[zhawkeye]

I disassembled a GFCI outlet some time ago to see how it worked...figured it out and also found the test button applied a shunt resistor to ground to trip the circuit.
Lately I've been blasted for stating the test button won't work without a ground...ok...I can see where the current can be shunted back to neutral prior to the pickup coil and trip the circuit without a ground.
This brings up a couple of questions...

One...I don't remember any statements on the GFCI packaging stating whether a ground needs to be present or not for proper test button operation. If they are manufactured in both styles, have they started placing this information on the packaging?

Two...it seems to me that the manufacturer's may have liability issues with the style that doesn't require a ground for proper test button operation...after all, it's a "ground fault" interrupter...without a ground connected to the outlet, it's just a "circuit fault interrupter".
And the fact that it "looks" like a grounded outlet that "tests" good, but ground is not present could be misleading for anyone using the outlet.

Anyway...any other information or comments are very welcome...

 

[bobkenney79]

The 'ground Fault' is when you become part of the circuit. The GFCI senses the change in current (hot to neutral) and interrupts the circuit, thus preventing your electrocution (burn). All of this takes place without the need for an actual safety ground which is not required for proper operation. Installations without a ground wire present are required to have a label affixed stating that there is no circuit ground (not always done - or the label comes off). Manufacturer is not liable for faulty installation (no label).

 

[surferdude2]

The ground fault protected against need not be to the equipment grounding conductor but to any ground, including (most importantly) the one through your body to some available ground you may be in contact with.

I suspect the earlier GFCI's had the test referenced to the EGC but that has since changed... didn't affect the protection, just the test method.

When using the GFCI on a non-grounded circuit, you are supposed to place a warning sticker that states "This receptacle is not grounded."

SD2

 

[bobkenney79]

The GFCI without a ground would show 'no ground' with the cheapest three light tester. With a voltmeter a test between 'hot' and ground would show no voltage which identifies the outlet as having no circuit ground (ground and neutral are bonded (connected) at the main panel).

 

[Oso954]

You are confusing grounding (bonding) the gfci with an EGC vs current escaping from the circuit/not returning (a ground fault).

The EGC does not need to be present for the test button (or the gfci) to work. That why we use them to protect the 2 wire (no ground) circuits that were installed prior to the NEC requiring the EGC.

The thing that doesn’t work is the 3 light plug in receptacle/GFCI tester.
Unlike the test button on the GFCI, the plug in tester only shunts the current to the EGC. Without the presence of the EGC (or grounded conduit system) the tester cannot trip the GFCI..

 

[bobkenney79]

The thing that doesn’t work is the 3 light plug in receptacle/GFCI tester.
Unlike the test button on the GFCI, the plug in tester only shunts the current to the EGC. Without the presence of the EGC (or grounded conduit system) the tester cannot trip the GFCI..

The GFCI manufacturers all state that proper testing can only be done with the built in test button. The three light tester button is convenient, but as already stated will not trip a GFCI without an equipment ground connected to the outlet.

 

[zhawkeye]

Thanks for all the input...I do understand the operation of GFCI circuits as I'm in the midst of GFCI testing with equipment we manufacture (it doesn't play nice if connected to a GFCI breaker).
I understand now from surferdude that they're all manufactured to work without a ground these days...thanks.
I guess my concern now is just with semantics...if there's no ground to the outlet, to me that's a fault as you WILL be the ground if there's an issue...and the test button should show that...but again, that's just me...

 

[CodeMatters]

GFCI isn't what I would I name it, but (A) I wasn't asked, and (B), it's water
under the bridge. Not sure my preferred name "unknown/unintended current
path interuptor" would help the average user either...........

I'm just guessing here, but, GFCI is perhaps the name given by NEMA or IEEE
or some other body. If so, the manufacturers are safest to simply follow the
standard naming convention.

Here's an internal dwg for anyone interested:

 

[Oso954]

GFCI isn't what I would I name it,

Do you like residual current device ?

How about Fehlerstromschutzschalter ?

