Induced voltage on ground wire

D

dummycheck

Member
Location
Texas
Occupation
Engineer
We are having a problem at one of our sites that has left a lot of our engineers, electricians, and others I've talked to about this scratching their heads. Want to run put it out there to see if anyone here has any thoughts.

Here is our setup: We installed a 600A 208V panelboard "PNLBD-1" in a parking lot. It is fed from the building service entrance by parallel runs of 500KCMIL phase conductors and #3/0 AWG GND. The PNLBD-1 feeds two vendor specific 400A 208V EV panelboards "EVPNL-1" and "EVPNL-2" that have branch circuits out to EV chargers. The run between PNLBD-1 and EVPNL-1 is parallel runs each with #4/0 AWG phase conductors and #3 AWG GND (we were limited by lug size the vendor provided us on the EVPNL so had to run parallel sets...) and the same for EVPNL-2. The run to EVPNL-1 is <10ft above grade in a trough since they're located right next to each other. The run to EVPNL-2 is ~100ft underground. The EVPNLs have power conductors and a low voltage signal wires that run from each EVPNL out to the EV chargers.

Here is our problem: The low voltage signal wire is seeing noise on the signal that is causing issues with the EV chargers. We have traced it down to an issue with the ground in EVPNLs. After some investigations and lots of measurement with a multimeter focusing mostly on EVPNL-2, we have found a very odd problem - when we disconnect the #3 AWG ground wires (in the parallel run between PNLBD-1 and EVPNL-2) on both ends so that it is hanging free at both panels but still run in the conduit, we are measuring voltage (~9V) on it with reference to the ground bus/enclosure/AG conduit/etc in PNLBD-1 (one lead on the ground wire and one on the other objects). When we measure the voltage between both ground wires still hanging free on both ends and each in their individual run in the BG conduit, we are measuring ~18V. When we trip the 400A feeder breaker feeding EVPNL-2 the voltage measurements goes away. When we pull the #3 AWG ground wire out of the conduit there is no damage to the wire, came out dry, and it pulled out easily. We made the ground connection by just running the ground wire laying on the asphalt and the noise on the signal wire went away.

We have a survey of existing underground utilities and the only other thing near our underground conduit run is a water line that runs parallel ~5ft away to a fire hydrant located near EVPNL-2.

So for some reason, when the ground wire is run in the conduit below grade we are getting a voltage induced on it and can't figure out why. Any ideas? Things we can test? Other things to check? We are all stumped...

We still want to megger the phase conductors in the conduits just to double check them, but I suspect they are fine.

Let me know if there are other details I can provide for additional clarity.
 
ok, so you're having a bit of induced voltage on your grounding wire. this is common. especially with highly capacitive loads. this is because the capacitance does not allow the sine waves to cancel each other out, as they normally would. the most common example of this i can think of is with simple circuits like electronic ballasts tripping GFI breakers. their electronics do not allow the sine waves to stay 180° apart, and cancel each other out. if you would like to verify this, it can be checked by simply placing an ammeter around all of your current carrying conductors, and taking note of the imbalance. but all of that is besides the point.

your issue is with your low-voltage signal wire. what you do not explain is, where is this low-voltage signal cable installed? inside the conduit with the feeders? buried next to it? is it shielded by either a ferrous conduit or cable shielding, to protect itself from the magnetic fields induced by your unbalanced fields, caused by the highly-capacitive loads? i hope the signal wire is not in the conduit with the feeders.

if shielded, are you grounding ONLY ONE END of this shield? fyi - if you are grounding/bonding BOTH ends of that shield, you're putting that shield in parallel with the EGC. that is a bad idea as it will inducing some current flow on it. attaching only one end, will allow it to dump it's magnetic fields, and provide shielding to the inner conductors, without turning into a parallel path for the EGC to flow across.
 
dummycheck said:
Here is our problem: The low voltage signal wire is seeing noise on the signal that is causing issues with the EV chargers. We have traced it down to an issue with the ground
Click to expand...
So the only issue is on the coms? Did you reclassify a Class 2 or Class 3 communication circuit as a Class 1 circuit so it could be pulled in the same conduit as the power?

I have seen this many times before and it is a classic code vs physics issue. While there ways to make this legal under the NEC the physics of the signal does not care about the Code, and the Code does not care about the data signal, just safety.
By putting those data / signal wires in a conduit with parallel 500KCMIL feeders carrying 400A, you have essentially built a transformer. My guess is the magnetic induction and capacitive coupling are going to create far too much electrical noise for a low voltage communication or a CT circuit to function correctly or without a very high error rate. Even if it is code compliant, the signal integrity will almost always degrade or fail in the same raceway with power wires, even using 600V rated shielded twisted pair cable.

You should try the asphalt test with the signal wires themselves. Lay some signal cable on the ground for a temporary run. If the noise disappears and the chargers start communicating properly, you have confirmation that the proximity to the parallel feeders in the conduit is the problem. It is a quick way to prove that you need a separate raceway for comms. There is likely nothing wrong with your power system its your data system.
 
I'm surprised that the EV equipment manufacturers haven't gone to fiber for their signal circuits to avoid this type of issue.
 
Master Nater said:
your issue is with your low-voltage signal wire. what you do not explain is, where is this low-voltage signal cable installed? inside the conduit with the feeders? buried next to it? is it shielded by either a ferrous conduit or cable shielding, to protect itself from the magnetic fields induced by your unbalanced fields, caused by the highly-capacitive loads? i hope the signal wire is not in the conduit with the feeders.

if shielded, are you grounding ONLY ONE END of this shield? fyi - if you are grounding/bonding BOTH ends of that shield, you're putting that shield in parallel with the EGC. that is a bad idea as it will inducing some current flow on it. attaching only one end, will allow it to dump it's magnetic fields, and provide shielding to the inner conductors, without turning into a parallel path for the EGC to flow across.
Click to expand...
Thanks for the reply, I think I need to clarify the comms wire setup.

