Ungrounded Delta

[dm9289]

I have occasion to run across 480v and 240v ungrounded systems in manufacturing facilities that I work at.

240v ungrounded delta if I take a meter phase to phase I get 240v.
These systems do have an earth connection (not to any phase) and if I measure phase to egc I do get voltages that are unequal. I’m told this has to do with capacitance that’s created.

Question if I connected a wire from a breaker to that egc will that cause overcurrent to trip the breaker???? In my mind with no actual connection it should not.

Question if I hook for example a light bulb from the circuit breaker to egc will it light????

I guess the fact that I get a reading to ground it confuses me since it’s an ungrounded system.

 

[electrofelon]

There will always be "leakage" and capacitive coupling to some degree. The amount will depend on the size of the system effectiveness of the insulations.

1. No the breaker would not trip, all you did is make it a grounded system.

2. How much you can "power" off the leakage and capacitive coupling will depend (see first sentence). A DMM of course works because the impedance is so high. Ive never tried a solenoid tester phase to ground on an ungrounded system.

 

[winnie]

Remember that to measure a voltage or run a load you need a complete circuit, something going from the transformer, through some wires, through your load, through some wires again, and back to the transformer.

With a normal grounded system, during a fault, the circuit is transformer -> wires -> fault -> egc system -> neutral bond -> transformer.

But what is the path in an ungrounded system? As @electrofelon says, you always have things like leakage (tiny amounts of current flowing _through_ your insulation, or flowing in dirt on the outside of your wires, or even through the air) and 'capacitive charging' (the currents flowing because the insulated wires and surrounding conduits are themselves actually capacitors.) These 'parasitic current paths are very high impedance.

Your fault circuit is thus transformer -> wires -> fault -> parasitic ground connection -> wires -> transformer.

To understand your measurements, you have to look at the series circuit, your meter (or other load) in series with the impedance of the parasitic connection. Like any other series circuit, the load with the higher impedance will 'see' the larger proportion of the voltage, and the total current flowing through the series circuit depends on the total impedance of the two (load and parasitic impedance combined).

Big ungrounded systems can have relatively low parasitic impedance and can have many amps of 'charging current'. You could probably light a lamp on one of these systems. They can certainly give you a severe shock.

Small ungrounded systems are used in operating rooms. The system can fault to ground through the patient without causing injury.

Jon

 

[augie47]

There were a number of ungrounded systems in this area at one time. The advantage being if one phase has a fault to ground it does not open a OCP device thus not impeding production, however, if you don't have some type ground fault monitoring there is little advantage as eventually you may well develop a second phase to ground which is a short circuit. In addition, many electronic drives will not function on ungrounded systems.
(this has prompted many of our facilities to change to grounded systems)
Under today's Code, ground fault monitoring is required.

 

[dm9289]

There will always be "leakage" and capacitive coupling to some degree. The amount will depend on the size of the system effectiveness of the insulations.

1. No the breaker would not trip, all you did is make it a grounded system.

2. How much you can "power" off the leakage and capacitive coupling will depend (see first sentence). A DMM of course works because the impedance is so high. Ive never tried a solenoid tester phase to ground on an ungrounded system.

Thank you your info is helpful. I will try to find an old school solonoid type tester.

 

[dm9289]

Remember that to measure a voltage or run a load you need a complete circuit, something going from the transformer, through some wires, through your load, through some wires again, and back to the transformer.

With a normal grounded system, during a fault, the circuit is transformer -> wires -> fault -> egc system -> neutral bond -> transformer.

But what is the path in an ungrounded system? As @electrofelon says, you always have things like leakage (tiny amounts of current flowing _through_ your insulation, or flowing in dirt on the outside of your wires, or even through the air) and 'capacitive charging' (the currents flowing because the insulated wires and surrounding conduits are themselves actually capacitors.) These 'parasitic current paths are very high impedance.

Your fault circuit is thus transformer -> wires -> fault -> parasitic ground connection -> wires -> transformer.

To understand your measurements, you have to look at the series circuit, your meter (or other load) in series with the impedance of the parasitic connection. Like any other series circuit, the load with the higher impedance will 'see' the larger proportion of the voltage, and the total current flowing through the series circuit depends on the total impedance of the two (load and parasitic impedance combined).

Big ungrounded systems can have relatively low parasitic impedance and can have many amps of 'charging current'. You could probably light a lamp on one of these systems. They can certainly give you a severe shock.

Small ungrounded systems are used in operating rooms. The system can fault to ground through the patient without causing injury.

