Diagnosing three way circuit - odd voltages

Status
Not open for further replies.

JHZR2

Member
Location
New Jersey
Occupation
Power Systems Engineer
Hi,

I'm diagnosing a pair of three way circuits that we found to be odd, because its operation started tripping a gfci. I had another look at it, and it appears that there are two sets of hots from two different breakers. In other words, the upstairs pair of three ways are fed via the gfci I mentioned, while the downstairs switches have another 120v source. I don't quite understand it.

What led me on to it is that one of the switches operates a light fixture, while the other operates an outlet. The outlet still has 56V on it when switched "off". Further, some legs seem to show 35V when switched in certain configurations. I've been reading about carter and California type three ways, but I'm not sure this is it. I'm trying to diagram it but even that's tough given the number of conductors. Any recommendations on how to deal with this??

Thanks!
 

growler

Senior Member
Location
Atlanta,GA
Hi,

I'm diagnosing a pair of three way circuits that we found to be odd, because its operation started tripping a gfci. I had another look at it, and it appears that there are two sets of hots from two different breakers. In other words, the upstairs pair of three ways are fed via the gfci I mentioned, while the downstairs switches have another 120v source. I don't quite understand it.

What led me on to it is that one of the switches operates a light fixture, while the other operates an outlet. The outlet still has 56V on it when switched "off". Further, some legs seem to show 35V when switched in certain configurations. I've been reading about carter and California type three ways, but I'm not sure this is it. I'm trying to diagram it but even that's tough given the number of conductors. Any recommendations on how to deal with this??

Thanks!

It may be tough to diagram the circuit but that's normally the only way.

Face it, wires are just wires until you figure out what they are doing.

I would start by locating the feed to the light and then power and then the travelers. Watch that they don't connect the neutrals from two different circuits togather.
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
Before anything else you need to make the voltage measurements with a low impedance meter or with a small 120V light bulb in parallel.
Phantom voltages from capacitive coupling can be complicated and confusing, so rather than trying to explain it all it is better to avoid it.
 

don_resqcapt19

Moderator
Staff member
Location
Illinois
Occupation
retired electrician
It was not unusual, around here, to find a hot and neutral from on circuit connected to the 3 way where the travelers should be and the hot and neutral from another circuit connected to the other 3 way in the same manner. You really had to watch the phasing on a service upgrade:)
 

J.P.

Senior Member
Location
United States
Those low voltage readings often times don't mean anything.

Isolate the hot and neutral feeding each circuit.

Isolate the travelers, make sure your neutrals from both circuits are not all tied together.

Old school used to power the light boxes, new school powers the switch boxes.

After that you should be all set.

Some lights don't get along with GFCI's. Don't run the switch from the gfi protected side unless you need to.

Often times when figuring stuff like this out its better to leave the power on to everything instead of running back and forth.

3 ways are dead simple, but there are a few ways to do them that can throw people. Any way they are done, the switches alternately break one leg and power the other.
 

JHZR2

Member
Location
New Jersey
Occupation
Power Systems Engineer
Here's a basic diagram:

0672DB55-CDEE-4548-BA33-77FC56D29743_zpsy5mnd77y.jpg


what sparked interest in this was the fact that when I made outlet #1 into a GFCI, and put the second set of conductors on the load (outgoing, protected) side, after that every time that the light was switched on, the GFCI would trip. I troubleshot the circuit, and learned that the power for the light and outlet #2, both of which are switched by three way switches with the hot conductor coming up from circuit #7, would trip the GFCI.

Well, notice that the neutral associated with the load common for the light, and the neutral for outlet #2 are twisted and soldered in the box that contains switch 2B and 3B. There is no traveling neutral on the runs between the two thre way switches.

I did identify the traveling conductors between switch 2 and 2B, and 3 and 3B by disconnecting everything, and using a DC source to identify conductor pairs and their routes.

So so my gut is that the GFCI in outlet #1 would trip because the light neutral somehow ties in to the neutrals for the downstream outlets and ceiling lights fed on the load side of outlet #1. Whe I put those outgoing load side conductors onto the incoming line side of the GFCI, along with the feeding conductors, then all works fine.

