Ground Overcurrent Problems

Status
Not open for further replies.
R

Richard Franklin

Member
Perhaps someone can help me by explaining about (or) the remedy for ground overcurrent for a Power Distribution Unit (PDU). My past experience in this type of problem was as follows: With single phase circuits the amperage load returning on the neutral conductor must match the same amperage on the hot phase conductor, or there is a leakage to ground causing a Ground Overcurrent situation. This can cause an alarm on monitoring equipment when the preset load is reached. Thus the problem I am identifying/repairing in following the work performed by others.
The PDU is a Separately Derived System with a 480 volt Delta and a 120/208 volt Wye Step Down Transformer, having three 225 ampere-42 circuit loadcenters. The PDU itself is used for Data eqipment within the area. The PDU has a system monitor for Voltage, Amperage, KW, Ground Overcurrent, etc. The alarm threshold for Ground Overcurrent is set at 5 amperes and currently the Ground Overcurrent displayed is 9.4 ampere. When a ampere reading is taken on the Grounded or Neutral Bussbar, which contains the grounding electrodes and the grounded conductor bonding point the reading is 8.9 ampere. From one of the PDU loadcenters there is a 100 ampere breaker feeding a 3 phase-4 wire with ground 120/208 volt sub-panel, using 2 AWG-CU THHN conductors with a distance of about 100 feet. The amperage is as follows: A phase-15.9, B phase-20.0, C phase-8.9, and the Neutral is 32.8. On the grounding conductor there is a ground overcurrent of 4.1 ampere, returning back to the PDU. This sub-panel has branch breakers which supply single phase 110 volt duplex receptacles, using 10/2 MC cable as the wiring method. The utilization equipment for the loads are Flat Screen LCD Televisions, which use a two wire male cord connector (no grounding conductor). The televisions have a power supply which reduces the supply voltage from 110 VAC to 13.5 volt DC and the load on each LCD television is 0.5 ampere.
When reading amperage on each brach circuit grounding conductor supplied by the Sub-panel there is Ground Overcurrent present. It appears that as more LCD Televisions are installed on each circuit, the Ground Overcurrent increases. When the grounding conductor is lifted from the Equipment Grounding Bar of the Sub-panel, there is little or no change of Ground Overcurrent readings at the PDU Loadcenter Monitor. I contacted the manafacture of the LCD Televisions and was advised that this Ground Overcurrent is not the fault of the manafactures product being used.
Could this Ground Overcurrent be caused by "Harmonics on the Neutral(s)" of the circuits? Could this be caused by "Inductance" of the MC cables in close proximity of a higher voltage of 277/480 volts AC? Could there be other Grounding issues not addressed upon the initial installation process? Other than identifying each problem circuit cable and tracing each Ground Overcurrent to achieve a zero ampere reading, does anyone have a suggestion how to approach this problem?
Thank you for your help.
RMF
 
After reading that you've got a bad neutral - somewhere.

I realize old buildings can be a challenge but they're present and will always be there. Pull a socket and measure neutral - ground when that circuit is under load.
 
I'm not sure what the problem is, but Marshall has a good place to start.

What are the voltage readings from line-to-neutral on the different phases, at the source, at the panels, and at the outlets?
 
The Equipment Grounding Conductors and bonded metal can easily form rings and loops. These rings and loops will pick up stray magnetic fields and develop induced currents. The currents that you measured on the EGCs could easily be caused by inductance with such stray fields.

However the SDS power supply should only be bonded to the EGC system at _one_ point; it should not be part of a loop. So any of the induced currents above should not show up on the Neutral to Ground bond, and should not show up as Ground Fault Current.

I concur with the others that you are looking for a neutral connected to ground somewhere.

Since you are using a current clamp to take measurements, you might consider measuring the net current on the various circuits. Take all conductors that are part of a single circuit (a neutral and all hots that share that neutral) and put them together in the clamp. The net current reading should be zero. Any current measured is imbalance current that is going over to ground or to another circuit.

-Jon
 
Follow up for Ground Overcurrent Problems (Reply#1)

Follow up for Ground Overcurrent Problems (Reply#1)

Thank you everyone for replying to my posting today...

