Flickering lights in a residence

Status
Not open for further replies.
A

ayerforce

Member
Location
Hubertus, Wisconsin
Occupation
Electrical contractor
I received a service call today for "flickering lights" in a house which is usually a matter of finding the poor connection, repairing it, and going home, but this one was different. It happens all throughout the house. To begin isolating the problem, I put a Fluke 87 meter across the two ungrounded conductors measuring AC voltage with min/max points. I put a second Fluke 87 across one ungrounded conductor and the grounded utility conductor. These measurements were taken at the utility LINE side of the meter (underground service) over a 15 minute period:
Line to Line min 233volts max 253.5 volts avg 252.4 volts
Line to GND min 109.9 volts max 127.2 volts avg 126.2 volts
The min points are very fast dips in the voltage.
The POCO had been called to the house last week, looked at the panel, saw some rusted breaker terminals, and said "the flickering lights are not our problem - the problem is in your panel".
The only problem now is that the POCO will not return to the site until the homeowner has replaced a number of their breakers including their main to rule out a corrosion problem. Fine, but didn't I just prove it to be a POCO problem by measuring at the line side of the meter? I hate to tell my customer to spend money when it won't fix the problem. Am I missing something here?
 
What does the other ungrounded wire read for voltage to ground? Does it go down as well, or does it go up? If it increases, you've got a loose neutral.
 
The line to ground voltage on the other conductor varies slightly from 126 volts to 123 volts, and doesn't rise when the other leg falls - I understand your point. I just wanted to verify that since I took readings at the line side of the meter, no poor connection or other problem in the house could cause such readings. I suspect insulation failure either in one ungrounded conductor in the earth or at the transformer, or a poor connection to the transformer secondary.
 
080519-1820 EST

ayerforce:

Can you re-run this test with the main breaker open? Also monitor incoming line current which should be essentially zero.

With no house load and you prove it with current measurements to be near 0, then the problem is the power company. Assuming you still see the voltage variation.

If the problem was from the residence, then there has to be a very large current variation to cause this voltage variation. If this is a 200 A service and the source is a 5% impedance, then we might expect about a 10 V change at the input to the meter for a 200 A load change. Are my guesses about correct.

If the house load variation was maybe 50 A, then there is still a source impedance and/or voltage problem. If you can correlate a specific load current change with a voltage change, then you can calculate the source impedance.

Your source voltage variation at the meter is about 10%.

Seeing you new comments it is probably a bad connection on the one hot side.

.
 
Last edited:
080519-1846 EST

Seeing the new data it is cearly a problem with the one hot side.

Which might mean only 1/2 the circuits in the residence have a flickering problem.

.
 
Thanks guys. I ran this test once with the main off AND the meter removed. No dips - just small fluctuations. I ran the test under load and got the dips to happen sometimes when I started a motor, and sometimes not. I would assume that only lights connected to this one ungrounded conductor will flicker, I just didn't want to charge my customer for a whole day of troubleshooting a problem that I can't fix anyway, so I've stopped for now.
 
I agree that it seems like it is a poco problem. Are there other houses connected to the same xformer and ask them if they are experiencing problems also. If the dip is only on the line side and has there been any digging done recently on the ground likely in the path to the transformer. Reguardless corroded breakers should be replaced to assure they will work when called upon to save a life or a house from fire there is no question this should be done first then call poco back. I have seen faulty mains cause similar problems water and breakers with al buss connections and little spring steel parts dont mix well.
 
Last edited:
There are other houses served from that XFMR secondary, and I'll need to ask permission to do the same tests at the neighbor's meters if the POCO is still reluctant to investigate further on their end. I'll be replacing ALL corroded equipment eventually - I want my customer to understand that they are replacing it for the right reasons; so it works when it needs to as you say - not because it makes their lights flicker. Thanks again to all!!
 
deja vu:

deja vu:

I had this problem once on only one leg. Kept interrupting my web work. Poco comes out and wants to put on a monitor for a week or so. I urged the tech to take a bucket ride and check the xfrmr. Loose bushing! New transformer in about an hour after that.
 
080519-2017 EST

ayerforce:

So with no load it appears you have no significant fluctations. This implies the source to the power company transformer primary is not the origin.

The original data you have provided is as follows:
Line to Line min 233 volts max 253.5 volts, difference = 20.5 V.
Line to Neu min 109.9 volts max 127.2 volts, difference = 17.3 V.

New info:
Other hot line to Neutral 123 to 126, difference = 3 V and would be reasonable to expect, but depends upon current load.

The principal variation is associated with one hot wire. If your voltmeter leads were directly touching the incoming wires to the meter, then the voltage variation is at the transformer secondary or a fracture in the incoming wire. Wire fracture is extremely unlikely. Thus, a bad connection, maybe even arcing at the wire termination to the transformer terminal.

What kind of load variations were occuring in the home while you ran the test? Were these variations across the hot lines, maybe an oven, or from neutral, maybe a flat iron? If no load variation, but a moderate load, then there was likely arcing at the transformer, or at least resistance variation. Arc voltage drop might be in the range of 10 V.

