Problem with GFCI breaker on transformer

[hurk27]

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.

.

That would be great, I often wanted to drag out my old quad trace Tectronics, to do this myself, but have had so much going on, I could never find the time.

The one thing I noted is it seems the transient spike seems to be after the switch from the load, as putting the MOV across the load seems to have no effect on the transient (GFCI's still tripped as before), but putting across the line and neutral ahead of the switch does seem to lower the amount of the spike(GFCI's don't trip as often or not all of the GFCI's in the apartment tripped as before) but doesn't remove it totally, but for some reason placing it across the switch seem to work (none of the GFCI's in the apartment tripped).

I posted about this a long time ago, about putting a MOV across a switch for a ceiling fan in a master bedroom, back when we were still allowed to put the master bath receptacle on with the general bedroom circuit, and the results were the same, it stopped the bathroom GFCI from tripping, but that went away in 1999 or was it 2002? but then came the arc-fault breakers which until Indiana eliminated them in 2005, ceiling fans caused these to trip also, I had to find away to get around this problem as the GFCI's we were getting back then was not very good in blocking this.

 

[reddy kw]

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

Sounds like a solid install, code-compliant all 'round, maybe GFCI isn't needed for this one outlet.
Is it on branch circ. with other outlets that do require GFCI protection?

 

[ozark01]

heres the one I used: V130LA20A

and can be found here:V130LA20A

And don't forget the heat shrink for the leads

It just don't seem right that a miracle cure is only $1.99!

 

[hurk27]

Gar here is the data sheet for the above MOV, if you like reading, these are made by Little Fuse for Radio Shack.

V130LA20A datasheet

 

[ELA]

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.

.

Gar, hurk27,

I never want to discount someones direct practical experience with what works and what does not. What works- works.

EMC theory dictates that the MOV or snubber be at located at the source of the energy which means at the coil.

Of course there are scenarios and variables that affect the result in the real world application.

One reason not to locate the MOV at the switch is that this results in a transient current flowing through the 60 foot of wiring between the transformer and the switch. This section of wiring will look like an higher impedance to the high frequency transient, thus a larger voltage can still develop at the load. In addition this 60 foot of cable will radiate EMI like an antenna. Currents can be induced into nearby or parallel run cabling.


In a real world application (w/MOV located at switch) this additional cable impedance can act as a filter to help slow down the rising edge of the transient to help make
the MOV more effective (if the leading edge of the transient is faster than the MOV can react to)
This could be what hurk has experienced.

A lot depends upon where the load that is being negatively affected is located in relation to the offending transient generator along with routing of the interconnecting wiring. What works in one situation may not in another.

I would go with a snubber at the load. This both suppresses the transient leading edge and reduces the overall amplitude of the transient.

 

[hurk27]

Gar, hurk27,

I never want to discount someones direct practical experience with what works and what does not. What works- works.

EMC theory dictates that the MOV or snubber be at located at the source of the energy which means at the coil.

Of course there are scenarios and variables that affect the result in the real world application.

One reason not to locate the MOV at the switch is that this results in a transient current flowing through the 60 foot of wiring between the transformer and the switch. This section of wiring will look like an higher impedance to the high frequency transient, thus a larger voltage can still develop at the load. In addition this 60 foot of cable will radiate EMI like an antenna. Currents can be induced into nearby or parallel run cabling.


In a real world application (w/MOV located at switch) this additional cable impedance can act as a filter to help slow down the rising edge of the transient to help make
the MOV more effective (if the leading edge of the transient is faster than the MOV can react to)
This could be what hurk has experienced.

A lot depends upon where the load that is being negatively affected is located in relation to the offending transient generator along with routing of the interconnecting wiring. What works in one situation may not in another.

I would go with a snubber at the load. This both suppresses the transient leading edge and reduces the overall amplitude of the transient.

