tworst

Move to NJ series worth it?

37 posts in this topic

At my company, we are bringing in some new lines and have the chance now to RFQ the PLC's.  Our current standard is CJ and CP series PLC's and I am wondering if going to the new NJ series would be worth it.  I don't know what all the advantages or cost differences may be at this point, although it does appear the NJ uses an IDE called SYSMAC, which is a selling point to me already.  I was hoping some members on this forum might have some input on their satisfaction from moving to the newer NJ series PLCs and the advantages it brings.  The bean counters where I work are pretty stingy, so any significant off the shelf cost of the NJ, along with any cost differential due to software, displays, I/O, etc would need concrete justification.  I am looking through the Omron website to learn more about NJ, but there is nothing like other people's experience with a new vs old product to get more confidence in making a decision.

Share this post


Link to post
Share on other sites

I will kick it off with one large advantage of the NJ over the CJ / CP:  motion control.  It is much more simple to program the motion control side of the NJ than it is with any of the CJ motion control modules or the CP pulse train control.  From simple positioning to high end synchronization, the NJ has the CJ and CP beat hands down in the motion control category.  If you have any servos and are not simply making the same type of machine over and over, the motion control programming time savings is substantial.  The EtherCAT network provides a high level of capability along with large amounts of feedback from the servo drive.  The EtherCAT network also offers high levels of performance on the I/O side of things with the NX slice I/O (which is very cost competitive).  You can also integrate your safety control into your EtherCAT network with the NX safety controller that attaches to slice I/O nodes.

Share this post


Link to post
Share on other sites

I'll back Michael's assessment. Our first experience was a very large machine with three sections. Our machine builder wanted to use three PLC's until they learned of the NJ with the NX I/O. We ended up with a very elegant solution where the three sections are connected by only ECAT. Additionally there are close to thirty ECAT nodes of servos, third party I/O, and protocol converters. Using the NJ really streamlined the wiring and complexity.

More recently, we built a machine with eight axis. We used Omron G5 series servos with ECAT. It's almost silly how easy it is to get them connected and moving.

If you're using the CP series, they are probably very small applications. I think it would be very difficult to justify that cost although I've heard that Omron will eventually market NJ or NX in a 'brick' style PLC (MAC) that may be cost effective.

Share this post


Link to post
Share on other sites
1 hour ago, IO_Rack said:

If you're using the CP series, they are probably very small applications. I think it would be very difficult to justify that cost although I've heard that Omron will eventually market NJ or NX in a 'brick' style PLC (MAC) that may be cost effective.

The "Brick" style NJ, the NX1P2 is available now if I am not mistaken.

Share this post


Link to post
Share on other sites

I did notice that the NJ is classified as an automation controller while the CJ is classified as a PLC, not sure what the difference is.

We do use a combination of CJ and CP PLC's, but I was thinking in the case where we are going to use a CJ anyway to upgrade it to an NJ.  Usually, a CJ is the main controller on a line and CP's are then used for low cost I/O.  

Share this post


Link to post
Share on other sites
Just now, tworst said:

Usually, a CJ is the main controller on a line and CP's are then used for low cost I/O.  

If you are going to switch out the CJ for an NJ, I would just switch out the whole thing.  NJ controller with NX I/O. 

Share this post


Link to post
Share on other sites

Depending on your requirements, there are now many models in the Sysmac family.  

If you already know and use the CJ Series extension modules, you can use them directly on a new NJ.  There are 3 levels of CPU, NJ101, NJ301, and NJ501.  They have some special features like direct database connection on some models.

If you are designing a complete new system, NX is even newer than NJ.  But right now, there are only 2 levels of NX, the NX1P and the NX701.

In either case, having EtherNet/IP and EtherCAT built into every CPU, combined with the NX I/O system and its safety controller, makes it a really simple system to work with.

The one thing you won't like is it programs in the IEC standard methods, including ladder and structured text.  No more PLC addressing, it's all tag based (though you can fake the addressing when required).   Keyboard shortcuts stayed the same for quick learning curve, but it's not the standard Omron command set you are used to from CJ.

