Safety with Ungrounded Secondaries

[Leor]

I recently saw a 3 wire transformer connected to a 120/208 3 phase supply. There is about 488 volts between phases on the secondaries and zero volts to ground. The equipment the transformer supplies does not use a neutral conductor. Even though there is no voltage to ground, my research suggests that there are still safety concerns with the secondary conductors, because "the system is capacitatively coupled to ground." This appears to mean that the 480 secondaries can be hazardous to personnel. Furthermore, based on my understanding, if one of the 480 volt secondary conductors accidentally contacted the system ground, it may not blow a fuse or trip a breaker. However, it would still be hazardous to both people and equipment. I need a better understanding of the safety considerations for this sort of installation.

 

[electrofelon]

I recently saw a 3 wire transformer connected to a 120/208 3 phase supply. There is about 488 volts between phases on the secondaries and zero volts to ground. The equipment the transformer supplies does not use a neutral conductor. Even though there is no voltage to ground, my research suggests that there are still safety concerns with the secondary conductors, because "the system is capacitatively coupled to ground." This appears to mean that the 480 secondaries can be hazardous to personnel. Furthermore, based on my understanding, if one of the 480 volt secondary conductors accidentally contacted the system ground, it may not blow a fuse or trip a breaker. However, it would still be hazardous to both people and equipment. I need a better understanding of the safety considerations for this sort of installation.

An ungrounded system isnt more dangerous than a grounded system. All the grounding myths have made many think "grounded=good" "ungrounded=bad, danger". Note that a first fault basically just makes it a grounded system - nothing "bad" happens. The only danger I can see in an ungrounded system is someone who doesnt really know what they are doing and working on it live and assuming it is still ungrounded when there has been a first fault that wasnt detected,or assuming "you cant get shocked because its ungrounded".

 

[gar]

171126-1500 EST

Why would an ungrounded 480 V delta be hazardous to you? Think about it.

(1) If you take any 480 V supply, grounded or not, and place yourself across it, then in almost every case that will be very hazardous to you. The current thru you will be a function of the 480 voltage and the resistance of you between the points of contact.

(2) If you take a floating from ground (meaning leakage current to ground is nil) 480 V supply, and grab tightly one output wire and also touch ground, then you will feel nothing, virtually no current flows thru you.

(3) Next ground the other end of the 480 V supply in (2). You will be zapped and probably killed.

(4) Connect a million volts between one end of a 480 V secondary and earth (assume the transformer insulation is good for more than a million volts). Put yourself between any point on the 480 V secondary and earth, and you will probably be killed.

.

 

[Leor]

171126-1500 EST

Why would an ungrounded 480 V delta be hazardous to you? Think about it.

(1) If you take any 480 V supply, grounded or not, and place yourself across it, then in almost every case that will be very hazardous to you. The current thru you will be a function of the 480 voltage and the resistance of you between the points of contact.

(2) If you take a floating from ground (meaning leakage current to ground is nil) 480 V supply, and grab tightly one output wire and also touch ground, then you will feel nothing, virtually no current flows thru you.

(3) Next ground the other end of the 480 V supply in (2). You will be zapped and probably killed.

(4) Connect a million volts between one end of a 480 V secondary and earth (assume the transformer insulation is good for more than a million volts). Put yourself between any point on the 480 V secondary and earth, and you will probably be killed.

.

Thank you, electrofelon and gar, for the clarifications. I’m learning, and what I’m learning suggests the following: If there is a “first fault” in this system (I assume this means one secondary conductor is contacting the system ground,) it then becomes a grounded system. Once this happens, then it is my understanding that you would receive a severe or fatal shock if you were to contact either of the OTHER TWO secondary conductors while you were also in contact with the system ground. The other thing I gather from your helpful explanations, is that if there is one of these first faults in this system, and then either of the other two secondary conductors contacts the system ground, this would constitute a short circuit between those two phase conductors. I am guessing that this is what you might call a second fault. Apparently, the only way to trip the overcurrent devices, under normal operation of this system, is when the equipment the system is supplying is drawing too much current, or there is a second fault in the system. What do you think? Please correct me if I have gotten any of this wrong.

 

[electrofelon]

Thank you, electrofelon and gar, for the clarifications. I’m learning, and what I’m learning suggests the following: If there is a “first fault” in this system (I assume this means one secondary conductor is contacting the system ground,) it then becomes a grounded system. Once this happens, then it is my understanding that you would receive a severe or fatal shock if you were to contact either of the OTHER TWO secondary conductors while you were also in contact with the system ground.

