GFCI Protection For VFD

[tortuga]

While the product standard does allow that long time to trip for lower level ground faults, in reality they trip pretty much instantaneously. If they actually tripped at the maximum permitted trip time you would have to hold the test button in for over 5 seconds.

Well thats completely non-standard and can't be relied upon. It definitely cant be expected as definition of Class A. If they want that they should call it something else.
Besides I am not sure the test button on GFCI's uses a 18k ohm resistor, I have tested some GFCI's with a known 6.6ma load along with an Ideal Sure Test and they always take 7 seconds or more (the sure test times out at 6999 ms) . Often I have to adjust the varistor load up to get them to trip.
I just wonder if GFCI's are indeed built over sensitive and out of spec, then they trip too often and manufacturers are starting to build them to spec (hi frequency or appliance GFCI's) what is the point of having the dead band 6-29 ma range in the spec?

 

[kwired]

Well thats completely non-standard and can't be relied upon. It definitely cant be expected as definition of Class A. If they want that they should call it something else.
Besides I am not sure the test button on GFCI's uses a 18k ohm resistor, I have tested some GFCI's with a known 6.6ma load along with an Ideal Sure Test and they always take 7 seconds or more (the sure test times out at 6999 ms) . Often I have to adjust the varistor load up to get them to trip.
I just wonder if GFCI's are indeed built over sensitive and out of spec, then they trip too often and manufacturers are starting to build them to spec (hi frequency or appliance GFCI's) what is the point of having the dead band 6-29 ma range in the spec?

Isn't it kind of a trip curve? The seven seconds might be maximum it can hold before tripping when current is at 6 mA, the higher the current the faster it will trip, up to some level like maybe your mentioned 29mA?

 

[don_resqcapt19]

Well thats completely non-standard and can't be relied upon. It definitely cant be expected as definition of Class A. If they want that they should call it something else.
Besides I am not sure the test button on GFCI's uses a 18k ohm resistor, I have tested some GFCI's with a known 6.6ma load along with an Ideal Sure Test and they always take 7 seconds or more (the sure test times out at 6999 ms) . Often I have to adjust the varistor load up to get them to trip.
I just wonder if GFCI's are indeed built over sensitive and out of spec, then they trip too often and manufacturers are starting to build them to spec (hi frequency or appliance GFCI's) what is the point of having the dead band 6-29 ma range in the spec?

The test resistor must be able to trip the device at 85% of rated voltage and not draw more than 9 mA at rated voltage.

What dead band? The time to trip in seconds is always the quantity (20/current in mA), raised to the 1.43 power.

 

[tortuga]

What dead band? The time to trip in seconds is always the quantity (20/current in mA), raised to the 1.43 power.

When a GFCI takes longer than 400ms to trip its not really protecting anyone from a touch voltage that could cause significant harm and this is the 'dead band'. All the major manufacturers know this.
Most ground faults dont stay at 5-29ma they ramp up to at least 45ma often 10A or more on a 15A branch circuit with an equipment ground.
You can see in this chart of the UL formula that a dead band starts around 35ma for a class A device, while a RCD will protect to 30ma.

ma​	GFCI trip time ms​
45​	313.6​
40​	371.1​
35​	449.2​
30​	560​
25​	726.8​
20​	1000​
15​	1508.9​
10​	2694.5​
9​	3132.6​
8​	3707.3​
7​	4487.3​
6​	5593.9​
5​	7260.2​
4​	9989.1​