Share this post


Link to post
Share on other sites

If your current applications uses any motion then the new Automation Controllers are well worth considering for upgrading.

The basic NX1P controllers are also available with 0 axes and are probably the most suited for CJ2 replacements with no motion.

More advanced NX1P controllers add multi-axis capabilities while the NJ's far surpass the capabilities of any existing CJ's.

OMRON tech folks have a transfer utility [though not currently available to end users] that converts programs from the CJ [CxOne] to NJ/NX platform [Sysmac Studio].

I am told it is not 100% perfect but that the majority of the code is transferred OK.

https://industrial.omron.us/en/products/nx1p#specifications_ordering_info

That said, the NJ/NX platforms using Sysmac Studio are the way of the future. Tag based programming coupled with the speed of EtherCat for I/O and servo motion is amazing.

The biggest hurdle is learning the new software, getting to grips with EtherCat network mapping methodology and accepting that tag based programming rather than fixed I/O mapping is the future.

My 2 cents

 

 

1 person likes this

Share this post


Link to post
Share on other sites

Any new NJ series would be used in a new installation.  I have seen a few comments about the undesirability of tag based programming.  I kind of like the idea of using tags instead of memory locations.  What is the disadvantage of using tags over memory locations?

Share this post


Link to post
Share on other sites

Too slow - I program by numbers - far quicker - old DOS guy. Also you do not really know where the memory is allocated I believe - could be wrong though. I just like to know where it all is and then I can just punch in the numbers instead of looking through silly tag names. It has to happen now with me not when I finally find the tag.

Share this post


Link to post
Share on other sites

I agree with you. Bob.  NJ- love the hardware; can't stand the software.:shrug:

Share this post


Link to post
Share on other sites

Must be getting old like me Jay! :-2

Share this post


Link to post
Share on other sites

I know how old you are.  Saw your last birthday.  I quit and you're still doing projects.:clap:

Share this post


Link to post
Share on other sites

Be bored without it! Always take on challenges as well that no one else wants. :lookingaround:

Share this post


Link to post
Share on other sites

If you stick with logical naming conventions, the tag methodology seems faster to me.  Of course, you could use Symbols in CXP in much the same fashion and I am a "middle of the road" -timer.  :-)

Share this post


Link to post
Share on other sites

I actually prefer the tag based.  I don't need to think about where the controller will store my data, I simply give it a name that means something to me.  And if you look at the capabilities created with custom data types like structures, you will never want to go back.

And in Sysmac Studio, the ones connected to physical I/O all show in the I/O Map screen.  If you burn up an output, all you need to do is move the variable name in the I/O Map to another output.  No program code search and replace.

But I have seen some idiots who make variable names like the old PLC addresses specifically so they can put them on a wire label.  Put the dang wire label in the comment and give the device an intelligent name!

Share this post


Link to post
Share on other sites
On 2/15/2017 at 1:47 AM, tworst said:

I did notice that the NJ is classified as an automation controller while the CJ is classified as a PLC, not sure what the difference is.

We do use a combination of CJ and CP PLC's, but I was thinking in the case where we are going to use a CJ anyway to upgrade it to an NJ.  Usually, a CJ is the main controller on a line and CP's are then used for low cost I/O.  

We call that "Marketing Strategy". What does a PLC do again, really, if not being a controller to automate machines? :D

Share this post


Link to post
Share on other sites

I am absolutely obsessed with making clear, consistent, and logical tag names.  Most people don't give it a seconds thought and come up with the first thing that pops in their head and to the point where the tag names become more of a point of confusion rather than helping you understand the code.  I have one line where the offsite programmer just called all the digital inputs on a card switch 1, switch 2, etc.  I just want to scream when I see this.  Anyway, I much prefer using tag names over memory locations as long as sufficient effort is put into the tag names to make them useful.

Marketing.  All they do most of the time is confuse everybody.

