ProSoundWeb Community

Sound Reinforcement - Forums for Live Sound Professionals - Your Displayed Name Must Be Your Real Full Name To Post In The Live Sound Forums => AC Power and Grounding => Topic started by: Mike Sokol on November 08, 2017, 07:38:13 AM

Title: GFCI Theory
Post by: Mike Sokol on November 08, 2017, 07:38:13 AM
PSW just republished my article on GFCI theory. If you're not totally up to speed on how they operate, this is a good primer.

https://tinyurl.com/yb74lh92
Title: Re: GFCI Theory
Post by: Tim McCulloch on November 08, 2017, 01:19:11 PM
PSW just republished my article on GFCI theory. If you're not totally up to speed on how they operate, this is a good primer.

https://tinyurl.com/yb74lh92

Good description of how it works and the water pump analogy/illustrations are perfect.

Now it's all about getting people to believe you.  I routinely encounter folks that blame GFCI (or RCD - residual current devise - to our UK friends) tripping on all manner of things: higher/lower line voltage, "excessive" current draw, phase of the moon and failed Voodoo rituals...

What none of these people will blame, though, is the 40 year old guitar amp with the polarity switch and "death cap" neutral coupling, or the "repaired" extension cord or the ground/neutral reverse wiring... or the 1-strand short inside a quad box or connector or chain hoist (the latter happened to us, the GFCI kept the short from energizing the entire hydraulic stage via the hoist chain and hang point).

Yeah, sometimes a GFCI/RCD will trip for reasons we cannot immediately determine.  That doesn't mean they're defective or a nuisance.

/rant
Title: Re: GFCI Theory
Post by: Jonathan Johnson on November 08, 2017, 03:46:19 PM
Yeah, sometimes a GFCI/RCD will trip for reasons we cannot immediately determine.  That doesn't mean they're defective or a nuisance.

Happened to me. Probably due to MOV surge protectors.

Bypassed the installed GFCI and used a couple of portable GFCI to get around the problem (distributing the leakage current).
Title: Re: GFCI Theory
Post by: John Roberts {JR} on November 08, 2017, 05:40:23 PM
I have had multiple random trips with no common theme.... One time coincident with a nearby lightning strike, another time the GFCI outlet did not come up after a power interruption. The most recent time trip was during a load of laundry, but no more trips from several weeks of doing laundry....

Note: I buy the cheapest GFCI I can find... the trip thresholds seem repeatable, random trips not enough to be a problem yet... I have 3 GFCI outlets and only one trips from any single event.

FWIW I feel pretty confident it will trip during a fault if it is so skittish without a fault.  :o

JR

PS: For theory of operation, how it actually works is pretty damn elegant...  Two identical windings on a magnetic core generate equal and opposite magnetic flux as long as line and neutral currents are identical.... A high turn secondary can give a very sensitive fault output if the line and neutral flux do not completely cancel out.  8)
Title: Re: GFCI Theory
Post by: Peter Morris on November 08, 2017, 08:05:12 PM
I have had multiple random trips with no common theme.... One time coincident with a nearby lightning strike, another time the GFCI outlet did not come up after a power interruption. The most recent time trip was during a load of laundry, but no more trips from several weeks of doing laundry....

Note: I buy the cheapest GFCI I can find... the trip thresholds seem repeatable, random trips not enough to be a problem yet... I have 3 GFCI outlets and only one trips from any single event.

FWIW I feel pretty confident it will trip during a fault if it is so skittish without a fault.  :o

JR

PS: For theory of operation, how it actually works is pretty damn elegant...  Two identical windings on a magnetic core generate equal and opposite magnetic flux as long as line and neutral currents are identical.... A high turn secondary can give a very sensitive fault output if the line and neutral flux do not completely cancel out.  8)

Whats I find interesting is that most other countries set the GFCI (RCD) trip current at 30ma and the RCD is placed at the switch board so that all the wiring in the building is protected.

