A little Knowledge Can Be A Dangerous Thing!
jgrassel said:
...
1. If I install something like 1769-SDN into a new AB Compact Logix system, can I then use it to communicate to another existing master system that talks DeviceNet?
2. Could my new CLX system with the card installed be a slave system where I send and receive bits and transmit some analog data values to the master system? (while doing its normal job of controlling the local instrumentation, pumps, etc)
3. Would this be relatively simple to setup a connection between DeviceNET memory locations and PLC controller memory locations such that I can then read/write values from my ladder logic?...
jgrassel,
Q1: Yes
Q2: Yes
Q3: Yes
I always say selections should be application based, and not necessarily preference driven. I'm as a big a fan of Ethernet/IP as the next, but I would not be too easily dissuaded from using DeviceNet, if it's the best fit for your specific application. While Industrial Ethernet, and the Ethernet/IP protocol are fast becoming an industry norm, DeviceNet is still a viable option, fully conforming to the ODVA's CIP standard.
Of course there is the experience factor here. If you feel you can pick it up reasonably quickly, the speed of which depending only on your ability to learn, nothing more, then go ahead. If you feel "out of your depth" then there are many here who can guide you if needs be.
The above being ok to proceed...
If you need to introduce a CompactLogix system that is required to communicate directly with an existing CompactLogix system, that is already using DeviceNet, then I would recommend you use the 1769-SDN and implement Slave Mode, as has been suggested. This makes the Scanner both a Master scanning its own Scanlist, while also producing Slave data which can be added to the existing Master Scanner's Scanlist. This setup is quite possible and fully supported. Multiple scanners can exist on a DeviceNet network, as masters, or slaves. Also, multiple SDNs can even co-exist in the same chassis, depending on which controller is used. For instance, the higher end controllers of the newer 5370 CompactLogix family can potentially have up to 23 SDN on their backplane, either producing or consuming data to the DeviceNet network.
Holmux said:
...I am not a big fan of Ethernet my self, well not when it comes to reliable controls, I have spend to many hours trouble shooting systems where the customer has chosen to use a $ 50 hub from Walmart.
You might need a few extra hours setting up devicenet "true", but when it runs it rock solid...
...Very often i find the choice of ethernet equipment out of my hands, I can come with a list as long with good idea's but it's down to the customer, I agree with you if the connection was made directly between the two PLC's.(cross over)
But if it has to be true the customers network, I would push for another solution.
Why is it that they are able to select a "$ 50 hub from Walmart"?
Why is it so often "down to the customer"?
Why is it that you often find the "choice of ethernet equipment out of (your) hands"?
While I agree that both are robust, reliable and deterministic, once set up correctly, I "think" perhaps what Holmux was alluding to (but failed to clarify?) was the fact that many customers/users know "something", or even "a lot" about Ethernet "equipment" compared to DeviceNet, and so they will, with "apparent expertise", use incorrect equipment, cabling, etc., or tell you, the hired "expert", what can/will/should be used, instead of the other way around...
..."What's the difference anyhoot?...
..."A switch is a switch?"...
..."Cat-6 cabling is your only man for the job, Go Cat-6!"...
..."RJ45 jacks are cheaper by the bucket load, as is my trusty el-cheapo crimper, free with every 50 Jacks! I'll buy that!"
..."Let's go Ethernetting!"...
Whether that was Holmux's intentended point or not, I'll make the point anyway. This is not in relation to which is better or more efficient, or easier to set up, but more about which is more prone to being set up incorrectly.
It is much easy for customers to introduce spurious, run-of-the-mill, office-rated or generic Ethernet equipment, not designed for robust and deterministic industrial networking, into their industrial systems. On the contrary, DeviceNet, and it's equipment, tends to be less familiar to many customers/users, to the point where they are perhaps not even comfortable in selecting which equipment to use, let alone put them to task. It's equipment is mostly purpose built, proprietary, and more difficult to configure for the untrained, so they will tend to stay clear of it. It is
less likely that a customer would as easily introduce DeviceNet equipment that would not be suitable to an Industrial Control System (ICS). Not impossible, just
less likely. Therefore, it is also less likely that they will attempt to instruct you, the "expert", as to which DeviceNet equipment is to be used.
DeviceNet was not designed, developed, or intended for anything other than providing robust, time-critical, deterministic and efficient data handling as an industrial fieldbus network i.e it is very specifically designed for robust industrial use.
A little knowledge being a dangerous thing...
Industrial Ethernet, on the other hand, has the misfortune of being developed from the all to well known IT world of Commercial, or Office grade Ethernet infrastructure. This is the primary reason why users will haplessly intermingle Commercial grade Ethernet equipment and cabling with Industrial grade Ethernet equipment, in either a blasé fashion, or unbeknownst to themselves. This may introduce potential issues, depending on the requirements and the environment in which they are used.
For Industrial Ethernet, the equipment, cable and jacks used should be of a high enough grade to suit their environment. I've mentioned here before, but the M.I.C.E. method of measuring the harshness of the environment that the Ethernet network resides in should be used...
M = Mechanical - Shock, Vibration, Crush, Impact
I = Ingress - Liquid Particles
C = Climatic/Chemical - Temperature, Humidity, Contaminants, Solar Radiation
E = EMI - Electro-Static Discharge (ESD), Radiated and Conducted Transients, Magnetic Fields
Office/Commercial = M.I.C.E. Grade 1
Light Industrial = M.I.C.E. Grade 2
Heavy Industrial = M.I.C.E. Grade 3
This method is what separates the Office/Commercial grade equipment from the Light/Heavy Industrial grade often required on the plant floor. Office/Commercial grade categories of cabling are not suitable for noisy Industrial environments. Standard switches and hubs are likewise not sufficiently immune to noise, or are often unmanaged, when they should be managed. Standard RJ45 jacks are not graded for resistance to vibration, ingress of moisture/dust, etc. These are some of the things that "IT Joe", while proficient in Office/Commercial Ethernet setups, is oblivious to when stepping into the realm of Industrial Ethernet networking.
The question one has to ask is whether the trade-off in using an already well established physical network layer strategy, in an attempt to facilitate its easier implementation, through familiarity, outweighs having created a unique and proprietary networking strategy, to mitigate the potential for mismatched grading within the network infrastructure, while possibly facing initial resistance to its use for fear of the unknown?
There are of course rafts of information on the web to distinguish between the two, or three types of Ethernet networks available, but how many are reading up on it?
Regards,
George
