything.

Consistent data transfers are important to motion controllers. Consistent data transfers mean the command and parameters are all sent together. This way there is no chance that a command can be executed with the wrong data.

Rockwell's EthernetIP is not consistent. The updates are asynchronous to the PLC scan. It is possible that our motion controller could get an updated command and with old data if the command and data are not updated all at once. This can be a disaster. To fix this we had to add an extra register to the Etherent/IP I/O map. This is the sync register. PLC issues the command with the required parameters. When done the PLC increments the sync register. The motion controller must wait until the sync register changes. Then the motion controller can be sure that the command and data were intended to be used together. We had to do the same thing for some Profibus DP masters because they had no functions like the DPRD_DAT and DPWR_DAT that synchronizes an asynchronous data transfer.

 

[surferb]

The idea is similar to database "transaction" - although the focus there tends to be more on an autonomous set of queries that execute all together or not at all. It's still the closest technical term that I could use. Or maybe "batch/array data transfers", which shoot for contiguous data, but may or may not guarantee all the points you mentioned.

I would call it 'all at once' or 'all together' data transfers. If someone knows of a English word that means all at once or 'all together' I would really like to know. I looked in the thesaurus and didn't find anything.

Looks cool - I'll have to give it a shot.

surferb, have you seen?
http://www.islayer.com/index.php?op=item&id=28

 

[LadderLogic]

I don't think we have a good substitute for 'consistent data' in the English language.

I would call it 'all at once' or 'all together' data transfers. If someone knows of a English word that means all at once or 'all together' I would really like to know. I looked in the thesaurus and didn't find anything.

OK, I see. Yes, this is a very important concept.

The term "data integrity" immediately comes to mind, but I have trouble making an adjective out of it - such as not to be confused with "integral data". A quick search via Google fetches a few examples of "integrity-assured data". That seems to be the most correct rendering, but, unfortunately, it is also a bit too long for English...

 

[cjd1965]

I don't think we have a good substitute for 'consistent data' in the English language.

I would call it 'all at once' or 'all together' data transfers. If someone knows of a English word that means all at once or 'all together' I would really like to know. I looked in the thesaurus and didn't find anything.

What about Simoultaneous ? (maybe my spelling if off)

 

[Brownhat]

I don't think we have a good substitute for 'consistent data' in the English language.

I would call it 'all at once' or 'all together' data transfers. If someone knows of a English word that means all at once or 'all together' I would really like to know. I looked in the thesaurus and didn't find anything.

How about concurrent? Or "en masse"; but that is more of a French word.

 

[rootboy]

No, but I can imagine. I hated LM90 too. It wasn't the LM90 as much as the WASI board that was required. I learned to hate non-standard interfaces starting with the WASI board.

I think that you are thinking of the Series Six, it was the one that required the WASI board. The Series One required the DCU, and the 90's had a serial (RS-422) port (or ethernet). I've no idea as to what the Series Three and the Series Five used.

http://www.cimtecautomation.com/ge_fanuc.htm

I feel your pain since you are trying to fix the mess. From a company standpoint it is a self inflicted wound. I would hold the head engineer responsible too.

Yeah, but what are you gonna do? And I stand corrected, he wasn't fired, he was retired with a sweet buyout package. Hmmm, this gives me an idea....

That is big. What are you doing?

Well, the program controls a Ube injection mold machine, and I've been tasked with getting the PID loops tuned in so that the barrels are up to temperature in an hour rather than two hours.

Because the programmer used the same PID for all of the barrels, I'm having a bit of trouble in figuring out where the Setpoint, Process Variable, Gains, etc, are coming from.

And since VersaPro wasn't giving me every result for the register that I was looking for, I decided to print it to file and just search for it there.

BTW, the file ended up being 1.14 Gigs in size, and itself refuses to load in Wordpad, so that didn't help much.


And I just figured out why VersaPro couldn't find it. The guy who wrote the program named the register that I'm looking for "R19900". The actual register that he is using is "%R23900". I had been searching for %R19900 instead of %R23900!

That's why the references that I came up with made no sense. What a Jerk.


That's ok, I'm on the right track now.

