Why does fluorescent require ground

Status
Not open for further replies.
I

ironphill

Member
Location
Conroe, TX
I've been told a million times that fluorescent fixtures require an EGC to work, and I've actually seen it happen, but nobody can ever explain why. Shouldn't the EGC just be a ground fault path and nothing else? And if it does send something on the EGC, which is the only reason I can think of for it to require it, wouldn't that mean that there's current on every piece of bonded metal between the fixture and the panel?
 
The metal shroud of the fixture provides a capacitive coupling to help the arc to establish itself in the lamp. Once current is flowing, that coupling is no longer needed. Most of the fixtures require the lamp to be within ?" of the grounded metal of the shroud.

It has also been my experience that the reason for failure has also been one of the wires that provide filament voltage has come loose from the tombstone. One more thing is the silicon powder should be left on the lamps when re-lamping. The silicon powder causes any moisture on a cold lamp to bead up and prevents the arc from traveling on the outside of the lamp during starting. :smile:
 
Last edited:
so after startup it no longer needs the ground? If you were to come across wiring with no EGC, would it be feasible to jump the neutral to the enclosure like you would a GFCI? And what about the screw shell type fluorescent bulbs that go in regular fixtures? The ground shouldn't have anything to do with those, right? I appreciate the info by the way:smile:
 
ironphill said:
so after startup it no longer needs the ground? If you were to come across wiring with no EGC, would it be feasible to jump the neutral to the enclosure like you would a GFCI? And what about the screw shell type fluorescent bulbs that go in regular fixtures? The ground shouldn't have anything to do with those, right? I appreciate the info by the way:smile:
Click to expand...

ironphill, Except at the Main Service Disconnect, you never, ever, ground the

neutral to anything !!! I don't know what you mean by " like you would at a

GFCI ", but if you are jumping the neutral to an enclosure, stop doing it. OK ?

I see you are just starting out in the trade, if you think you are going to

stay with it, then, go to this site as often as you can it will really help.
 
I've seen some guys who've been around for a while do this. When we came across some old wiring that only had a hot and a neutral at a location that needed to be changed to a GFI, they jumped the neutral down to the ground lug to make it function properly. Why it works that way since it'll trip if you touch the neutral to ground, I don't know, but it does. And in 250.140 (in the 2002 code, since that's what I have handy), it says for ranges and clothes dryers you can "ground the frame to the grounded circuit conductor" on existing wiring provided a few conditions are met. What would be the best fix for a fluorescent that's acting up with old wiring with no EGC?
 
250.140 only applied to ranges and dryers..............this is no longer permitted for new construction, it only applies to existing range and dryer circuits.

As far as bonding the neutral to ground, it is only permitted at the SERVICE DISCONNECT.....no where else. If someone is teaching you to bond else where they are wrong and probably not a licensed electrician.
 
charlie said:
The metal shroud of the fixture provides a capacitive coupling to help the arc to establish itself in the lamp. Once current is flowing, that coupling is no longer needed. Most of the fixtures require the lamp to be within ?" of the grounded metal of the shroud.

It has also been my experience that the reason for failure has also been one of the wires that provide filament voltage has come loose from the tombstone. One more thing is the silicon powder should be left on the lamps when re-lamping. The silicon powder causes any moisture on a cold lamp to bead up and prevents the arc from traveling on the outside of the lamp during starting. :smile:
Click to expand...
With some fixtures this capacative coupling is enough to trip a GFCI.
 
ironphill said:
When we came across some old wiring that only had a hot and a neutral at a location that needed to be changed to a GFI, they jumped the neutral down to the ground lug to make it function properly.
Click to expand...
The jumper is not needed, never mind it's being dangerous and illegal.
 
ironphill said:
I've seen some guys who've been around for a while do this. When we came across some old wiring that only had a hot and a neutral at a location that needed to be changed to a GFI, they jumped the neutral down to the ground lug to make it function properly. Why it works that way since it'll trip if you touch the neutral to ground, I don't know, but it does. And in 250.140 (in the 2002 code, since that's what I have handy), it says for ranges and clothes dryers you can "ground the frame to the grounded circuit conductor" on existing wiring provided a few conditions are met. What would be the best fix for a fluorescent that's acting up with old wiring with no EGC?
Click to expand...
This one is right up my alley.:smile:
 
080826-2126 EST

Numerous times in the past I have used fluorescent lamps with only neutral and hot and they started fine. Keep in mind that three prong wiring did not come into general use until about the mid 50s and lots of fluorescents were in use before this.

See the history on fluorescent lamps:
http://inventors.about.com/library/inventors/bl_fluorescent.htm

I could not find a quick reference as to when the 3 prong plug became common. It was not before 1953, but may have been on instruments by 1955.
http://en.wikipedia.org/wiki/Domestic_AC_power_plugs_and_sockets

.
 
