Ethernet Cable Grounding

Status
Not open for further replies.
J

jakeparsons03

Member
Location
USA
Occupation
Engineer
We are having a few network issues at the facility I am working at. Specifically we are using Allen Bradley Powerflex drives that keep getting "Net IO Loss" faults on them. My fear is that we are using the incorrect insulation on our CAT-6A cables, however when looking at the datasheet I saw that the cables required grounding. The cables we are using are GenSPEED 10 Cat 6A F/UTP cables. I have never seen a grounded ethernet cable before in my limited experience. Has anyone ever grounded ethernet cables in an IDF rack? Would I just put in a 1RU ground plane below the patch panel?
 
Z

zbang

Senior Member
Location
Roughly 5346 miles from Earls Court
There's nothing to ground on Unshielded Twisted Pair (UTP) however that isn't really UTP, it's really STP and should be automatically grounded at the patch panel if the correct connectors and panel are used.

Before even going there, were the cables tested/certified? Lots of people think they can bend the spec's for Ethernet and it's not very forgiving at those speeds. (You might get 350' out, but don't expect 500 to work; too many people seem to.)
 
J

jakeparsons03

Member
Location
USA
Occupation
Engineer
Yeah they certified the cable. The certifier will throw a fit if anything is even slightly out of spec from what I heard. I don't know if it tests for grounding though. From what I have seen of the connectors they used (metal phoenix contact connectors) I don't think think they grounded the foil jacket though. The switches are industrial, I would assume they would have grounding in them.
 
J

jakeparsons03

Member
Location
USA
Occupation
Engineer
They go from stratix switches to meraki switches. I know the stratix ones are industrial grade. Meraki probably not but I would think you would only want grounding on one side
 
P

petersonra

Senior Member
Location
Northern illinois
Occupation
engineer
I don't see that there's any real benefit to use shielded twisted pairs over unshielded twisted pair in most ethernet applications. People like to overspecify because they think they're getting something better when often they don't understand what it is they are specifying. Shielded twisted pairs use metal connectors.

If you plug them into a switch that has some kind of connection to ground on the plug at both ends this does not seem like a good idea to me. In any case the signal is isolated from ground so it shouldn't make all that much difference.

I don't recall ever using shielded twisted pair cable with power flex drives. What part number are you using for the drive?
 
S

suemarkp

Senior Member
Location
Kent, WA
Occupation
Retired Engineer
How are these wired? Only a cable from the switch to each drive? A cable from the switch to a patch panel to fixed cable to a jack and then a patch cable to the drive?

Many 10G cables are shielded, and the length limit is rather short (less than 100'). For shielding, all cables need to be shielded using patch panels and jacks that have metal around the perimeter of the jack.

Grounding of the shield should be done by the switch unless its a cheap one with no grounded plug, no metal around the RJ45 jacks, and no ground stud. At underfloor patch panel distribution, we'd twist together all the drain wires and put a ground bar in the distribution rack.

Is all power in the room from the same transformer (or transformers that are all grounded to the same thing in that room)?

Do these drives really require cat6A cable and run at 10G? Or did someone just "buy the fastest" where cat5 or cat5A would have been fine? 10G is fussier than 1G.

You could temporarily convert a run to fiber using short cat6A jumpers at the source and destination with a wire/fiber converter and see if your problems go away.
 
P

petersonra

Senior Member
Location
Northern illinois
Occupation
engineer
Net IO Loss??

Are you sure that's what the drive fault is claiming and not net.io timeout?

If you are using managed or semi-managed switches you might want to turn off auto negotiate on the ports talking to the drives and the PLC. If I remember correctly with auto negotiate turned on the port drops out for about a hundred milliseconds periodically to see if whatever is attached to it might be able to run at a higher speed than what it's currently set at. This doesn't really matter most of the time because you will never see it, but an IO link might. However my memory is fairly faulty about these kinds of things so you might want to call the drives tech support people at Allen Bradley first and ask them if this is an appropriate thing to do as a next step.
 
tom baker

tom baker

First Chief Moderator
Staff member
Location
Bremerton, Washington
I agree give AB tech support a call. When I did I&C and had issues, I would read thru the manual, make sure all was wired and programmed OK, and then give tech support a call. When you get to the right group, I they may have had similar calls and know right was the issue is. Its not always your issue.
 
tallgirl

tallgirl

Senior Member
Location
Great White North
Occupation
Controls Systems firmware engineer
jakeparsons03 said:
We are having a few network issues at the facility I am working at. Specifically we are using Allen Bradley Powerflex drives that keep getting "Net IO Loss" faults on them. My fear is that we are using the incorrect insulation on our CAT-6A cables, however when looking at the datasheet I saw that the cables required grounding. The cables we are using are GenSPEED 10 Cat 6A F/UTP cables. I have never seen a grounded ethernet cable before in my limited experience. Has anyone ever grounded ethernet cables in an IDF rack? Would I just put in a 1RU ground plane below the patch panel?
Click to expand...
I've never seen it either.
What is the shield supposed to connect to? Do they have metal plugs which are bonded to the shield, and then supposed to go into the switch with the plug making electrical contact? Given the potential differences in ground from one building to another, that sounds like a disaster waiting to happen.

TL;DR: WTAF?
 
tallgirl

tallgirl

Senior Member
Location
Great White North
Occupation
Controls Systems firmware engineer
jakeparsons03 said:
They go from stratix switches to meraki switches. I know the stratix ones are industrial grade. Meraki probably not but I would think you would only want grounding on one side
Click to expand...
That's correct. "Ground" is only 0V relative to itself and everything within a reasonable distance bonded to the same ground reference.
 
S

suemarkp

Senior Member
Location
Kent, WA
Occupation
Retired Engineer
tallgirl said:
I've never seen it either.
What is the shield supposed to connect to? Do they have metal plugs which are bonded to the shield, and then supposed to go into the switch with the plug making electrical contact? Given the potential differences in ground from one building to another, that sounds like a disaster waiting to happen.

TL;DR: WTAF?
Click to expand...
The shield on shielded patch cables connects to a metal wrap on the RJ45 plug. Likewise, the switch or patch panel should have a metal perimeter around the RJ45 jack which is grounded to the switch. Shielded feed thru's have the same metal perimeter on both sides. Punch down blocks you usually need to pull out the drain wires, wrap them together, and ground them to something yourself.

Higher end switches will have a ground stud. I never used those, as I considered the grounded power cord to be sufficient. Cheap little switches with power bricks don't usually support shielded cables if you want them grounded (unless you only want to ground at one end which could actually be better).

The signal wires themselves don't care. They are differential signals and usually work fine with unshielded triwsted pair. A shielded cable that is grounded at both ends with a difference in ground potential (e.g. separate buildings) is asking for trouble as current will flow on the shield. I would really try to avoid shielded cables between buildings and use fiber instead. Fiber is really easy with ethernet.
 
Status
Not open for further replies.
Top