Another electrical mystery

Reply

  #1  
Old 04-07-06, 07:04 PM
Member
Thread Starter
Join Date: Jan 2005
Posts: 135
Another electrical mystery

Here's one that baffled a few people today. Any thoughts would be appreciated.

Old romex wiring with no ground in an old house. 50 amp service. A 15 amp breaker serves the bathroom.

Power comes into the receptacle in a receptacle/switch combo box. There's a jumper between the receptacle and the switch. The red wire in a 14/3 cable runs from the switch to a junction box on the wall (it's an old, odd setup).

According to a multimeter, there's power everywhere- at the panelboard, receptacle, switch, junction box, and the light and fan. When anything, even a lamp with a 60 watt bulb, is plugged into the receptacle, the power is gone! Unplug whatever was plugged in, and the power returns- or shall I say, the hot wires read "hot."

We checked the neutrals with a plug-in tester. It showed "no ground" (of course) but not "open neutral."

The receptacle, switch, and fan/light were hooked up to another circuit and worked fine.

The homeowner knows the best thing to do would be to run new wiring. But until his budget allows, does anyone have any suggestions as to what to check? I left out some information because this post was already too long, but would be happy to provide more details.

Thanks very much.
 
Sponsored Links
  #2  
Old 04-07-06, 07:38 PM
Banned. Rule And/Or Policy Violation
Join Date: Feb 2006
Location: North of Boston, MA.
Posts: 2,113
-The receptacle, switch, and fan/light were hooked up to another circuit and worked fine. -

How did you acomplish this?

-According to a multimeter, there's power everywhere-

What is/was your referance point/s?


*****I left out some information****

This may be the most important part.


For humor: it may be a series ckt.
 
  #3  
Old 04-07-06, 08:31 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
Originally Posted by findtheriver
According to a multimeter
Yawn.

A multimeter is a high impedance voltmeter.
It is perfectly capable of reading phantom voltages and voltage carried by coal dust.

> When anything, even a lamp with a 60 watt bulb, is plugged
> into the receptacle, the power is gone!
> Unplug whatever was plugged in, and the power returns-
> or shall I say, the hot wires read "hot."

Correct. Take readings only while the light is turned on.


> does anyone have any suggestions as to what to check?

Everything while the light is plugged in.
Start from the panel. The first place that has no voltage, there's your problem.

With no load on the circuit, your voltmeter is to be used for entertainment purposes only.
 
  #4  
Old 04-07-06, 08:45 PM
Banned. Rule And/Or Policy Violation
Join Date: Feb 2006
Location: North of Boston, MA.
Posts: 2,113
*With no load on the circuit, your voltmeter is to be used for entertainment purposes only.

How do you figure? Would you use a tic-trace and/or wiggies instead? should I put the fluke on E-Bay?
Please, advise.
 
  #5  
Old 04-07-06, 09:04 PM
Member
Join Date: Mar 2006
Location: Wet side of Washington state.
Posts: 18,388
While digital multimeters are a wonderful tool they can indeed give false readings on unloaded circuits.

An analog meter is better for this type of troubleshooting and a wiggie is even better. Best is what bolide suggested, test the circuit while it has a load.
 
  #6  
Old 04-07-06, 09:58 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
it's a fluke

Originally Posted by lectriclee
How do you figure?
Oh, sure, make me say it.

> Would you use a tic-trace

No.

> and/or wiggies instead?

Yes.

> should I put the Fluke on eBay?

Yes. Seriously, it is fine provided that you know how to interpret the results.

In this case, the only mystery comes from thinking that the no-load reading means something that it obviously does not.

If you know that the voltmeter reads 120V for more reasons than just that you have a good circuit, that's fine, no mystery.

Otherwise, switch your multimeter to ohms x 1000 and use it for a lie detector.
 
  #7  
Old 04-08-06, 12:05 AM
Member
Thread Starter
Join Date: Jan 2005
Posts: 135
Thank you for the very interesting replies.

lectriclee:


-The receptacle, switch, and fan/light were hooked up to another circuit and worked fine. -

How did you acomplish this?
We first tested for continuity. Then for kicks we removed the items and connected them to a known working circuit.

