autotransformer sanity check

[electrofelon]

They are not the same point if not bonded together. Your conditions 2 and 3 each leave one point or the other "floating". When things are balanced they will be at/very near same potential. Throw some unbalance in there and one or the other will drift potential wise from the other. Bond them together and you force them to be same potential regardless of balance condition.

In a wye-wye autotransformer the HO and XO are the same point. They are not two points that may or may not be bonded together.

 

[kwired]

In a wye-wye autotransformer the HO and XO are the same point. They are not two points that may or may not be bonded together.

If you don't connect the 480/277 supply neutral to the wye point of the autotransformer, they are not the same point. They may run at same potential if the load is balanced. Your conditions 2 and 3 both leave out a solid link from supply neutral and wye point of that transformer which allows voltage at that point to drift in reference to the supply neutral if load is not balanced.

 

[electrofelon]

If you don't connect the 480/277 supply neutral to the wye point of the autotransformer, they are not the same point. They may run at same potential if the load is balanced. Your conditions 2 and 3 both leave out a solid link from supply neutral and wye point of that transformer which allows voltage at that point to drift in reference to the supply neutral if load is not balanced.

There has been some confusion about how many "points" we have. I generally assume we are talking about transformer terminals, but if we call the supply neutral and load neutral "points" then yes we have three points. Apparently we get voltage instability with my numbers 2 and 3. I am not disputing that.

 

[kwired]

There has been some confusion about how many "points" we have. I generally assume we are talking about transformer terminals, but if we call the supply neutral and load neutral "points" then yes we have three points. Apparently we get voltage instability with my numbers 2 and 3. I am not disputing that.

Three points become equivalent to one when they are bonded together. But not all of your scenarios were bonding them together.

 

[Besoeker]

In a wye-wye autotransformer the HO and XO are the same point. They are not two points that may or may not be bonded together.

Yes, agreed. One would have thought that obvious from the diagram posted earlier.
And I'm not sure it should be called wye-wye. It's just wye with taps at different points on the same windings. Isn't it?

 

[electrofelon]

Three points become equivalent to one when they are bonded together. But not all of your scenarios were bonding them together.

Right. That is the discussion, about what happens in the scenarios where all three points are not bonded.

And I'm not sure it should be called wye-wye. It's just wye with taps at different points on the same windings. Isn't it?

Yes you are correct, that is bad terminology

 

[kwired]

Yes you are correct, that is bad terminology

Maybe. If supply were from a 480 volt delta secondary (no neutral conductor) you could still connect this transformer to it and have wye voltages including 277 from each outer end to the center point. You could not connect that center point to ground if the supply were a corner ground system though, that would put 480 volts across two of the 277 volt coils. Can't say without further reading if NEC would allow such, but it will function that way.

 

[winnie]

No experience with such a beast, but sure seems to me that attaching the 480/277 supply neutral to the wye point and the ungrounded supply lines to the outer ends will force 277 volts across each coil, less any voltage drop in the supply circuit. Neutral can't "drift" when you put such a supply conductor there, unless there is unbalance severe enough to throw the supply voltage off balance.

This is something of a 'hand waving' explanation, but consider:

A pair of resistance loads in series, connected to a 240V supply. The junction between the two loads will be 'neutral' if the two loads are balanced, but if they have different values then the connection between the two loads will not be a good neutral point.

Now consider two matched 120V:120V single phase transformers with their primaries in series and with resistive loads on the secondaries, again connected to your 240V supply.

The primary current will be the same resistive current to the loads, plus a bit of magnetizing current. With no load connected to the secondary the primary (and thus secondary) voltages are balanced and you can derive a neutral. Similarly with balanced resistive loads you can derive a neutral, but if you have unbalanced loads then the 'neutral' voltage can float to any value between the two supply terminals.

Now consider a single phase auto-transformer with a 240V coil and a center tap, again connected to your 240V supply. You have a _single_ coil on a single core with a mid point tap. Any 'line to neutral' load is magnetically coupled to the _entire_ supply coil. A single coil does a pretty good job of deriving a single phase midpoint, which will be approximately neutral when supplied with a split phase source.

A 'wye' autotransformer has 3 separate coils with their various taps. Current loading on _one_ coil is not well distributed to the other coils. This means that if you put a large load on one of your coils (say by connecting a single phase load between a tap on a coil to 'wye point') then high current will want to flow on that single coil, but not on the other coils.

If your transformer were a bank of 3 entirely separate single phase autotransformers in a wye configuration, then the above analysis would be solid. However with a shared three phase core the analysis gets more complex, with load on one coil somewhat 'seen' by the other coils.

