I've scoured all of my resources for help in fixing my problem with a home electrical problem ... but so far I haven't found a clean answer. I probably just need to start crawling through the attic and that's the plan today ... looking for anything unusual along the circuit from breaker box to light switch.

Anyway ... I have a light switch that reads 120V between the ground and hot but only 106V between hot and neutral. The lights on this circuit will not come on. Here's the "additional info part" ... They worked for years but some months ago the lights started flickering a bit. I thought maybe a bad switch, bad bulb or bad connection. A couple of nights ago the lights went completely out. That's when I tested the voltages and replaced switches (yes, a hip shot ... praying for an easy fix). Switch replacement did nothing ... and all of the connections in the box look fine.

I'm sitting there pondering how to isolate the problem when suddenly the lights come on. I grab my Fluke and measure hot to neutral ... 120V. What the heck? Then the lights flicker and go out. I measure hot to neutral again ... this time it's 106V. It's very rare the lights will come back on but on a couple of occasions I have been able to flip the switch several times and they come on and flicker and then go out.

At the breaker box I'm getting 120V across hot and ground. I'm going to check hot to neutral today. And then I guess I'm going to trace that line all the way back to the light switch. It's a hot and humid day here in south Texas ... a few hours crawling in the attic will probably do me some good mentally ... assuming I find the problem.

If anyone has some recommendations on what I may be looking for I would greatly appreciate it ... I love electricians but I'm hoping to do most of the grunt work before calling one out.

Thanks for any help.

 

you have a bad connection on a neutral wire somewhere. look for connections that have or are now hot (temprature wise) the hot and neutral are looped through all the circuits. it might be near the wireing to the light nearest the breaker panel, where it ties to the line rom the pannel. a junction box in the attic maybe.

 

At 106 volts the lights should come on. Because of that, I assume you are testing without a load.

Check it under load and "hot to neutral" at breaker box.

Check neutral connection in the main service box. It will be easy there, as you know which circuit it is. (by seeing which breaker interrupts power to switch)

It's rare to be outside a box. mice etc.

The connection will be discolored or even melting insulation when you find it.

Using a meter or lamp, find all the outlets and lights on that circuit. (by switching breaker on and off)

If neutral is ok at panel, check each "box" between panel and breaker until you find the place where neutral is lost. (they will be in a line, roughly) It's usually in the last box that works, not the first one that's dead. Don't know why.

edit:
Is neutral available at switch box?

 

Alfacliff / Inwo ... thank you both for your responses. At least I know I'm on the right path trying to track down a faulty neutral line.

A couple notes ... 106V was a test without a load ... I think it falls to about 60 or 70 when I flip the switch ... regardless the lights don't come on. If the lights don't come on when I flip the switch I know I'm going to read 106V at the switch ... if I get lucky and the lights come on ... I know I'm going to read 122V.

Update! The lights are shining bright and no flickering ... and they've been on for over an hour which is the longest they've been on in the last five days.

I took one last look at the outside breaker box and tested voltage between the circuit breaker for the light switch and the neutral bar. 123V all the way across. I happened to notice a wire nut connecting two white wires together (neutral) and gave it a firm twist to make sure the connection was good. I then went inside and measured hot to neutral expecting to see 106V (my problem) before heading up in the attic. The meter read 123V ... I thought hmmm ... and flicked the switch ... lights came on ... strong ... no flickering.

I went back outside ... twisted that wire nut again a little tighter ... trying to determine if it was the source of my problem ... went back inside ... no lights.

I went back outside ... took the wire nut off ... looked at the wires ... nothing unusual ... put the wire nut back on ... snugged it down. Came back in the house ... 123V on the switch ... lights were on full bright no flickering.

Did I fix the problem? I doubt it. I am convinced however that the problem is the feed between the breaker box and the light switch. I've disconnected every wire in that box and never could get a reading of 120V across hot and any neutral. Now that everything is working I get 123V across hot and neutral. I also believe the problem is most likely in the breaker box area. Is it that wire nut connecting two neutral wires? That would certainly be consistent with the symptoms.

I'm scared to touch that wire nut again! My wife will love me when she gets home and the lights are working. I'm just not feeling a solid conclusion here. I'm attaching my wire nut in the breaker box picture just in case there's an interest

. Again ... thanks for the responses!

 

look for a bluish discoloration on the terminals, thats the first sign of heat and corrosion. also, plastic wireuts tend to discolor or melt with bad connections. white wires also start discoloring with heat.

