Upgrade

[Buckaroo 0]

I am sure my situation is not new to most, that's the reason I would like some input from others on how to proceed. Some backround information: I work in a food manufacturing facitlity. Currently we have a small amount of processors networked via a Controllogix bridge with an ethernet module and a DHRIO module. On this DH+ network there are basically 3 nodes I need to be concerned with regarding the near future. Node 1 is CIP (clean in place) 1 system. Node 2 is CIP (clean in place) 2 system. Node 3 is a silo monitoring system. The silo monitoring system is also passing data to an RSVIEW32 application. Now what happens is I get alot of questions of the variety "Can you do.........." or "Can...... be done". Realizing the nature of this facility, nothing is ever an easy task. The layout alone presents many problems with routing wires, cables and the like, not to mention the actual location of everything. Current question I have been asked is to add some I/O to the CIP1 system. This would include VFD, as well as some flow meters and maybe some other I/O that has not been defined yet. The processer (5/04) is located some distance from the location of the desired new I/O. I have started doing some light reading regarding remote I/O and this is were I would like some input. Does it make sense to utilize the 1747-SN and 1747-ASB modules to install remote racks? Or would it be better in the long run to upgrade to a Contollogix based system thereby using remote I/O with that system. I have not used remote I/O in the SLC platform and used it for a couple of very small projects with a Contrologix processer. I do not want to recommend the SLC based remote I/O if it were to become obselete in the near future. I also do not have an unlimited budget so that is a concern. I would like to someday integrate all the above nodes but realize it might take a few years to get that accomplished. I realize I may have used vague or general terms in this post. If there are any questions I would be glad to answer to help the flow of information. So with that said, what would be the best way to proceed? Thanks in advance of any ideas. Tom

[Armadillo852 25]

Have you looked at using 1794 flex I/O? Just a thought.

[Ken Roach 194]

You should be able to get 1746 I/O for another ten years from A-B, and on the aftermarket for a long time after that. If you plan to stick with SLC-500 controllers and don't have a ControlLogix migration strategy, there's no problem using ordinary 1747-SN / 1747-ASB networking and ordinary 1746 I/O, especially if you already have spares. If this is part of a ControlLogix migration plan, I would recommend using 1794 FLEX I/O with the 1794-ASB adapters. When you migrate to ControlLogix in the future, you can switch out the 1794-ASB for a 1794-ACNR15 or a 1794-AENT adapter and run those with the ControlLogix. Run some CAT5e cable along with the RIO cable when you do this wiring, and you'll be a step ahead when it's time to migrate.

[Buckaroo 0]

Thanks for the input. We currently do not have a Controllogix migration plan. This is an area I have been trying to explore with my company. However when I get price quotes for some of the ideas that they come up with I get what I like to call "Deer caught in the headlight" look and then asked if it can be done cheaper. Sure it can be done cheaper but then what happens is I get stuck later done the road when asked if I can now add ........... and can't due to limitations of previous modifications. I am trying to come up with a master plan and would like it to include a Controllogix migration. Again thanks for the ideas. I will investigate more. Tom

[paulengr 44]