However a wye autotransformer has a relatively high impedance neutral point, and thus the neutral point voltage is not very stable when subjected to unbalanced loading. When wye autotransformers are used the supply neutral is connected to the 'wye point' in order to stabilize the line to neutral voltages.

When it is desired to intentionally derive a 3 phase neutral from a 3 phase source (say for grounding an ungrounded delta) then a 'zig-zag' transformer is commonly used. This sort of transformer places 2 coils on each leg of the core, with 2 different phases connected to each coil.

-Jon

 

[ggunn]

Assuming the inverter needs a neutral, with an autotransformer you run a neutral from the inverter to the transformer and from the transformer back to the service. With an isolation transformer you run a neutral from the inverter to the transformer but not from the transformer back to the service. Easy peasy.

Sorry to reply to myself, but is there any reason why this should not be the correct strategy? If the neutral at the inverter is only for voltage reference, the neutrals can be the same size as the EGCs.

As to the concept of using the EGC as the inverter neutral, although it may be functionally possible (I have seen it as a wiring option in at least one inverter installation manual), I'll wager that many AHJ's would balk at it.

 

[electrofelon]

Sorry to reply to myself, but is there any reason why this should not be the correct strategy? If the neutral at the inverter is only for voltage reference, the neutrals can be the same size as the EGCs.

As to the concept of using the EGC as the inverter neutral, although it may be functionally possible (I have seen it as a wiring option in at least one inverter installation manual), I'll wager that many AHJ's would balk at it.

I agree with you on the transformer connections. As for isolation transformers, we could throw a wrench in the works and talk about the supply neutral if we had a wye-wye instead of a delta-wye. For autos, I am still curious what would happen if you didn't land the neutral on the autotransformer and just ran it right to the load. Just a curiousity though.

As for the EGC, perhaps that is best to discuss in my "neutral update" thread. Solectria and chint do say the neutral is for volrage reference only and "if acceptable" to the AHJ you can jumper the neutral to the egc. I don't know why they don't just skip the neutral terminal then, tag the voltage circuit to the egc internally, and take all the doubt out. I don't know if it's some UL/red tape/beaurocracy thing, but it is stupid.

 

[jumper]

Very interesting comment/discussion! "stabilize???" the neutral???

Do we need to start a new thread to discuss this or can we morph this into that discussion since the OP was fully answered?

WHERE can we further discuss this fascinating topic? Do we have administrator approval to continue here, or?

It is e-felon’s thread and project. The thread goes where he wants it to go.

 

[kwired]

I agree with you on the transformer connections. As for isolation transformers, we could throw a wrench in the works and talk about the supply neutral if we had a wye-wye instead of a delta-wye. For autos, I am still curious what would happen if you didn't land the neutral on the autotransformer and just ran it right to the load. Just a curiousity though.

As for the EGC, perhaps that is best to discuss in my "neutral update" thread. Solectria and chint do say the neutral is for volrage reference only and "if acceptable" to the AHJ you can jumper the neutral to the egc. I don't know why they don't just skip the neutral terminal then, tag the voltage circuit to the egc internally, and take all the doubt out. I don't know if it's some UL/red tape/beaurocracy thing, but it is stupid.

I think you might not notice anything if the load is balanced. If not balanced you are likely to see unbalanced voltages to neutral, that will vary as load varies. I suppose if severe enough unbalance even voltages between 208 volt taps can change some.

 

[mike_kilroy]

...However a wye autotransformer has a relatively high impedance neutral point, and thus the neutral point voltage is not very stable when subjected to unbalanced loading...

Please add to this statement. Like relative to what?

It is me experience that auto transformers by necessary design have approx. 1/2 the impedance of equiv iso xfmr.

But you suggest higher Z, so compared to what? If you compare to a POCO 200kva wye supply, sure, this 2-3% Z auto xfmr would be higher, but so would the equiv 5-6% Z iso xfmr, no?

 

[mike_kilroy]

Right. The transformer would - in true autotransformer fashion - derive the neutral with the three X phases, then you could use it for the H/load. That was the thinking.



In my case, although I didnt state it, The "load" is a three phase PV inverter. I am 98% sure that the neutral is just used for phase monitoring and carries little to no current and that the output is very closely balanced. Thus perhaps we didntt need to bring the supply neutral from the 208 source for this application.

It is my opinion that not tying to the 208 wye neutral, which we know is also tied to EGC, you will leave your PV inverter susceptible to blowing up on input voltage spikes.

