Deriving a neutral: musings

Status
Not open for further replies.

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
In thread http://forums.mikeholt.com/showthread.php?t=87981 I hijacked a useful thread with a theory discussion. The original poster was looking to run some lighting with a 480V three wire feed, and had asked about deriving the neutral.

Vern responded that you can't pull a neutral out of thin air, and I threw in a comment about 'zig-zag' transformers. Rather than hiding answers which are actually useful to the original poster, I figured that I should direct any further discussion of this approach to a different thread.

In doing more reading, I've learned that there are _two_ different things called zig-zag transformers. One type is a regular transformer with separate primary and secondary, but where each secondary phase is made from a pair of coils driven by two primary phases. This phase shifts the output, and is useful for some harmonic mitigation situations.

The other type of 'zig-zag' transformers is essentially an autotransformer; there are no separate primary and secondary circuits. Each phase has two coils, electrically in series. The two coils are placed on different legs of the core. See http://www.electrical-contractor.ne...flat/Forum/15/topic/000072/Number/0/site_id/1

This second type of transformer is used to derive a neutral in a delta system. The most common use is to derive a neutral that is used to ground an otherwise ungrounded delta, generally with a rather _undersized_ zig-zag transformer and a method of high resistance grounding. Such small 'grounding transformer' is not sized to the full load of the supply delta; it is just large enough to provide the neutral and to tolerate fault current when it occurs; it is not sized to carry load on a regular basis.

There are other types of grounding transformers; I don't understand the reason for selecting one or another type.

Cody K said:
Would you not bond the GEC to the ungrounded conductor on the secondary? Primary is irrevelant.

In the situation that I was describing, a zig-zag autotransformer is sized to carry neutral loads in a system that uses a neutral. Because this transformer creates a neutral, and the neutral is being used as a current carrying conductor, the system is required to be grounded. Since this is an autotransformer, with no isolation between primary and secondary, you can't simply bond the secondary and ignore the primary. You have a single system (not a second separately derived system) and you have to figure out where to ground it.

This is just me musing on how such a transformer would be used. I see three scenarios:

1) Ungrounded delta. Zig-zig added to the transformer to create a grounded system. Treat like a transformer bank which provides a wye secondary.

2) Ungrounded delta, zig-zag somewhere down a feeder. Can't make this work legally. You have created a system that needs to be grounded, but don't have a good location for the bond.

3) Grounded supply, zig-zag somewhere down a feeder. Now you have a system which needs to be grounded, and which is grounded, but you also have a derived neutral which cannot be connected to ground, and is thus an ungrounded conductor even though it is at 0V.

-Jon
 
Zig-Zag and Wye-Delta

Zig-Zag and Wye-Delta

A zig-zag transformer has essentially no impedance during a line-ground fault so must be resistively grounded to limit current. Their use is to provide a ground fault source for OCPD operation on an otherwise ungrounded system and to limit line-ground voltage during line-ground faults.

A grounded wye to closed delta transformer will do the same when the wye connection is added to a delta supply. I know of a padmounted bank made of three line-ground transformers serving a 480V delta motor load. Even when the load is disconnected, there is 50A in the 480V because the voltage on the transformer primaries is unbalanced causing the bank to attempt to correct the unbalance by forcing line current.

A grounded wye service has no use for a zig-zag grounding transformer because the service is already referenced to ground.

A zig-zag transformer is not really an auto-transformer (one voltage in and one voltage out), it is a magnetic device connected into an electric system for system protection. Application of grounding transformers must be tied intimately with protection systems.
 
winnie said:
This is just me musing on how such a transformer would be used. I see three scenarios:
...
3) Grounded supply, zig-zag somewhere down a feeder. Now you have a system which needs to be grounded, and which is grounded, but you also have a derived neutral which cannot be connected to ground, and is thus an ungrounded conductor even though it is at 0V.
Wouldn't the derived system neutral be at 277V to ground? That is, in order for the supply to be grounded, it would have to be a corner-grounded delta system and the grounded corner would be considered the 0 volt reference. The derived neutral connot operate at 0 volts as this would mean there is a short circuit to ground.
 
Smart $ said:
Wouldn't the derived system neutral be at 277V to ground? That is, in order for the supply to be grounded, it would have to be a corner-grounded delta system and the grounded corner would be considered the 0 volt reference. The derived neutral connot operate at 0 volts as this would mean there is a short circuit to ground.

I partially agree.

If the source is a corner grounded delta, and you connect a zig-zag transformer, then the neutral of the zig-zag would be at 277V (presuming a 480V system).

But if the source is an ungrounded delta, then the neutral of the zig-zag will be at an undefined voltage.

IMHO if we create a neutral and use it as a circuit conductor, then we are required to bond the neutral to ground.

-Jon
 
beanland said:
A zig-zag transformer has essentially no impedance during a line-ground fault so must be resistively grounded to limit current. Their use is to provide a ground fault source for OCPD operation on an otherwise ungrounded system and to limit line-ground voltage during line-ground faults.

