3 way switch 90 Volts when off

[JMahaney]

Why am I getting 90V with three way switch off.
I had a customer today wanting to be able to control his kitchen light by two switches. Should have been easy I’ve done it thousand times. So I started with running a three wire from his existing switch to a new switch location. Came off the existing hot and neutral for power. When I was done wiring it up and turned the circuit back on I got 90V from hot to neutral at the light fixture with the switch off. 120V when the three way was on.
I thought maybe pinched wire or induction.
I ran a new three wire on the ground this time just to test it to make sure my wire wasn’t damaged and I got the same thing 90 V in the off position. I’ve then ran a new home run from the panel to the switch, And also a new switch leg. Still had 90V from hot to neutral?

 

[ptonsparky]

What meter are you using?
High impedance?

 

[JMahaney]

What meter are you using?
High impedance?

Klein A/C DC Auto- Ranging. Basic $80 amp clamp meter

 

[ptonsparky]

Ran out of edit time. Dang

Verify the polarity of the existing switching.

 

[ptonsparky]

Is the original box a hot & switch leg from fixture?

 

[Flicker Index]

Well, the pair of travelers are usually in the same jacket and one of the two is always hot. So, it's just like energizing one prong of an extension cord to hot and measuring the other prong against ground. Because of the long parallel run, there's a fair bit of capacitive coupling.
For example, line side switch is linked to red. light side switch is linked to black. There is a capacitor (the lead capacitance between red and black).

Take an entire roll of Romex. Making sure that the inside end is not shorted, measure the uF between black and white. You'll get a measurable reading.

 

[gar]

210201-24051 EST

JMahaney:

You may have wired thousands of similar circuits, but you basically lack an understanding of electrical circuit theory.

First, do you have a neutral at the first switch box that differs little in voltage from earth, metal water pipes, or other things that may connect to your power transformer center tap and earth or whatever when any of the switches change state? Generally I would expect a volt or less, but a few volts might occur sometime.

Assume the neutral looks good at the first switch box, then test it at the destination load.

Assume it looks good at the load.

Second, connect a 100 W incandescent bulb at the first switch box location between the incoming hot and neutral. Bulb should be full brightness for either switch position of the first switch.

Third, leave the bulb connected to neutral, and connect the bulb hot to one of the switch outputs. You should be able switch the bulb on and off. Then move the bulb hot to the other hot output. Now the bulb should change state (on-off), but with opposite phasing of the switch position.

Fourth, this should be enough information for you to move on to the next switch position, and see what happens.

You also need to tell us where you see this 90 V. Is that across the load? If so, then the load should probably glow at a lower intensity.

Something is lacking in your description of the problem.

.

 

[don_resqcapt19]

Take a look at this document from Fluke....it applies to all high impedance meters. The link will open a PDF.

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwiW5e7b2cvuAhULlKwKHVTeCB8QFjABegQIAxAC&url=https%3A%2F%2Fdam-assets.fluke.com%2Fs3fs-public%2F2105317_A_w.pdf&usg=AOvVaw2wets5ydzi1sZAdB_bJcSJ

 

[hbiss]

I got 90V from hot to neutral at the light fixture with the switch off. 120V when the three way was on.

So I assume the switches and fixture work properly. Why did you even bother checking voltages? What kind of fixture- incandescent? Don't you think that with 90V at the fixture it would light pretty brightly when "off"?

-Hal

 

[LarryFine]

That's why I use a solenoid tester for troubleshooting, and use a voltmeter only when I need to know the exact voltage.

 

[hbiss]

I just use a light bulb in a pigtail socket.

-Hal

 

[gar]

210202-1946 EST

In the first post by JMahaney he is extremely unclear on how he wired the circuit, or exactly where he measured the 90 V.

If he actually measured the voltage directly across the load when he switched power off, then that light would still be glowing to some extent.

If the switching circuit was of an electro-mechanical contact type of switches in good condition, then the series impedance of that circuit should be more than 100 megohms when off. A load to read 90 V from a 120 V source would have to be many times that series impedance. Furthermore the input impedance of a Fluke 27 or 87 meter is only about 10 megohms. So something is fishy about the description of the problem.

Following is some useful information about a #14 copper 2 wire Romex cable with EGC centered between the two wires:

DC resistance of 100 ft of wire in series in the cable, 50 ft cable with white and black shorted at one end, making the loop length 100 ft, equals 0.28 ohms measured with a DC bridge.

With all wires in the 50 ft cable disconnected from each other, meaning EGC is just floating. The capacitance from black to white was 14 pfd per foot. Thus, 50 ft is 700 pfd, and Xc = about 3.6 megohms at 60 Hz.

With EGC and white tied together, and black isolated, the capacitance from black to the pair white and EGC was 22.6 pfd per foot. Thus, 50 ft is 1130 pfd, and Xc = about 2.1 megohms at 60 Hz.

If 120 V is applied to black relative to neutral, EGC wire in cable is connected to neutral (produces a partial shield between black and white), then with a Fluke 27 white to neutral reads 34 V.

If EGC is floated (EGC provides little shielding), and 120 V is applied to black, then white to neutral reads 103 V on a Fluke 27.

These experiments provide some idea of how cable capacitance can fit in.

Note: a Fluke has a nominal input impedance of 10 megohms. I did not try to lookup its input capacitance.

We need more accurate detailed information from JMahaney on his actual circuit, and measurements.

.

