3-phase transformers and common core

Also do common core transformers tend to balance the phases? By this, I mean if the secondary loads were unequal, would it appear (more) equal at the primary? Or if, say the primary was fed with unequal voltage (say, 208/208/200 or something because whatever was feeding it was somehow flawed) would the secondary voltages be more equal, percentage wise?

One reference is IEEE Standard C57.105-1978 IEEE Guide for Application of Transformer Connections in Three-Phase Distribution Systems. TOC at

--s falke

"Michael Mor> Also do common core transformers tend to balance the phases? By this,

in article snipped-for-privacy@news4.newsguy.com, snipped-for-privacy@ipal.net at snipped-for-privacy@ipal.net wrote on 4/8/04 7:23 AM:

Sorry, but to my weak mind, I do not understand you.

Bill

It does. A three phase, three leg core type transformer has different zero sequence characteristics than three single phase units or a three phase (shell type) core that has a flux return path.

The flux will not have a return path and will saturate the core. As a result, the shunt impedance will be much lower than for 3 independent cores of the same rating.

On Thu, 08 Apr 2004 14:28:53 -0700 Paul Hovnanian P.E. wrote: | snipped-for-privacy@ipal.net wrote: |> |> 3-phase distribution transformers are mostly manufactured using a common |> core. That is, there are 3 separate winding sets, but the cores are |> connected together as one big piece. Presumably this is to reduce costs. |> |> I have seen transformer designs where there are 2 separated windings on |> each side, but the windings are still magnetically coupled design not |> being wound over top each other, so presumably the circularity of the |> core pass the magnetic field completely or nearly enough so for this to |> be a pratical design. |> |> Questions: |> |> Why would this not affect how a 3-phase transformer works, given that it's |> construction is just like the 1-phase transformer with separate windings |> sets, with just one more winding set added? | | It does. A three phase, three leg core type transformer has different | zero sequence characteristics than three single phase units or a three | phase (shell type) core that has a flux return path. | |> Do the 3 winding sets interact in some special way that is integral to the |> design? |> |> What if I were apply single phase power to each of the 3 windings of an |> integrated 3-phase transformer, in parallel? | | The flux will not have a return path and will saturate the core. As a | result, the shunt impedance will be much lower than for 3 independent | cores of the same rating.

The designs I see in 3-phase transformers have the 2 other phases in parallel with respect to any one chosen phase, in terms of the flux path. So I have two questions about an alteration of design:

1. What if, instead of a single bar across the 3 cores along the top, and another across the bottom, what if the 3 cores were arranged in a triangle and the bar on top and bottom replaced with a triangular block?

1. What if the 3 windings were arranged in series with respect to the flux path, rather than with 2 others in parallel, such as each winding set being at a 120 degree point on a toroid?

Older designs used single pots and then combined the secondaries for a service. It was cheaper to build smaller single phase transformers than big ones. Like the single phase transformers on poles. The designs changed a lot because of the PCBs to air cooling the 3 phase core was invented. (I am sure there were other reasons) It by its nature of being a single mass, for all purposes is more efficient that the grouping of single phase pots. Just my view, waiting for the gurus to chime in

Not quite 'parallel'. While that is certainly the way it looks, think about how the flux levels vary in each leg with time. When 'A' is peaked with the 'N' at the top, the 'B' and 'C' are about equal with the 'S' at the top. So the 'B' and 'C' are in 'parallel' and the pair of them are in 'series' with 'A'. At other parts of the cycle, other combinations occur. Because of the time displacement in each phase, the relative orientation of the MMF constantly shifts back and forth amongst the phases.

This would probably work, but you will have increased the cost of the core for no apparent reason.

No, If I understand what you're suggesting, this would not work.