I understand that they are used as grounding transformers...but is there any physics reason that a suitably sized zig-zag transformer could not be used to derive a neutral and create a 4 wire system from a three wire system?

beanland said:
A zig-zag transformer is not really an auto-transformer (one voltage in and one voltage out), it is a magnetic device connected into an electric system for system protection.

The single phase equivalent of a zig-zag transformer would be a single phase coil with a center tap, without a separate primary or secondary coil. Such a device looks exactly like a 2:1 autotransformer. I contend that a zig-zag transformer is essentially an autotransformer, in that you have a set of coils which develops a voltage not present at the input, but where the output is not electrically isolated from the input.

-Jon
 
winnie said:
I partially agree.

If the source is a corner grounded delta, and you connect a zig-zag transformer, then the neutral of the zig-zag would be at 277V (presuming a 480V system).
This sounds like a full agreement to me :wink:

...as my post was regarding only OP scenario #3

But if the source is an ungrounded delta, then the neutral of the zig-zag will be at an undefined voltage.
The situation here is different than the one for which I made a comment... but I agree with the statement as it is, prior to grounding the neutral.

IMHO if we create a neutral and use it as a circuit conductor, then we are required to bond the neutral to ground.
That is correct... 250.20(B)(2). Best forget using a zig-zag xfmr on a grounded delta system. Short of changing the existing supply or adding a new one, the only viable option would be to use a delta-wye xfmr.

So in summation, scenario #3 would not be a consideration... leaving only scenarios #1 and #2. Would you please elaborate on your scenario #2 statement of "Can't make this work legally. You have created a system that needs to be grounded, but don't have a good location for the bond." Why? I understand it may not be a "good location" for bonding, but how does that make it "illegal"?
 
Smart $ said:
Would you please elaborate on your scenario #2 statement of "Can't make this work legally. You have created a system that needs to be grounded, but don't have a good location for the bond." Why? I understand it may not be a "good location" for bonding, but how does that make it "illegal"?

Consider the situation: you have a delta secondary on a transformer (no bonding connections have been made yet). These conductors go to a disconnect panel, and then from the disconnect panel you have a feeder to another panel. Just upstream of the second panel, you add a 'zig-zag' transformer and derive a neutral, which you intend to use as a circuit conductor. IMHO this neutral needs to be bonded, but it needs to be bonded somewhere between the transformer and the first panel...but the neutral has been created 'somewhere down a feeder'; so it seems to me that the neutral would need to be carried back to the main disconnect and bonded to ground there.

Writing the above up, I realize that I made the assumption that 'somewhere down a feeder' meant between the main disconnect and some random panel, but it could just as easily have been between the supply transformer and the main disconnect.

-Jon
 
winnie said:
Consider the situation: you have a delta secondary on a transformer (no bonding connections have been made yet). These conductors go to a disconnect panel, and then from the disconnect panel you have a feeder to another panel. Just upstream of the second panel, you add a 'zig-zag' transformer and derive a neutral, which you intend to use as a circuit conductor. IMHO this neutral needs to be bonded, but it needs to be bonded somewhere between the transformer and the first panel...but the neutral has been created 'somewhere down a feeder'; so it seems to me that the neutral would need to be carried back to the main disconnect and bonded to ground there.

Writing the above up, I realize that I made the assumption that 'somewhere down a feeder' meant between the main disconnect and some random panel, but it could just as easily have been between the supply transformer and the main disconnect.

-Jon
"IMHO this neutral needs to be bonded..."
Not an opinion... it is required to be grounded.

"...seems to me that the neutral would need to be carried back to the main disconnect and bonded to ground there."
Actually it would not be carrying the neutral back to the main, it would be bringing the GEC to the zig-zag xfmr (or its load-side disconnect). I believe this is true regardless of where you put the zig-zag in the distribution scheme.

Essentially, this converts your scenario #2 into a scenario #1.
 
winnie said:
I understand that they are used as grounding transformers...but is there any physics reason that a suitably sized zig-zag transformer could not be used to derive a neutral and create a 4 wire system from a three wire system?

The single phase equivalent of a zig-zag transformer would be a single phase coil with a center tap, without a separate primary or secondary coil. Such a device looks exactly like a 2:1 autotransformer. I contend that a zig-zag transformer is essentially an autotransformer, in that you have a set of coils which develops a voltage not present at the input, but where the output is not electrically isolated from the input.

-Jon

Not really. There is no physics reason why it cannot be done. If I can take a delta-wye transformer to convert an ungrounded service to grounded, surely I can do the same with a zig-zag transformer. Would be interesting to examine voltages and flux in the core.

The magnetic characteristics of a zig-zag are nothing like an autotransformer, regardless of what the windings look like. What you are describing is more like a balancing reactor for single-phase or phase-ground loads fed from two sources. The flux in the core of the zig-zag is intentionally canceled for unbalanced voltages. The flux in an autotransformer is not.
 
Status
Not open for further replies.
Top