I've been right there several times over the last decade. The problem with what you have is that it is very expensive, support is dwindling, and it's really slow. With CLX, the issue is that the first initial install is going to run you about $10K just for hardware/software if you do it "on the cheap". At this point the pricing on SLC and PLC-5 processors is so high that ONE Ethernet processor for either one is the same as the entire CLX chassis, processor, some IO, and programming software. So the first processor swap you do if one goes bad pays for the project. Once you've got one in place, the rest aren't so bad and the initial software price (roughly half of the cost) is already paid for. The processor is VERY different from what you are used to, so there's a big learning curve, too. And although Rockwell does offer a "code converter" for both SLC and PLC-5 that will basically rewrite your code from Logix 5/500 format to 5000 format, the IO doesn't directly convert at all and there's a few fixups you have to do by hand (real time clocks are not the same). This is all to say nothing of the training issues that come up in doing this. Logix 5000 is tag-based and you are MUCH better off switching from the old (N7:0, F8:1, etc.) notation to the new one (CIP_Step, Motor_Amps, etc.). CLX encourages development of lots of small modular ladders (or things that aren't even ladders) rather than just a half dozen very, very large ones. The result is that although the code converters let an outside contractor "convert" one for you on the cheap, it is usually MUCH better simply rewriting all the code from scratch. If you do a good IO checkout job and code design, the debugging stage is usually not any worse with either approach. In addition if you have a PLC-5 (you don't), they make these cool adapters that plug into the CLX IO modules and allow the original 1771-chassis swing arms to attach DIRECTLY to the 1756 IO modules so you don't even have to rewire any IO. With the SLC, you don't have that option. Still, moving all the wiring isn't that bad. My suggestion is to approach this two fold. First, recognize that ultimately the lowest cost system from a maintenance perspective is to have lots of remote IO and not the huge single-chassis based approach because if you go with a single chassis you spend most of your time troubleshooting running wiring out to the IO and back, never mind the sensors, as all the conduit/wire tray gets beat around over time. Going with remote IO eliminates this and results in a lower total installation as well as maintenance cost. I try to put remote IO panels no more than 50-100 feet from the sensors whenever possible. If your IO density is very low (like on a big conveyor system) then this rule gets violated. Your choices are basically to go with DeviceNet or ASI bus, or just put in the long cable runs anyways and put IO panels where it's convenient to maintain them. Although conceptually ASI and DeviceNet have a lot going for them, in practice at least DeviceNet is a real pain to troubleshoot when something goes wrong, so I make it a practice to avoid it whenever possible. That brings up the issue of IO density. Chassis-based IO in the SLC world means ASB modules. This only becomes useful when you have extremely high density IO in one small area, on the order of 200+ IO points. Otherwise, you can use 1794-Flex IO. Flex has two advantages for you. First, the 1794-RIO module works on your existing RIO scheme and talks directly to a 1746-SN. Second, if you later decide to switch over to ControlLogix, you can either leave the 1746-RIO module in place and talk to it with a 1756-DHRIO module, or else switch it out for a 1794-AENT and adopt Ethernet as your IO platform for relatively low cost. Flex IO makes sense for the upgradeability reason. Previously, there was also the 1792 "Standard Block" IO but all of that stuff is now on life support only. This was a low density system (usually 16 points) but included the communication adapter as part of the block (always RIO). The newer equivalents are in the "CompactBlock" family. I'd stay away though because when you upgrade this stuff, the entire module (IO and communication) has to be replaced. The other option is "Point IO". This is advertised as a low density solution and looks almost identical to Siemens IO. In practice, the big advertisement (lower total cost/IO point) doesn't actually pan out in practice so I have no use for this stuff. Another strange option for low density IO is "PLC as IO". Micrologix PLC's are so inexpensive that you can use them outright as IO in low density scenarios, AND be able to do Ethernet (if you desire) that quite frankly, I've put in several of them