DIY Home Improvement Forum banner
21 - 36 of 36 Posts

·
Banned
Joined
·
5,990 Posts
Something odd happened with this GFCI today! I was hammering in Romex staples near where the GFCI and it tripped from the vibrations.
So now, I'm wondering if the GFCI is defective?
Are these usually sensitive to any sort of physical vibration etc?
I suppose they shock-test these things to some level of G force. Dropping a wrist watch on a hard floor is 5000 Gs, I think.

Does it reset?
 

·
Banned
Joined
·
5,990 Posts
Oh yes! And test's OK as well! :huh:
I'm inclined to think it shouldn't have done it.

You can test it independently for a dollar or so in parts by running resistors from the hot outlet slot to ground.
If you use 120v in Canada
for 1 mA use a 120k 1/4w resistor
for 2 mA use 62k 1/2w
3, parallel these two resistors.
4mA, use 30k 1w
5mA, parallel #1 and #4
 

·
Master Joatmon
Joined
·
28 Posts
I finally got around to picking up a 15A GFCI to replace the 20A. Unfortunatley, the store only had Levitron. So, the box says 15A and moulded into the plastic is 15A. However, there is a white sticker on it that says 20A. WTF? :eek: I examined the 20A GFCI I just removed. Same thing, 15A moulded into the plastic, but with a similar white 20A sticker. The only differences between these two GFCI's is that the new one is brown, the old one is white (brown goes better with the brown glass tiled backsplash); the new one is "Tamper Resistant" and the old one is not. I am going to go ahead and install the brown one this weekend and see if simply having a different unit from a different lot will produce different (hopefully favorable) results....
 

·
Just call me Andrew
Joined
·
2,279 Posts
Discussion Starter · #25 ·
A 20A outlet has a "T" shaped slot on it. It allows you to plug a 20A device (which also has a "T" shaped plug) into it.

A 15-Amp outlet is rated for 20A "passthrough", meaning you can put it on a 20A circuit but not plug a 20A device into it. If your old outlet didn't have a T slot, then it was a 15A outlet.
 

·
Master Joatmon
Joined
·
28 Posts
Bingo secutanudu!

Yep, both GFCI's I have do not have a T-shaped slot.

I installed the new, brown GFCI and apparently it has a different trip point. Now I can turn on and off my upstream fans to the GFCI at will and it does not trip. :thumbup:

I tossed the white one in the trash.
 

·
Master Joatmon
Joined
·
28 Posts
Yeah. Both are Levitron. I really didn't have a choice. It's all the hardware store had. I rolled the dice hoping that a different lot from the same manufacture would produce different results. I guess I hit the GFCI lottery. :laughing: If it does it again, I will put more effort into looking for non-Levitron. Perhaps buy online.
 

·
Registered
Joined
·
38 Posts
Must be Levitons...

What manufacturer gfci do you have? Mine is a Leviton 15A from home depot. I have some in the house from lowe's - I think cooper. I'll try swapping them.

Also - I think it only trips when flipping the switch slowly.

Maybe a different switch - like a fancy timer switch - would help.
I am having the same issue in my daughter's bedroom with a ceiling fan. I replaced a non-grounded outlet with a Leviton SnapLock (might not be the right name, threw out the package). The fan is on the load side of the circuit from the GFCI. When the fan is turned off the GFCI trips. I've replaced the GFCI with another Leviton and same issue. Moved the old GFCI to the bathroom and had no issues with it. I also replaced the switch for the fan and this also made no difference.

Its not consistent either. It seems to not trip if the lights in the fan are turned on when I turn off the switch. This is leading me to believe when I turn of the switch with the fan on, the fan is essentially becoming a generator and putting a voltage on the wire to the switch which the GFCI is somehow picking up and tripping. With the lights on, the voltage is dissipated.

(first post on the forum, so sorry if its long... been searching this forum for a while looking for a couple of issues I have. This is one of them so if the Cooper idea works, I know where I can get one).
 

