John T, et. al. Weird DC electric problem ??

[LJD]

I just encountered something I would not have believed if I had not seen it myself.

I have a cabin in the Adirondacks with off-grid solar electric. It's a 12 volt DC system. There are also two 3000 watt inverters to make AC. The cabin has two wire systems throughout. One part has outlets and lights that are 12 volts DC. The other part has outlets and lights that are 120 volts AC.
I mention this to explain why there are three service panels. One for DC, one for AC with mod-wave inverter #1, and another panel for sine-wave inverter #2.

Panel one is tied to the ground system (two 6 foot rods ten feet apart). The other two do not have the neutral bonding screws installed. So, I guess panel one can be loosely referred to as the "main panel."

This system passed inspection although the inspector did scratch his head a bit.

Now - here's what just happened.

My 12 volt DC lights are wired up with standard AC components. 10 gauge Romex NM cable, standard boxes, HD snap switches, etc. The only thing slightly special is the panel is Square D type QO. I had to use QO since it's one of the few AC panels that is also rated for DC (along with the breakers.)

To make this long story shorter . . . one of my lights had a 12 volt CFL bulb that worked fine. But - I just got some 12 volt LED bulbs and decided to try one. The reason being that LEDs can handle being turned off and on much better then CFLs.

So, I screwed in the new LED bulb. I then noticed it glowed dimly even when turned off. So - my first assumption was that the AC wall switch did not perfectly breaking current when used for DC. So, I unhooked the switch. Guess what? The LED bulb was still glowing dimly with only one wire hooked to it ! I did not think it was possible to make a bulb work at all with only one lead hooked to it. It was the negative lead, by the way. I checked voltage at the bulb when glowing and read 169 millivolts. Seems an LED bulb can glow a bit with extremely low voltage.

To fix the problem, I did this - but it was a lucky guess. I went down to the DC electric breaker panel and installed the bonding screw into what would be the "neutral" bus bar if it was hooked for AC. For DC, the "neutral" bar is for negative. So, shorting the negative bus bar to ground made the bulb go out. I checked amp flow at the bonding screw and it was .005 of an amp.

Here's the bulb with only one wire hooked to it:

<a href="http://s104.photobucket.com/albums/m162/jdemaris/?action=view&amp;current=100_0242s.jpg" target="_blank">

</a>

<a href="http://s104.photobucket.com/albums/m162/jdemaris/?action=view&amp;current=100_0235s.jpg" target="_blank">

</a>

DC wired panel (Square D rated DC and AC)

<a href="http://s104.photobucket.com/albums/m162/jdemaris/?action=view&amp;current=100_0243s.jpg" target="_blank">

</a>

<a href="http://s104.photobucket.com/albums/m162/jdemaris/?action=view&amp;current=100_0227s.jpg" target="_blank">

</a>

RMS voltmeter showing 169 millivolts to light when "off" (ignore reversed polarity)

<a href="http://s104.photobucket.com/albums/m162/jdemaris/?action=view&amp;current=100_0240s.jpg" target="_blank">

</a>

Amp reading through bonding screw that shorts negative to ground - showing .oo5 if an amp.

<a href="http://s104.photobucket.com/albums/m162/jdemaris/?action=view&amp;current=100_0228s.jpg" target="_blank">

</a>

Switch box with switch removed and light glowing as shown:

<a href="http://s104.photobucket.com/albums/m162/jdemaris/?action=view&amp;current=100_0239.jpg" target="_blank">

</a>

 

[Lou from Wi.]

Jd,as always, interesting post. I have a question for you though,, two states that I have lived in the electric company required two 5/8 X 8 ft long copper ground rods driven in the ground, spaced 10 ft apart using # 4 bare continious to connect them to the panel box. One state required a bonding screw in the panel box, the other required the # 4 bare going to the neutral bar in the panel box, Seems to me, the 8 ft would give you a better ground,than 6 ft.But it must be different strokes for different folks.

When I wired my daughters house with a 50 circuit Square D panel box, I drove a 5/8 X 8 ground rod and put it on the neutral lug in the panel box.We're on the same service as hers, our lights will flicker,but hers doesn't.I wonder why that is? Any suggestions??.
I enjoy your posts, as always.
Regards, LOU

 

[JMOR]

LDJ, I am not there to walk around with test equipment and make measurements/experiment with changes/etc., but will give a shot in the dark &amp; something for you to try. Guessing that the LED needs so little current to make some light, that capacitive or maybe inductive coupling is putting enough current into the DC circuits from the AC lines that it illuminates. Don't forget, LED stands for Light Emitting DIODE &amp; diode is a rectifier of AC.
Try it with inverters turned OFF. Does it still illuminate?

 

[MarkB_MI]

That's an interesting one, LJD.

First off, I think that 169 millivolt DC indication is a red herring. It takes about six-tenths of a volt to forward bias a diode, and I think you'd need a least a couple of volts ac or dc before the electronics in the bulb would turn on the LEDs. I suspect that's a side effect of having the LED bulb in the circuit, and that it's actually ac noise that's powering the bulb. That pretty much has to be the case, since the bulb turns on as soon as you hook just one wire to it.

Assuming the cause is ac noise, shorting the neutral to ground at the panel reduced the amount of noise available to power the light.

So why doesn't everyone notice this effect? My guess is it's because LED bulbs are seldom used by themselves on circuit. Usually there are other lights, either halogen or LED on the same switch, and that's enough of a load to prevent a single LED bulb from turning on.

 

[LJD]

The AC noise is an interesting idea. I wish now I'd thought of turning off the inverters to see if that light went out. No inverters, no AC. Next time I'm there I'll try it.

 

[LJD]

I could not get 8 foot rods into the ground at this location due to solid rock. I used two 5/8 galvanized rods hooked in series to a vertical steel well casing that goes in the ground 80 feet deep. It is an approved and inspected ground system.

Size of the ground wire is dependent on the size of the service entrance cable. A #2 service cable allows a #8 ground wire IF in conduit. 1/0 service cable allows #6 ground. 2/0 or 3/0 calls for #4 ground wire, etc.

Here in NY we are still using 2007 code. Not sure what the newest and greatest code now calls for.

 

[Dean]

I expect induced current caused by inductive coupling with the AC wireing is the cause.

I will ve VERY surprised if the lamps glow without the inverters operating unless there are HVAC transmission lines overhead.

Dean

 

[glennster]

i had a similiar situation at the old farm. it was an egg farm that had been shut down for about 10 years. sat on 160 acres. northwest end of the property sat 2 40 x 408 ft buildings that housed chickens, and at the north east end was 2 more 40 x 408 buildings. power had been cut at the pole feeding the west buildings, then underground ran thru the field to the east buildings. couple 100 yards to the north running parallell were steel power lines, 345kv and 186kv transmission lines. altho no power was to any buildings, the wiring showed 26 volts. could not figure where the voltage was coming from. electrician came out and found that the static from the transmission lines was supplying the power to the buildings. pretty weird.

 

[MarkB_MI]

Your inverters kick out a lot of high-frequency harmonics. There's a good chance LED bulbs are sensitive to them.

 

[souNdguy]

and i wonder what his meter would show on ac vs dc setting...