 

[mike_kilroy]

Yes they are the same point, no question about that. The discussion is about running a neutral with the supply, or not and letting the transformer derive the neutral for the load. We have three possibilities (#3 is new, I dont think it has been discussed yet):

1. supply neutral and load neutral both get landed to the common XO/HO terminal
2. Only the load neutral gets landed on the XO/HO terminal and no neutral is run with the supply.
3. Run the supply neutral to the load without hitting the XO/HO terminal.

What would happen for #3?

I would think the system would perform exactly the same as #1, since the Z from each phase of 208 to neutral will be nearly identical to Z from the 480v taps to same neutral.

 

[Ingenieur]

It is my opinion that not tying to the 208 wye neutral, which we know is also tied to EGC, you will leave your PV inverter susceptible to blowing up on input voltage spikes.

agreed
I would bet the inverter mfg requires this

just looked at the solaredge inv manuals from the other thread and they do

 

[mike_kilroy]

agreed
I would bet the inverter mfg requires this

just looked at the solaredge inv manuals from the other thread and they do

This is really a lot simpler to answer than guessing or getting all our opinions. There is a manufacturer who makes the inverter. They will for sure require a certain wiring of nuetral and EGC. REQUIRE. No guessing from us.

For example, we have applied, repaired, and sold Refu inverters for motors for over 20 years. Other brands before that. They REQUIRE certain power source wiring. They are also a major PV inverter supplier worldwide. They REQUIRE that nuetral and power sourcd EGC be tied solidly together. Where is the guess work?

[pix not printing here so attached instead]


I assure all on this thread that the PV inverter mfgr in question here also has detailed REQUIREMENTS for neutral and power supply EGC. Time to stop guessing and ASK THEM for the facts in my opinion.

 

[winnie]

Please add to this statement. Like relative to what?

It is me experience that auto transformers by necessary design have approx. 1/2 the impedance of equiv iso xfmr.

But you suggest higher Z, so compared to what? If you compare to a POCO 200kva wye supply, sure, this 2-3% Z auto xfmr would be higher, but so would the equiv 5-6% Z iso xfmr, no?

We need to make sure we are comparing apples to apples.

Normally impedance of a transformer is given by shorting the output terminals, and adjusting the primary voltage until the secondary current is the full load current. The % impedance is simply the test primary voltage/rated primary voltage. This is _not_ the impedance that I am talking about.

In this discussion we are asking about the stability of a 'derived neutral'. You have a 3 phase source, and only bring the 'hots' to the transformer, and use the transformer to create a neutral. The impedance at this 'neutral' terminal of the transformer is simply the voltage change/current flow, essentially an expression of E/I=R from Ohm's law. The greater the voltage change for a given current, the higher the impedance.

Forget about transformers for a moment, and use a set of resistors in a wye configuration to derive a neutral. If the resistors are balanced than the neutral point voltage would be pretty close to neutral of the connected hots. But if the resistors are unequal the wye point would not be near the neutral voltage. Any _load_ on this 'derived neutral' would unbalance the resistor network.

A wye autotransformer created from a bank of 3 single phase autotransformers would have very high impedance, acting much like the resistor based 'derived neutral'. Load between any 'hot' and neutral would only alter the loading on a single coil, and would not increase primary current on the other two coils. Since current must balance into the wye point, what would really happen is that the voltage on the loaded coil would decrease and the voltage on the other two coils would increase, until you got to equilibrium where the currents balanced.

In a more common 3 phase autotransformer with 3 coils on the same steel core, the impedance is better, because current flow on one coil changes the magnetic flux seen by the other coils...but still relatively high impedance in the sense that the neutral voltage change will be quite large for a given neutral current flow.

A very common use of a derived neutral is to change an ungrounded delta system into something that looks like a grounded wye system. This is done with a 'grounding transformer'. Grounding transformers are either zig-zag transformers, or they are wye transformers with delta coils that are not externally connected. Straight wye autotransformers are not used as grounding transformers. (The wye/unconnected delta arrangement uses circulating current in the delta to reduce the impedance of the neutral.)

-Jon

 

[GoldDigger]

A slightly different way to look at the wye with an unconnected internal delta is that the delta forces the three wye coil voltages to add up to a closed triangle, but does nothing to deal with any current imbalance at the wye point caused by an imbalanced load.
If you have a load only on one wye phase, an unconnected delta does nothing to allow that to happen with zero net current at the wye point.

In some earlier threads there was a clear recommendation to leave the primary wye open on a wye-delta transformer. That is a different situation because any single phase load on the delta (no neutral) can still be satisfied with non zero current in all of the wye primary coils, allowing a zero current solution for the wye point.

Sent from my XT1585 using Tapatalk

 

[electrofelon]

Thanks for the explanations. I feel like I understand the neutral stability issue much better now.