Baffling Home Power Problem

[Joseph Wolf]

One circuit in my house intermittently goes out 3 to 5 times per day. After about 5 to 10 minutes, the circuit re-energizes itself. No breakers trip. It just comes back on.

The lamps on the circuit will flicker for a second or two, the power delivery stops. By accident, we discovered that if we turn on the air conditioning, the power is immediately restored.

I’ve consulted with two electricians. The first one thoroughly examined the box and all connections and confirmed that everything looked good. The power delivery is solid and within spec. We could not cause or replicate the problem.

Both electricians I consulted said it must be a loose wire in an outlet on the circuit. I pulled every outlet, light switch and fixture. Connections are solid and there was no scorching or smell. Yet the problems continues. It happens 5 to 10 times a day. Lasts for a few minutes, then power restores. I can speed it up if I turn on the AC.

Help!

 

[Snoonyb]

Welcome.

It sounds like you have a loose connection.
In saying the connections are firm, are the recep. back-stabbed or are the conductors secured under the screws, have all the wire nuts in the boxes been removed, the conductors, all respectively, been twisted together and the wire nuts replaced?

 

[Joseph Wolf]

Welcome.

It sounds like you have a loose connection.
In saying the connections are firm, are the recep. back-stabbed or are the conductors secured under the screws, have all the wire nuts in the boxes been removed, the conductors, all respectively, been twisted together and the wire nuts replaced?

Thanks for the response, but YES. It was very tedious. I pulled out every fixture and redid every connection.

 

[hornetd]

Snoonyb is correct!

Every connection must be checked. If you do find that you have any of the push in spring contact receptacles they are the most likely culprits by far. If you do find any of that type consider replacing those receptacles and switches with the rear insertion screw tightening variety. If you do that it will absolutely eliminate any intermittent open connection which is occurring at the device terminals. If you do have the push in only receptacles or switches I would strongly suggest that you replace them as soon as you can manage it.

There are several different failure modes with the spring grip push in type. Some of them were badly made and have too much space in the insertion hole. That will allow the conductor to move out from under the spring which is is the only contact that the wire has with the wiper straps in the receptacle. This is partially due to the manufactures trying to make the terminals suitable for both 14 and 12 gauge conductors. The manufacturers claim that such shifting is very unlikely but the electrical testing laboratories will only list that type of connection for a single wire gauge now. That allows the insertion pathway to be a close fit to the particular wire size which prevents the wire from having any room to move.

Another failure mode is that the insulation on the conductor was not stripped according to the directions so that insulation extends too far into the entry hole and holds the spring contact out of full contact with the wire. This is the second reason that the laboratories will only list the connections for a single gauge of wire now. Because of the tight fit on this newer type of push in spring grip terminals any insulation will not fit into the hole. If that leaves too short a stripped length of wire into the terminal it will fail the "tug test" which is a part of the instructions included in the listing and labeling and therefore required by the National Electric Code in the United States.

Most people are unaware that wires change slightly with the changes in the current flowing through them. As the current increases the conductor becomes slightly larger do to the heating of the metal caused by the current flow. When the current decreases or stops the wire contracts thus becoming slightly smaller. That cyclic change in dimension of a wire is normal but it does apply mechanical strain to the device used to terminate the wire to the device. If the connection is on the edge of conductive contact that could lead to arcing at a low level which will heat the connection quite a bit and possibly open it. When the connection cools contact between the iffy termination and the wire is reestablished and the gradual expansion of the wire resumes. The connection opens fully or partially and the cycle repeats itself.

The quality of connections in the 3 major types of device terminals is:
Push in spring contacts are the least reliable;
Screw terminals are reliable if the installer had the skill and the time to make a 2/3 or more wrap around the screw in the clockwise direction from the unstripped portion of the wire to the wire tip. That way when the screw is tightened it will pull the wire in tighter to the screw shaft thus making a better connection.
The most reliable; according to testing done at the Underwriters Laboratory; is the Screw and Sems washer pressure connection. The wire is placed and held under the captive washer or pressure plate and the screw is tightened to clamp the wire between the screw threaded stationary plate and the captive washer or pressure plate. Of this latter type the ones with the square washer which has the corners turned down is the best. The pressure plates used by some manufacturers are the absolute minimum thickness to pass laboratory inspection but even those are still better than the connections made by wrapping the screw if that is done by unqualified helpers and especially better than the spring contact push in type.