The 'flux' from a winding must find a closed circuit to follow from one side of a coil back to the other side of the same coil for best operation. If another coil is along that path (as in a normal 3-phase transformer), the flux generated by it will either aid or oppose the first coil's. If it opposes, then it will reduce the amount of flux generated for the same MMF. Just the same as an air-gap or other magnetic 'restriction' and the overall flux levels drop. The second two coils *would* be the right polarity for

If your suggesting putting three single phase windings on one toroid fed from a single phase supply, then the toroid would have to be large enough that all three coils working together did not saturate the core.

daestrom

On Thu, 08 Apr 2004 17:21:05 GMT Repeating Rifle wrote: | in article snipped-for-privacy@news4.newsguy.com, snipped-for-privacy@ipal.net at | snipped-for-privacy@ipal.net wrote on 4/8/04 7:23 AM: | |> 3-phase distribution transformers are mostly manufactured using a common |> core. That is, there are 3 separate winding sets, but the cores are |> connected together as one big piece. Presumably this is to reduce costs. | | | | Sorry, but to my weak mind, I do not understand you.

Look at the image files in this web directory:

It is still common practice to build a three phase overhead bank from three single phase transformers. Its easier to balance (physically) the cans on a pole with a bracket that keeps the bank's CG nearer the pole center line.

For other installations, its all a matter of cost, although three single phase padmounts are rarely made into a bank because of the added clutter

I see lots of delta connections. These would tend to mitigate problems caused by the lack of a zero sequence flux return path.

On Thu, 08 Apr 2004 21:45:15 -0700 Paul Hovnanian P.E. wrote: | snipped-for-privacy@ipal.net wrote: |> |> On Thu, 08 Apr 2004 17:21:05 GMT Repeating Rifle wrote: |> | in article snipped-for-privacy@news4.newsguy.com, snipped-for-privacy@ipal.net at |> | snipped-for-privacy@ipal.net wrote on 4/8/04 7:23 AM: |> | |> |> 3-phase distribution transformers are mostly manufactured using a common |> |> core. That is, there are 3 separate winding sets, but the cores are |> |> connected together as one big piece. Presumably this is to reduce costs. |> | |> | |> | |> | Sorry, but to my weak mind, I do not understand you. |> |> Look at the image files in this web directory: |> |>

They are intended to show the structure of the core frame. And yes, they do show delta connection on the primary. But that's the most common.

|> 1. What if, instead of a single bar across the 3 cores along the top, and |> another across the bottom, what if the 3 cores were arranged in a | triangle |> and the bar on top and bottom replaced with a triangular block? | | This would probably work, but you will have increased the cost of the core | for no apparent reason.

It wouldn't have to be a solid block at the top and bottom. It could be that part of the vertical core goes across one way, and part the other way, to the other 2 cores. Everything would be symmetrical.

|> 2. What if the 3 windings were arranged in series with respect to the | flux |> path, rather than with 2 others in parallel, such as each winding set |> being at a 120 degree point on a toroid? | | No, If I understand what you're suggesting, this would not work. | | The 'flux' from a winding must find a closed circuit to follow from one side | of a coil back to the other side of the same coil for best operation. If | another coil is along that path (as in a normal 3-phase transformer), the | flux generated by it will either aid or oppose the first coil's. If it | opposes, then it will reduce the amount of flux generated for the same MMF. | Just the same as an air-gap or other magnetic 'restriction' and the overall | flux levels drop. The second two coils *would* be the right polarity for | *part* of the cycle, but not the entire cycle. | | If your suggesting putting three single phase windings on one toroid fed | from a single phase supply, then the toroid would have to be large enough | that all three coils working together did not saturate the core.

What I had hoped might be possible, but it does not see to be, is to make a toroid core with 6 windings placed at 60 degree intervals, with all the even ones primary and all the odd ones secondary. Then power it in three phases. The idea was to get a phase shift in the secondary because each winding is in between 2 of the primaries.