-According to a multimeter, there's power everywhere-

What is/was your referance point/s?
Not really sure if this is what you mean, but in testing for voltage, we put the probes at each point and had readings of 119, etc.


*****I left out some information****

This may be the most important part.
Okay, I'll post back with more information.


Bolide:


A multimeter is a high impedance voltmeter.
It is perfectly capable of reading phantom voltages and voltage carried by coal dust.
I didn't know this. Kind of changes everything.

> When anything, even a lamp with a 60 watt bulb, is plugged
> into the receptacle, the power is gone!
> Unplug whatever was plugged in, and the power returns-
> or shall I say, the hot wires read "hot."

Correct. Take readings only while the light is turned on.
In that case, there is no voltage reading.


> does anyone have any suggestions as to what to check?

Everything while the light is plugged in.
Start from the panel. The first place that has no voltage, there's your problem.
Makes sense.

With no load on the circuit, your voltmeter is to be used for entertainment purposes only.
Back to the books for me.


Furd:

While digital multimeters are a wonderful tool they can indeed give false readings on unloaded circuits.
When a non-contact tester indicates the wires are hot, but when you plug something in the wires go dead, is a false initial hot reading also a possibility?

An analog meter is better for this type of troubleshooting and a wiggie is even better. Best is what bolide suggested, test the circuit while it has a load.
Okay. But when the circuit has a load there is no power. What I mean is: when something is plugged into the receptacle, there is no sign of power coming in. I guess that solves the mystery?
 
  #8  
Old 04-08-06, 12:28 AM
Member
Thread Starter
Join Date: Jan 2005
Posts: 135
Probably should have included this info before. Sorry about that.

About the junction box: There are 4 cables coming into the box: 1 goes to the fan/light, 1 is from the 14/3 which is connected to the receptacle and switch, and 2 cables go to unknown destinations. That's right- they do not connect to any known light, receptacle or switch anywhere in the house.

Everything's been working fine for many years. About six months ago the fan unit was replaced, but there's been no other changes or additions made to the wiring.

Given all of this, what would an electrician's thought process be here (besides updating the wiring, that is)? One day the switch is flipped and there's no light or fan action. The bulb is okay. A non-contact tester shows the correct wires are hot. A multimeter shows that the voltage is correct (I realize this doesn't mean anything). When a lamp is plugged into the receptacle, nothing shows power. When the lamp is removed, the ostensible indications of power return.

Thanks for wading through this, and for sharing your insight.
 
  #9  
Old 04-08-06, 01:28 AM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
> When a non-contact tester indicates the wires are hot, but
> when you plug something in the wires go dead, is a false
> initial hot reading also a possibility?

It's the same problem. You have some type of extremely weak connection.
An open neutral will read 125V, light up your test light, tick your non-contact tester, etc.
That doesn't mean that useful power is available.


> I guess that solves the mystery?

Well, there's no mystery.
It's just a matter of tracking down where the faulty connection is.
From what you described, it could be on the hot or neutral side, though probably on the hot side. You didn't provide enough readings to ascertain this.
You need to measure to ground also.

Did you know that old police radar guns would clock trees as going 70mph?
It's not because there was anything wrong with the radar gun.
The problem in that a tree doesn't provide a good echo.
When using an instrument you can trust it only within its limits of reliability.


> When the lamp is removed, the ostensible indications of power return.

It's not ostensible. As I said, it's for entertainment purposes only - just like when your ohm meter shows that someone is lying. It's okay for parties or to amuse friends who really haven't a clue about electricity.

That's why when you say that your multitester says 120V, I shrug it off. I know that you are hoping that this is some big clue, but it isn't. An ostensible indication of power is running lights brightly.

Actually, it's readings like 68V that are interesting.
Just remember to take readings under load if you are looking for power on the hot side.

You can test a neutral that is connected, it will read 0V to ground.
You take the neutral apart, touch the correct end, and it can kill you.
The neutral doesn't care what your multitester says either.
 

Last edited by bolide; 04-08-06 at 01:59 AM.
  #10  
Old 04-08-06, 01:41 AM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
> There are 4 cables coming into the box:
> 1 goes to the fan/light, 1 is from the 14/3
> which is connected to the receptacle and switch, and
Okay.