with no code at all. I just wanted them as an IO module. A non-Rockwell option to consider is going to either Modbus or Modbus/TCP. Rockwell does NOT support Ethernet-based IO for SLC/PLC-5. You CAN do it using Modbus/TCP and/or Modbus which are both standards and there are plenty of third party cards that support it (and curiously the Micrologix PLC's also support Modbus). This can be successful and get you "over the hump" and both reduces costs and gets out of the "RIO" rut. I like cards from Acromag for this, and others have recommended Phoenix Contact for the same purpose. OK...now as to the "getting to CLX" hump...like I said, it's a $10K "nut". No way you are getting there any time soon without a project big enough to buy another PLC. Using Flex IO you can at least have an available upgrade path. So let's look at the PLC/HMI point. Right now on DH+ you are going to run into bandwidth issues (if you haven't already). Upgrading to a SLC 5/05 is prohibitvely expensive (and as I said...might as well go CLX at that point). So here's another option. Buy a Digi One IAP. This little module has 1 Ethernet port, and two serial ports. Plug one serial port into your SLC. Program the card as a bridge. Now if you need to access the SLC serial port, use the second serial port. But you can talk directly to the SLC over the serial port using the card as a converter and access it via Ethernet from RS-View 32. Just tell RS-View 32 that the unit is a "SLC 5/05". It won't know the difference. This lets you upgrade all 3 units to Ethernet for not much money. Individually, you'll be limited to serial port speed (19.2 kbps) but once onto Ethernet, total capacity is now 100 Mbps or more, so total bandwidth increases to 19.2 kbps x 3 = 57.6 kbps, but you can expand your network with no further bandwidth limits. Then whenever you do go CLX, the HMI side of things is already installed (just switch to an ENBT card and talk directly to the processor instead of using the converter). The first "Ethernet" conversion gets your feet wet, and the others are easy enough to do. You can also program via this interface scheme simultaneously so don't be afraid to use Logix 500 to do that (it will tolerate the confusion about what the processor is). It just takes some tricks to get RS-Linx to go along with this (configure one driver per node, rather than one driver and multiple nodes). This is much cheaper than what the AB guys usually try to get you to do (buy a CLX backplane as a "bridge" and buy the processor later). It sounds good but it's so darned expensive that it's just not practical. You might also want to push for the Ethernet option for the drive as well. Depending on the size of the motor, a Powerflex 40 series drive is MUCH cheaper than the PF 70 or PF 750 series drives that the AB reps usually want to push. The PF 40 can be wired to run off analog/digital controls rather than networked, and the COMM-E card is cheap and easy to install and gives you Ethernet-based monitoring and programming. This gives you an excuse to go to Ethernet, and when the ControlLogix upgrade does come, you can just dump all the analog/digital controls and go to direct control over the drive as remote IO via Ethernet. You CAN of course drop in the DH+/RIO card and do the same thing with SLC but you've killed your upgrade path. You can also consider using the "DS1" or whatever the port is called which is essentially a combination Modbus/proprietary protocol port which can be used if you want to go the Modbus route for control, leaving the adapter port free for a different card. The DS1 ports can also be daisy chained if you want to try to drive multiple drives from a single network port. There is one additional issue to consider on your quest especially with what you laid out. With ControlNet and RIO, you MUST shut down a PLC (sometimes with Controlnet there are options but it's limited) and reprogram everything. You can't simply add IO on the fly with the system running. With Ethernet/IP (CLX), you can do just that if it's a 1756 chassis or a drive...which means your system can run 24/7 if you are careful with the design AND add/change IO without actually shutting everything down. This cannot be done with 1794 Flex IO, or RIO, or in many cases, ControlNet or DeviceNet. Hence in plants with continuous systems there is a tremendous advantage to using rack IO and ControlLogix systems. Note also that the cost hit for doing chassis-based IO with ControlLogix is not as bad as doing it with 1746 and 1771 IO. The racks are considerably cheaper and the IO makes up for SOME of the cost for the power supply, but CLX power supplies are still very overpriced in my opinion. Edited 26 Feb 2011 by paulengr