For the follow up to the Ground Overcurrent Problem I checked the voltage, current and power output on the PDU.
The readings are as follows:
Input Voltage:
A-B/477 VAC
B-C/480 VAC
C-A/483 VAC

Output Voltage:
X-Y/213 VAC
Y-Z/214 VAC
Z-X/213 VAC
X-N/123 VAC
Y-N/123 VAC
Z-N/123 VAC

Output Current:
X-73 AMP
Y-87 AMP
Z-64 AMP
N-60 AMP
G-9.4 AMP

Output Power:
KVA-27
KW-24
PF-.86
FREQ-60.0
%LD-12

The voltage readings for the sub-panel are as follows:
A-B/212 VAC
B-C/210 VAC
C-A/211 VAC
A-N/123 VAC
B-N/122 VAC
C-N/123 VAC

This building is about 30 years in age, brick and concrete construction and has a large floor area with 11 floors. There are several Main Services and several Main Distribution Panels with sub-panels to supply the power.

This PDU is located on the first floor and the sub-panel is located on the third floor. "Jon" may be on the right track by suggesting measuring the current readings for the circuits being used. Taking the net current readings at the sub-panel for a "zero" reading is not possible.

Today I went to several floors following the path of the conduit and junction/splice boxes. I seem to have found a potential problem checking the branch circuits on the sixth floor. When clamping the Amprobe onto the hot and neutral conductors together of a branch circuit, the reading was not "zero". In fact the measured load on this circuit was 5.3 ampere. "Jon" in your message you mentioned that the imbalance current is likely going over to ground or to another circuit. Is there a second test procedure which would likely confirm these findings with in the branch circuits for the sub-panel?

I checked the PDU bonding point and there is only one system grounding point. I also checked the sub-panel and the neutral and grounding bars are seperate points. If this is inductance induced currents, is there a way to "cancel out" the stray magnetic field that is producing this Ground Overcurrent condition?

Thank you again for your help...
 
Last edited:
If you find a circuit that has significant imbalance current, then shut that circuit off and see if this changes the ground current reading at the PDU. Also try to measure net current on the circuit with the circuit off.

Remember that when you have multi-wire branch circuits (multiple hot conductors sharing a neutral), you need to check all associated conductors _together_.

You can buy 'clamp on' amp meters where the 'clamp' is actually a flexible cable that you can stretch around entire feeder sets.

I would not worry about any 'magnetically induced' currents in the EGC. These cannot enter the ground bond wire unless there are _multiple_ connections between neutral and ground. So you need to find the neutral-ground faults anyway.

As a formatting request, please use paragraphs to separate your various thoughts. It is very hard to read what you have written.

-Jon
 
There are several ways to locate this problem.

1. Perform zero sequence readings on each branch circuit (if possible) use a True RMS amp clamp and measure the phase/phases and grounded/neutral coonductors (depending on the type of branch circuit) at the same time the circuits with any amperage reading are the culprits.

2. Measure grounded/neutral conductor voltage at the panel (should be close to ?0? millivolts) circuits with any loads should have an increase in measured voltage (between the grounded/neutral conductor and the Equipment Grounding Conductor (copper conductor or conduit makes no difference). If the voltage is low, similar to the reading in the PDU, that is a suspect circuit, this is not the best method but is useful.

3. During a schedule outage turn off all the branch circuit breakers, unplug loads if possible and remove the grounded/neutral conductors from the neutral termination bar. Utilizing a low voltage megger 50VDC or 100 VDC (I use a low voltage megger to avoid damaging equipment with 500 or 1000 VDC meggers we normally use. The neutrals that megger bad, ?0? megohms, need to be traced out and the ?SHORTED? grounded/neutral conductor will need to be replaced, repaired or lifted depending on the exact nature of the problem.
 
That's a lot of ground current for such a small load.
I will guess that there is a neutral to ground connection made somewhere downstream of the PDU (other than the one at the step down transformer itself in the PDU).
 
Using the clamp-on to find circuits with net current is the way I go. After finding a net current circuit, measure the individual conductors so you know what is going on.

