phantom voltage

[Rattlesnake308]

I am a student at rosedale technical institute i wired a project for my instructor as a lab grade. iwired a three way switch to a half switched receptacle my project is wired in mc cable with metal boxes on wooden studs the electricity is coming from a gfi receptacle we have an extension cord with the female end cut off and three alligator clips in place for hooking up to our projects. the bottom receptacle is hot all the time the top is switched bottom read 113.7 volts the top reads 113.7 volts when travelers energize the receptacle on top but when one of the three ways turns off the receptacle i see a voltage of 33.5 volts the load is not coming from the neutral i checked that with my meter replaced all three components checked again still seeing the voltage when switches are supposed to have top receptacle off can anyone tell me where my voltage is coming from my instructor is as baffled as i am my original thought was that the voltage was being induced in the mc cable but all boxes switches and receptacle are grounded the circuit is wired properly as the plug tester shows any ideas would be appreciated

 

[Buck Parrish]

Chances are the circuit is not wired properly. Is it possible that the cords male end is plugged in wrong. You might flip it over and try it. You know this but the big hole is the neutral in the outlet.

Can you be more specific how you wired it, Do your travlers go from one switch straight to the other. Or in a light box.

It's a slow night here , come on back with more details. We will fix it so you will get an A+.
Then may be your instructor will get a video from Mike Holt on Harmonics. That will make for some interesting class discussion.

Like the big lady told me. When I asked her what was wrong. She says " Their be a SHORTAGE goin on" "We's got a SHORTAGE" in hare.

 

[iwire]

Chances are the circuit is not wired properly.

I think chances are the circuit is wired fine and the 33 volts that is being read is just capacitance coupling with the energized conductors.


If you had a load connected to this circuit the reading would disappear.


Phantom Voltages

Regardless of the name it is a real voltage with extremely low current.

 

[iwire]

Voltage on switch leg. Switch off

 

[brian john]

Chances are the circuit is not wired properly. Is it possible that the cords male end is plugged in wrong. You might flip it over and try it. You know this but the big hole is the neutral in the outlet.

Can you be more specific how you wired it, Do your travlers go from one switch straight to the other. Or in a light box.

It's a slow night here , come on back with more details. We will fix it so you will get an A+.
Then may be your instructor will get a video from Mike Holt on Harmonics. That will make for some interesting class discussion.

Like the big lady told me. When I asked her what was wrong. She says " Their be a SHORTAGE goin on" "We's got a SHORTAGE" in hare.

99.9% Bob has it (especially based on the posters comments) there are several threads in Mike Holt regarding this topic, very common and very common for electricians to be unaware of a this.

 

[Buck Parrish]

Hey , I was wondering where you we're at Iwire.

Rattlesnake you should get an A+ now.
Read the link Iwire provided . Print that out and bring it to school. Then try a low impedence meter.
I too , shall remember that. Thanks Iwire.


I hope Rattle snake comes back. That's very helpful information.


But did you like my part about the " Shortage going on" that really happened to me. It was a main phase out.

 

[K8MHZ]

I think chances are the circuit is wired fine and the 33 volts that is being read is just capacitance coupling with the energized conductors.


If you had a load connected to this circuit the reading would disappear.


Phantom Voltages

Regardless of the name it is a real voltage with extremely low current.

I attributed it to induction, not capacitive coupling.

Either way, a solenoid tester is the only way to go when checking 3 way switches. In fact, any time a DVOM shows unexpected voltage readings a low impedance test should be performed.

Home Depot has Ideal solenoid type testers for 20 bucks right now.

 

[gar]

090509-1238 EST

Rattlesnake308:

Almost certainly you have capacitive coupling or leakage resistance causing the reading in the off state.

It would be much easier to follow your train of thought with some paragraphs to separate the various components of your discussion.

From your first single pole double throw switch, the one connected to the hot wire of the power source, to a second SPDT switch that connects to the hot terminal of the switched receptacle, you have two wires in close proximity to each other. When one of these wires is hot the other is floating. There is capacitive coupling between the two wires. About 20 pfd per foot for Romex. Not much different for other similarly spaced wires.

Suppose you had 5 feet at 20 pfd per foot. Then the total coupling capacitance is 100 pfd. The capacitive reactance of 100 pfd at 60 Hz is Xc = 1/(2*Pi*f*C) = 10^12/(6.28*60*100) = 10^9/6.28*6 = 1000/37.7 megohms = 26 megohms.

If the meter you are using is a 10 megohm input resistance, then as a real ball park figure the voltage at the meter input is 120*10/(10+26) = 33 V. This is an invalid calculation for several reasons. I am assuming the reactive component is in phase with the resistive component. This is not so. Further I assumed the meter input was pure resistance. It is not. I really need to know more about the meter. Then I could draw the vector diagram and get more accurate results. But really the point here is you want an explanation for the cause and not precise calculated values.

To prove it is capacitive coupling or high resistance leakage add some shunt resistance across the meter terminals. Put 100,000 ohms in parallel with the meter and it will probably drop to something in the range of 1 V (0.3 V).


K8MHZ:

If by induced you mean inductively (magnetic field) coupled, then it is not.

