In the picture below you can see an operation of a transformer station and consumers. So in this case 3A flow back via the PEN to transformer station. enter image description here picture 1 - standard operation

In the picture below you can see a "phantasy" installation where the ground of transformer station and plant is separated. enter image description here picture 2 - phantasy operation

1. Is the standard operation more efficient than the phantasy installation, I mean do the coils of the transformer have to perform less, because of the current backflow? From my point of view its not because in each case - standard or phantasy - the current flow comes at potential of 0V
2. What is the reason that the N is usually carried back via PEN from plant to trafostation? Why not ground plant and trafostation separately?
Still you are invited to add comments.
Assuming from all your answers the main reasons are:
  1. Cost - When you lead back PEN to transformer station it takes only one grounding point
  2. Safety, load balance - With only one grounding point and same N on trafo and consumer side, you assure that all current which goes into consumers also goes out. With different grounding points, with different resistances you could have different potentials on N where you don't want one
  3. In "phantasy installation" you would have high current loads on groundings, this would need bigger diameters on the grounding lines and could lead to undefined behaviour during time, e.g. corrosion
  • \$\begingroup\$ Please simulate it with a few different (orders of magnitude) impedance on your ground return. Next up, fault cases. Line short to ground or loss of one phase. Compare those results to metallic return. \$\endgroup\$
    – winny
    Dec 8, 2023 at 10:33
  • \$\begingroup\$ I assume there will be a lot disbalances, but Iam not capable of doing those calculations \$\endgroup\$ Dec 8, 2023 at 10:41
  • \$\begingroup\$ Perhaps, but any simulation software will happily do it for you :-) My point is that unless the soil is very moist and conductive all the time, the added expense of the metallic return will be overcome by the benefits for the grid operation. \$\endgroup\$
    – winny
    Dec 8, 2023 at 10:48
  • 3
    \$\begingroup\$ "Trafostation" seems to be a German word; what is that in English? I did a quick Google search and it seems like "Trafostation" may be the German word for "substation," but I'm not certain. \$\endgroup\$ Dec 8, 2023 at 18:03
  • \$\begingroup\$ Thanks for comment and sorry, for not being precise. Trafostation is a german abbreviation for transformer station or like you mentioned substation. Like the ones you have in the streets transforming medium voltage down to low voltage, so it can be used in residential or industry applications. I will update pictures with correct words... \$\endgroup\$ Dec 9, 2023 at 8:51

4 Answers 4


Is your town built on a giant copper nugget?

What you're trying to do is replace PEN with "the dirt". Dirt does not have the capacity for the task.

What is "the task"? It is to return imbalance currents. Remember, you have drawn the loads as three single-phase loads. As such it is not guaranteed that current is exactly the same on all three. Thus, you will have return current which needs to flow.

Now, Luiz Olivera has shown a method it can be made to work, by pushing voltage so very high (and thus current so very low) that the current is negligible and therefore the voltage drop across the dirt is bearable. This is called SWER = Single Wire Earth Return.

An example of fail

However, you want to do that at normal (safe in buildings) distribution voltages and currents. That's not going to work. In fact, we know exactly what happens when you try, thanks to the (North) American approach There, a hybrid system is used; domestically it's viewed as "Bonded TT" but the rest of the world considers it PEN with local earth spikes.

enter image description here

Homes in low density areas only get 2 phases from a split-phase transformer.

You see that we have set the stage for your theory to be tested. All that is needed is for something unfortunate to happen to that neutral wire. That's not a problem: Many homes have pole-to-home overhead supply wiring with triplex or quadruplex, and the bare wire responsible for carrying the weight is neutral as well. That whips in the wind for 30 years (aluminum has no fatigue limit) and SNAP! The wire breaks, neutral is disconnected, and we are putting your theory to the test.

And here's what we get. Yes, this happens often enough that it has a name: Lost Neutral. With no center neutral, the three (or two, often) line-neutral voltages wander all over the place. It's much easier to see in North American "2-phase" (really split-phase), because the two 120V legs are not 120V anymore, *but they still sum to 240V. They might be 110V and 130V. Or 100V and 140V. Or 150V and 90V. In the North American experience, the voltage doesn't wander terribly far, because that earth spike is making a difference. It is able to handle return current, just with very high resistance, which is tilting the voltages out of whack.

Obviously with split-phase, the voltage is a "see-saw". With 3-phase, it's more of a triangle-shaped see-saw with the pivot in the center of the triangle. More like a tug-of-war, honestly, with loads pulling on each end, and neutral where the ropes meet, but also tied to the center by a "bungee" stretch rope.

So it is not a good thing at all, no. The American experience is it can go for days or weeks undetected, simply because the earth spike is keeping voltages centered enough that things are not outright breaking. Loads are acting strangely, but they're still working, by and large. However this is still damaging to equipment when one phase pulls much heavier than another phase, causing the other phase(s) to have exceptionally high voltage. And that's what the local earth spike does for you.

Without it, you get a classic broken PEN, as discussed here by John Ward.

When it happened to us at my winter cottage, we lost a Crock Pot dish because it didn't get warm enough, and my sweetie said "I'm sorry, the toaster is very SLOW today". Toasters don't have speeds, so I was up out of my chair checking voltages and sure enough. The utility came out in an hour, on a Sunday, and fixed us and a neighbor with a lost phase.

  • 1
    \$\begingroup\$ Many thanks for your answer. Your essential and accurate sentence: "What you're trying to do is replace PEN with "the dirt". Dirt does not have the capacity for the task." is for me key \$\endgroup\$ Dec 9, 2023 at 9:09

My experience in systems with return path through ground is limited to rural areas in Brazil. Here, many remote locations make use of a system consisting of just one cable around 36kV. The transformer has only one medium voltage (MV) bushing and a grounding point while the LV side has three bushings with center tap configuration. As the picture below (taken from here: https://www.youtube.com/watch?v=MFunnbM_Oyo).

enter image description here

I'm sure it's used due to it's very low cost. During the operation, there is some problems to deal with. The voltage regulation is severe, making it very difficult for electric motors to start. In the LV side there is two voltages available, as 127V (phase to ground) and 254V (phase to phase). When the 127V is used, the ground return in LV (high current) makes the regulation even worst, so it can only be used for very light loads.

One last safety issue is the MV grounding, if the grounding cable is broken somehow, the transformer tank will be energized at 36kV.

So, the high voltage regulation means that, in fact, there is voltage drop through the grounding path and I think it also sheds some light to your questions.


First of all, the load on L1, L2 and L3 is mostly balanced on the consumer side. Since the phase between them is 120° and assuming 1 Ampere is flowing through L1, you would have -0.5 Ampere travelling through L2 and L3 each and no current through N.

current through L1, L2 and L3  https://en.wikipedia.org/wiki/File:3_phase_AC_waveform.svg

Image source: User:J JMesserly modification of original svg by User:SiriusA, Public domain, via Wikimedia Commons

Assuming now you have a consumer who uses just L1, all current has to travel through N back to the transformer. In this scenario and your phantasy installation all current would travel through the ground, since the resistance of earth and rock is not so great as a cable, there would be a significant voltage drop. All metal pipes, street signs and waterstreams between the consumer and the transformer could be on some potential and therefore be electrified, this is a public safety hazard!


The main reason is safety. Metallic ground costs more. The expected currents are low due to load balance. Triplen harmonics will however flow.

  • \$\begingroup\$ Thanks for your answer! The cost argument makes sense, there is only one grounding point if I lead back N and the diameters could be smaller Could you please explain your statement "The expected currents are low due to load balance"? \$\endgroup\$ Dec 8, 2023 at 9:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.