Good, but it isn't quite the same situation. In your case only you are screwed. In my case all the people that want to configure Profibus DP must waste time figuring out how it is done. The first time Profibus DP was configured for our motion controller it took a Siemens engineer 3.5 hours to get it to work. That is way too long. Evetnaully a call to Johnson City, TN got the project on the way.

Ah, Siemens, they're currently on my sh*t list as well.

Tomorrow the Siemens rep and I are going to go over what replacement parts that we need from them, but so far they haven't been able to get me a size 4 (or whatever the IEC equivalent is) starter to replace the one that burned up a week ago.

On a postive note. I think I have got past the steep learning curve. I sent my example to my customer. I didn't implement LD [AR2,#P0.0]'s suggestion about making a record for the status and error bits but that will come. I will be OK as long as I keep using the Step7 enough so I don't forget the magic handshakes and mouse clicks.

Good luck, from what little that I've had to deal with Siemens, I see that it's going to be an uphill struggle all of the way.

 

[Peter Nachtwey]

Most of my response is mindless BS except for one question.

I think that you are thinking of the Series Six, it was the one that required the WASI board.

The 90-70 did at one time too. We don't have a series 6. We have a 90-70 for testing our very old VME hydraulic motion control cards. The 90-70, LM90 and VME board belong with the dinosaurs in a museum. The 90-70 wasn't really VME compatible in that its bus drivers could not sink or source the current that a VME bus master should.

Well, the program controls a Ube injection mold machine, and I've been tasked with getting the PID loops tuned in so that the barrels are up to temperature in an hour rather than two hours.

This is the question. What makes anyone think that this is a tuning problem? If the control output is getting close are at 100% there is nothing you can do to heat the barrel up faster except to get more or more powerful heaters. Unless you have the gains very low and the control output is at about %50 at max then you will not get double the heat rate. You should start another thread because it is a valid problem and question.

BTW, the file ended up being 1.14 Gigs in size, and itself refuses to load in Wordpad, so that didn't help much.

Hopefully a good junk of that is memory image.
That is too much code for an injection molding machine. The injection part can be done in a day. Tuning the temperature PIDs just requires a little fussing with the gains but 20 or so zones shouldn't take that many rungs.

And I just figured out why VersaPro couldn't find it. The guy who wrote the program named the register that I'm looking for "R19900". The actual register that he is using is "%R23900". I had been searching for %R19900 instead of %R23900!

That's why the references that I came up with made no sense. What a Jerk.

Aren't there symbols or tag names.

That's ok, I'm on the right track now.

I think I am too.

I took a break from the Step7 today. I worked on writing example programmers for our controller. It was like goofing off by comparison.

 

[Apel]

.

Now my symbols aren't enabled and I don't know why.


Try to use key combination:Ctrl+Shift+Q->with sysblol/without symbol

 

[Peter Nachtwey]

.
Now my symbols aren't enabled and I don't know why.

I have partially figured out my problem. The SFCs don't come with labels. I assumed they did because the SFC have IN0 and IN1. The on-line documentation had LADDR and ANY. We are just supposed to know that IN0 is really LADDR. I was trying to get the LADDR label to show up in my program.

Try to use key combination:Ctrl+Shift+Q->with sysblol/without symbol

I have that working I think. That was mentioned above posts.

 

[S7Guy]

I have partially figured out my problem. The SFCs don't come with labels. I assumed they did because the SFC have IN0 and IN1. The on-line documentation had LADDR and ANY. We are just supposed to know that IN0 is really LADDR. I was trying to get the LADDR label to show up in my program.

Actually, they do come with labels. The only time I've seen them without labels is if someone uploaded them from the PLC into their project. Have you tried deleting them from Simatic Manager, and then calling them again in your code? The should then appear in your block folder again and the the parameter names should show up. If they don't then it's possible that your SFC library is corrupted, or maybe someone uploaded the SFCs into your library.

 

[Apel]

Dear Mr.Peter AND ALL,

"The SFCs don't come with labels. I assumed they did because the SFC have IN0 and IN1. The on-line documentation had LADDR and ANY. We are just supposed to know that IN0 is really LADDR. I was trying to get the LADDR label to show up in my program."

All SFC from S7 library.So you can get nice help from simatic manager.Just Select any SFB/SFC/or any block that belongs to library and "PRESS F1".THEN YOU ARE HAPPY.