Problem with GFCI breaker on transformer

Status
Not open for further replies.
O

ozark01

Senior Member
Need expert opinions....please! :)

I finished wiring a pool today that has four 12v LED lights fed from a 300 watt Intermatic 120/12V transformer. The tansformer is being fed with a 15 amp GFCI breaker. The 120v is being broken by a single pole switch 60 feet away from the transformer.

The setup works as it should but if someone cuts the switch on and off in rapid succession (four or five times) the GFCI breaker kicks out. I know this is not the normal way to use the switch but the family has kids and it is possible for this to happen.

My question is "Is it normal for a GFCI breaker to kick out from in-rush current from a transformer when the power is cycled rapidly." and if is normal "Is there a way to fix the problem?"

Thanks for any help you guys can provide.
 
raider1

raider1

Senior Member
Staff member
Location
Logan, Utah
How close to the pool are the LED lights?

Remember that GFCI protection on the primary side of the LV transformer will not provide GFCI protection for the secondary.

As for the inrush current causing the GFCI breaker to trip, I doubt it as a GFCI breaker looks for an inbalance of current from the hot to the neutral and if this exceeds 4 to 5 milliamps it will trip.

Chris
 
O

ozark01

Senior Member
They are no-nitch lights mounted in the side of an inground fiberglass pool...below water level.
 
O

ozark01

Senior Member
I was just reading 680.23 (8) and it looks like I don't need a GFCI breaker if the transformer is listed and the secondary is 15 volts or less. Is this the case?
 
G

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
100902-1445 EST

ozark01:

It is not the inrush current. Instead it is the inductive kick from turning off the transformer.

If a current is flowing in an inductor and you attempt to instantaneously change the current, then the inductor will not allow you to do that. The inductor will generate whatever voltage is required to maintain that current. Thus, if no other means is provided a spark or arc will occur somewhere to maintain that current flow. It is this high voltage transient that is tripping the GFCI.

What is the GFCI brand and how recently was it purchased?

.
 
reddy kw

reddy kw

Member
Location
Mudville, upstate New York.
ozark01: If transformer is code-compliant it has 120 volt side/LV sides isolated. I'm not sure low-voltage induced current kick-back is problem here. HOWEVER, here's 2 suggestions: (1) Install dc-rated snap switch for the light switch. Modern 'quiet toggle' switches are "quick make-quick brake", don't suppress arcs enough, may permit kick-back current to be sensed by GFCI. (2) Add LV diodes to supply wires of LV luminaires. LED lights dissipate energy when power is cut. Diodes on LV line to lights, of suitable voltage/wattage (i.e. for automotive use) would stop kick-back current when power is cut.
Down side: It must be 100% idiot proof-we can't have customers getting electrocuted. It may be hard to eliminate GFCI protection for LV lights AND keep this safe for civilian use AND code-compliant. Here's a few reasons:
Isolation of primary is critical safety issue. Is transformer of the type spec'd in Art. 680.23 (A)(2)?
Primary/secondary barrier and tranny enclosure need to be grounded, so do non-current carrying metal parts of LV luminaires (unless LISTED for use ungrounded, see Art. 680.6[1]).
Non-GFCI power outlet can only supply power to <15 volt LV lighting. Tranny would have to be hard-wired within 20 ft. of pool. If tranny is plugged into DUPLEX recept., it would have to be GFCI protected since other equipment could be connected. ALL receptacles within 20 ft. of pool must have GFCI protection. See Art.680.22 (A)(1),(2),(3),(4). I'm pretty sure there's other applicable code issues, too.
All things considered, perhaps it's safer, easier, and more cost-effective to keep the GFCI and (a)change switch to 'fast make-fast break' arc-quenching type, (b)relocate switch to better location, (c)make switch tamper-proof, and (d)not let kids play with electrical equipment.
 
Last edited:
O

ozark01

Senior Member
gar said:
100902-1445 EST

ozark01:

It is not the inrush current. Instead it is the inductive kick from turning off the transformer.

If a current is flowing in an inductor and you attempt to instantaneously change the current, then the inductor will not allow you to do that. The inductor will generate whatever voltage is required to maintain that current. Thus, if no other means is provided a spark or arc will occur somewhere to maintain that current flow. It is this high voltage transient that is tripping the GFCI.

What is the GFCI brand and how recently was it purchased?

.
Click to expand...

It's a Cutler Hammer 15 amp GFCI breaker and it is new as of yesterday.
 
