Detecting a ground fault on the secondary of a transformer

[Scott Pratz]

I work on a ship, and we have a 440V ungrounded distribution system. We have load centers that have transformers bumping that 440V down to 120V for recepticles and other services. This standard ship setup always has ground fault detection on the 440V side (at the switchboards) and at each individual load center. I have two major questions about this:

1. I understand that you "cannot detect a ground over a load center". Can someone explain why this cannot be done? Is it because you can only see a "load increase" from the primary side? if everything shares a ground, why can't I detect a ground from either side of the transformer? Please provide visual aids if you are able, I think seeing the path would help, I am having trouble visualing this process.

2. How does this ground fault detection system operate? I know I haven't given enough information to fully answer the question, however this is an electrical ground fault detection system build for delta configuration in an ungrounded system. The system is a passive detection system. I understand how measuring potential difference between the hull and system makes sense, but there are 5 load centers- and they each measure individual grounds in their system. If two load centers have the EXACT same phase on the 120 side, why couldn't a ground in LC1 be detected on another load center?

Thanks!

 

[tech99]

I work on a ship, and we have a 440V ungrounded distribution system. We have load centers that have transformers bumping that 440V down to 120V for recepticles and other services. This standard ship setup always has ground fault detection on the 440V side (at the switchboards) and at each individual load center. I have two major questions about this:

1. I understand that you "cannot detect a ground over a load center". Can someone explain why this cannot be done? Is it because you can only see a "load increase" from the primary side? if everything shares a ground, why can't I detect a ground from either side of the transformer? Please provide visual aids if you are able, I think seeing the path would help, I am having trouble visualing this process.

2. How does this ground fault detection system operate? I know I haven't given enough information to fully answer the question, however this is an electrical ground fault detection system build for delta configuration in an ungrounded system. The system is a passive detection system. I understand how measuring potential difference between the hull and system makes sense, but there are 5 load centers- and they each measure individual grounds in their system. If two load centers have the EXACT same phase on the 120 side, why couldn't a ground in LC1 be detected on another load center?

Thanks!

The transformers act as isolating devices, so the primary and secondary circuits are separate with no metallic connection (they are linked by magnetism alone). If you look at the isolated circuit formed by a secondary and its load, you could connect a ground to it at any point and everything would carry on working. The primary side does not know you have done this. But we do not want a ground on the 120V circuit because it could allow an accidental path via someone's body. In order to detect an unwanted ground path, one method is to automatically compare the voltages to ground of the two wires of the circuit, which should be equal.

 

[Scott Pratz]

The transformers act as isolating devices, so the primary and secondary circuits are separate with no metallic connection (they are linked by magnetism alone). If you look at the isolated circuit formed by a secondary and its load, you could connect a ground to it at any point and everything would carry on working. The primary side does not know you have done this. But we do not want a ground on the 120V circuit because it could allow an accidental path via someone's body. In order to detect an unwanted ground path, one method is to automatically compare the voltages to ground of the two wires of the circuit, which should be equal.

You didn't completely move to answer the second question, if you were intending to. I understand the potential measuring abilities, but if I have 2 LC's in phase with one another, let's say one hot is grounded instantaneously with a -60V. Now let's say that another LC's ground fault detection (connected between hull and it's "instanteously also -60V) is sitting near the orig LC's ground fault. Wouldn't the hull be charged -60V and the potential difference between the leg register as a ground?

 

[jim hardy]

1. I understand that you "cannot detect a ground over a load center". Can someone explain why this cannot be done?

Probably we could if we had a little more detail.
What if any connections to hull exist at each 120 volt transformer ?
You mention 60 volts to hull. That suggests there's some sort of ground indication at least.
If so, the answer to this question

If two load centers have the EXACT same phase on the 120 side, why couldn't a ground in LC1 be detected on another load center?

is my
" Poor Man's Version of Kirchoff's Current Law : Current wants only to get back to where it came from. "

Current that leaves one end of a transformer winding must get back to the other end of the same transformer winding,
a separate but identical winding will not do.

That's why the ground indicators on different transformers won't interact unless somebody has fouled up the field wiring, criss-crossing feeds from different LC'c.

Are you good with paint?
If so, draw two flashlights and observe there is no single connection that'll make them share current.

 

[Averagesupernova]

Schematic would be helpful.

 

[nsaspook]

The ground fault system on the ships (commissioned in 1962) floating utility power was just a set of lamps on each leg to the ships hull ground. If one lamp went dark that indicated a ground fault. The location was found by switching off main distribution circuits until the lamp came back on. We then just followed the branch switches until it was isolated.

There are much better options depending on what system is installed.
Line To Ground Voltage Monitoring