--
Tom Horne
Retired Electrician after 55 years in the craft.

 

[Snoonyb]

Thanks for the response, but YES. It was very tedious. I pulled out every fixture and redid every connection.

Thanks Tom, good info, as usual.

Thanks, this may be a potential option for you to try, and they are available at almost all hrdw. vendors, in one form or another; https://www.homedepot.com/p/Gardner-Bender-Outlet-and-GFCI-Tester-GFI-3501/202867890

Could be an indication of where the fault lies.

 

[Joseph Wolf]

Thanks Tom, good info, as usual.

Thanks, this may be a potential option for you to try, and they are available at almost all hrdw. vendors, in one form or another; https://www.homedepot.com/p/Gardner-Bender-Outlet-and-GFCI-Tester-GFI-3501/202867890

Could be an indication of where the fault lies.

I appreciate your reply, but how would a circuit tester be of use? Unless I’m missing something. When the power is on, it will show no fault. When the power is out, it won’t show anything. Don’t see how it would assist in isolating the issue described....

 

[Snoonyb]

Generally, the circuit will not lose both sides of the hot pair at the same time, even if the breaker fails, the tester will tell the condition of the failure.

Even if the breaker, at the moment tests good, the same condition that MR. Horne described in the recep. may be occurring there.

We're just looking for answers, and as obtuse as it may seem, these products are developed by man, and will fail.

 

[JoeD]

Are more than one circuit affected? It sounds to me like one of the hot wires feeding your house has a loose connection. Turning on the AC back feeds the missing leg. When you turn on the AC does it run properly?
The electrician won't be able fix that. It is a POCO issue.

 

[Steve123]

Could also be a hidden junction (which would be against code ). Newer house or older with signs of DIY work ?

Loose connection can generate heat. Not sure if it would be possible to pick up that heat with an IR camera ?

 

[Michael Armstrong]

Map the circuit that fails from the service panel to the end, then note which outlets fail. The problem lies either in the last one that doesn't fail or the first one that does. If those 2 are OK, then you've missed a junction. This is another tedious job, but I've done it in every house I've owned, and the maps have been invaluable.

 

[MrMiz]

figure out what the first connection (outlet) is on the circuit. Plug in the outlet tester and watch it. If the outlet test goes out/flickers too at the first connection (outlet). Then congratulations you haven't really found the first connection OR ... even worse the last person that did electrical has hidden a connection from you. Either in the wal (which really sucks and is against code because of fire danger, but I've seen it MANY times) or if you lucky there may be a "junction" box with a blank cover plate somewhere our missing. If you do find a junction box open it and if it's in circuit and doesn't look like a whole other circuit is spliced in then stick a contact tester in there and wait for it to flicker. If flickers there then again there is another connection your not finding between that and the circuit breaker (probably hidden). If it DOES NOT flicker then everything after it is suspect. work your way through the outlets in a linear fashion and between the one that doesn't flicker and the one that does you've got a nasty surprise somewhere.
Best of luck with this. I see this a lot in basement remodels as they sometimes get done without a permit and nobody walked through to explain why splicing a wire in a wall/ceiling/floor is a terrible idea. The few that I have fixed I've had to pull new wires to the circuit (between the one that doesn't flicker and the one that does) and in one case I had to remodel a whole room only to find that they had a splice in the ceiling, behind the drywall, for what can only be described as the laziest of repairs.

I'm afraid there is no simple, cheap, or easy resolution to the problem your describing. Edit- basically Michael Armstrong explained it better and with few words in the comment above this.

 

[MrMiz]

Could also be a hidden junction (which would be against code ). Newer house or older with signs of DIY work ?

Loose connection can generate heat. Not sure if it would be possible to pick up that heat with an IR camera ?

You would think a big box store would rent out thermal cameras. I can only imagine with their return policies that they would be sick of people buying them, then returning them the same day.

Edit- I just noticed that HD doesn't carry them in store anymore. They are online order only, but their return policy is in place. Of course with the wall this problem is described I would be able to buy and nice one in the cost of service calls.

 

[bud16415]

It is unlikely but possible that you have an open/close condition in the wire itself. I had a friend that had this issue in a house he had built and it drove him nuts tracking it down. He finally figured it had to be a break in one of the white wires running thru a finished wall. To test it he used the bare ground wire to replace the white wire. I might get some negative feed back suggesting that, but it is what he did. when he did that the problem went away and then he opened the wall and ran a new cable. Sure enough a nail or screw had broke the wire and was making on and off contact.


