Your Intuition of Interfacing Two Pieces of Equipment.

[Chris Elston 79]

Traditionally, I am sure most of us will probably agree that if you have equipment A, and equipment B and you need to start equipment B, based on when equipment A is ready, you probably would design a form C dry contact closure to interface these two pieces of machinery. Example: A conveyor conveys a part into position, once the conveyor stops, the conveyor signals a dry contact closure to start an auxiliary piece of equipment to operate on part at conveyor position. The auxiliary equipment may even have dry contacts back to the conveyor to indicate "ready" status, "busy" status. Point is, you're using dry contacts. Now in today's age with control logix processors, are interface relays considered stone ages? At what point, if any have you trusted that your Ethernet network is "good" enough to interface two pieces of machinery? For me, I think I am almost there, except when I know if the network goes down and I can't do without that interface signal. For example I rate the following: Micrologix 1000 (conveyor) to a Micrologix 1000 (aux equipment) = I use hard wire dry contacts. Micrologix 1100 (conveyor) to a Micrologix 1100 (aux equipment) = I use hard wire dry contacts. Micrologix 1100 (conveyor) to a Logix Processor (aux equipment) = I use hard wire dry contacts. Logix Processor (conveyor) to a Micrologix 1100 (aux equipment) = I use hard wire dry contacts. Logix Processor (conveyor) to a Logix Processor (aux equipment) = I use networked messaging. (depending) Anyone else "think" like me or have any thoughts? Are you one of those people that if you had two Ethernet processors, that you would use an interlocked message instruction to interface two machines no matter what, like a SLC 5/05 and a Micrologix 1100? Have you had any problems with that or regretted using network messages to interface? Just curious what the industry gurus still do or if you feel you have evolved.

[servicer22 0]

By far my preference is dry contacts. 1. Easy to troubleshoot 2. Easy to setup. (ie tie relay coil to downstream "Auto" lamp) 3. No need for "network troubleshooting". 4. Cost. (hardware and labor) Of course certain processes need more than a "GO/NOGO" signal, but I dont think thats what we are talking about. If you have a network guy looking to justify his excistance, we all know what he would choose. thats my 18 years of 2 cents. Greg

[Chris Elston 79]

I hear ya.

[Ken Moore 23]

With two Logix Processors I would go with the ethernet in most cases, but I would not use messages. I would either use produced consumed tags, or just have one processor monitor the I/O of the second one. If it were two CLX's with controlnet, then hands down I'd use the controlnet network for the connection, and produced/consumed tags.

[Clay B. 1]

I would network but for a different reason. When you use hardwired systems upgrades are a pain. With networoking you can do it without extra wire, just your laptop. FYI... I always use a hub and create one network for the cell that is then sent across our plant network. This way I plugin right where I need to be. To ease troubleshooting I desginate interger blocks for sending and receiving. Example: Machine A Machine B Machine C Machine A send interger block is N16 recieve is N17 Machine B send interger block is N18 receive is N19 Machine C send interger block is N20 receive is N21 Each PLC has these blocks labeled as such so I know exactly where I need to go when checking. Since I starrted using whole blocks I have never run out of room to send info back and forth. I rememeber the day when dry contacts was ALL you could do. I do NOT miss those days.

[Sleepy Wombat 27]

I would do the network thing if the processors are on a network. It is also quite easy to do some simple watch dog timing logic for both ends.

[TimWilborne 108]

I would definitely network I have a really hard time with this argument... Your are networking two PLCs. It doesn't matter the brand, type, etc. The most basic setup hardwired would be the PLC receives an input and sends an output. Basically 4 wires. Troubleshooting can be tough, is it a 24VDC system, a 120VAC system, is the voltage referenced to ground or is it floating, etc. Then you have the hand signals with the guy looking at the one PLC while your looking at the other. With Ethernet you can plug you laptop or a media checker in and know right away whether the wires are in tack or not. Actually, you don't even have to go on the floor, can you ping both devices... So at the bare minimum hardwire would be 4 wires versus Ethernet's 8 wires. I'm not seeing the troubleshooting being any harder. Next argument...our technicians are trained to troubleshoot hardwiring. Ethernet would be too difficult...You don't even need a laptop. The make hand held testers that show whether the connection is good or not and even where the break is in the wire if there is a problem. I think they can handle this...

[rpraveenkum 5]

I prefer with network with producer & consumer. since more signals can be crossedlinked & future expansion is easy with out harware change

[paulengr 44]

Relays are good for about 1 million cycles, give or take. Depending on the speed of your assembly line, this means they fail every few months. PLC's + industrial networking equipment are generally good for 10-20 years. Who is more reliable? When relays fail, they can fail both opened and closed. Networks are inherently self checking. So if there is a failure, which one is easier to diagnose & fix? The trick here is that the onus is on you, the controls/networking guy to make this stuff easy to troubleshoot. If you put it all out there and don't supply idiot lights, it becomes very difficult to troubleshoot. The electricians are going to need diagrams of the network as well as some way to narrow down the search for the bad connection. I put indicators all over the place for this purpose. Don't rely on the "link/activity" lights on Ethernet ports either. By themselves, they might help 80% of the time. They are not a panacea though. The biggest problem I've noticed is that people tend to be very care free with their network cabling, and they don't practice good isolation (speaking in the logical sense...keeping PLC signalling away from general plant networks). Quite often the network stuff is maintained by IT people. With IT, "24/7" networking means that it is acceptable for the equipment to fail once in a while and that a five minute outage is no big deal as long as it doesn't happen too often. Thus, from a controls point of view, IT networks are totally unreliable. If you are going to network, don't rely on theirs. The second problem is that electricians don't understand that networks, almost without regard to speed (RS-232 is perhaps the only exception), are high enough speed that all electrical connections are electrically "long". This means that the wiring is a transmission line...inductance, capacitance, etc., become a major issue. You can't be free to wire things up like you can with relays. Any only wire out of the junk box isn't going to work. When you strip the end of CAT 5E to put it into an RJ-45 connector, don't strip off 5 inches! Once the RJ-45 is crimped on, the edge of the jacket should be down inside the RJ-45. You can't crimp the wires in doors, undo the twists in a twisted pair, or practice poor grounding practices (separating signal and earth grounds). The best way earlyon to train electricians what to watch for is to have training on EMI. Once they have a good grasp on EMI, they will be in a better position to understand the do's and dont's with networking. Once they have a grasp on what you can and can't do with network cables, generally speaking, it will be much more reliable than your old form C relays. Making it over that "hump" is the difficult part.