 

[zhawkeye]

I'd like "Keepsufromshoken"...

 

[curiousB]

GFCIs don’t measure current. The current level doesn’t matter. GFCIs measure a difference in current between hot and neutral lines. If the difference exceeds 5mA then the GFCI trips.

The button on the GFCI routes a 5mA current through the torrid core with one line and bypasses the core with the other. This sets up an intentional
imbalance used to test the differential current trip circuitry. It does not need a ground.

External GFCI testers have no means to pass one wire through the torrid core while the other bypasses. As a result they leak a current to ground so that the current in the hot is Iload +5mA and the neutral is Iload + 0mA. The GFCI sees the imbalance and trips. External GFCI testers do require a ground.








Sent from my iPhone using Tapatalk

 

[dmxtothemax]

QUOTE - .I don't remember any statements on the GFCI packaging stating whether a ground needs to be present or not for proper test button operation.



in most electrical instalations neutral is effectively ground.

 

[dmxtothemax]

I think a more accurate name for them is "core balance relay"


:glasses:

 

[zhawkeye]

It's not that I don't know how they work...I do...on the one I disassembled, there was a pickup coil on each leg...these feed into both the inverting and non-inverting inputs of an op amp. The op amp amplifies the difference between the two inputs, so as long as they're the same, the output of the op amp remains inactive (whatever level that may be). Whenever the inputs vary enough, the output of the op amp goes active and trips the circuit.
And yes, the test button shunts current to unbalance the circuit so that it trips...

I can throw together a block diagram of the control electronics if interested...

 

[curiousB]

Incorrect.

There is one torrid core both hot and neutral wires pass through. This is the primary differential transformer. If the currents are equal then they exactly cancel out and there is no flux in the torrid. Once there is a ground leak the currents don’t cancel and a differential signal is produced on the secondary windings. This goes to the electronics where it is filtered and compared to a trip threshold level. If a trip is detected an SCR is fired to trip the internal contactor.

The other torrid is for the neutral/ground short detection feature. This lesser known feature injects a 120hz signal
into both lines as a scheme to detect if they are shorted.



www.onsemi.com/pub/Collateral/RV4141A-D.pdf







Sent from my iPhone using Tapatalk

 

[zhawkeye]

Obviously things have changed since I first dissected a GFCI outlet...that one used an op amp which can be used for a multitude of applications...the onsemi part above is a dedicated IC for GFCI applications.
Note that the diagram omits the reset function...and latching contacts (as shown) generally require additional circuitry to reset the contacts...just cycling power doesn't reset latching contacts...

 

[curiousB]

GFCIs have used chips like the above for 20 years or so. The chip designs are well travelled now.


Figure 3 of the attached shows the trip solenoid. The contacts are mechanically latched to closed and the firing solenoid trips the spring loaded mechanism to open the contacts.


Sent from my iPhone using Tapatalk

 

[jbfan]

GFCI isn't what I would I name it, but (A) I wasn't asked, and (B), it's water
under the bridge. Not sure my preferred name "unknown/unintended current
path interuptor" would help the average user either...........

I'm just guessing here, but, GFCI is perhaps the name given by NEMA or IEEE
or some other body. If so, the manufacturers are safest to simply follow the
standard naming convention.

Here's an internal dwg for anyone interested:

Way back when what ever you want to call them were first introduced, they weren't cheap, and you only needed one to feed all the required locations.
I called them money makers, because people didn't know why their bathrooms weren't working when the garage receptacle tripped.:biggrin2:

 

[curiousB]

Way back when what ever you want to call them were first introduced, they weren't cheap, and you only needed one to feed all the required locations.
I called them money makers, because people didn't know why their bathrooms weren't working when the garage receptacle tripped.:biggrin2:

I totally agree. Actually they are so inexpensive now that NEC should put some rules that they only serve devices in the room or area where they are installed. Saving $10 on a GFCI so a homeowner can wonder why the GFCI hidden behind the wood pile in the garage is tripped causing a bathroom to be out of power.... false economics in today's world.