But I do see around 56VAC at the outlet #2, and in the light socket when I pull the bulb. The 56V is enough to dimly light the bulb in my outlet wiring tester, but it is not enough to make the incandescent bulb in the light fixture illuminate. I was using a fluke 179 and 87V, somtheynare pretty sensitive, but the (tiny) bulb in the outlet tester was indication to me that it is in some way, real. If I switch the outlet #2 three way to the position so there is 56v instead of 120v there, the GFCI I installed there does not illuminate it's "ready" light.

So right now I have a GFCI in the outlet #1 location, both the incoming and outgoing conductors are on the line side... As a result, switching the light on does not cause the GFCI in outlet #1 to trip the way it did when I had the outgoing conductors attached on the load side before.

Thoughts??
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
151012-2043 EDT

JHZR2:

Your odd voltage problem is a result of capacitive coupling to a floating wire (mean a wire that under certain switching conditions is connected to neither hot or neutral). GoldDigger has indicated this as your strange voltage condition, and he has described a means to eliminate the strange voltage readings.

Your original question is confusing as I read it.

Do you have two separate circuits with their own 3-way switches and loads? This would mean one set of 3-way switches controls one set of loads (could be sockets and lights), and is fed from one breaker, and a different set of 3-way switches controls a different set of loads, and is fed from a different breaker. And the two different circuits are in no way interconnected. Meaning that one set of 3-way switches is not intended to have any control of the load(s) of the other 3-way circuit. If this is the case, then concentrate on troubleshooting one circuit at a time by turning the other circuit's circuit breaker off.

Doing this, then what is the problem?

If these are separate circuits, then why confuse the issue by intermingling the discussion of the two circuits at the same time, they would be different entities and should be treated separately until something proves otherwise?

.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
151012-2133 EDT

JHZR2:

I had not seen your last post until I had made my post. You have done a good job of finding the problem.

Clearly the original wiring is wrong because the neutral for the switched loads is not run in close proximity to the associsted hot(s) for the switched loads. As wired this causes an unbalanced current in the GFCI when a light or socket load is on. But even before the GFCI was installed it was wrong because this creates a large unwanted magnetic field from the separated neutral.

One way or another you need to rewire the 3-ways correctly.

Until the neutral and hots for the 3-ways are wired correctly (meaning the hot wire and its balancing neutral current must be in close proximity) you will have a large unwanted magnetic field. The hot and neutral in a Romex cable with equal, but opposite currents, provides good magnetic field cancelation.

.
 

J.P.

Senior Member
Location
United States
If your not planning on replacing the 12/2 with 12/3 I don't see what you can really do besides not running out of the load side of that GFI....
 

JHZR2

Member
Location
New Jersey
Occupation
Power Systems Engineer
151012-2043 EDT

JHZR2:

Your odd voltage problem is a result of capacitive coupling to a floating wire (mean a wire that under certain switching conditions is connected to neither hot or neutral). GoldDigger has indicated this as your strange voltage condition, and he has described a means to eliminate the strange voltage readings.

Your original question is confusing as I read it.

Do you have two separate circuits with their own 3-way switches and loads? This would mean one set of 3-way switches controls one set of loads (could be sockets and lights), and is fed from one breaker, and a different set of 3-way switches controls a different set of loads, and is fed from a different breaker. And the two different circuits are in no way interconnected. Meaning that one set of 3-way switches is not intended to have any control of the load(s) of the other 3-way circuit. If this is the case, then concentrate on troubleshooting one circuit at a time by turning the other circuit's circuit breaker off.

Doing this, then what is the problem?

If these are separate circuits, then why confuse the issue by intermingling the discussion of the two circuits at the same time, they would be different entities and should be treated separately until something proves otherwise?

.

Sorry, I wasn't clear enough in my OP, because I was still trying to figure out the situation. The diagram above trumps all earlier definitions of the circuits by me.

End of the day, the following are the case:

1) the set of three way switches (#2 and 3) are powered by the hot that is switched by #1. There is no traveling neutral.

2) the loads coming off the common terminals on switches #2B and #3B have a neutral from someplace else.

3) if one connects the outgoing conductors from outlet #1, which is fed directly from the panel, to the load terminals of a GFCI outlet, the GFCI will trip when loads are turned on via any three way switch (2 or 2b; 3 or 3b), indicating that the neutral must be shared with the conductors fed from outlet #1, and that some of the current is returning that way, showing an imbalance in the GFCI and causing it to trip.