Share this post


Link to post
Share on other sites
18 hours ago, BobB said:

Too slow - I program by numbers - far quicker - old DOS guy. Also you do not really know where the memory is allocated I believe - could be wrong though. I just like to know where it all is and then I can just punch in the numbers instead of looking through silly tag names. It has to happen now with me not when I finally find the tag.

If you are comfortable with fixed memory mapping then STICK with it!

WIth Tag based mapping you RARELY have to worry about overlapping memory areas.

IMHO a big +

As Mike Walsh says, " If you stick with logical naming conventions, the tag methodology seems faster to me".

Just a matter of being consistent with your naming strategy.

Furthermore GLOBAL and LOCAL tags allow for more efficient programming and the ability to create Libraries that can be used across multiple projects.

This was one of the key concepts behind IEC 61131-3 and works for me.

Though I still love the CJ2 with CX-One for certain projects, the NJ/NX for more advanced and complex projects is a win win.

 

 

Share this post


Link to post
Share on other sites

With CX symbols, when I have no need to know location of data, I use symbol name and address auto allocation. If, however, I need/want to use physical addresses for some reason, why should anybody deprive me of this possibility? Let ME decide!

Edited by Sergei Troizky
2 people like this

Share this post


Link to post
Share on other sites
2 hours ago, Sergei Troizky said:

With CX symbols, when I have no need to know location of data, I use symbol name and address auto allocation. If, however, I need/want to use physical addresses for some reason, why should anybody deprive me of this possibility? Let ME decide!

Sergei brings up an excellent point as we had this discussion at our plant some time ago. I don't remember the exact reason we needed it but even a Union would not help us because we were using Structures. 

Back on topic. Not sure how old everyone is here but I've been programming (Assembly and Basic) since the mid 80's (in school) and started programming PLCs in the later 80's. I took to 'tag based' programming very well and prefer it in many ways. My only inhibition is being able to access physical memory, as mentioned above. I think this is a rare case.  If your first programming experience is with PLCs then your learning curve may be more difficult. In my opinion it's really just a matter of understanding that your 'tag data types' are pre-defined and understanding how to deal with arrays.

Omron has done a great job implementing the IEC standard although Sysmac Studio could use more maturity. I wouldn't let that stop you from migrating, they update frequently with software improvements. The hardware is the same quality you've seen in the CJ series with more capability.

Edited by IO_Rack

Share this post


Link to post
Share on other sites

Sergei you can still set CJ-style addresses on your variables in the NJ.  The capability is there, you just have to use it.

As for the comment about why take it away?  Think about the years of people overlapping double word data or real number data.  Personally, I am of the mindset why should I waste time making a memory map when it's a PC and it can do it for me?  Does it make one bit of difference in the operation of my program which tiny spot in the chip stores a value?  Nope, as long as it stores it and gives me a way to access it.

Share this post


Link to post
Share on other sites

Those  "CJ-style addresses on your variables in the NJ" are not addresses, but only variable aliases, and must be created manually.

People overlapping double word data or real number data learn on their errors and develop expertise. I am ready to pay such cost to be able to move long integer into regular integer if I want so, or compare values of different representation. PC doing things for you is great, but not if it starts cutting your own abilities. With CX symbols, programmer is a master, with tags he is a slave.

2 people like this

Share this post


Link to post
Share on other sites

Here here Sergei - I like to know where everything is and access it as I wish. Bloody tags are a pain!

Just went through this with a friend of mine - he likes the auto allocation/tag based method - he did not come up the hard way with hand held programmers and DOS software. I showed him what I am doing with a job at the moment with a CJ2M and Ethernet I/O and he was quite astounded at how quickly I was able to write a program and get it running. He does not use Omron normally but was enormously impressed withe CX-P and how quickly I was able to set up alarm routines and the like and he had never seen the function BCNT before - he liked that one. Easiest way for alarms, mutes and the like.

1 person likes this

Share this post


Link to post
Share on other sites

I think we hi jacked this thread, but I still feel like it is important.  Maybe we should move it to another topic. :shrug:

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now