·
Banned
Joined
·
5,990 Posts
When the fan is turned off the GFCI trips.
It seems to not trip if the lights in the fan are turned on when I turn off the switch.
GFCIs are supposed to trip when they sense a current difference. The trip time depends on the current, according to T = (20 / I)^1.43, so 20 mA of leakage current will trip in less than one second and 5 mA will trip in less than 7 seconds.

Anyway, this current difference is a 'differential mode' signal.
When the current in both lines changes equally, like when you turn something on or off, this is a 'common mode' signal and is supposed to be rejected by the GFCI. The higher the 'common mode rejection ratio' the better the device does its job.

The current change in both lines is larger when there are no other loads on the GFCI, so if you have a steady load of a few amps the fan turning off won't make that much of a difference.

I was unable to find GFCI patents relating to methods of getting around these spurious responses to large, abrupt, common mode signals. If you can find some, like by Googling
GFCI
the assignee company is a probably a good place from which to buy one of these.

Good addt'l search keywords would be
abrupt
change
current
nuisance
false
rejection
large
load
differential

BTW, if you have a good idea, others will steal it. It's the way of the world!
http://ewweb.com/mag/electric_leviton_wins_gfci/
 

·
Super Moderator
Joined
·
14,959 Posts
GFCIs are supposed to trip when they sense a current difference. The trip time depends on the current, according to T = (20 / I)^1.43, so 20 mA of leakage current will trip in less than one second and 5 mA will trip in less than 7 seconds.
You might want to compare your "facts" vs the published specs provided by Hubbell. Hubbell state the response times at .025 seconds. I guess you could say that .025 seconds is less than 7 seconds, but I wouldn't want to depend on your "stats". Are you even aware of the amount of damage that could be done if UL allowed trip levels like you allege?

Here is the link, check the bottom of page 3.

https://www.hubbellnet.com/max_htm/PDF_Literature_Library/h4511.pdf
 

·
Tool Geek
Joined
·
2,590 Posts
--Trip Times--

Yoyizit is correct about GFCI's not tripping instaneously.
While GFCI's will typically trip in 25 ms or so at fault currents exceeding 20 to 30mA, they are permitted by UL to take several seconds to trip at fault currents in the 6mA range.


Source:
http://ecmweb.com/mag/electric_think_gfci/
Ref: 2nd paragraph after Why GFCI's

The Hubbell Spec qualifies their 25mS number by adding Nominal
 

·
Registered
Joined
·
1,149 Posts
this is an interesting topic Yoyizit. where dose 1.43 come from? at first i thought you meant peak current, which is 1.41, but you seem firm on 1.43.

as for minding against spurious faults at turn on, what i think is happening is the large inrush of motor in the fan (5-7x nominal) is creating a high di/dt (rate of change of current) which is parasitically coupling to ground. since some of the current is going to ground, instead of returning to the source, it triggers the GFCI as a differential fault. if this is the problem, then solution is to buy a fan with an advanced motor drive inside (variable speed) and hope it comes with electronic protection against inrush. or maybe you could just coil up some of the source wire to create an inductance.. what do you think?

as for common mode trips, easiest solution here is to get some high-mue torroidal cores and just run the power wires through it. the black & the white through the donought center of the core. the common mode flux will circulate in the core.

Knucklez
 

·
Banned
Joined
·
5,990 Posts
where dose 1.43 come from?
UL 943, which probably used Dalziel's research. This formula is also in the online manual for Ideal's house wiring tester/analyzer, 61-165 or 65-165. It plots out as a straight line on log-log paper and I guess they used a best fit routine to match the data. The curve is a maximum.

Like a circuit breaker's trip curve is designed to prevent wire from reaching excessive temperatures this curve is designed to protect most people from injury or death. As I recall, Dalziel had trouble getting reliable data on kids because instead of reporting their discomfort level they would cry.

I guess a variable speed motor may help, if the speed control doesn't cause additional transients.
 
21 - 36 of 36 Posts
Top