Two reasons:

First, the meter circuit has no closed conductive path unless you include the capacitive path.

Second, the induced voltage in a coil can be determined from the equation:
v = K*N*dp/dt
v = induced voltage
K is a constant that is dependent upon on the shape of the coil
N is the number of turns
dp/dt is the rate of change of the flux coupling the coil

Just using a meter with a pair of test leads connected to something creates a 1 turn coil. If this coil is in proximity to a wire carrying an alternating current then a voltage will be induced. The important question is how big is the induced voltage.

I have a 2200 turn coil with a 0.5" x 0.5" window. By getting close to my main panel I get 0.012 to 0.015 V induced. If the coil was one turn, then the voltage would be around 7 microvolts.

Sometimes one does have to be concerned about the 1 turn measurement loop. I had this problem many years ago in measuring the current and opening time of a breaker in a short circuit test. The source was a number of 6 V car batteries in series with heavy cable and the breaker. Upon closure of this circuit the current rose very rapidly. Therefore, dp/dt was very large and the induced voltage produced an error many times the voltage from the shunt being used to measure the current. This required designing a shunt to eliminate the induced voltage.

I can create a large one turn coil, maybe 50' by 10', put this on the ground below my primary lines with some 30' to 35' between the primary lines and the coil and get fractional volt induced in the one turn coil.

Therefore it would be quite difficult to induce 33 V at 60 Hz into a 1 turn coil of reasonable size.

.

 

[K8MHZ]

If by induced you mean inductively (magnetic field) coupled, then it is not.

I don't want this to go OT, but at 60 Hz, I have to disagree with you. At 150 MHz, I would agree with you about capacitive coupling as opposed to inductive coupling, not at 60 Hz.

However the coupling is caused, I see it all the time in the field.

Using the proper equipment is the key. For testing AC voltages, a solenoid tester should be used unless the lower impedance poses a danger to the circuit in some way.

can anyone tell me where my voltage is coming from my instructor is as baffled as i am

Without mincing words, your instructor needs to find a job he or she is better suited for. Voltage readings taken with high impedance meters on three way switch circuits should only baffle a layperson, not an instructor.

 

[brian john]

Without mincing words, your instructor needs to find a job he or she is better suited for. Voltage readings taken with high impedance meters on three way switch circuits should only baffle a layperson, not an instructor.

I would hope the instructor knew this. BUT I make a good living explaining stuff I would think most electrician should know. Many in the trade do not stay up to date with many changes and or common issues in the trade.

 

[gar]

090518-1030 EST

K8MHZ:

Your comment

I don't want this to go OT, but at 60 Hz, I have to disagree with you. At 150 MHz, I would agree with you about capacitive coupling as opposed to inductive coupling, not at 60 Hz.

is wrong.

I have explained this in other threads but will do it again.

To have a voltage generated by inductive coupling from another conductor there must be current flow in the other conductor. The magnetic flux produced around a wire is proportional to current. Zero current zero flux. The magnetically induced voltage in a conductor is v = K*df/dt where v is the induced voltage, K a constant, df/dt the rate of change of flux.

Capacitive coupling will be a function of the capacitance and resistance of a series circuit and the associated voltage divider result.

An experiment to illustrate why we are concerned with capacitive coupling in this "stray voltage" question, and not inductive. The meter is a Fluke 27, and the circuit is a 3 wire 50 ft extension cord. Test load about 12 A.

First: the EGC is left floating. No load current thru the circuit. The hot and neutral wires of the extension cord are connected to a 120 V source. The voltage between the floating EGC and neutral is 52 V. The meter input is 10 megohm shunted with some capacitance, maybe 100 pfd. This voltage is not inductively coupled because there is negligible current flowing. The cord is in a coil of about 20" diameter.

Second: the EGC is still floating, and the same coil configuration. The meter is connected from EGC at the plug end to the EGC at the socket end. Now an 11 turn 20" diameter coil exists connected to the meter. The residual induced voltage is 0.001 V.

A load current of about 12 A was connected to the socket end of the extension cord. The magnetically induced voltage is 0.002 V. Clearly a voltage level much lower than the capacitive coupled value of 52 V from the first experiment.

Third: same experiment as the second except uncoiled the cord into an undefined 1 turn loop. The no current reading was 0.001 V and with 12 A the reading was 0.002 V.

Fourth: a 2200 turn coil on a 0.5" x 0.5" coil form showed little induced voltage until it was in close proximity to the extension cord 20 " coil. Then the reading was about 0.015 V.

You do not magnetically induce large voltages without a large number of turns in a coil and high rates of change of flux coupling that coil.

These concepts are fundamental to high school physics and are extensively covered in lower level electrical engineering courses. One reference is "Electric and Magnetic Fields", Stephen S. Attwood, John Wiley & Sons, 3rd edition, 1949.

The approximate capacitance of my extension cord from hot to EGC is 1400 pfd. The capacitive reactance at 60 Hz is about 2 megohms. This would predict a greater voltage on the meter than I read, but the circuit is not a single capacitor and resistor.

.