> 2 cables go to unknown destinations.
Take the wire nut off the hot side and start killing breakers until it goes dead.
Leave the breaker off a while and perhaps you'll find out what it powers.
I just hope it doesn't go to a smoke detector or it might take a while to figure that out.

Unknown circuits are your enemy when you have a problem.

Of course, if it's already dead, cap both black ends with wire nuts. Put a note in the box that says, "Unknown circuit opened 2006-Apr-08".


> Given all of this, what would an electrician's thought process
> be here (besides updating the wiring, that is)?

Trace out the circuit to the point where the hot voltage drops out.


> One day the switch is flipped and there's no light or fan action.
> The bulb is okay.

Open circuit.

How about backstabbed wiring?
Are the wires on screws or just poked into the back of the switch or receptacle (as some people do with 14 gauge)?

Pull it out an do it right. It's probably the culprit.
If it isn't the culprit, at least you corrected a potential problem.
 

Last edited by bolide; 04-08-06 at 02:00 AM.
  #11  
Old 04-08-06, 05:52 AM
HandyRon's Avatar
Member
Join Date: Dec 2001
Location: New York
Posts: 1,365
There is a good bulletin bout phantom voltages.
http://www.nema.org/prod/wire/build/...088%202003.doc
 
  #12  
Old 04-08-06, 08:09 AM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
sell your Fluke on eBay; get a wiggy

That's a good resource, Ron. I like it when someone brings factual material into the discussion.

Originally Posted by NEMA ENGINEERING DEPARTMENT (Bulletin no. 88, October 1998 / rev. February 2003)
In order to help minimize the likelihood of reaching a wrong conclusion from this phenomenon, NEMA recommends the use of a UL Listed or OSHA-approved low impedance multimeter in place of a high impedance multimeter or other high impedance measuring device for testing on open conductors where there is no hard electrical connection.
Without a low impedance measuring device, a high voltage reading is an inconclusive indication of possible faults in the cable.
 
  #13  
Old 04-08-06, 09:52 AM
Join Date: Mar 2006
Location: Fort Collins, Colorado
Posts: 159
Originally Posted by findtheriver
When a lamp is plugged into the receptacle, nothing shows power. When the lamp is removed, the ostensible indications of power return.
To clarify, the bad connection in the "hot" wires can carry a tiny amount of load. That's why the plug tester lights up when you plug it in, but when you connect an incandescent light bulb to the circuit, the bad connection burns itself open.

When the big load goes away, the tiny connection comes back together.

Given the fact that the non-contact voltage tester goes dark when the load is applied, you have isolated the problem to the hot wires. Do look at your neutrals as you wade through the circuit, but they are not causing this particular problem.

I'd say calling a digital multimeter a "novelty" is waxing melodramatic, but I guess the point was made in a way the poster will remember.
 
  #14  
Old 04-08-06, 10:33 AM
Member
Join Date: Oct 2005
Location: USA
Posts: 1,152
test lights

Does anyone still use the old time edison base incandescent with a pigtailed socket? I recall even being able to use one on limited 240v testing if you pick a 20w bulb. These put an end to phantom voltages/high resistance connections chasing you around.
 
  #15  
Old 04-08-06, 12:34 PM
Member
Thread Starter
Join Date: Jan 2005
Posts: 135
Thank you to everyone for helping me figure this out.

Bolide:

It's the same problem. You have some type of extremely weak connection. An open neutral will read 125V, light up your test light, tick your non-contact tester, etc. That doesn't mean that useful power is available.
What fooled me is that these testers never failed to indicate the correct condition before. You've opened my eyes.


Well, there's no mystery. It's just a matter of tracking down where the faulty connection is. From what you described, it could be on the hot or neutral side, though probably on the hot side. You didn't provide enough readings to ascertain this.
You need to measure to ground also.
The connections in the bathroom and at the breaker were checked- pulled apart and reconnected. I'll measure the ground.