On Thu, 08 Apr 2004 14:28:53 -0700 Paul Hovnanian P.E. wrote: | snipped-for-privacy@ipal.net wrote: |> |> 3-phase distribution transformers are mostly manufactured using a common |> core. That is, there are 3 separate winding sets, but the cores are |> connected together as one big piece. Presumably this is to reduce costs. |> |> I have seen transformer designs where there are 2 separated windings on |> each side, but the windings are still magnetically coupled design not |> being wound over top each other, so presumably the circularity of the |> core pass the magnetic field completely or nearly enough so for this to |> be a pratical design. |> |> Questions: |> |> Why would this not affect how a 3-phase transformer works, given that it's |> construction is just like the 1-phase transformer with separate windings |> sets, with just one more winding set added? | | It does. A three phase, three leg core type transformer has different | zero sequence characteristics than three single phase units or a three | phase (shell type) core that has a flux return path. | |> Do the 3 winding sets interact in some special way that is integral to the |> design? |> |> What if I were apply single phase power to each of the 3 windings of an |> integrated 3-phase transformer, in parallel? | | The flux will not have a return path and will saturate the core. As a | result, the shunt impedance will be much lower than for 3 independent | cores of the same rating.

So what happens when one leg opens to a primary wired delta? As much as I know is that winding unaffected by the failure still gets full voltage, but the two remaining windings are now in series with that voltage from the two remaining hot wires. So on the secondary, assuming WYE, it "collapses" to single phase because of the single phase energizing the primary, along with half voltage on two of the hot-to-neutral connections and 86.6% voltage on two of the hot-to-hot connections. But what happens to the flux in the core under this situation?

Good point. I have not done many overheads in the last 20 years. Everything I have done has been ground mount.. Gad, forgot about the loads on the poles.

Since the delta primary winding configuration will not support zero sequence currents, the core flux will not see any either so saturation will not be a problem.

|> It is still common practice to build a three phase overhead bank from |> three single phase transformers. Its easier to balance (physically) the |> cans on a pole with a bracket that keeps the bank's CG nearer the pole |> center line. |>

|> For other installations, its all a matter of cost, although three single |> phase padmounts are rarely made into a bank because of the added clutter |> 3 boxes would create. |>

|> -- |> Paul Hovnanian mailto: snipped-for-privacy@Hovnanian.com |> note to spammers: a Washington State resident |> ------------------------------------------------------------------ |> Ask not for whom the tolls. | | Good point. I have not done many overheads in the last 20 years. Everything | I have done has been ground mount.. Gad, forgot about the loads on the | poles.

I guess the "three phase in one tank" pole pigs are not that good of an idea considering the pole balance; they'd have to be used on some cross member between 2 poles unless they are rather spamm.

I've driven by some rather large pole pigs arranged in three's serving commercial buildings. I didn't get a good look at one I saw serving a shopping plaza, but I'd guesstimate the tank volumes are about 4 to 6 times the volume of the 75 kVA single phase serving our neighborhood.

One thing I did see on a few of these is that on the two side transformers, there were 3 secondary terminals (much like a 240/120 volt transformer for home service), but the connections were made only to the center terminal and just one of the end terminals. It's like they are using 240/120 transformers to put together a 208Y/120 configuration. But wouldn't that force derating the transformer due to the current limit on the secondary? Or are these actually split secondary (e.g. 240x120 instead of 240/120) that have been rewired somewhere non-obvious to get the secondary windings in parallel with 120 volts?

BTW, the biggest "pole pig" I ever saw, which is certainly not going to be side mountable on a single pole, was about 7 feet high and 3.5 feet wide. Three of them were situated on a pad behind a building. I could not see the nameplate from outside the fence, but I could see the tap changers. In addition to the warning to de-energize before changing, there were a few voltages around 4160 volts. I don't know what the secondary voltage was, but the wires going into the building were as thick as my arm (including the insulation). The thing that struck me odd was that these units were entirely shaped as pole pigs, although with lots of large fins all around. They also looked very old, including being almost entirely a sea-green color tank that looked like copper or brass oxidation.