O

ozark01

Senior Member
reddy kw said:
ozark01: If transformer is code-compliant it has 120 volt side/LV sides isolated. I'm not sure low-voltage induced current kick-back is problem here. HOWEVER, here's 2 suggestions: (1) Install dc-rated snap switch for the light switch. Modern 'quiet toggle' switches are "quick make-quick brake", don't suppress arcs enough, may permit kick-back current to be sensed by GFCI. (2) Add LV diodes to supply wires of LV luminaires. LED lights dissipate energy when power is cut. Diodes on LV line to lights, of suitable voltage/wattage (i.e. for automotive use) would stop kick-back current when power is cut.
Down side: It must be 100% idiot proof-we can't have customers getting electrocuted. It may be hard to eliminate GFCI protection for LV lights AND keep this safe for civilian use AND code-compliant. Here's a few reasons:
Isolation of primary is critical safety issue. Is transformer of the type spec'd in Art. 680.23 (A)(2)?
Primary/secondary barrier and tranny enclosure need to be grounded, so do non-current carrying metal parts of LV luminaires (unless LISTED for use ungrounded, see Art. 680.6[1]).
Non-GFCI power outlet can only supply power to <15 volt LV lighting. Tranny would have to be hard-wired within 20 ft. of pool. If tranny is plugged into DUPLEX recept., it would have to be GFCI protected since other equipment could be connected. ALL receptacles within 20 ft. of pool must have GFCI protection. See Art.680.22 (A)(1),(2),(3),(4). I'm pretty sure there's other applicable code issues, too.
All things considered, perhaps it's safer, easier, and more cost-effective to keep the GFCI and (a)change switch to 'fast make-fast break' arc-quenching type, (b)relocate switch to better location, (c)make switch tamper-proof, and (d)not let kids play with electrical equipment.
Click to expand...

The transformer is an Intermatic PX-300. It is hardwired about 25' from the pool. The LED lights came with a 50' two-wire cord made onto the fixtures.


http://www.poolandspacontrols.com/products/transformers/safety transformers.aspx
 
G

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
100902-2235 EST

ozark01:

Originally I had thought that Leviton had done a good job on their new receptacles relative to line transients. However, recently I did some testing of an inductive load and could trip the GFCI, but not easily. I looked further at their circuit board layout, and in one area I can see a layout change that could further reduce their sensitivity.

I have no experience with your Cutler Hammer breaker.

There are techniques near your transformer primary that could reduce the problem.

.
 
H

hurk27

Senior Member
Location
Portage, Indiana NEC: 2008
Also using a two pole switch by running the transformer load in series across both contacts the way they do for DC, this breaks the load faster and provides more distance between the contact, but is not always fool proof as even having a longer gap, at some point in time, the contacts will still be close enough to arc, so it doesn't always work.

Another way to stop inductive kickback is to put a 130v MOV across the switch, this bleeds off the inductive kick slower into the load, and doesn't cause the GFCI to trip, but it will cause a very small leakage current to the load so the breaker must be turned off if the transformer needs to be worked on, I get them from Radio Shack for about $2.00 for two in a pack.

Now would it be code compliant????? the NEC doesn't say I cant put a MOV device on a switch?
 
ELA

ELA

Senior Member
Occupation
Electrical Test Engineer
Or put the MOV or snubber at the transformer, which is the source of the inductive kickback.
 
O

ozark01

Senior Member
ELA said:
Or put the MOV or snubber at the transformer, which is the source of the inductive kickback.
Click to expand...

If I put it at the transformer how would the MOV be wired? I understand putting it across the switch as Hurk27 said but not sure how it would be wired at the transformer.
 
G

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
100903-1623 EST

ozark01:

The MOV goes in parallel with the transformer primary at the transformer. This may eliminate the problem, or just reduce the problem. Additional filtering may be required.

You need to select an MOV rated for use on a 120 V supply. Additionally you have to choose an energy rating.

.
 
H

hurk27

Senior Member
Location
Portage, Indiana NEC: 2008
gar said:
100903-1623 EST

ozark01:

The MOV goes in parallel with the transformer primary at the transformer. This may eliminate the problem, or just reduce the problem. Additional filtering may be required.

You need to select an MOV rated for use on a 120 V supply. Additionally you have to choose an energy rating.

.
Click to expand...

Believe it or not Gar, putting it across the switch was the most effective in stopping inductive kickback, I had a low voltage lighting transformer in a kitchen that was kicking most of the GFCI's in an apartment, the runs were very short to the load center, and it would even kick GFCI's on the opposite leg in the panel, I tried putting the MOV across the load, across the line, which would reduce it across the line, but not eliminate it, finally in desperation (or just messing around) I put it across the single pole switch, and like magic the tripping GFCI's stopped.

The only explanation I can come up with is, it seems to bled off the charge more slowly across the switch and keeps it from building up, also I noticed that the spike seems to only occur on the line side of the switch and not on the load side, so putting it across the load seems to do nothing, while putting it on the line side will lower the peek value, but placing it across the switch (parallel with the contacts) will eliminate it all together.

Give it a try?
 
O

ozark01

Senior Member
I have a JB where the switch legs, 120v supply from the GFCI breaker, and the primary side of the transformer all come together. I'll get me a 130v MOV and give it a shot. It will be easy to try different configurations in this box.
 
G

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
100903-2138 EST

hurk27:

Interesting about the location across the switch. Sometime I will see if I can perform an experiment to look at the transient. Maybe ELA will also try some experiments.

.
 
Status
Not open for further replies.
Top