Did you know that old police radar guns would clock trees as going 70mph? It's not because there was anything wrong with the radar gun. The problem in that a tree doesn't provide a good echo. When using an instrument you can trust it only within its limits of reliability.
Another excellent point. Especially when dealing with electricity, where a mistake can be very dangerous.


That's why when you say that your multitester says 120V, I shrug it off. I know that you are hoping that this is some big clue, but it isn't. An ostensible indication of power is running lights brightly.
Got it.


Actually, it's readings like 68V that are interesting.
Just remember to take readings under load if you are looking for power on the hot side.
Understood.

You can test a neutral that is connected, it will read 0V to ground. You take the neutral apart, touch the correct end, and it can kill you. The neutral doesn't care what your multitester says either.
I'm learning the importance of knowing what a tester does and does not reliably indicate. I appreciate your time.
 
  #16  
Old 04-08-06, 12:44 PM
Member
Thread Starter
Join Date: Jan 2005
Posts: 135
Bolide:


> 2 cables go to unknown destinations.

Take the wire nut off the hot side and start killing breakers until it goes dead. Leave the breaker off a while and perhaps you'll find out what it powers. I just hope it doesn't go to a smoke detector or it might take a while to figure that out.
The hot side of the wire-nutted connection for the unknown cables goes dead when the bathroom breaker- the same one which shuts off the receptacle, switch, and fan/light- is turned off.

Unknown circuits are your enemy when you have a problem.
Ha! I see you've been to my neighborhood )


Of course, if it's already dead, cap both black ends with wire nuts. Put a note in the box that says, "Unknown circuit opened 2006-Apr-08".
Okay.


Trace out the circuit to the point where the hot voltage drops out.
It appears that the problem is on the cable that's feeding power. Unfortunately, this means removing much lathe and plaster. If this is the case, maybe the homeowner is better off updating the wiring rather than fixing the old one.


How about backstabbed wiring?
Are the wires on screws or just poked into the back of the switch or receptacle (as some people do with 14 gauge)?
The receptacle was indeed backstabbed. It was replaced and the wires were connected to the screw terminals. Thanks for the suggestions.
 
  #17  
Old 04-08-06, 12:48 PM
Member
Thread Starter
Join Date: Jan 2005
Posts: 135
HandyRon: I appreciate the link.

RockyMountain:

To clarify, the bad connection in the "hot" wires can carry a tiny amount of load. That's why the plug tester lights up when you plug it in, but when you connect an incandescent light bulb to the circuit, the bad connection burns itself open.
This seems to be the case here.

When the big load goes away, the tiny connection comes back together.
That would explain our confusion.

Given the fact that the non-contact voltage tester goes dark when the load is applied, you have isolated the problem to the hot wires. Do look at your neutrals as you wade through the circuit, but they are not causing this particular problem.
This narrows down the possibilities, at least.

I'd say calling a digital multimeter a "novelty" is waxing melodramatic, but I guess the point was made in a way the poster will remember.
It made a strong impression. Thanks very much.
 
  #18  
Old 04-08-06, 02:23 PM
ray2047's Avatar
Group Moderator
Join Date: Mar 2006
Location: USA
Posts: 32,803
Originally Posted by telecom guy
Does anyone still use the old time edison base incandescent with a pigtailed socket? I recall even being able to use one on limited 240v testing if you pick a 20w bulb. These put an end to phantom voltages/high resistance connections chasing you around.
My father used a tester with two series wired light sockets mounted on a small board and a pair of probes ( same wattage bulb in each). Half bright meant it was about 120 v. Full bright meant 240V.
 
  #19  
Old 04-08-06, 09:13 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
> The hot side of the wire-nutted connection for the unknown
> cables goes dead when the bathroom breaker-

These days the circuit that feed the bathroom may not feed any other loads.
But since you know that it feeds off the bathroom circuit, you can mark that.


> It appears that the problem is on the cable that's feeding power.

Via a GFCI?


> Unfortunately, this means removing much lathe and plaster.

Why? Do you have no idea where it goes?


> The receptacle was indeed backstabbed.

Perhaps there is yet another backstabbed device somewhere upstream that is supposed to be feeding your dead circuit.
 
