How to wire up a 240VAC receptacle

[Spork Schivago]

The three 16-gauge should be fine.   The wires that come with the PSU have three 17-gauge.   Unless those cords really are made for 240VAC running down L1...

How would I calculate the current draw that the PSU will be pulling from the NEMA receptacle for each wire?    Would it be 1400 watts / 120 volts or would it be 700 watts / 120 volts?

Or would it just be 1400 watts / 240 volts?   I'd think not the second, because L1 and L2 will each carry 120V.   But then again, we're still dealing with 240V so the current overall shouldn't change....but for the individual conductors, do we treat L1 and L2 as just having 240V total between them and the current draw per conductor (in the physical plug) will be roughly 5.9 amps, or would be roughly 11.7 amps?

1400w/240v

So just under 6 amps.

Sent from my Moto x4 using Tapatalk

Even though L1 and L2 are each 120VAC in reference to ground?    Just wanted to make sure.

Now one thing I will mention, the cord I'm sacrificing is NOT one of the cords that came with the PSUs and does NOT have the four pads on the bottom of it.   So I won't be plugging into those 4 pads at all, I'll just be by-passing them, but that shouldn't be an issue I wouldn't think for the simple fact that my PDU is NOT an "Intelligent" PDU and does not have the four pads per input....those four pads I believe are ONLY used for the Intelligent PDUs.

[Bratster]

See the post right above your reply, stop thinking about 120 volts. Your load is 240 volts.
The ground wire carries no current it is only for safety.

Sent from my Moto x4 using Tapatalk

[Spork Schivago]

The three 16-gauge should be fine.   The wires that come with the PSU have three 17-gauge.   Unless those cords really are made for 240VAC running down L1...

How would I calculate the current draw that the PSU will be pulling from the NEMA receptacle for each wire?    Would it be 1400 watts / 120 volts or would it be 700 watts / 120 volts?

Or would it just be 1400 watts / 240 volts?   I'd think not the second, because L1 and L2 will each carry 120V.   But then again, we're still dealing with 240V so the current overall shouldn't change....but for the individual conductors, do we treat L1 and L2 as just having 240V total between them and the current draw per conductor (in the physical plug) will be roughly 5.9 amps, or would be roughly 11.7 amps?

They don't each carry 120V. There is 240V between them and that is all there is to it. The two wires make the circuit so the current running through them will be equal.

You won't be hitting the full power draw anyway. Assemble your lead for this test (only. it won't be suitable for continued use as the breaker protecting it is rather too large), and give it a go. It'll be fine, it'll power up, it will complain about lack of redundant supplies.

We will be drawing more than 5.7 kW from the server, therefore, we need at least 2 PDUs.

I really hope you don't mean one server.

Also remember they use redundant supplies and you're very unlikely to draw the full 1400W from a single machine.

I gotcha.   Gonna wire it up and see what happens.   Wish me luck!   Broke my DMM (just the stand).   Knocked it over when I was looking for something.   That sucks.   Had that thing since I was a kid.   It was expensive at the time, but maybe I can buy a used unit to replace the stand.

[IanB]

Even though L1 and L2 are each 120VAC in reference to ground?    Just wanted to make sure.

Let's try this a different way. There are two wires, L1 and L2. Your multimeter has two probes, red and black. There are only two ways you can connect your meter probes to L1 and L2: either red-L1 and black-L2, or red-L2 and black-L1. Either way the meter will read 240 V AC. There are no other ways you can connect a meter to 2 wires, therefore there is no way to get a voltage other than 240 V from two wires.

If you had three wires there are more options, but since there are only two wires only one voltage is possible. You cannot get more than one voltage from 2 wires (other than zero volts by leaving one or both wires disconnected).

Whenever you are thinking about this you must ignore the ground wire as it is not allowed to carry current and must not be part of a circuit. If ever your electrical inspector finds the ground wire as part of a circuit you will fail inspection. The ground wire is forbidden to be used for anything except safety.

[Spork Schivago]

See the post right above your reply, stop thinking about 120 volts. Your load is 240 volts.
The ground wire carries no current it is only for safety.