My question would be, does anyone manufacture a pre-packaged snubber with this type of design in this area?
Most of the snubbers I have used in the past were for general transistor switching designs in which the snubber was placed across the transistor in a similar fashion as placing the MOV across the switch, also snubbers have been used in this fashion in many DC circuits all except for the good old diode used across a relay coil, which just reverse shunts the inductive kick back, to the most simple snubber used in the distributor of most cars in the years of ignition points, it was a simple capacitor again across the points, that minimized the arcing across the points to allow them to last longer and to provide more of a fly-wheel effect for the coil.

I just don't remember any placed across the load except the diode or a RC type snubber, which was also placed across the switching method in some designs?

 

[reddy kw]

I still say the easiest solution is to keep children away from the switch.:grin:

 

[ozark01]

I got my two Radio Shack specials today and will try to go by the house Tuesday for my test. They are spending $80k on the pool, have been waiting since May to get it installed, and this weekend is the first weekend they are able to use it so I would not dare go by before Tuesday! (I got my sh&t out of the way the same day the pump was set! The EC was not going to be the bad guy here!) I'll post back as to what happens.

 

[ELA]

Yes prepackaged Snubbers are available. This came up in another thread. A person can do a search on snubbers.

One of the common uses are across the coil of contactors.

 

[ELA]

Heres one I found using "RC snubber":

http://www.branom.com/literature/snubber.html

See the application note at the bottom.

 

[hurk27]

Heres one I found using "RC snubber":

http://www.branom.com/literature/snubber.html

See the application note at the bottom.

Thanks for the info.
It is an RC type, fairly easy to make, all you have to remember is to use a carbon type resistor, not a wire wound, and again they even offer an optional wiring method of it across the switch, but it does say the preferred method is across the load, the only thing I see as a draw back is it would allow a higher voltage pass through, where the MOV would clamp at 375 volts@100 amps, not sure which would be more effective in stopping inductive kick back from tripping a GFCI, but it seems both methods should. like you said each case might be different, so a little experimenting in each case would be the best way to find out.
the three methods of installing one is across the load (line to neutral), across the line ahead of the switch (line to neutral) and across the switch( line to switch leg) to see which one works the best.

The link you gave in the previous post, also have the same MOV listed for the same use, but only says to wire it across the load, they list it as ILS1 HERE
which looking at the specs it is the same one I posted above, but it is listed for inductive kick back. so I'm just going to have to get me a 0.1?f cap, and a 47Ω carbon resistor, and see how well the RC snubber does for this problem?

 

[ozark01]

Here is a follow-up as to what I found when I installed a MOV across the switch contacts and across the hot and neutral that feed the primary on the transformer.

I first installed the MOV across the switch contacts and saw an increase in the tolerance of the GFCI breaker. Without the MOV the breaker would kick-out after two or three operations. With the MOV installed it would kick-out after fifteen or twenty switch operations. This was a tolerable condition in my opinion and in the opinion of the home owner.

With the MOV installed across the transformer primary leads the four LED lights started acting weird. They are the kind of lights that slowly change colors while in operation. With the MOV installed they started flashing randomly so that is as far as I got with this setup. I'm not sure if it would have helped with the breaker kicking out or not.

When I left I told the home owner to give it a try and if the breaker kicking became a problem let me know and I would come back and replace the GFCI breaker with a regular breaker since the code does not require the GFCI breaker in this application. I hope it works OK for her as is because I like the the GFCI breaker on the circuit due to the location of the switch. I also told her I would keep looking for a better...more reliable solution.

 

[gar]

100907-1700 EST

What are they for and where are these colored lights?

.

 

[ozark01]

Please refer to post # 1 & 3.

They are four 12V LED pool lights mounted in a fiberglass pool.

 

[gar]

100907-1830 EST

I thought you were talking about some indicator lights.

There is some other problem. An MOV of the proper voltage rating across the primary of the transformer should not make the load lights flash.

An MOV of the proper voltage rating looks like a moderately high resistance when the voltage applied to the MOV is below a threshold value. Only when the peak of the line voltage goes above about 1.5 times the peak of the normal AC waveform does the MOV start to conduct very much.

