DeviceNet and it's future ? Would like to learn more about DeviceNet.

I have been in the industrial automation field for a long time but there is so much I still have to learn . One is understanding , troubleshooting , DeviceNet communication . Is DeviceNet ( or others ) being phased out , do you think ? If so , what would be replacing it ? Would you folks have any good educational links that would explain understanding and troubleshooting DeviceNet? Also , I understand that there are some good network tools to test the DeviceNet networks and cables , such as Woodhead , etc.

Thanks in advance

Technote 44883 "DeviceNet Book of Knowledge".

I'm not sure DeviceNet is being phased out, it's still pretty prevalent so good to learn. I would not use it for Drive netowrks at this point. Ethernet/IP for that. If you need a field bus for IO DNet is still an option. I personally have been leaning more towards AS-Interface.

It's my understanding, if there is a problem w/DNet, it's almost always an installation issue. It is also high risk when adding new components to the DNet scanner. I've heard stories of plants being down for hours/days because someone didn't upload/save the DNet configuration and they downloaded a new configuration to add a device only to wipe out entire drive networks. First time I was adding a drive to a DNet network I was told many times over to upload and save, and do it multiple times. Thankfully I didn't it right the first time.

Keeping track of mapping can be a bit of a hassle as well, so that can trip people up too.

Every new system I see with a Rockwell PLC these days is using EtherNet/IP instead of Device Net. That's the direction things are going. Obviously there is a huge DNet install base, and so it'll continue to be out there for a long time.

Devicenet and its* future

I believe it will still be around for quite some time, because it is robust, if not fast, and has a massive installed base already.

RoTaTech said:

I believe it will still be around for quite some time, because it is robust, if not fast, and has a massive installed base already.

Click to expand...

DeviceNet isn't fast. It is half the speed of CanOpen. The packets are small and half the packet is overhead.

Peter Nachtwey said:

DeviceNet isn't fast. .

Click to expand...

My comment was mean to say that it *is not fast.

DeviceNet was a good solution for the problems of its time, but that time has passed. The main issue was that in industrial networking, you were highly likely to run into a situation where the network cabling was going to be in the same space/raceway as power wiring, so the insulation would need to be rated for 600V even though the network voltage was only 24VDC. DeviceNet cabling solved that problem. It also solved some noise immunity problems by "brute force" signals based on current rather than shielding, which created issues of voltage drop, leading to a somewhat complicated system of trunk cables, branch cables and maximum distance limitations to power supplies. Once you understood all of that and got the right design tools it wasn't hard to implement, but the learning curve was steep for beginners, often resulting in costly mistakes. Still, it was the best solution for a host of other industrial networking issues for a long time. In Europe, the same issues had lead to Siemens developing Profibus, but that never got a good foothold in North America. There are some installations sure, but nowhere near what Dnet has.

What has changed all of that now is the development and release of good quality Ethernet cable that has 600V rated insulation. Ethernet /IP is the Ethernet protocol standard released and supported by the same people in charge of Devicenet standards, the ODVA (Open Devicenet Vendors Assoc.). If you are going to start learning more about industrial networking, I would recommend going that route. Devicenet is not going anywhere for a long time because, as mentioned, the installed base in NA is huge. But the future is Ethernet, and most likely Ethernet /IP.

That said, because the installed base is so big, and because most people will be migrating to Ethernet, there is some intrinsic value in learning all of the complexities of Dnet, because as my generation ages out if the workforce, you young whippersnappers will need at least a few people who understand it 20 years from now...

Hey! you kids get off of my lawn!

jraef said:

That said, because the installed base is so big, and because most people will be migrating to Ethernet, there is some intrinsic value in learning all of the complexities of Dnet, because as my generation ages out if the workforce, you young whippersnappers will need at least a few people who understand it 20 years from now...

Click to expand...

Hardware obsolesce planning and migration will be huge over the next 10 years. Decoding Dnet as part of a migration path will be a necessity. I'm in a facility that has loads of Dnet, and as we migrate away from PLC5s, update our MCC panels any Dnet is getting replaced with Ethernet/IP. I'm commissioning a replacement VFD panel as we speak. Field IO will stay until a major equipment change provides funding otherwise. Basic IO over Dnet is hard to justify replacing, especially if you have to go through any quality validation process.

We have loads of valves controlled over Dnet, not sure what that strategy will be moving forward.

Paully's5.0 said:

...

We have loads of valves controlled over Dnet, not sure what that strategy will be moving forward.

Click to expand...

Probably IO Link once it gets fleshed out a little more. There is a lot of synergy going on between IO Link and Ethernet/IP already.