Lately I have been looking at my own home source impedance. I am not real happy with the results, but I do not know the transformer size. My calculated source impedance from current and voltage change for the 120 loop is about 0.05 ohms and only 0.0074 ohms of this is the service drop wire component, roughly 20 %. I used 11.7 A for a test current (an electric heater). My voltage change measurements were after the meter but at the input to the main fuses. These changes were 0.6 and 0.5 V for the 11.7 A load change.

These are just some figures for a frame of reference. Your service cable resistance from transformer to the meter is most certainly higher than mine by a considerable amount.

You need to calculate an estimated resistance of the service wires. And you need to know what kind of load current variations are occuring in the home while the fluctations are occurring. Primarily on the side where the biggest voltage changes occur.

Since there is not a time correlation between the three different voltage measurements one can not say much about the neutral connection at the transformer. My guess is that all three connections may not be real good, but the principal joint problem is the hot wire where the large voltage change occurs.

At our shop which has a fair sized 3 phase service we had random voltage change problems for a long time. Sometimes it was several months with no problem. In the end it turned out to be arcing at both the neutral and one of the hot legs, but not the wild leg. Obviously in the beginning it was just bad connections at the pole transformer, and only at the end was arcing occurring.

.
 
Regarding load variations at the time of test:
no one home
15 incandescent fixtures on, 9 eight foot two lamp slimline fixtures in the basement, various power supplies (alarm, computer, etc.). Only line - line load is was the well which wasn't running. The load was relatively steady and you could actually see the effect of the dips in the fluorescent fixtures - they wouldn't go out but light output was reduced for a fraction of a second. The dips were enough for my Flukes to catch them and I think they only sample 4 times per second. I think I'll call someone from the POCO tomorrow who isn't reading protocol off a computer screen to maybe get a little more attention focused on the problem!
 
gar - thanks for the detailed reply. Regarding calculating the impedance of the service wires, I don't know if that's possible. We have:
1) 170 feet of what I believe to be 2/0 aluminum direct burial cable between my meter ped and the nearest pole.
2) a POCO connection at the top of the nearest pole which serves 1 more residence
3) another connection at the top of the nearest pole to the overhead wires running back to the pole with the transformer (yes, multiple splices due to people switching from overhead to UG services over time, etc.)
4) 80 feet of lord knows what running from pole to pole
5) a connection at the transformer pole which gets us to the transformer and also a third house
Whew! I'm starting to see why they don't want to come out and troubleshoot this!
Is this perhaps an application for a loop impedance tester? I'm itching to buy one, but my wife thinks otherwise!
 
ayerforce said:
I've got a better chance of getting a copy of that magazine than getting my loop tester!!
Click to expand...

I just thought you needed some levity, after all that troubleshooting.

Best of luck, sounds like you're on the right track. I usually don't have that trouble getting POCO to come out to check thier gear. They want to know, and if I ask them they come running. It's about the customer, and if there's one less thing to worry about they do that for us. Then we know for sure.

Our POCO works good with us.
 
080520-0707 EST

ayerforce:

When you had no load there was very little fluctuation, and with a moderately constant load there was fluctuation, and you measured the voltage between the wires or there was no arcing at the meter terminals and you measured the voltage on the terminals and not the wires, then the problem is between the meter and the pole. It is not at the pole primary.

Just your 170 ft of 2/0 Al is about:
Copper at 340 ft is 0.078*340/1000 = 0.027.
Assuming a 100 to 62 ratio, then
Aluminum is 0.027*100/62 = 0.027*1.61 = 0.044 ohms.

A 10 A change thru this portion of the path produces 0.44 V drop. A refrigerator might produce 5 times that for inrush, but how many times would that occur in 15 minutes, and that would still be only 2 V for this part of the line.

The likelihood of a neighbor causing the variations is low because with your customer's load removed the variations were low.

I am going to guess the problem is at the termination of the 170 ft cable at the pole. There are a number unknowns, but voltage variations with a constant load and constant source voltage means that there is a varying impedance in the path. If the problem was on the primary side of the transformer you would see the same variation on both sides of the secondary.

.
 
080520-0742 EST

On the loop tester:
Just use a 1500 W heater and a good digital meter that can read to 0.1 V at 240.

Maybe two 1500 W units in series for 240.

.
 
K8MHZ said:
I would like to know how the OP tested the line side of the meter without breaking the tab to open the meter socket. Most POCO's frown on 'load side' electricians (inside wiremen) cutting their tabs.
Click to expand...

Around here, there's a $10,000.00 "incentive" to stay out of the meter channel. :mad: .................. :smile: :smile:
 
For underground services around here, we see the Milbank U3358 metering pedestal. Yes, there is a POCO seal on the screw(old type) or hasp(new type) used to remove the cover, but that's the only access to this ped - no separation. I couldn't test my load side connections or access my service entrance conductors without removing the seal. Done this way for the last 20 years. Always using proper PPE and always 2 electricians on site minimum W/CPR training.
 
Status
Not open for further replies.
Top