Sent from my Moto x4 using Tapatalk

I actually wasn't thinking about 120VAC.   I had successfully calculated the load to be around 5.6 amp with a full load (which we won't have, because the upgrades aren't installed yet), so I'm good.   I just wanted to have you guys double check my work and provide a few scenarios.   I said, "...we're still dealing with 240V so the current overall shouldn't change...".   I just type how my mind thinks sometimes.   Gotta work on that I guess.

[Spork Schivago]

Even though L1 and L2 are each 120VAC in reference to ground?    Just wanted to make sure.

Let's try this a different way. There are two wires, L1 and L2. Your multimeter has two probes, red and black. There are only two ways you can connect your meter probes to L1 and L2: either red-L1 and black-L2, or red-L2 and black-L1. Either way the meter will read 240 V AC. There are no other ways you can connect a meter to 2 wires, therefore there is no way to get a voltage other than 240 V from two wires.

If you had three wires there are more options, but since there are only two wires only one voltage is possible. You cannot get more than one voltage from 2 wires (other than zero volts by leaving one or both wires disconnected).

Whenever you are thinking about this you must ignore the ground wire as it is not allowed to carry current and must not be part of a circuit. If ever your electrical inspector finds the ground wire as part of a circuit you will fail inspection. The ground wire is forbidden to be used for anything except safety.

That last part is still something I gotta get used to, because in DC, we use the ground for more than just safety.   It's the ground that completes the circuit.   So I have to keep switching back and fourth in my head between AC and DC (I'm currently working at the moment, and while the compiler is running, I bounce back here, then when it's done, I upload the code to a BGA component and watch my logic analyzer for a bit and then come back here, bounce back there, etc).

[IanB]

That last part is still something I gotta get used to, because in DC, we use the ground for more than just safety.   It's the ground that completes the circuit.   So I have to keep switching back and fourth in my head between AC and DC (I'm currently working at the moment, and while the compiler is running, I bounce back here, then when it's done, I upload the code to a BGA component and watch my logic analyzer for a bit and then come back here, bounce back there, etc).

Actually, no, that's not really true. Ground is used interchangeably with two different meanings, but really there should be two different words used. (In the UK there actually are two different words, ground and earth.)

The first meaning of ground is as the current return path, or the reference point for voltages. This could be called "signal ground" in some contexts. This is what ground means in the UK and what you may be familiar with in low voltage DC circuits.

The second meaning of ground is for electrical safety, as a protection mechanism. This ground must not carry current and must not be used in circuits. This is what earth means in the UK (sometimes called "protective earth" or "PE").

[Spork Schivago]

That last part is still something I gotta get used to, because in DC, we use the ground for more than just safety.   It's the ground that completes the circuit.   So I have to keep switching back and fourth in my head between AC and DC (I'm currently working at the moment, and while the compiler is running, I bounce back here, then when it's done, I upload the code to a BGA component and watch my logic analyzer for a bit and then come back here, bounce back there, etc).

Actually, no, that's not really true. Ground is used interchangeably with two different meanings, but really there should be two different words used. (In the UK there actually are two different words, ground and earth.)

The first meaning of ground is as the current return path, or the reference point for voltages. This could be called "signal ground" in some contexts. This is what ground means in the UK and what you may be familiar with in low voltage DC circuits.

The second meaning of ground is for electrical safety, as a protection mechanism. This ground must not carry current and must not be used in circuits. This is what earth means in the UK (sometimes called "protective earth" or "PE").

In DC, we have a bunch of words, depending on what you're dealing with, that essentially serve the same purpose.   For example, the term High-Voltage Differential Signaling could mean many things, depending on who you're talking to or what you're dealing with.   But then you have Low-Voltage Differential Signaling, which to my recollection, is actually a standard set by TIA/EIA.   And then, you gotta be careful about the low-voltage signal line sharing the same ground as a DC power supply line because you can end up getting some current returning to ground.   It's a bit of a different world than AC.   I mean, there's similarities it seems, but AC now seems kind of simple compared to DC.   I just never had the time to properly learn it.