 

[gar]

090518-2021 EST

I should be able to approximate an equivalent circuit for my 50 ft extension cord that would more closely predict the voltage I measured with the floating EGC.

I have now measured the capacitance of each pair combination.
Neutral to EGC is 1390 pfd
Hot to EGC .... is 1460 pfd
Hot to neutral . is 1430 pfd

Hot to neutral is simply a shunt across the supply voltage. The other two capacitors form a voltage divider from the source voltage. Thus, the equivalent circuit for the source, excluding the meter input capacitance, becomes a voltage of 123* 1460*1390/(1390*(1460+1390) ) = 123*1460/(1460+1390) = 63 V. The source impedance for this equivalent circuit is a capacitor of 146+1390 = 2850 pfd. Xc for 2850 pfd at 60 Hz is 930,000 ohms. With this as the equivalent circuit the meter reading would be close to 63 V.

There is input capacitance to the Fluke and this would lower the voltage a little.

This is probably a fairly good equivalent circuit.

If I applied power to only the hot wire and measured the floating voltage of EGC, then I do have something closer to the inter-wire capacitance between hot and EGC plus the series capacitor from hot to floating neutral and from floating neutral to EGC. This circuit would predict something closer to full line voltage. Experimentally I get 109 V.

.

 

[Rattlesnake308]

thanks for all the info i have megometered all the conductors checked with a low impedience multimeter megger says no shorts between the conductors and the outer jacket of the mc also the light on a wiggy for continuity lights up al the posts were great though and i had an A on the project before i originally posted learned lots from you guys or ladies whichever the case maybe thanks again for taking the time to help a newcomer in the field

 

[K8MHZ]

gar,

Thanks for taking the time for such a detailed explanation. I think I see where you are going....

But, I still have questions. They would be OT so I will save them for another time.

 

[gar]

090520-2019 EST

K8MHZ:

If you have some additional questions, then go ahead and ask? The questions might help others understand what happens under similar conditions.

Some more measurements I made. Anaconda NM 12-2/12 GRD 40 or so years old. About 18" long. Hot to EGC or Neutral to EGC 25 to 26 pfd. Hot to neutral 20 pfd. This is to be expected. More insulation and greater spacing between Hot and Neutral. Wires not in a cable will have less predictable capacitance values.

If you consider a cable such as Belden 8723, two twisted pairs, each pair individually shielded, then there are quite different results, and one has to define the test and operating conditions. If the shield is grounded and the excitation to the twisted pairs is referenced to ground, then there is virtually no capacitive coupling from one pair to the other. If the shield is floating, then there is capacitive coupling from one pair to the other.

.

 

[K8MHZ]

If you have some additional questions, then go ahead and ask? The questions might help others understand what happens under similar conditions.

Thanks, but I think I will sit and read for a while first, do some research on my own and then if mud hasn't cleared I will try to pose an understandable question.

I will tell you one thing, my main interest is test equipment and exactly how each device works. I have been collecting electrical test equipment for years. I even have a working Geiger counter and a mho meter. I will probably never use them, but I think they are cool to have. I have 'home brewed' test equipment for years, including battery testers for cell phones, logic testers for automotive applications, a de-gausser, short circuit tracers, radio direction finding equipment, test lights and the like.

The more I have the more I want.

I've got it bad, don't I....:roll:

EDIT: I remember the day, some 25 years ago, that my beloved Triplett VOM was flicked into the afterlife by an unfortunate fall off a 12 foot ladder. She was a beaut. I still have an analog meter but my Triplett was my first love and will always have a place in my heart. Sniff...

 

[76nemo]

Thanks, but I think I will sit and read for a while first, do some research on my own and then if mud hasn't cleared I will try to pose an understandable question.

I will tell you one thing, my main interest is test equipment and exactly how each device works. I have been collecting electrical test equipment for years. I even have a working Geiger counter and a mho meter. I will probably never use them, but I think they are cool to have. I have 'home brewed' test equipment for years, including battery testers for cell phones, logic testers for automotive applications, a de-gausser, short circuit tracers, radio direction finding equipment, test lights and the like.

The more I have the more I want.

I've got it bad, don't I....:roll:

EDIT: I remember the day, some 25 years ago, that my beloved Triplett VOM was flicked into the afterlife by an unfortunate fall off a 12 foot ladder. She was a beaut. I still have an analog meter but my Triplett was my first love and will always have a place in my heart. Sniff...

What kind of equipment do you own to measure Siemens?

 

[K8MHZ]

What kind of equipment do you own to measure Siemens?

A Hickok Micromhometer which is real old and I don't know if it works properly or not. It's just one of my collection.

I suppose you could also do the inverse of an ohm reading. For instance 10 ohms would equal .10 mhos or Siemens.

 

[gar]

090523-0549 EST

76nemo:

What do you want to measure?

You can measure an 0.001 ohm resistance with a 1500 watt heater and a Fluke 27. This would give you a resolution of better than 1 foot with #12 copper wire.

.

 

[e57]

Phantom Voltages

Regardless of the name it is a real voltage with extremely low current.

Thank you Bob - I was about to have a kiniption... It has been said, I needn't deal with the rest....