  #20  
Old 04-09-06, 10:13 AM
Member
Thread Starter
Join Date: Jan 2005
Posts: 135
Bolide:



These days the circuit that feed the bathroom may not feed any other loads. But since you know that it feeds off the bathroom circuit, you can mark that.
Okay. The speculation is that a long time ago the cables may have fed ceiling and hallway lights, and hallway receptacles, but no one is really sure. We had 5 pairs of eyes looking around for what these cables could be feeding but found nothing. All the other lights, switches and receptacles are controlled by different breakers.


> It appears that the problem is on the cable that's feeding power.

Via a GFCI?
It's not a GFCI. It will be soon.


> Unfortunately, this means removing much lathe and plaster.

Why? Do you have no idea where it goes?
I believe the power runs from the breaker box in a closet, across the living room to the bathroom. The homeowner thinks that the wires run beneath the ceiling. The unknown cables leaving the bathroom junction box head up, which supports this conjecture.

The crawlspace on this side of the house is very low. It may be possible to feed wire without breaking the ceiling


> The receptacle was indeed backstabbed.

Perhaps there is yet another backstabbed device somewhere upstream that is supposed to be feeding your dead circuit.
It's possible that a device was covered up. But we've had 5 pairs of eyes checking every nook and cranny and we can't find anything else that seems to be on this circuit.

Is my reasoning sound here: since the wiring appears to be good in the bathroom from the receptacle--> switch-->j-box--> fan/light, then the problem seems to lie in either the cable feeding power from the breaker box to the bathroom receptacle, or in one of both of the unknown cables exiting the j-box.

If we disconnect the power feed cable and hook it up to a device without connecting it to anything else (like the fan unit), then we can determine whether the problem is on the power feed cable.

If the power feed cable is good, then perhaps we could disconnect the unknown cables in the j-box. Then we connect the power cable to a GFCI, the GFCI to the switch, and the switch to the fan unit. If it works, we're.... done?


Thanks for your help, Bolide. I realize that a proper re-wiring job is called for, but I'm just dealing with the reality of the situation.
 
  #21  
Old 04-09-06, 12:44 PM
bolide's Avatar
Member
Join Date: Jan 2006
Location: PA
Posts: 1,909
> It's possible that a device was covered up.
Or that a box was concealed.

Both actions are prohibited.


> Is my reasoning sound here: since the wiring appears to be good in
> the bathroom from the receptacle--> switch-->j-box--> fan/light,
> then the problem seems to lie in either the cable feeding power from
> the breaker box to the bathroom receptacle, or in one of both of
> the unknown cables exiting the j-box.

Could be. It seems obvious that you have an upstream problem.


> If the power feed cable is good,

If you don't have power at the box, how could it be good?
Do you have power at the panel?

The problem has to be between where you do have power and where you do not.
 
  #22  
Old 04-09-06, 08:05 PM
Member
Thread Starter
Join Date: Jan 2005
Posts: 135
Bolide,


> It's possible that a device was covered up.
Or that a box was concealed.

Both actions are prohibited.
True. But you can imagine the kind of things that have been done on old houses over the years. If I told you some of the knob and tube splice and dice jobs I've seen it would make you cringe.

> Is my reasoning sound here: since the wiring appears to be good in
> the bathroom from the receptacle--> switch-->j-box--> fan/light,
> then the problem seems to lie in either the cable feeding power from
> the breaker box to the bathroom receptacle, or in one of both of
> the unknown cables exiting the j-box.

Could be. It seems obvious that you have an upstream problem.
Okay. Further investigation will take place this week.


> If the power feed cable is good,

If you don't have power at the box, how could it be good?
I meant that the bathroom receptacle was never checked by itself on that circuit It was always hooked up to the switch, which was hooked up to all the rest.

If the power feed does come straight from the breaker to that receptacle, and that connection is tested independent of the switch et al, and it's good, then we could eliminate that run as the problem. Then we'd know that the problem was somewhere between the receptacle>switch>j-box>unknown cables, not between the breaker>receptacle.


Do you have power at the panel?
Yes.


The problem has to be between where you do have power and where you do not.
Right. The quest continues. Your help is invaluable, Bolide. Thank you.
 
Reply

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Thread Tools
Search this Thread
Display Modes
'