Jon

Thank you I was not seeing it as a series voltage divider. The systems I speak of, 1 is a 3000a 480v UG and the other is an emergency service 240v 200amp overhead.

 

[zbang]

I will try to find an old school solonoid type tester.

That's easy, your local distributor should have something or order from amazon/etc. Many of use use the Fluke T+Pro, Knopp also makes a good one, and Klien/Ideal also sell them. Every working electrician should have something like that.

 

[dm9289]

There were a number of ungrounded systems in this area at one time. The advantage being if one phase has a fault to ground it does not open a OCP device thus not impeding production, however, if you don't have some type ground fault monitoring there is little advantage as eventually you may well develop a second phase to ground which is a short circuit. In addition, many electronic drives will not function on ungrounded systems.
(this has prompted many of our facilities to change to grounded systems)
Under today's Code, ground fault monitoring is required.

Thank you and yes this stuff is from 1950. I have experienced some set up issues with drives as you mentioned.

 

[dm9289]

That's easy, your local distributor should have something or order from amazon/etc. Many of use use the Fluke T+Pro, Knopp also makes a good one, and Klien/Ideal also sell them. Every working electrician should have something like that.

Thank you I have a fluke that folks refer to as a wiggy but I think it’s more a modern dmm with an amp clamp.

So I guess what you are saying is with the T plus pro phase to ground I would not get a reading.

 

[ptonsparky]

Thank you I have a fluke that folks refer to as a wiggy but I think it’s more a modern dmm with an amp clamp.

So I guess what you are saying is with the T plus pro phase to ground I would not get a reading.

The T+Pro is not low impedance, unless they've changed them lately. It is lower than a Fluke 87 but not what you want. Your local Menards, Home Depot or Lowes should have a solinoid (sp) tester.

 

[zbang]

If you do, it should be smaller, but that depends on the actual leakage. In general, phase-ground readings on ungrounded delta systems are interesting but often not useful, that is, if they're all similar then the leakage is roughly the same over all the phases. OTOH if one is much lower that could indicate a problem somewhere, but maybe not.

Which fluke do you have?

The T+Pro is not low impedance

Claims to be, but now we're asking what does "low" mean? 100 ohm, 10k, or >1meg ?

 

[ptonsparky]

Measuring with two 87s and two T+ to an isolated metallic valve in a PEX water line.
87 - 124.3
87 - 124.2
Both 87s voltage was 123,8 & 124.0

T+ 81
T+ 81
Both T+ voltage was 46.9 & 47.1

A T+ & an 87, voltage was 81 & 80.9 respectively.

Claims to be, but now we're asking what does "low" mean? 100 ohm, 10k, or >1meg ?

Yes, who defines low impedance?

 

[winnie]

Yes, who defines low impedance?

That is always context dependent. What is the impedance of the system you are measuring?

IMHO if the impedance of the meter is > 10x the impedance of the system, it is a high impedance meter, and < 0.1x the impedance of the system, then it is a low impedance meter. In between the measurements will be confusing.

For some instrumentation applications, common practice is to use 'matched' equipment, with an impedance equal to the source being measured.

-Jon

 

[GoldDigger]

In addition, many electronic drives will not function on ungrounded systems.
(this has prompted many of our facilities to change to grounded systems)
....

Two of the mechanisms for problems are:

1. Protective surge suppression devices or transient filtering that are connected from each line to ground/EGC. These will not work properly on an ungrounded system.
2. Internal components which are designed to withstand only a limited voltage to ground and are subject to damage when exposed to line-line voltage. In addition, a re-striking arc fault to ground on an ungrounded system can pump the line-ground voltage up to several times the line-line voltage.

Of course if the equipment has a Neutral terminal, not just an EGC terminal, there is probably no way to use an ungrounded system without rewiring the device.

 

[dm9289]

If you do, it should be smaller, but that depends on the actual leakage. In general, phase-ground readings on ungrounded delta systems are interesting but often not useful, that is, if they're all similar then the leakage is roughly the same over all the phases. OTOH if one is much lower that could indicate a problem somewhere, but maybe not.

Which fluke do you have?


Claims to be, but now we're asking what does "low" mean? 100 ohm, 10k, or >1meg ?

Fluke 87

 

[zbang]

The 87 is definitely a high impedance meter.
From the manual-

Each of the five ac/dc voltage ranges presents an input impedance of approximately 10 MΩ in parallel with less than 100 pF. Ac voltage is ac-coupled to the 10 MΩ input.