Anyway, I hooked the BGA rework station backup, the server powered up just fine like you guys were saying, so I have to admit that I was wrong.   I will send an email to my Account Executive to see about a return.   We still have the box, which is important, because the PDU hasn't failed or anything.   So it doesn't get warranted and needs the original packaging or it'd be refused.   We've had it for a bit of time now because we were going by what HPE techs said and thought we really needed the transformer for whatever reason.   But I think this gives us some leverage to work with if they don't want to accept it because of the time.   As I see it, they're in the wrong here, not me.   What worries me is I installed it in the rack.   That's where I'm going to have issues, if there are any issues.

I'll remove it and see if they can tell it's been installed.   I don't like to lie, I don't think it's right, so if I'm asked, I'll tell them, yes, I installed it.   If they don't accept the return, I guess it's time to start thinking about dropping HPE and maybe switching to Lenovo.   I've heard their server side equipment is pretty decent.

[Gregg]

We gotta back up the train here a bit.   I understand the purpose of the neutral and ground in the house.   If we have a short, we want that current and voltage to go to the path of least resistance, hopefully back to the breaker, to trip it

OK, let’s try to go back to some basics.
I assume you know how a transformer works; I suggest you grab some paper and draw the circuits out for yourself as you follow along; visual reinforcement is a great learning aid and along the way you might have the “Ah Ha” moment where this becomes clear. 
Vin / Vout = Turns in / Turns out and the current capacity is more or less determined by winding size
If you have a 120VAC primary transformer with a 24VAC secondary, you get 24VAC output
Suppose you put this transformer on a metal chassis like old radios etc. and you plug it into the wall with a standard grounded 3 pin plug so that L1 is fused and connected to either wire of the primary and the neutral is connected to the other lead of the primary with the ground connected to the chassis.  All is well with the circuit here; the primary is a completed circuit via the neutral and L1.   

If one of the primary windings anywhere along the coils faults to the chassis, the ground that previously carried zero current suddenly carries the fault current back to the house breaker panel where it completes the circuit via the ground to neutral tie I showed in the previous post and the fuse should blow.   

At this point of the discussion, the secondary winding is completely isolated from the primary.  Let’s call them X1 and X2.  You could connect either lead to the chassis (which is grounded) and nothing would happen; no current would flow.  The other lead of the secondary would then be 24VAC to the chassis, often called chassis ground in this case which is also referenced to earth ground via the 3 prong plug. 

Let’s assume neither X1 nor X2 secondary winding are connected to the chassis for this paragraph.  If you take two 12 volt lamps of the same wattage in series and connect them to X1 and X2, they will both light equally.  They have the same resistance.  If you connect the midpoint between these two lamps to the chassis, nothing bad will happen, it is still isolated and, current will not flow to the chassis.  Now, connect a 5 watt and 10 watt both 12volt lamps in series and connect them to X1 and X2.  The 5 watt lamp gets very bright in a hurry and goes poof; the 10 watt lamp barely glows until the circuit goes open with the failure of the 5 watt lamp; this happens the same if the midpoint is connected to the chassis because it is isolated.

Now let’s change things a bit and connect a wire to the center of the secondary windings, a true center tap I’ll call C. Nothing on the secondary is connected to the chassis at this point.
X1 to X2 is still 24VAC
X1 to C is 12VAC
X2 to C is 12VAC 
Now, connect a 5 watt and 10 watt both 12volt lamps in series and connect them to X1 and X2 BUT with the midpoint connected to C.
They both light with their respective brightness and wattage; the 5 watt one doesn’t go poof!  The unbalanced extra 5 watts used by the 10 watt lamp is carried back to the transformer center tap C.  The 5 watt lamp and half of the 10 watt lamp’s current are the same as two 5 watt lamps in series without the midpoint connected to the transformer; the imbalanced 5 watts for the 10 watt lamp returns back to the transformer via the center tap.

Now if we connect the center tap C to the chassis (which is referenced to earth ground as previously stated) there is no harm, no foul and no current flow between C and the chassis.
And when we repeat the experiment, the results are the same.  The wire from the midpoint of the lamp back to C still carries the 5 watts; chassis ground is just a reference.

The transformer at the pole though....you said, "For instance, if you had 30 amps load on the 120volt side from L1 to neutral and 20 amps load on the other side from L2 to neutral, the neutral conductor would carry the 10 amp difference back to the transformer."