Then you have to keep tracing towards the loads. I measure circuits coming out of any junction boxes on the way. Sooner or later you will find the neutral/ground connection.

Karl
 
Karl pretty much summed it up.

1. Utilizing a current clamp of sufficient size measure the net current on each panel feeder and neutral conductor. I use a low current LEM style (flexible CT) for this or a leakage current clamp with a large opening.
2. Then move to the branch circuits I use a small opening leakage current clamp for this. Sometimes this is next to impossible due to PDU design.
3. Also check the current on the neutral to ground jumper (if this can be achieved safely). I had one job where the factory installed the CT in the wrong location.
4. Are there other PDU's in the room, cross connected neutrals is a common problem, current has to return to the other PDU through the ground.
5. Check that sub panel neutral termination bar for a bond connection. Actually check all panels for this.
6. Utilizing a digital voltmeter measure neutral to ground voltage at the loads, if you locate a load with “0” volts you are close to the culprit.
7. If all else fails, arrange for a power outage and measure resistance to ground on the branch circuit neutrals (disconnect the loads) We had a
site where the equipment had a manufacturing flaw and the neutrals were shorted to ground in the equipment
 
It sounds like someone or something connected a neutral to ground downstream of the service point.

Might be a failed component. More likely someone thought it was a good idea.

I do not buy the idea that this amount of current could be caused by "stray" fields.
 
karl riley said:
Using the clamp-on to find circuits with net current is the way I go. After finding a net current circuit, measure the individual conductors so you know what is going on.

Then you have to keep tracing towards the loads. I measure circuits coming out of any junction boxes on the way. Sooner or later you will find the neutral/ground connection.

Karl
Click to expand...

I was just curious of what the final problem turned out to be. We had a VERY similar situation that turned out to be an almost imperceptable nick in the neutral of the MC cable in a computer room underfloor branch circuit. We don't know what caused it. The nick was within the cable, not at the receptacle.

We had the luxury of transfering the load to another panel. So to find it we simply shut power down, lifted the neutral to ground bond and measured the resistance from neutral to ground with a Fluke digital VOM, which didn't show infinate resistance. (I don't know what the actual ohm value was.)

Then we moved downstream to the panelboard and started lifting neutrals until we found the culprit.

Thanks.
 
Last edited:
Follow-up of Objectable Current...

Follow-up of Objectable Current...

Hello,
Resolution follow-up to the objectable current on the grounding conductor...

I presented all the information which was given by each person and presented a document to my supervisors.

I followed the lead of using a True RMS clamp-on meter and was able to find the circuits which contained the fault. I isolated the problem circuit, disconnected the wiring and instructed the contractor as how to perform the repair per instructions by "Jon". I also included the other helpful solutions provided the by the forum. This information was very knowledgeable to diagnose and repair the problem.

The contractor is currently working on the removal of the objectable current, among other issues pertaining to scope of the job he is to perform.

I thank you for your help in this issue...

Sincerely,
RMF
 
RMF, good work. But what was the smoking gun?

Yes, nicks in the neutral are sometimes the cause. Often it is intentional connections of neutral to ground, say, in a receptacle. Sometimes mixing up neutral with ground conductors, sometimes a N/G short in a device or appliance. The simplest one you hope for is neutral bus grounded in a subpanel, or a neutral connected to a ground bus. Etc etc.

Karl
 
In my experience, the number one cause of ground current (besides leakage current) is wiring errors and second is nicked conductors. In Washington, D.C. MC and BX is prevalent, I have seen countless incidents were a nicked neutral from a hacksaw cut shorted to the metal jacket (and yes an anti-short was installed). I always explain to the customers nick the phase conductor trips the CB; nick the neutral, ground current. The condition may have existed for years.
 
Brian, similar experience.

Here's a variation: hot was nicked somewhere and shorted. Electrician didn't want to trace the nick so he used the white for a hot and the EGC for a neutral. I was wondering why a white was going into a breaker! It was the magnetic field reading which gave the alert originally.

Karl
 
Status
Not open for further replies.
Top