 

Unlikely that was the problem if it didn't show signs of heating.

Were the wires twisted inside the connector.

If installed properly wires will be twisted.

Also follow that white wire to the neutral bar connection.

There may be up to 12 other devices on a normal branch circuit.

Catch it not working and check voltage at each device without disturbing anything.

 

It could be inside the "nut", but could be the contact between the neutral wire and the neutral bus. While you tighten/untighten the "nut", you somewhat distturb this connecton and it might be coming on and off. I would check that too.

About the "nut". Sometimes they're just twisted on wires and it often causes bad contacts. When you take the "nut" off, the wires should be twisted together so that they stay well connected even when you try to pull them apart with the "nut" off.

 

Inwo ... I did take the wire nut off and the wires looked like a good connection ... twisted together firmly. I can't imagine that me tightening the wing nut did anything useful. I like the advice about catching it while it's not working ... that seems to be the only time I can isolate the failure.

I actually simplified this circuit when I posted because it made my head hurt after getting started with the problem. Now that I've lived it for three days I know it like the back of my hand. This "light switch circuit" runs three separate light groups. One switch controls 8 recessed lights in the living room .... another switch controls a single recessed light over the kitchen stove ... and a third, fourth and fifth switch controls a group of four recessed lights in the kitchen (two three-way switches and a single four-way switch). When the failure is present every hot-neutral check at the switch is 106V.

I found the switch where the power originates (a long crawl through the attic chasing the actual wires). It's pretty clean ... and the line coming from the breaker box to this first switch was easy to locate.

The one thing that continues to stump me is measuring voltage across a known hot (power coming in to the first switch) and the "individual" neutrals. The image below is where the power comes in ... and I measured across the hot and each individual neutral ... I was looking for the line coming in (a hot-neutral pair). I expected to get 106V across only one of the neutral wires but instead I got a variety of voltages ... one of them being 120V. I don't understand how I get any reading across the hot and a disconnected neutral that goes to another light bulb. The other interesting observation is that i believe the hot-neutral pair coming from the breaker box was the one reading 103V. I only think these two wires are a pair based on how they "feed" into the junction box ... it looks like each pair of the four wires coming into this junction box comes through it's own hole at the top. I wanted the superhero power of flying when I was young but now I think seeing through walls would be much more useful.

 

if the neutral were connected properly, there should be a 120 volt between any hot and neutral. a bad nutral connection will be indicated by less than 120 volts, due to leakage to ground on the neutral also capacitance to ground will give you a little leakage on an open wire.

 

Leave neutrals connected for testing. Take the wire nut off for access, but leave connected.

Individual readings with unknown loads will be confusing.

Check hot to neutral in each box under load while not working.
If you're right and first box has low reading, it narrows things a lot.

Looks like a DIY job. Not a good sign. I would not have passed this. ie. twisting ground wires together is not approved.

 

I do not like the Screw-on wire nuts, they were banned in the UK many years ago.
Max.

 

I do not like the Screw-on wire nuts, they were banned in the UK many years ago.
Max.

Wow!

I can count on my fingers and toes the number of failures Ive seen in 50 years.

The connection comes from the twisted wires. The pressure comes from the wire nut. Approved for aluminum to copper connections.

Not that I think aluminum should ever be used in branch circuit wiring.

When I started in 64, the boss only used Buchanan tm. method of 4-way crimp.

What do you use over there?

 

I *just* last week had this same problem, only it was throughout my entire house!

Seems when a huge tree branch fell off the top of a 50 year old maple it landed across the feed line from the pole to my house, and it “kinda-sorta” broke the neutral connection back up the pole. By “kinda-sorta” I mean the utility worker who went up my pole (after going up my house and redoing those connections “just to be sure”) found the neutral was damaged, and would sometimes open but then fuse itself back together until it wiggles apart again.

As this neutral is between the legs of a 220V service each leg would run around 110 depending on what the relative loads were on each leg. If say the washing machine motor went off the voltage would imbalance, the lights would flash (sometimes dim) and the neutral would weld itself again for a while.

One thing you may want to do is see how your meter responds when you just put one lead on the hot line: some digital meters will show whatever voltage they feel like. However, an old analog dial pointer meter will only read a real voltage and may be better in finding the open line.

I would worry about leaving these lights ON without tracking down the problem: the intermittently solid connection can heat up if it sorta breaks open.