Yes. But note in practice, no one who knows what they are doing is going to feel free to touch a phase and ground on an ungrounded system. Yes in theory you could do it, but you have capacitive coupling and the possibility that a first fault has already happened so you cant rely on not getting shocked phase to ground.

The other thing I gather from your helpful explanations, is that if there is one of these first faults in this system, and then either of the other two secondary conductors contacts the system ground, this would constitute a short circuit between those two phase conductors. I am guessing that this is what you might call a second fault. Apparently, the only way to trip the overcurrent devices, under normal operation of this system, is when the equipment the system is supplying is drawing too much current, or there is a second fault in the system.

correct. A second fault would result in high fault currents and activate an OCPD just like in a grounded system.

 

[Leor]

Yes. But note in practice, no one who knows what they are doing is going to feel free to touch a phase and ground on an ungrounded system. Yes in theory you could do it, but you have capacitive coupling and the possibility that a first fault has already happened so you cant rely on not getting shocked phase to ground.



correct. A second fault would result in high fault currents and activate an OCPD just like in a grounded system.

Great. Thank you electrofelon, I feel more comfortable with this setup now, and I can start thinking in terms of working with the system.

This system has no ground fault detection. During regular maintenance, what's the best way to check, to make sure there isn't a ground fault? Option 1: Should I turn off power to the transformer and check for continuity between the secondary conductors and the system ground? Option 2: Should I leave the transformer on and check for voltage between the secondary conductors and the system ground? Or should I consider other options?

This raises questions about test equipment. I don't think I can use my digital meter (Fluke T5-600) for option 2, because it always seems to detect voltage between the secondary conductors and the system ground when the transformer is energized. I think I can use my solenoid tester (Knopp K-60) for option 2, because it doesn't read any voltage between the secondaries and the ground with the transformer energized, when there is no ground fault.

What do you think?

 

[gar]

171126-2003 EST

Leor:

In a real world application there will be enough capacitive and resistive leakage that some balance will occur. This may not be an equal balance.

If you use a low impedance meter you will introduce an unbalance that was not already there. Use only a high impedance meter.

Within machines that have a floating delta, relay logic days, it was common to use three pilot lights as an unbalance indicator. Each pilot light was wired from a leg to ground. This resulted in a balanced load on the three phase delta, and if there was no other loading that produced an unbalance, then the three pilot light would have about the same brightness.

On the main 480 floating bus distribution system in a plant I don't know whether the three bulb method was used or not. It would work here as well as within a machine.

Your option 2 is the one to use. Option 1 won't work. For a balanced system the ratio should be about (L to Grd) * 1.7 = (L to L) .

.

 

[electrofelon]

First let me say I have only come across a few ungrounded systems. They were old 600V, no ground detectors. So I am sure someone will chime in that has more experience troubleshooting these. I think you are on the right track: You can use a high impedance meter like a DMM to measure the voltages to ground. If it reasonably in the middle, ie 277 on a 480 system or 347 on a 600 system, you should be in pretty good shape. IF you get close to 0 on one of the phases to ground, you probably have a fault. The trick is that middle ground and figuring out if its a fault or just some "uneven" capacitive coupling or some higher leakage current. That is when you can use your low impedance meter to see if its enough to make the solenoid move.

 

[Leor]

First let me say I have only come across a few ungrounded systems. They were old 600V, no ground detectors. So I am sure someone will chime in that has more experience troubleshooting these. I think you are on the right track: You can use a high impedance meter like a DMM to measure the voltages to ground. If it reasonably in the middle, ie 277 on a 480 system or 347 on a 600 system, you should be in pretty good shape. IF you get close to 0 on one of the phases to ground, you probably have a fault. The trick is that middle ground and figuring out if its a fault or just some "uneven" capacitive coupling or some higher leakage current. That is when you can use your low impedance meter to see if its enough to make the solenoid move.

OK gar, electrofelon, thank you both again. This all makes sense. I think I'm ready to go.

 

[jumper]

Here are two articles and a thread on the subject I found interesting. I have never worked on these systems and your thread made me curious.

When someone mentions ungrounded systems my mind seems to always go to two situations.

Surgery in a hospital and large cranes with heavy loads, both are times when suddenly losing power can be a very bad thing.

Anyways, the links:

http://www.ecmag.com/section/codes-...and-regulations-code-rules-ungrounded-systems

https://www.schneider-electric.us/e...-and-ungrounded-electrical-system-designs.jsp

http://forums.mikeholt.com/showthread.php?t=143948