You're not really bonding the ground and the neutral if there's no ground at your location. I understand you're not supposed to do that, but what happens if there's a ground fault with no ground? All it would do is energize the metal. Isn't the whole point of an EGC to give all the metal bonded to it a ground fault path through the grounded conductor? you're not bringing your ground and neutral together again, you're just giving the frame of a range, or in this case a fluorescent fixture housing a way to trip if it gets energized (not to mention giving it the only available path to ground).
 
Check out (2005) 250.130(C) for various options to get an EGC to go where one hasn't been before. Had to use it recently on a hotel remodel. The original structure was built in the early 40s from what I understand. Yeah, a whole lot of hot and neutral only and even some traces of knob and tube. Talk about good times...
 
ironphill said:
When we came across some old wiring that only had a hot and a neutral at a location that needed to be changed to a GFI, they jumped the neutral down to the ground lug to make it function properly.
Click to expand...

What am I missing here, why do you need a ground wire?

Are you using a plug-in GFI tester rather than the "test" button on the GFI?
 
ironphill:

Read the whole thread that follows and see if you can understand how a GFCI works and why the EGC has nothing to do with the operation of a GFCI. Also note the great deal of misunderstanding about what a GFCI senses.

http://forums.mikeholt.com/showthread.php?t=102022

If you think the ground terminal on a GFCI connects to anything in the GFCI circuitry, then

First, measure the resistance with an ohmmeter between the ground terminal and individually to each of the input terminals. Of course the GFCI needs to be in its reset state, and obviously the input and output terminals can not be connected to anything else when using the ohmmeter. If the GFCI is new, out of the box, then you may need to apply power to the input to reset the GFCI. I am assuming the meter will read near infinity.

Second, if the first test shows a high resistance, then:
Use a 10,000 ohm 1 W series resistor and a 100 to 200 V DC source to check for leakage between the ground terminal and the input terminals. Measure the voltage drop across the resistor to determine leakage current. 1 V across 10,000 ohms is 1/10,000 A = 0.1 MA. A DC test would measure resistive leakage. An AC test will include current from the capacitance between the power components and the ground components.

I have not run this test yet, but I know what to expect with the Leviton GFCI with which I have experimented.

.
 
080827-1127 EST

Here is some very rough information on a Leviton 7899-W GFCI:

All experiments with GFCI reset, neutral jumped to hot, and between ground terminal and the jumper wire. Current with 10,000 ohms and a Fluke 27.

Capacitance with Tektronix LC meter 21 pfd, I believe about 300 kHz.
Capacitance with GR 1650-A bridge 8 pfd, at 1 kHz.
Calculated capacitance at 60 Hz around 2 pfd.

DC leakage with 24 V less than 0.1 MV across 10 k. Less than 0.01 microamp.
AC leakage with 120 V about 0.9 MV across 10 k. About 0.09 microamp.

DC is greater than 2400 megohms
AC measured at 60 Hz is 1333 megohms from capacitive reactance.
Assuming all capacitive, then C = 1/(2*Pi*60*1.3*10^9) = 2 * 10^-12 or 2 pfd.

I do not have time to run other experiments, but it appears the dielectric constant changes substantially with frequency. This I did not expect.

.
 
You're right about the GFI. I think we were just trying to get around the open ground thing, so that when you put a plug in tester in it, everything appeared to be fine. It was a short cut that I apparently didn't know was a short cut at the time. I appreciate the help guys.
 
080827-2100 EST

Ran one more capacitance check. This time with a bridge that I developed over 50 years ago for liquid level measurement. This is at 50 kHz and the result was 11.5 pfd. The GR bridge is a standard LRC bridge, my bridge is a balanced capacitor bridge, and the Tektronix meter is an oscillator frequency shifted compared to a fixed oscillator.

Looks quite clear that the dielectric constant has a large change with frequency.

.
 
ironphill said:
You're right about the GFI. I think we were just trying to get around the open ground thing, so that when you put a plug in tester in it, everything appeared to be fine. It was a short cut that I apparently didn't know was a short cut at the time. I appreciate the help guys.
Click to expand...

I'm glad to hear that from you, Like I said before, stick around and you will

learn something new every day. I've been at it for 41 years and still learn new

stuff all the time here.
 
080830-0858 EST

I have rerun the Tektronix LC meter capacitance check and now read 14 pfd instead of 21 pfd. I might have misread the meter or improperly held it. Note capacitance in this range is very small. Holding a plastic screwdriver handle in my hand and touching the LC meter probe to the screw driver shank causes a reading of 8 pfd.

More experimenting indicates the problem was how I held the GFCI.

.
 
Status
Not open for further replies.
Top