Sorry I cannot make out what is going on in the junction box picture. Generally yes, each cable should be entering through it’s own port so I see 4 cables entering the top. Each cable should have a black (hot), white (neutral), and either green or bare (ground) wire. An optional red in a cable is used for a switched hot line (may be either hot or open per some switch).

If you are sure of the supply line cable is giving you 103 volts then the problem is before this box.

Breaks don’t have to happen inside junction boxes, it can also happen within the cable itself. Have you done any recent work on your house? Any drilling or nailing that may have compromises a cable?

(Aside to others: I have never heard of wire nut connections inside a breaker box. Is this to code?)

 

Breaker boxes cannot be used as junction boxes.

However wires can be extended if they terminate in the same enclosure.

If all were spliced it would be frowned on as poor workmanship!

And another sign of DIY. Cable sheath must enter box. Also cable must be fastened within a few inches or have cable clamps.

More than one cable in a hole is sometimes allowed.

When I wired homes, I would often put both cables in a hole next to the stud. Looked neater that way.

Now cables are not allowed next to stud at all. Screws and nails that missed stud were hitting the wire too often.
Or so they say.

 

The lights went out after working for about 4 hours. I did not operate any of the switches during those four hours and ... I know this is going to sound obvious ... the minute I did touch a switch that's when they stopped working. I just walked up to one of switches thinking I'd flick it on and off a couple of times making me feel better that things were working and instead ... when I flicked off the switch all the lights went off and would not come back on. I took off all of the light switch plates and tested between hot and neutral ... 106V without a load and around 2V with a load ... ugh.

I also checked the breaker box and tested that wire nut location with the two neutrals and the entire neutral bar ... 122V between hot and any neutral wire in the breaker box.

Inwo ... you mentioned twisting ground wires was not okay. That light switch box looks like metal but it's actually plastic. Is twisting still not a good idea? Are you expecting something like a wire nut on the bare ground wires?

 

I took off all of the light switch plates and tested between hot and neutral ... 106V without a load and around 2V with a load ... ugh.

Did you do test from neutral to ground at these points, see what the difference is at these points?
Max.

 

Hmmm ... with a load the neutral to ground voltage is around 118V ... without a load the meter is reading 15V. This is on every switch ... however, one switch the load voltage is around 110V ... a little lower than the others.

 

Wow!

I can count on my fingers and toes the number of failures Ive seen in 50 years.

The connection comes from the twisted wires. The pressure comes from the wire nut. Approved for aluminum to copper connections.

Not that I think aluminum should ever be used in branch circuit wiring.

When I started in 64, the boss only used Buchanan tm. method of 4-way crimp.

What do you use over there?

+1 to that.
The only Wire Nut (a trademark of Ideal Industries!) failures I've seen are when one of (too many) wires just misses connection and 'goes ohmic.'

Also +1 to 'What do you use over there?'

 

The lights went out after working for about 4 hours. I did not operate any of the switches during those four hours and ... I know this is going to sound obvious ... the minute I did touch a switch that's when they stopped working. Don't do that after you get it to show symptoms.! I just walked up to one of switches thinking I'd flick it on and off a couple of times making me feel better that things were working and instead ... when I flicked off the switch all the lights went off and would not come back on. I took off all of the light switch plates and tested between hot and neutral ... 106V without a load and around 2V with a load ... ugh.

I also checked the breaker box and tested that wire nut location with the two neutrals and the entire neutral bar ... 122V between hot and any neutral wire in the breaker box.

Inwo ... you mentioned twisting ground wires was not okay. That light switch box looks like metal but it's actually plastic. Is twisting still not a good idea? Are you expecting something like a wire nut on the bare ground wires? Absolutly!

Have things ready to test. Covers off. Don't touch!

Check hot-neutral voltages under load, carefully. Do not disturb.

Once you know the circuit path and voltage at all points, we can tell you exactly where the problem is.

For example:
All points show low voltage except breaker to neutral bar.
Then it's obvious, that the high resistance is between those two points.
From neutral buss to first device box.

 

Hmmm ... with a load the neutral to ground voltage is around 120V ... without a load the meter is reading 15V.

As others have confirmed, definitely points to a VERY high resistance in the neutral somewhere, a check from neutral to earth at different points, should pin point it, there should be almost no difference at any point, just a small difference allowing for current in the neutral if a load is present..
Max.