[Buckaroo 0]

I had to read this a couple of times to digest all the information. Thank you for all the input. I have some requests for quotes out with a couple of ways to go. One interesting (read frustrating) thing is when my company gets a quote from an outside source to do a large PLC based project it seems they just rubber stamp it with no real thought. When I am asked to get prices / look into the feasibiltiy of something I get micro managed to death. With this current question I am trying to do it as economical as possible but with an eye towards the future for upgrades. Not an easy task for sure. I will continue to explore options. Again thanks for the advice Tom

[Buckaroo 0]

If I go with a Controllogix migration upgrade what are the pros and cons to Flex I/O vs Point I/O for remote I/O? Thanks Tom

[Ken Roach 194]

In my humble opinion: Bulletin 1794 FLEX I/O Pros: - Mature and stable platform with wide network connectivity, wide product selection, medium density, and third party support. - RIO, ControlNet, and DeviceNet support connectivity to PLC/SLC and ControlNet and EtherNet/IP (including Device-Level-Ring) support connectivity to ControlLogix. - Spare parts may already be available if you have an installed base. Stocked by many RA distributors. Cons: - Mature platform (first released around 1994) may reach the end of its life in 10-15 years. RA has no schedule to obsolete FLEX, but because it's older it is further through its life cycle. - Cost is relatively high if density is unnecessary. - Configuration profiles in RSLogix 5000 are lightly supported. Data types and configurations are oriented toward backward (PLC-2/3/5) compatibility. Bulletin 1734 POINT I/O Pros: - Modern I/O platform under current development by RA. - Very flexible mix of discrete, analog, and specialty I/O. - Core RA protocols supported; DeviceNet, ControlNet, EtherNet/IP. - RSLogix 5000 profiles very similar to 1756 I/O on ControlNet and EtherNet/IP. Cons: - Large numbers of modules required for large I/O count. - High modularity means more contacts/corrosion/vibration/Bubba failure possibilities. - Visually similar to inexpensive third parties, confusing purchasing agents. - No support for RIO - Rapid development means keying/version and firmware issues are more likely. - High variety means low distributor stock levels. I would make my decision based on: 1. Do I have any FLEX installed ? 2. Do I plan to install RIO first, then migrate to ControlNet or EtherNet/IP ? 3. Does my distributor stock the products ? 4. Do I have high-density or low-density I/O panels ? I've used both platforms extensively and my instinct is to use FLEX unless I have low-density, high-variety locations or I'm using the inexpensive 1734D and 1734-PDN implementations on DeviceNet. I did have a customer use POINT with 4-point modules and he stacked up 52 of them in one panel !

[paulengr 44]

[Buckaroo 0]

Ken and Paul, Much thanks for all the information. It is nice to be able to utilize someone's vast epxerience and knowledge. Right now I am leaning toward a Controllogix system where I will install this in a location previously identified for remote I/O. The thought here is I eliminate some of the remote components and still access the 5/04 via the DH+ network. I wil then be able to in the future migrate that to the full Controllogix upgrade if that is what is desired. Also this location affords the opportunity to pick up a fair amount of additional I/O in the future to upgrade some other processes. Interesting, as the price quotes come in and I change the sytem around, the price difference between using the 1747-SN /1794-ASB modules for just adding Flex I/O to the 5/04 processer and then comparing that to a Contollogix upgrade, the prices get in the ball park where I think I can sell them on the Controllogix plan. Again thanks for all the information. It really helps. Tom

[paulengr 44]

An entire CLX processor plus programming software runs about the same as JUST the 5/05 or PLC-5/40E last I checked. So what I've done to sell it in the past is on the fact that the processor is about 1/3 of the cost of a single PLC swap out. So it pays for itself in terms of maintenance after the first PLC replacement. If your plant has good power quality, and you keep your PLC racks in ideal conditions (dry, cool, clean), then you probably won't see many processor changes. If you work in a foundry, steel mill, or mine, the story is very different and although they are still pretty darned sturdy, you can and do wipe one out once in a while. It's a tougher sell but as soon as you get into a semi-substantial amount of new IO (now just 2 or 3 here and there) without new processors and the like, selling it is a serious uphill battle. Another option is CompactLogix. The price difference between full ControlLogix and CompactLogix isn't that much, and AB will try to upsell you obviously. There are some obvious features deleted from it, but it's kind of like the differences between a PLC-5 and SLC....the old PLC-5 platform had a lot of additional features over SLC but the SLC was pretty good in it's own right and costs a little less. And just as it is with them, you can intermix the two like if you want a CompactLogix PLC but want to use a 1756 chassis for IO, except that you can outright access the 1756 chassis without any sort of PLC-5 bit twiddling concerns (going from decimal to octal).

[Lars_01 0]

The cost saving for Compact logix can actually be quite large, depending on the type of sytem that you are using. But you need to be carefull not to save too much. If you have a plant were you only use compactlogix and only program in ladder (And can deal without some other functions) you can buy a cheaper version of RsLogix 5000 (And there are quite a lot to save there, specially if you need a couple of licenses). If the process is divided up in modules or smaller sections with many processors rather than a main PLC, it is probably wiser to go with the compactlogix. What you were talking about sounded like 2 CIP stations and a level indicator / Controller of a tank on one controller. To me that sounds like CompactLogix, but it is hard to be certain about it with this limited amount of information. If there is a main plant PLC somewhere that should probably be a ControlLogix... But the tricky part is to not try to save too much, then it will be more expensive in the end... I have been there and it is hard to get out of it...