ControlBoss for DeviceNet problems

Our DeviceNet network is wired up using ControlBoss cabling.

Recently we have had a number of problems with either the CAN-Hi or CAN-Low opening up in the cabling. This can be proven by removing one of the terminating resistors and checking the resistance of the CAN leads. If you get 120 ohms, you are good to the remaining terminator. If not, work your way back until you find the break. Unplugging the connector and then plugging it back in fixes it for awhile.

So ideally, I would get rid of DeviceNet altogether, but that ain't going to happen anytime soon. Alternatively, I think that going with fiber would be nice, but the price of the media converters is outrageous. If any of you know of a reasonable source for converters, I am all ears.

What about thick cable in place of the ControlBoss? Adding repeaters as necessary I suppose…

What do you guys use, and are you having problems with it?


Thanks,

John

Is this a high-flex application, or something else where the wires are getting damaged ?

Are the open circuits occurring at the connectors, or actually in the middle of a network cable segment ?

I've used mostly A-B and Turck cables for DeviceNet over the years, never ControlBOSS.

Ken Roach said:

Is this a high-flex application, or something else where the wires are getting damaged ?

Click to expand...

No, and little to no vibration. Some of the bad connections have occurred over ovens, where it gets a bit toasty though.

Ken Roach said:

Are the open circuits occurring at the connectors, or actually in the middle of a network cable segment ?

Click to expand...

At the connectors, with the exception of one failure, the CAN-Low wire had gotten pushed out of the bulkhead connector at the panel somehow.

Ken Roach said:

I've used mostly A-B and Turck cables for DeviceNet over the years, never ControlBOSS.

Click to expand...

No problems? It might be the way to go.

Are the connectors frequently plugged and unplugged ?

I have one Turck cordset that gets unplugged once a day, and the operators are really harsh with it. The locking nut came undone and they just kept twisting it until the wires inside the cabinet twisted up and tore out of their terminals. I just remind the operators to be gentle, and replace it when they manage to break a pin or damage the knurled ring so it won't engage the threads.

There's not much you can do when operators abuse the equipment, except put it on a periodic inspection and replacement schedule.

That's very interesting that it's one specific pin that gets pushed out. I wonder if the assemblers put them together in a way that damaged the insertion latching tabs for that particular pin, or if the connectors are misaligned when being installed so that that pin gets pushed.

The cordsets I use are rubber and are probably not suited for high-temperature applications above ovens. All of mine run in ambient temperature or in refrigerated environments.

Ken Roach said:

Are the connectors frequently plugged and unplugged ?

Click to expand...

Only when the DeviceNet goes down, so yes.

Ken Roach said:

I have one Turck cordset that gets unplugged once a day, and the operators are really harsh with it. The locking nut came undone and they just kept twisting it until the wires inside the cabinet twisted up and tore out of their terminals. I just remind the operators to be gentle, and replace it when they manage to break a pin or damage the knurled ring so it won't engage the threads.

Click to expand...

No, that's not happening here, the connections are never touched unless there is a problem.

Ken Roach said:

That's very interesting that it's one specific pin that gets pushed out. I wonder if the assemblers put them together in a way that damaged the insertion latching tabs for that particular pin, or if the connectors are misaligned when being installed so that that pin gets pushed.

Click to expand...

It was just the one, and it became obvious when my scope showed that the CAN-Low was flatlining.

Ken Roach said:

The cordsets I use are rubber and are probably not suited for high-temperature applications above ovens. All of mine run in ambient temperature or in refrigerated environments.

Click to expand...

Fortunately, most of our runs are away from the ovens. So it should work for everywhere else without a problem.

A network where operators go plugging and unplugging connectors 'whenever the network goes down' sounds like it has poor workmanship from the beginning.

I've installed hundreds of DeviceNets and most of them are reliable right out of the gate, so much that the usual complaint is 'we don't remember how to fix anything because we haven't touched it in years'.

But the ones that had damaged connectors and loose wires and poor design, well, those are the ones where the bad effects continue for years.

I had one installation where they didn't just tear out the DeviceNet, but also replaced all the A-B controllers with Siemens controllers. They proceeded to hire the same contractor to pull and install all the Profibus cable.

Guess what network they have persistent and intermittent problems with now ?

Ken Roach said:

A network where operators go plugging and unplugging connectors 'whenever the network goes down' sounds like it has poor workmanship from the beginning.