There is a module in EVPNL-2 that connects a belden 3090A cable from the module out to the EV charger. So this is part of the branch circuit and not the feeder circuit that we seeing the voltage on the ground wire from. This module references ground from the EVPNL-2 ground bus, so I guess since this voltage is being induced on the ground wire feeding the panel, then it’s getting out of whack.

The belden 3090A is a shielded twisted pair, but the spec from the vendor is to only use one of the wires and they don’t have us grounding out the shield.
 
Didn’t read it all but if there’s a problem of comms in with power to me the simple thing to do would replace in its own race way even drag across parking lot to check and make sure that fixes it.

No matter what wire in a conduit un landed on both ends with phase conductors working will always have some capacitive/induced voltage on it. Many call ghost voltage but a true Low z meter is what’s needed for actually troubleshooting “real power” on power circuits.
 
dummycheck said:
Thanks for the reply, I think I need to clarify the comms wire setup.

There is a module in EVPNL-2 that connects a belden 3090A cable from the module out to the EV charger. So this is part of the branch circuit and not the feeder circuit that we seeing the voltage on the ground wire from. This module references ground from the EVPNL-2 ground bus, so I guess since this voltage is being induced on the ground wire feeding the panel, then it’s getting out of whack.

The belden 3090A is a shielded twisted pair, but the spec from the vendor is to only use one of the wires and they don’t have us grounding out the shield.
Click to expand...
i fail to see what good having shielded cable is, if you don't ground one end. ground/bond one end, not both, and see if this cures your communication issue.

i assume you have a grounding electrode at the EVPNL-2 station before things head out to the chargers, so any static does not have to track all the way back to the PNLBD-1 service enclosure to ground out...? if you do have a grounding electrode, do you know what it's resistance through the earth is?
 
Last edited:
dummycheck said:
The belden 3090A is a shielded twisted pair, but the spec from the vendor is to only use one of the wires and they don’t have us grounding out the shield.
Click to expand...
It has to be grounded at one end to drain the capacitive charge.
dummycheck said:
belden 3090A cable from the module out to the EV charger. So this is part of the branch circuit and not the feeder circuit
Click to expand...
So the Belden cable is run in the branch circuit conduit and its a re-classified as a class 1 system or is labeled as a class 1 system?
Or a separate conduit and class 2?
 
tortuga said:
You should try the asphalt test with the signal wires themselves. Lay some signal cable on the ground for a temporary run. If the noise disappears and the chargers start communicating properly, you have confirmation that the proximity to the parallel feeders in the conduit is the problem. It is a quick way to prove that you need a separate raceway for comms. There is likely nothing wrong with your power system its your data system.
Click to expand...
This ^^^
 
Master Nater said:
i assume you have a grounding electrode at the EVPNL-2 station before things head out to the chargers, so any static does not have to track all the way back to the PNLBD-1 service enclosure to ground out...? if you do have a grounding electrode, do you know what it's resistance through the earth is?
Click to expand...
We did not install a grounding electrode at EVPNL-2. I was under the impression that this would mostly be to provide a path for surges like lightning strikes. Can you tell me more about the static tracking back through the system?
 
tortuga said:
So the Belden cable is run in the branch circuit conduit and its a re-classified as a class 1 system or is labeled as a class 1 system?
Or a separate conduit and class 2?
Click to expand...
I think that's correct, I think there is some way the charger manufacturer is classifying this cable to make it compliant to run in the same conduit as the branch circuit current carrying conductors. I went down the code rabbit hole at one point to double check them on this and found that they're bending the rules a bit, but for the most part it's compliant.
 
Agreed with the group here that the un-grounded Belden 3090A is pointless, so we're going to try to ground it out to see if that fixes the problem. But the challenge I'm getting back on this from the vendor is that they have installations across the country that don't have this grounded out and they're all fine.

And they're bigger concern is that the module is seeing a ground that is at 9V or whatever, so that reference it's referencing against the low voltage signal. I don't know the ins and outs of what this module is actually looking for or reading, so I might need to dig into that with them more.
 
250.32 - Buildings or structures supplied by feeder(s) or branch circuit(s).

(A) Grounding Electrode System and Grounding Electrode Conductor - A building(s) or structure(s) supplied by a feeder(s) or branch circuit(s) shall have a grounding electrode system and grounding electrode conductor installed in accorance with Part III of article 250.

exception: A grounding electorde system and grounding electrode conductor shall not be required if only a single branch circuit, including a multiwire branch circuit, supplies the building or structure an the branch circuit includes an equipment grounding conducotr for grounding the normally non-current-carrying metal parts of equipment.

your distribution EVPNL-2 qualifies as a 'structure' and needs it's own grounding electrode.

if you have not given it a reference to the earth, any static that builds up will only have the path back through your EGC back to your service, until it finds a fault path. lightning is simply a large amount of static. this could very well account for some of your EGC voltage potential.

the way static tends to act, is like a capacitor. it builds it's charge, and continues to build until it's voltage potential between itself and the earth is capable of overcoming the resistance to jump to it's source (the earth). simply stated, static's "source" is the earth. we have to give it a fault to get home.

(insert image of Happy Gilmore yelling "are you too good for your home!?")
 
You must log in or register to reply here.
Top