ProSoundWeb Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: [1] 2  All   Go Down

Author Topic: GFCI Failure Study  (Read 3113 times)

Mike Sokol

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 3361
  • Lead instructor for the No~Shock~Zone
    • No~Shock~Zone Electrical Safety
GFCI Failure Study
« on: February 25, 2015, 06:53:31 am »

I found an interesting study on the percentages of failed GFCI's in homes. While this survey was done back in 2001, it does include a lot of forensic analysis of exactly what components failed under various moisture and lightning conditions. Yes, it's 112 pages of geek-speak, but there's numerous charts showing between 10% and 20% dead GFCI's in the field. I'll read the entire report in the next few days and snag a few screen shots that might be useful to us. See http://www.mikeholt.com/files/PDF/NEMA_GFCI_Study.pdf
« Last Edit: February 25, 2015, 07:48:10 am by Mike Sokol »
Logged
Mike Sokol
mike@noshockzone.org
www.NoShockZone.org

Stephen Swaffer

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2361
Re: GFCI Failure Study
« Reply #1 on: February 25, 2015, 08:58:58 am »

That roughly corresponds to my experience when I was doing an annual plant wide GFCI test as preventative maintenance-but I experienced that failure rate every year in spite of bad GFCIs being replaced (I did the replacement so I know the repairs happened).

The most disconcerting is control 0312-4-tested OK but did not trip with ground fault when under load due to mis assembly-I wonder how many more like that made it to the field?

Take away for readers doing audio relying on GFCIs for safety-it takes what? 5 sec? to hit test/reset when you set up.  Not a bad habit to get into. 
Logged
Steve Swaffer

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16608
  • Hickory, Mississippi, USA
    • Resotune
Re: GFCI Failure Study
« Reply #2 on: February 25, 2015, 09:44:51 am »

That is not good...

Stephen were any of the failures from GFCIs that were already replaced once? Are the replacements from the same batch as the original install parts?

That failure rate sounds uncomfortably high.

JR 

Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

Mike Sokol

  • Moderator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 3361
  • Lead instructor for the No~Shock~Zone
    • No~Shock~Zone Electrical Safety
Re: GFCI Failure Study
« Reply #3 on: February 25, 2015, 10:34:15 am »

That is not good...

Stephen were any of the failures from GFCIs that were already replaced once? Are the replacements from the same batch as the original install parts?

That failure rate sounds uncomfortably high.

JR

Yeah, this is a really interesting report that Stephen could add to.
Logged
Mike Sokol
mike@noshockzone.org
www.NoShockZone.org

Stephen Swaffer

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 2361
Re: GFCI Failure Study
« Reply #4 on: February 25, 2015, 12:15:03 pm »

We had an electrical supplier in weekly and didn't pursue quantity discounts for maintenance so these would be different batches bought a year apart.  This was a harsh environment-both ambient and electrically-lots of welders/drives, located close to POCOs feeders to town, so I am sure we were harder on them than a typical environment. 

The newer ones do a bettr job of self diagnosise on reset-and that failure rate does include "will not reset"-which is not an unsafe condition.

I did have an unusual experience with a customer-I installed a GFCI for a camper by a pond of his.  Name brand weather resistant and it failed twice in side of 18 months (both times will not reset).  Sent GFCIs to manufacturer but never received any feedback.

These experiences have led me to view GFCIs as backup protection only.  Everything needs to be safe even withut a GFCI, then if something unexpected/uncontrollable happens.  So for instance with your safety power strip, my feeling is receptacles still should be tested. 
Logged
Steve Swaffer

frank kayser

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1528
  • Maryland suburbs of Washington DC
Re: GFCI Failure Study
« Reply #5 on: February 25, 2015, 12:32:33 pm »

Stupid question time:
Is there a problem plugging a "portable" GFCI (like the one JR is modifying) into a (for lack of better description) installed GFCI? 


Last year, nosing around, I also saw some long-term test that was also quoting that 20% number... I was appalled. Figuring my small town center, I know of at least five GFCI devices - statistically one of those would be bad in some shape or form...


The idea of serial connection of GFCI is that I can carry one I have of known operation, maintenance, and quality, and plug it into just about anything, more or less bypassing any problematic/failed unit (and with the addition of a stinger cap, solving another possible death vector)


Comments?
frank
Logged

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16608
  • Hickory, Mississippi, USA
    • Resotune
Re: GFCI Failure Study
« Reply #6 on: February 25, 2015, 01:14:15 pm »

plugging multiple GFCI in series will provide higher probability one will trip from a fault, but makes "will not reset" loss of power more likely too.

 JR
Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

frank kayser

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1528
  • Maryland suburbs of Washington DC
Re: GFCI Failure Study
« Reply #7 on: February 25, 2015, 01:45:17 pm »

plugging multiple GFCI in series will provide higher probability one will trip from a fault, but makes "will not reset" loss of power more likely too.

 JR
To clarify, JR, the higher probability will be that of tripping due to a REAL fault - if all is working properly, a false trip is no more likely than a single GFCI, correct? - putting it differently one does not directly affect the other?


frank





Logged

John Roberts {JR}

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 16608
  • Hickory, Mississippi, USA
    • Resotune
Re: GFCI Failure Study
« Reply #8 on: February 25, 2015, 02:01:30 pm »

To clarify, JR, the higher probability will be that of tripping due to a REAL fault - if all is working properly, a false trip is no more likely than a single GFCI, correct? - putting it differently one does not directly affect the other?


frank

Correct.. The GFCI I took apart does not even connect to ground it just passes through.. They just null the line vs. neutral current feeding that GFCI device...

As has been discussed plugging multiple separate GFCI into a common GFCI (in parallel) could exceed the common GFCI threshold current even if none of the individual GFCIs trip. The GFCI device I used is 5 mA +/-1 mA so four of these with only 2 mA of leakage each while not tripping themselves, could exceed the common GFCI's  5 mA threshold.

In series each GFCI would respond only to it's local leakage.

JR   
Logged
Don't tune your drums half-ass. Listen to what a properly "cleared" drum sounds like.   http://circularscience.com/

frank kayser

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1528
  • Maryland suburbs of Washington DC
Re: GFCI Failure Study
« Reply #9 on: February 25, 2015, 03:05:51 pm »

Correct.. The GFCI I took apart does not even connect to ground it just passes through.. They just null the line vs. neutral current feeding that GFCI device...

As has been discussed plugging multiple separate GFCI into a common GFCI (in parallel) could exceed the common GFCI threshold current even if none of the individual GFCIs trip. The GFCI device I used is 5 mA +/-1 mA so four of these with only 2 mA of leakage each while not tripping themselves, could exceed the common GFCI's  5 mA threshold.

In series each GFCI would respond only to it's local leakage.

JR   
Hi again, JR,


I do remember reading (and quoting) plugging one surge protector (MOV) into another (into another) serially could present a case where the additive leakage of all the surge protectors could exceed the trip current on a GFCI -
I think I must have missed the multiple GFCI thing, and substituted in my mind "GFCI" for "Surge Protector", and "serial" for "parallel". 
Lets just say I was unaware of a leakage "problem" in GFCI devices.


I do appreciate your patience in this matter.  I really do want to get my mind right.


Color me embarrassed,
frank



Logged

ProSoundWeb Community

Re: GFCI Failure Study
« Reply #9 on: February 25, 2015, 03:05:51 pm »


Pages: [1] 2  All   Go Up
 



Page created in 0.056 seconds with 22 queries.