I've installed hundreds of DeviceNets and most of them are reliable right out of the gate, so much that the usual complaint is 'we don't remember how to fix anything because we haven't touched it in years'.

But the ones that had damaged connectors and loose wires and poor design, well, those are the ones where the bad effects continue for years.

I had one installation where they didn't just tear out the DeviceNet, but also replaced all the A-B controllers with Siemens controllers. They proceeded to hire the same contractor to pull and install all the Profibus cable.

Guess what network they have persistent and intermittent problems with now ?

Click to expand...

The workmanship is fine as far as that goes. And no one unplugs anything here unless it is Maintenance. The vast majority of our connections have not been touched since installation (2008). The only reason why they are getting unplugged now is to isolate the open connection in the network. How else are you going to do it?

I do remember that we had a lot of trouble with the ControlBoss connectors in the beginning because they were so heavy that they would deflect downwards and open up the DeviceNet pins. The "solution" for the problem was to tie-wrap them up.

The only issues concerning the installation of ControlBoss at our plant are:

1) We should have used 90 degree connectors instead of straight connectors to prevent the deflection problem. But the tie-wraps (as hokey as they are ) should have fixed this.

2) Due to the nature of using factory cables, we tend to have a considerable amount of cable wrapped up in the cable tray. For example, places where a 250 foot cable would do, might have received a 350 foot cable.

3) ControlBoss can be used to run 120VAC in the same cable as DeviceNet. We do this in a few places, but I don't agree with the policy.

Other than that, I can't see anything wrong with the installation.

It happened again right before lunch, and this time I went out to "strongly" suggest that Maintenance open up the connection and re-seat it, since then we haven't had any additional problems.

It might just be that opening these connections, spraying them down with contact cleaner, and then re-connecting them needs to be part of the PM plan.

Well, Maintenance is out there replacing the Tee and spraying contact cleaner on it. We will see how it goes...

I misunderstood your comment about 'only when the DeviceNet goes down'; you were talking about isolating the open circuit, and I was inferring that 'take it apart and plug it back in' was a frequent troubleshooting or 'resetting' technique that was causing increasing wear on the connectors.

Large lengths of coiled cables might be a concern in terms of crosstalk, especially if they are coiled with power cables. If you have a trunkline with a couple hundred extra feet inserted here and there, you might have exceeded the recommended trunk length without knowing it.

Does ControlBOSS actually recommend running 120V AC in the same cable as the DeviceNet signal wires ? That's a very basic violation of the principles of signal noise reduction.

Ken Roach said:

I misunderstood your comment about 'only when the DeviceNet goes down'; you were talking about isolating the open circuit, and I was inferring that 'take it apart and plug it back in' was a frequent troubleshooting or 'resetting' technique that was causing increasing wear on the connectors.

Click to expand...

No, only when it goes down (which is frequently enough). And only if I see that I have lost a terminating resistor. But I think that we are going to make taking it apart and spraying contact cleaner an annual event.

Since we talked, I worked on a DeviceNet problem in Body (not my normal stomping grounds). It was going Bus Off every so often (probably about 2 - 4 times an hour). We found that the 24 VDC was actually 22.8 VDC at the block that was giving us trouble, so we cranked up the power supply and it quieted down.

22 VDC shouldn't have given us that much trouble, so I may be out there again next week...

The block that was giving us trouble was a Guard I/O (1791 series I believe). It had already been replaced, which is SOP for this block because of its failure rate. Although I suspect that the block that got trashed was okay.

Ken Roach said:

Large lengths of coiled cables might be a concern in terms of crosstalk, especially if they are coiled with power cables. If you have a trunkline with a couple hundred extra feet inserted here and there, you might have exceeded the recommended trunk length without knowing it.

Click to expand...

Yup. When I see a particularly bad one, I will get the guys to replace it with a shorter cable. Since we are using ControlBoss cable, I don't have the option of cutting it to a reasonable length and splicing on a new end.

Ken Roach said:

Does ControlBOSS actually recommend running 120V AC in the same cable as the DeviceNet signal wires ? That's a very basic violation of the principles of signal noise reduction.

Click to expand...

I doubt it, and I agree with you 100%. I'll have to see where we do this and have 120 VAC run in separately. We don't do it everywhere, so it might not be that larger of a contributing factor.

Thanks.