Get a scope, a pulsed adjustable current source, and study the characteristics of an MOV. Your test could be done with a low voltage MOV.

Is the thing you are calling a transformer really something else that is used to convert the 120 V AC to the low voltage current limited power to the LED lights, and not a normal transformer? Even if it is something else an MOV across the input should not cause a problem. An MOV in its normal state is a high impedance across a very low impedance source voltage.

.

 

[ozark01]

Here is a link to the transformer. As far as I can tell it is a regular transformer designed for pool lights.

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

 

[gar]

100907-1950 EST

It does appear to be a standard transformer. Per the written description it implies a quality design. Two separate molded coils, and an electrostatic shield between primary and secondary. An overload device is also included.

The enclosure must be properly grounded. Should have nothing to do with the problem. The electrostatic shield is probably tied to the enclosure.

Nothing here would imply a problem with an MOV across the primary.

.

 

[ozark01]

I'm going to try and find out more about the lights themselves. They have some circuity in them that allows them to operate in different patterns. I'll post what I find in case anyone is still interested. This thread is kinda long and boring I know...

 

[hurk27]

I'm going to try and find out more about the lights themselves. They have some circuity in them that allows them to operate in different patterns. I'll post what I find in case anyone is still interested. This thread is kinda long and boring I know...

I was thinking that when you tried the MOV across the transformer primary(120 volt side) you might have accidentally switch the mode to a different pattern, other wise like Gar says the MOV would not have any effect on the load? some of these mode switches are electronic, and switching on and off the power could cause the mode to change. do you have a web site for the lights?

seems strange these lights don't come with there own power supply? But then again I find it strange the the NEC doesn't require them to be listed as an assembly like it does for other low voltage lights:roll:

 

[ELA]

Here is a follow-up as to what I found when I installed a MOV across the switch contacts and across the hot and neutral that feed the primary on the transformer. .

Using a 130V MOV?
What does the nominal line voltage read at the transformer?

I first installed the MOV across the switch contacts and saw an increase in the tolerance of the GFCI breaker. Without the MOV the breaker would kick-out after two or three operations. With the MOV installed it would kick-out after fifteen or twenty switch operations. This was a tolerable condition in my opinion and in the opinion of the home owner. .

I am afraid the customer may not ultimately be real happy. The random nature of when it trips and when it does not depends upon the point in the AC waveform at which you happen to switch the load. The customer may be unlucky and happen to have it trip more often on them.

With the MOV installed across the transformer primary leads the four LED lights started acting weird. They are the kind of lights that slowly change colors while in operation. With the MOV installed they started flashing randomly so that is as far as I got with this setup. I'm not sure if it would have helped with the breaker kicking out or not..

That is interesting. It would be great if you had a O'scope so you could view the input waveform. Any chance the peaks would be clipped?
This is pure speculation and a bit of a stretch but what if the MOV were conducting at the peak of the waveform and "starving" the load of peak charge currents into the load side switching regulator?
((As I said "a stretch" and I mention it only because what you experienced does not make sense otherwise.))

Personally I would not use a 130V (maximum usable AC) on a 120V line that could possibly get as high as (+10%) or 132V? I would use a 150V rated unit. I know this is a fine detail but you will normally see a 150V unit used in equipment rated for 120V operation.

When I left I told the home owner to give it a try and if the breaker kicking became a problem let me know and I would come back and replace the GFCI breaker with a regular breaker since the code does not require the GFCI breaker in this application. I hope it works OK for her as is because I like the the GFCI breaker on the circuit due to the location of the switch. I also told her I would keep looking for a better...more reliable solution.

Is the GFCI breaker newer or older? Newer GFCI breakers probably have MOVs in them already.
That is why I recommended an RC snubber. The MOV clamps the peak of the transient but does nothing to limit the fast rise time of the transient prior to clamping. The snubber would limit the rise time. Using both an snubber and an MOV might be good if you wanted to try that.

Or if the breaker is old have you tried a newer one?