It is an exhaustive process doing this stuff but I have also seen people replace cable lengths externally outside the wall as a test of the cable in the wall. Once you find the problem you then leave the old cable in the wall and find the best path to run the new one.


You want to rule out all the likely causes first, back stabs and wire nuts, and it sounds like you have done that.

 

[Eddie_T]

I had a problem with a breaker and the stab on the bus bar. The breaker didn't trip but the connection would sputter and open and then reconnect when it cooled. That kind of problem can damage the plating on the bus bar stab.

 

[hornetd]

I had a problem with a breaker and the stab on the bus bar. The breaker didn't trip but the connection would sputter and open and then reconnect when it cooled. That kind of problem can damage the plating on the bus bar stab.

Eddie T

When you said "stab on the bus bar" I immediately thought Federal Pacific Electric (FPE). Did I guess correctly?

--
Tom Horne

 

[Eddie_T]

Eddie T

When you said "stab on the bus bar" I immediately thought Federal Pacific Electric (FPE). Did I guess correctly?

--
Tom Horne

It was labeled Sears and Roebuck I don't know who made it. It has been replaced now.

 

[Michael Armstrong]

You might try replacing the existing circuit breaker in the service panel with a combination AFCI breaker. I'm only familiar with Square D's QO120CAFI, but I think most work this way. They watch for several faults, including Fault to Ground (but NOT a GFCI-type fault), Arc Fault, and Short Circuit. So, install this breaker, wait for it to trip, and the CB's internal diagnostics can be used to determine the type of fault. I'm betting on an arc fault, which might be of interest. I'm sure there are tools that can determine how far such a fault is from the panel, which would really be of interest.

 

[Michael Armstrong]

More on the "tools" I alluded to. The technology is called "time domain reflectometry". The tool sends a pulse down a wire (I've used it on a transmission line, but I don't see why it wouldn't work on an electrical circuit, which is basically a low-impedance transmission line), and listens for a return, which could be caused by a break in the line, or any other fault that changes the characteristic impedance. Tom Horne would know more about this, and might suggest a tool.

Also, if you want to try the thermal camera thing, There's a nifty gadget that turns your cellphone into a thermal camera, but I doubt it would tell you much:



https://www.flir.com/products/flir-one-gen-3. A bargain at $200.

 

[hornetd]

You might try replacing the existing circuit breaker in the service panel with a combination AFCI breaker. I'm only familiar with Square D's QO120CAFI, but I think most work this way. They watch for several faults, including Fault to Ground (but NOT a GFCI-type fault), Arc Fault, and Short Circuit. So, install this breaker, wait for it to trip, and the CB's internal diagnostics can be used to determine the type of fault. I'm betting on an arc fault, which might be of interest. I'm sure there are tools that can determine how far such a fault is from the panel, which would really be of interest.

Michael Armstrong

You are correct that AFCI breakers that are not of the dual service type do not have the 5 mA trip point Ground Fault Circuit Interrupter (GFCI) built in. The dual service type AFCI/GFCI, which provides both functions, are readily available at not much more money than the single service AFCI only type. AFCI Breakers do have Ground Fault Protection of Equipment built in. The trip point for Ground Fault Protection of Equipment is 30 mA of leakage current.

I am puzzled about how you would use the "CB's internal diagnostics can be used to determine the type of fault." Those breakers do not have any event memory built in an no port of any kind for recording it's operation on a fault.

--
Tom Horne

 

[hornetd]

More on the "tools" I alluded to. The technology is called "time domain reflectometry". The tool sends a pulse down a wire (I've used it on a transmission line, but I don't see why it wouldn't work on an electrical circuit, which is basically a low-impedance transmission line), and listens for a return, which could be caused by a break in the line, or any other fault that changes the characteristic impedance. Tom Home would know more about this, and might suggest a tool.

Also, if you want to try the thermal camera thing, There's a nifty gadget that turns your cellphone into a thermal camera, but I doubt it would tell you much:

View attachment 22444

https://www.flir.com/products/flir-one-gen-3. A bargain at $200.

A Time Domain Reflectometer does not work on circuits that have branching pathways. It would report each one as a fault. On point to point cables such as Coaxial, Twisted Pair and similar cables it can give you the distance to a fault along the cable within a foot in many cases. The additional challenge is you have to know how the cable was run to measure the distance the TDR had indicated.

--
Tom Horne