GFCI vs RCD

[wptski]

I'm in the US where we use GFCI which from all I've read and diagrams are for the most part except tripping current, the same as a UK RCD.

I have use of a Fluke 1653 which the manual states will trip a RCD when L-PE loop test is performed. I can't verify that because we don't use them in the US.

The Fluke 1653 "doesn't" trip a US type GFCI! I wonder if anybody knows why? The RCD tests do!

 

[Adam_151]

Yes, as far as I'm aware an amercian GFCI is similar to an RCD except:

6ma devices are the most common
There is no distinction between an RCD and an RCBO as it were, GFCIs that fit in the panel (DB) are RCBOs, while those in outlets are just RCDs (which incidentally have a set of terminals to protect anything downstream on the same radial circuit)

Not sure a fluke 1653 is suitable for testing in the US really, don't they make test kit for the us market?

And as for the origional question, the fluke 1653 will try not to trip an RCD (can't always manage it though), it works on the DC biasing principle (If you want to see the RCD to trip, put the neutral probe on the earth)

 

[wptski]

Yes, as far as I'm aware an amercian GFCI is similar to an RCD except:

6ma devices are the most common
There is no distinction between an RCD and an RCBO as it were, GFCIs that fit in the panel (DB) are RCBOs, while those in outlets are just RCDs (which incidentally have a set of terminals to protect anything downstream on the same radial circuit)

Not sure a fluke 1653 is suitable for testing in the US really, don't they make test kit for the us market?

And as for the origional question, the fluke 1653 will try not to trip an RCD (can't always manage it though), it works on the DC biasing principle (If you want to see the RCD to trip, put the neutral probe on the earth)

We have CB's with GFCI's built in too. GFCI minimum trip current is 4-6ms and maximum trip time is 25ms.

The Fluke 1653 manual states that it will trip a RCD during the L-PE loop test and I've seen posts in other forums confirming this.

I have a similiar tester that reads our L-N-G impedance here in the US. It trips a GFCI during the G impedance test and another test similiar to a PSC test. Its manual states that this is normal too. It has a seperate GFCI trip time test also.

Correct, the Fluke 1653 is not sold in the US but I was offered use of one just to tinker with.

One difference between a RCD and GFCI is that at least in the receptacle type there is also a N-G monitor also. I think this was added to the original GFCI designs.

The question remains. What different in the operation/design of a US GFCI and UK RCD that would cause the Fluke 1653 to trip one but not the other during a L-PE loop test? It could be a question of why does the DC biasing of the Fluke 1653 not trip a US GFCI.

 

[ColJack]

isn't the voltage in the US 120v?

for this reason the leakage through the meter when doing earth loop tests would be less...

230v = 23mA x 10Kohms

120v= 12mA x 10Kohms

resistance of the meter will be constant..

 

[Spark123]

Not sure on that particular tester, some of the testers as Adam says have DLok which locks the RCD with DC to stop it tripping when the L-E Loop test is carried out.
The most basic RCD we have works on ballancing, the phase and neutral are connected as current transformers on a torroidal transformer. If the phase and neutral are ballanced then they cancel each other out. If there is an imballance, the resulting magnetic field is picked up by the sense coil and causes the RCD trip cutting off the power.
The most common size used here for additional personal protection is a 30mA device. It must trip within 200/300ms (depends on BS) at the rated trip current and within 40ms at 5x the trip current.

 

[wptski]

isn't the voltage in the US 120v?

for this reason the leakage through the meter when doing earth loop tests would be less...

230v = 23mA x 10Kohms

120v= 12mA x 10Kohms

resistance of the meter will be constant..

Right, we are 120V. Is 10Kohms just a value that you picked? Ours trip at 4-6ma.

Using the lowest trip current of 10ma on the 1653, a GFCI with trip in 3-4ms but at 30ms it's 0ms.

 

[wptski]

Not sure on that particular tester, some of the testers as Adam says have DLok which locks the RCD with DC to stop it tripping when the L-E Loop test is carried out.
The most basic RCD we have works on ballancing, the phase and neutral are connected as current transformers on a torroidal transformer. If the phase and neutral are ballanced then they cancel each other out. If there is an imballance, the resulting magnetic field is picked up by the sense coil and causes the RCD trip cutting off the power.
The most common size used here for additional personal protection is a 30mA device. It must trip within 200/300ms (depends on BS) at the rated trip current and within 40ms at 5x the trip current.

That's the point. Our GFCI's work the same way using balanced current but with a lower limit than a RCD.

As I mentioned, 4-6ma is the lower limit but to be within US UL standards, it can have a trip time up to 5600ms. The maximum trip time of 25ms might be at >200ma as it's all on a graph.

 

[securespark]

Do the GFCI's have a connection to ground?

 

[wptski]

Do the GFCI's have a connection to ground?

Yes.

EDIT: The manual states that in impedance loop test at 400V=20A and at 230V=12A. Using that same ratio, it appears at 120V, it would be about 6A.