 

[gar]

210206-1516 EST

I doubt we will hear back from JMahaney. But I will continue.

On the previous experiments I ran on the 50 ft hunk of Romex there were results that clearly indicated the capacitive effect of the cable. But, the Fluke AC meter readings did not correlate well with the LRC bridge measurements on the cable.

Next, I rolled out the cable into an approximately straight line. Reran the LRC measurements. Negligible change. So the coiled wire was not a significant factor.

My bridge is a General Radio 1650-A from early 1960s. The fundamental design was probably created in the 1920s or 1930s. But long before this time bridge circuits were understood. The 1650-A has two available internal excitation sources, 6 V battery and a 1 kHz oscillator, also a 1 kHz filter or no filter. The capacitance values were measured with this instrument, and dissipation was also noted, very low. Results not much different than for the coiled wire.

Using nominal wall power 120 V 60 Hz excitation to the cable, assuming it is nearly a pure capacitor, and measuring the voltage with the capacitance in series with the meter I got estimated results for the capacitance of the cable considerably larger than the bridge measurements. Why?

I know the input impedance of the meter is about 10 megohms shunt resistance, but I don't know the input parallel capacitance, and don't want to bother trying to find out. In the end the input capacitance would have been of no use.

So the next step was to use a 1 K carbon resistor as a current shunt to measure capacitive current. One K is small compared to a couple megohms and thus should not introduce much error, and this low value compared to the meter input impedance washes out the meter input impedance as a major effect. Got pretty much the same result as with the meter as the load on the capacitor. Why?

Thinking about this I realized that I have a small amount of AC voltage waveform distortion near voltage peaks. This is from all the computer capacitance input power supplies hanging on the power lines today. This can be visually seen but is not large. But by by comparison with power line sine waves I had scope plots of from the early 1960s there was distortion not seen in the 1960s.

So do I do an FFT? That is too much trouble and does not provide a direct measurement of what I want. Still using the 1 K shunt for current measurement I got one of my capacitor boxes and put this in series with the cable capacitance. With the capacitance box shorted I got a shunt reading of 90 mV. Then I unshorted the capacitor box, and adjusted for 45 mV. Box read 1100 pfd. Good correlation.

.

 

[JMahaney]

Thanks for trying to help.

 

[ptonsparky]

Thanks for trying to help.

Did any of this make sense?

 

[James L]

Thanks for trying to help.

Y'know, this right here really perplexed me...

So I started with running a three wire from his existing switch to a new switch location. Came off the existing hot and neutral for power.

If you came off an existing single pole switch, why are you coming off hot and neutral for power? I think therein lies the problem.

If you have a functional single pole switch, and you add a second switch, all you do is run a 3-wire to the new location. Designate 2 conductors for travellers (red + white usually) and the third for common (black usually), which you tie in to either of the two that are connected to the single pole

Edited to add...
Unless you made a California 3-way, with hot + neutral connected to traveler screws on each end, then one conductor to the light from each switch. I could see some wonky voltage there with the switches off. That's not a good way to wire a 3-way

 

[LarryFine]

If you have a functional single pole switch, and you add a second switch, all you do is run a 3-wire to the new location. Designate 2 conductors for travellers (red + white usually) and the third for common (black usually), which you tie in to either of the two that are connected to the single pole

Actually, the correct way is to feed the new switch hot on the white (re-colored), and make the black and red the travelers back.

 

[James L]

Actually, the correct way is to feed the new switch hot on the white (re-colored), and make the black and red the travelers back.

I said usual, not correct

 

[JMahaney]

Y'know, this right here really perplexed me...



If you came off an existing single pole switch, why are you coming off hot and neutral for power? I think therein lies the problem.

If you have a functional single pole switch, and you add a second switch, all you do is run a 3-wire to the new location. Designate 2 conductors for travellers (red + white usually) and the third for common (black usually), which you tie in to either of the two that are connected to the single pole

Edited to add...
Unless you made a California 3-way, with hot + neutral connected to traveler screws on each end, then one conductor to the light from each switch. I could see some wonky voltage there with the switches off. That's not a good way to wire a 3-way

Y'know, this right here really perplexed me...



If you came off an existing single pole switch, why are you coming off hot and neutral for power? I think therein lies the problem.

If you have a functional single pole switch, and you add a second switch, all you do is run a 3-wire to the new location. Designate 2 conductors for travellers (red + white usually) and the third for common (black usually), which you tie in to either of the two that are connected to the single pole

Edited to add...
Unless you made a California 3-way, with hot + neutral connected to traveler screws on each end, then one conductor to the light from each switch. I could see some wonky voltage there with the switches off. That's not a good way to wire a 3-way

Y'know, this right here really perplexed me...



If you came off an existing single pole switch, why are you coming off hot and neutral for power? I think therein lies the problem.

If you have a functional single pole switch, and you add a second switch, all you do is run a 3-wire to the new location. Designate 2 conductors for travellers (red + white usually) and the third for common (black usually), which you tie in to either of the two that are connected to the single pole

Edited to add...
Unless you made a California 3-way, with hot + neutral connected to traveler screws on each end, then one conductor to the light from each switch. I could see some wonky voltage there with the switches off. That's not a good way to wire a 3-way

Did any of this make sense?

Yes your making perfect sense. I was just trying to say I was using the existing functioning wire from the old single pole switch to convert to a 3 way set up.
I do also believe a California three way was causing my issue. Further down on the circuit was the garage with an old California three way setup.