Separate ground rod?

Status
Not open for further replies.
W

wireman

Inactive, Email Never Verified
I'm involved in the classic conflict between electronics people and adhering to the NEC's grounding rules.

Here's the problem:
Our customer has an Ethernet communication problem between several PLC's. The PLC equipment is located in several different panels throughout the plant. The PLC rep feels it might be a noise problem wants us to drive a separate ground rod near each panel and run a ground wires from the panel to it.

I'm not disagreeing with him that there is a noise problem but I know that driving separate ground rods is not correct. I feel what should be done is to make sure the existing system ground is solid and then pull separate ground wires back from each panel and attach them properly. What are some specific sections of the code that will point this out?
 
Re: Separate ground rod?

250.96 and 250.46(D) should get you started.
There are alot of other Article s that say you can't do it their way.

Make that 250.146(D)

[ January 27, 2004, 07:09 PM: Message edited by: russ ]
 
Re: Separate ground rod?

IMHO driving ground rods and connecting them to the panel or anything else is performing electrical work and falls under the NEC AND should be performed by a qualified or licensed electrician.

Mike P.
 
Re: Separate ground rod?

Look at the ethernet cable run lengths - in a largish plant environment its not too hard to exceed to specs and get signal problems - particularly if they're trying to run at higher speeds like 100 or 1G.

Also, what kind of cable is being run? Unshielded in a factory environment would be asking for problems.

If the stuff is long and unshielded, a quick fix might be to simply run the gear at 10mbps if the application won't suffer from slower speeds. The slower speeds allow longer run length.
 
Re: Separate ground rod?

I think there getting what we call a ground loop.
This is the same noise that causes a 60hz hum in audio equipment. with Ethernet you have a connected system between multiple power sources and if the grounding on any of them have any current via the grounding this current will share the path of the Ethernet wire ground ref. the best way to avoid this is to use optical connections like fiber optics but to localize the source of this current it will take a lot of current tracing to locate where the current is being drawn on the system. another way is to use a balance power supply to isolate the grounding. The NEC allows for this system in article 647 of the 2002 NEC the other way is to use a UPS system that always runs the load from the inverter (isolated) and all the utility supply's is the makeup power for the batteries. But just driving ground rods will have no effect on the noise problem. To check for current on the Ethernet cable just put a clamp type amp prob around the Ethernet cable (cat5) if it shows more then a .250 of an amp you will have the noise. All it takes is to have one receptacle with the neutral touching the grounding conductor or conduit to cause this. and/or a sub-panel with the grounds and neutral not kept separate. this allows current of the neutral to share the return path with the grounding. In an industrial plant you might have to turn off breakers to isolate which circuit is causing this current then go through this circuit with a fine tooth comb to find what is causing this current. Any questions just ask.
 
Re: Separate ground rod?

Tony
Unshielded in a factory environment would be asking for problems.
Click to expand...
Ethernet cable (cat5) is a twisted pair cable and is self shielding as with each twist the the offending signal is placed 180 deg out of phase and is canceled. this is why it is now used for phone wires as you wont have cross talk between the different pairs. then to top it off all four pairs are in a full twist. If shielding was required then there would be separate shielding between the pairs that Carry's the clock signals and the pairs that Carry's the data signals.

[ January 27, 2004, 08:21 PM: Message edited by: hurk27 ]
 
Re: Separate ground rod?

I also forgot to mention that using a single point ground reference is also a good way to avoid this but would be next to imposable in an industrial environment. as this would require installing a insulated green grounding conductor in the conduit with the circuit feeding all the PLC's and having the PLC's fed from one source, and by using isolated grounding receptacle's.
 
Re: Separate ground rod?

First ethernet is a balanced signal transmission method, and ground is not involved (sorry wayne). So besides from violating code, driving seperate grounds rods is not going to solve any problem, in fact it could make it worse.

tonyi and hurk have hit on the two solutions. Of the two, optical cable and modems is the best solution. By using optical modems it does not matter is the e-net is 10-Base-T or 100-Base-T.

Shielded cable might work and keep common mode noise from being injected (EMI), but to make it work correctly every PLC would have to be supplied from a common transformer/UPS/etc employing a single ground point system method. I doubt all the PLC's are arranged this way.
 
Re: Separate ground rod?

What you want to reference for these gentlemen is 250.6 and 250.54.

250.6 Objectional Current over Grounding Conductors. Paying particular attention to (A), (B), &(C).

250.54 Supplementary Grounding Electrodes.
these grounding electrodes are permitted, as long as they are bonded to the building grounding system. You are not permitted to install these supplementary groundrods and isolate them from the system grounding.

Pierre
 
Re: Separate ground rod?

Dereck
Look like I'm doing some learning here LOL :eek:
A while back I had a problem with my own network here at home I have 6 computers and the one in my bedroom would loose connection with the hub. after checking a few things I was about to leave and I turned off my TV set just after I did this AOL IM jumped up on the monitor this let me know that it was now seeing the hub. So I turned the TV back on and again I lost the connection. After some deep thought, I pulled the receptacle the TV was plugged into it was OK but the next one down the line the ground wire was touching the neutral. I shoved the ground back in the box and tried the TV again and it didn't bother the connection this time. so I just figured that the Ethernet used a ground reference through one of the pairs. Maybe the current on the computer chaise was the cause.? :confused:

[ January 28, 2004, 02:08 AM: Message edited by: hurk27 ]
 
Re: Separate ground rod?

Thanks for providing all the code references and other input. I'll start reading up on them.

Just to add a little more info on the subject, the PLC rep has pointed out several things to check and the grounding was only one of them. I'll let you know how things turn out.

Thanks again.
 
Re: Separate ground rod?

The twist canceling effect is nice in theory, but in practice it doesn't always cut it.

At the IBM Boca plant we had miles and miles of shielded ICS token ring cable. The stuff was silent as a crypt. At the slower speeds, 1000+' runs through manufacturing areas were perfectly fine.
 
Status
Not open for further replies.
Top