4) an incandescent bulb in the light socket that shows a 56V stray voltage does not illuminate, but the low impedance short that is the resistive element in the bulb is also not enough to mitigate these stray voltages.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
151013-0747 EDT

JHZR2:
4) an incandescent bulb in the light socket that shows a 56V stray voltage does not illuminate, but the low impedance short that is the resistive element in the bulb is also not enough to mitigate these stray voltages.
I assume this 56 V is measured relative to ground (earth), or a known good neutral, or a known good EGC. If the GFCI (what appears to be the source of the neutral at the light bulb) is tripped, then both HOT and NEUTRAL on the ouput side of the GFCI are open relative to the input side of the GFCI. This means that under the tripped condition there is no neutral at the bulb. Thus, whether the bulb is screwed in or not should not make much change in your stray voltage reading.

If you set the 3-way switches to the light bulb to the off state and with the GFCI tripped you may not see the 56 V value (meaning 0) or some different value, and also with no load at the socket, then reset the GFCI and with the bulb screwed in there should be no voltage on the hot or neutral side of the bulb. Unscrew the bulb and the bulb neutral should show no stray voltage, but the bulb hot side may show some voltage. All of these voltages measured with a 10 megohm input impedance meter.

.
 

kwired

Electron manager
Location
NE Nebraska
I did not read all posts thoroughly so sorry if I repeat something.

GFCI will not hold if you introduce pathways to other circuits whether it be another "hot" from same bus in the panel or another neutral or tying a protected neutral to a grounding conductor. I think you figured that out early on that more then one circuit is somehow tied together in what is there.

Voltage still present at a switched load (assuming the switch does open) is either capacitive coupling (make tests with a low impedance meter and you will eliminate capacitive coupling) or possibly the neutral conductor is what is being switched, or a little of both conditions may be present.

Voltage reading of 56 volts is about 1/2 of nominal 120 volts - equal division of applied voltage means you are measuring at/near mid point of the voltage drop of whatever you are measuring across.
 

JHZR2

Member
Location
New Jersey
Occupation
Power Systems Engineer
If your not planning on replacing the 12/2 with 12/3 I don't see what you can really do besides not running out of the load side of that GFI....


The wiring won't be replaced. I did jumper both conductors to the line side, and of course all is well.

It's annoying when there's different neutral conductors used, from different sources. This isn't knob and tube!!!
 

JHZR2

Member
Location
New Jersey
Occupation
Power Systems Engineer
I did not read all posts thoroughly so sorry if I repeat something.

GFCI will not hold if you introduce pathways to other circuits whether it be another "hot" from same bus in the panel or another neutral or tying a protected neutral to a grounding conductor. I think you figured that out early on that more then one circuit is somehow tied together in what is there.

Voltage still present at a switched load (assuming the switch does open) is either capacitive coupling (make tests with a low impedance meter and you will eliminate capacitive coupling) or possibly the neutral conductor is what is being switched, or a little of both conditions may be present.

Voltage reading of 56 volts is about 1/2 of nominal 120 volts - equal division of applied voltage means you are measuring at/near mid point of the voltage drop of whatever you are measuring across.

I definitely verified with the travelers and commons that the hots are switched. No issues there. My only concern is that when I have an actual bulb in the fixture, or an outlet tester in the socket (which does glow dimly), the voltage is still observed.
 

JFletcher

Senior Member
Location
Williamsburg, VA
Thanks, nice reference!

You're welcome. today, we had a call for troubles with 2 sets of 3 ways (separate circuits). One set wouldnt switch on a ceiling fan light, the other set would work a hallway light then after half a dozen or so switch cycles on both sides, the hallways light would stay on, regardless of either 3 way position.:blink: We found all 4 3 ways bad, as they would 'half switch' or not switch when barely toggled or shifted sideways a hair, causing, well, a slew of problems that were easier fixed by replacing switches than measuring voltages, continuity or tracing wires.

The HO had removed two of the four switch box covers, but gotten no further. This one almost stumped us, as seeing 4 3 ways failed is not an everyday thing. I guess it makes sense tho in that if the switches are used evenly, then it stands to reason they'll fail fairly close to one another, especially after the first fails, the second is going to see 20x the use it did ("why isnt this working? *up down up down up down up down....*), until it too fails.

I dunno how 3 ways are constructed internally, however, apparently, they can fail in any number of ways causing head-scratching problems. After today, I will *almost* always suspect a switch (a wear item) over a wiring issue, unless the HO has recently replaced a light or ceiling fan.
 
Status
Not open for further replies.
Top