The point of ventricular fibrillation with an electric shock is a current time relationship. In general 30ma combined with a short trip time is effective at saving lives and in practice 30ma RCD's have minimal nuisance tripping.

Its also worth noting that the GFCI/RCD does not really limit the shock current you receive just the time you receive it. My experience has been with 240V systems, perhaps with 120V system you often find the touch current to be quite low, so to ensure you get the tripping time quick enough you need a much lower trip current (?) or perhaps your regulators have just been very conservative. (?)

For critical applications such as operating theatres the trip current is reduced to 10ma.

Here are the curves that the RCD trip currents and times are based on.   
Title: Re: GFCI Theory
Post by: Mike Sokol on November 09, 2017, 06:23:19 AM
PS: For theory of operation, how it actually works is pretty damn elegant...  Two identical windings on a magnetic core generate equal and opposite magnetic flux as long as line and neutral currents are identical.... A high turn secondary can give a very sensitive fault output if the line and neutral flux do not completely cancel out.  8)

It really works very similarly to a balanced XLR input on a mixing console. The CMRR (Common Mode Rejection Ratio) of both of these systems cancel out any common mode signals, but pass differential signals.
Title: Re: GFCI Theory
Post by: Keith Broughton on November 09, 2017, 08:17:27 AM
PSW just republished my article on GFCI theory. If you're not totally up to speed on how they operate, this is a good primer.

https://tinyurl.com/yb74lh92
Thanks Mike. Great explanation :)
Title: Re: GFCI Theory
Post by: John Roberts {JR} on November 09, 2017, 09:46:25 AM
Whats I find interesting is that most other countries set the GFCI (RCD) trip current at 30ma and the RCD is placed at the switch board so that all the wiring in the building is protected.

The point of ventricular fibrillation with an electric shock is a current time relationship. In general 30ma combined with a short trip time is effective at saving lives and in practice 30ma RCD's have minimal nuisance tripping.

Its also worth noting that the GFCI/RCD does not really limit the shock current you receive just the time you receive it. My experience has been with 240V systems, perhaps with 120V system you often find the touch current to be quite low, so to ensure you get the tripping time quick enough you need a much lower trip current (?) or perhaps your regulators have just been very conservative. (?)

For critical applications such as operating theatres the trip current is reduced to 10ma.

Here are the curves that the RCD trip currents and times are based on.
Good luck relitigating GFCI design thresholds...

The 5mA+/- threshold is safely below the muscular contraction range.

It seems practical to have both higher threshold GFCI breakers at the panel and normal sensitivity GFCI outlets near water. I suspect future power panels to include all kinds of instrumentation, some we probably haven't thought of yet.

JR
Title: Re: GFCI Theory
Post by: Riley Casey on November 09, 2017, 11:38:37 AM
What I have seen is that GFCIs in portable use are much more likely to go in for nuisance tripping.  I've dismantled the versions with the breaker built into the outlet and the parts are tiny as befits something that is able to move physically in response to milliamps of current.  Once again we are being bitten by products made for a different market than our applications.  Dedicated GFCI circuit breakers appear to be somewhat less likely to nuisance trip when used in portable service perhaps because PD racks don't get slammed around quite as much as quad boxes.
Title: Re: GFCI Theory
Post by: Stephen Swaffer on November 09, 2017, 12:18:28 PM
Whats I find interesting is that most other countries set the GFCI (RCD) trip current at 30ma and the RCD is placed at the switch board so that all the wiring in the building is protected.


Probably the difference between a standard being set by lawyers and insurance companies rather than an engineer?

Dedicated GFCI breakers cost roughly $30-$40 more per device than GFCI receptacles.  Unfortunately these leads to receptacles being the more likely choice.  My guess is that it is easier to build a robust device with tighter threshholds that are still safe when you have more $$ to work with?  Recpetacles do offer convenience for resetting that may or may not be there with breakers.