Flickering Lights - N-G Bond?

[Baeljb]

Some code and bonding questions. I am troubleshooting flickering lights in a large historic home. Renovated 3-years ago. We reviewed the electrical distribution, which has lead to some controversy on grounding and bonding. Served by a 50kVA 120/240 split single phase Xmer. Shared by three homes. All phase conductors from Xmer to house main panel are 4/0 AL. All Neutrals are 2/0 AL. 150’ overhead run from Xmer to next pole, then underground, across alley, to meter and 200amp. outdoor panel on detached garage. Then 150’ buried 3-wire run to 200amp. disconnect located in yard. Then 100’ buried 4-wire run to home main panel. Then two sub-panels. Both outdoor panel and disconnect have two #6 bare wire grounds originating from underground. Both have neutral-ground bonds. We see about a 5-volt drop when large 120-volt home loads turn on. We see a corresponding 2-volt surge on the opposite phase. I have seen voltage readings on loose and missing bonds. Usually more severe voltage shifts. The main copper cold water pipe is right by the main house panel. Why would they put disconnect in middle of yard? Is it legal to bond both outdoor detached garage panel and disconnect? Can we move the disconnect inside home and bond using copper pipe as one GEC? Will poor existing grounds explain voltage shift with 120-volt loads? Other suggestions? Thank you.

 

[GoldDigger]

Neutral-Ground bonds will have little or no effect on the voltage drop and shift combination that you describe.
More likely is a high resistance neutral somewhere between you and POCO.
But for a long run, a 2+% voltage drop in the hot wire and a corresponding 2+ degree offset in the neutral voltage is consistent with normal VD on a long run of wire that is marginally too small.
What size are the 150' and 100' run, and how big (in kW) are the large 120V loads you speak of?

FWIW, farms and rural residences once often had a central pole with meter and disconnect and lines running from there to each outbuilding.

 

[Baeljb]

The 150' and 100' runs are 4/0 phase and 2/0 neutral. The loads are large residential dryers, dishwashers, sub-zero fridge, treadmill, etc. So on the 120 volt loads, all are on 20amp. circuits. I am still concerned why a voltage drop on Phase A triggers a voltage increase on Phase B and visa-versa?

 

[romex jockey]

We see a corresponding 2-volt surge on the opposite phase

check the noodle

~RJ~

 

[synchro]

We see about a 5-volt drop when large 120-volt home loads turn on. We see a corresponding 2-volt surge on the opposite phase.

I am still concerned why a voltage drop on Phase A triggers a voltage increase on Phase B and visa-versa?

The voltage drop you're measuring on phase A to neutral is the sum of the voltage drops this extra current causes on both the line A and neutral conductors. The voltage drop this current produces on the neutral conductor is in-phase with the Phase A voltage because it's being "pulled" that way by the Ph A to N load, and so this neutral voltage is 180° from the Phase B voltage. In your example it causes 2 VRMS on the neutral that's is in phase with Phase A, and so that adds 2V to the Phase B-to-neutral voltage. This is similar to how the LA-N and LB-N 120V voltages at 180° from each other add up to 240V across them.

 

[ptonsparky]

You can have a typical VD, with a corresponding voltage swing because long distance and reduced neutral as GD explained....or....a compromised neutral because of poor connections or faulted underground as RJ suggests.

I wouldn’t be worried about 2 volts but 10 is a concern.

 

[Baeljb]

Thanks all. Any ideas why they would have put the home disconnect out in the middle of the yard vs. having a breaker in the home main panel?

 

[electrofelon]

.....then underground, across alley, to meter and 200amp. outdoor panel on detached garage. Then 150’ buried 3-wire run to 200amp. disconnect located in yard. Then 100’ buried 4-wire run to home main panel. Then two sub-panels. Both outdoor panel and disconnect have two #6 bare wire grounds originating from underground. Both have neutral-ground bonds.....

3 wire feeders between buildings/structures are no longer compliant, since the 2008 code IIRC. Presumably that 3 wire feed was installing before that changed. Maybe the 4 wire 100' run was done during that recent renovation. Basically, the end of a three wire feeder should have a N-G bond, the end of a 4 wire feeder should not. Incorrect installation in this department wouldnt cause light flickering, just clarifying that if you are looking over things.

 

[Baeljb]

i am presuming if we recommend that they do any work requiring a permit, the LA may require them to replace the 3-wire feed? To solve the flickering problem, I was going to recommend that they isolate the lighting branch circuits onto new sub-panels fed from true online double conversion UPSs. Has worked great in the past.