Yes, that is true as in the lamp scenario above; but to carry it further:  It would be very helpful for you to draw this on paper.
Suppose you have another transformer connected at the primary exactly as above but with two 12VAC secondary windings, completely isolated from each other.  So as not to confuse this with the above the first winding leads will be labeled X3 and X4; the other will be X5 and X6.
If you connect your 5 watt 12V lamp to X3 and X4 it will light; circuit completed.  If you then connect the 10 watt lamp to X5 and X6 it too will light; a completely separate circuit.  With both lamps on it you have 4 wires to the transformer secondary 
With the above 4 wire scenario in mind, if you connect X4 to X5 there will be 24VAC between X3 and X6; and the X4&X5 connection is the center tap and behaves EXACTLY the same as the single winding with the center tap mentioned above.  Only three wires are needed for your two lamps and the center tap only carries the imbalanced current.

On a side note, not relevant to this discussion, if you connect X3&X5 as well as X4&X6 both making two parallel windings (they have to be wound the same direction or phasing if you prefer) you will get 12VAC at double the current capacity. 

Where are these loads coming from?    Some place in the transformer, or in the house?   Why would there ever be an excess of current?   A 60 watt bulb will always draw 60 watt, even if you have it hooked to a 200-amp breaker.   So how could you ever have an excess of current?   Even if you were to directly wire hot to neutral or hot to ground, if neutral and ground are at 0V, wouldn't there be an instant volt drop where they connected and a lot of heat generated?

The loads mentioned are hypothetical.  Every time you turn on a light or 120 volt appliance the situation changes.  Think of the transformer as the source of your power however you want to think about it; volts, amps, watts etc. If you have a 60 watt bulb connected from L1 to neutral and you turn on your 1200 watt 120V toaster connected from L2 to neutral, there will be an imbalance.  The 200 amp breaker doesn’t care unless it sees an overload on one or both legs L1 and/or L2.  The breaker doesn’t see the neutral, nor does it have to as the neutral amps can never be greater than either L1 or L2 in normal circumstances. 
If you directly wire L1 or L2 to neutral or to a properly connected ground it will go boom.  If you can stand them, Electroboom’s YouTube videos are happy to demonstrate this so you don’t have to.
And to answer your question, there would be heat generated but it would be governed by the available voltage and current as well as the time before something (like the breaker or wire vaporizing) breaks the circuit.  If you turn off the breaker and connect line to neutral, the breaker should pop back off before any real damage is done and you won’t experience any real heating.
You would be better not thinking of ground and neutral as zero volts but as a reference point.  Zero volts is like absolute zero temperature, unattainable in real life but a good reference point.  Calling it ground or neutral is a lot more specific, purposeful and doesn’t mess with your head as much.

I had an old camaro and the ignition went bad.   I had to hot wire it.   I'd use the fuse panel.   I'd hook to a 12V source, then use a wire to tie into the fuse for the Ignition.   Then I'd hook into the 12V source and tie into the starter, but just long enough to get running.   One day, I was a little careless and grounded that 12V source.   The wire evaporated in my hand and left a white indent.   At first it didn't hurt, then it hurt real bad.   But there wasn't no excess current that just traveled safely back to the cells in the battery.   Wouldn't an excess current be called a short?

I guess you found out the hard way why it is called “hot wiring” a vehicle.  FYI a “short” is basically any unwanted return to source.  In your case the wire you were holding made a circuit from the +12VDC source (basically the battery or unfused path to the battery + terminal) to the engine block or chassis which are very well bonded to the – terminal of the battery.  A good car battery is easily capable of producing over 1000 amps at about 10 volts for a short time.  That is 10 kilowatts available but as your wire that is the weak part of the circuit heats up, its resistance goes up significantly which probably saved you from being burned a lot worse and / or welding something you didn’t want welded.  In the Camaro case the current wasn’t excess, it was just available from the source.  Your wire was the same as a filament in a headlamp as far as the battery was concerned; it was connected between the battery poles and the current took the path of least resistance in the newly made circuit.  The wire you used did have resistance, all wire does.  If you had used a really big wire, you could have blown up the battery. 

Also, how come our neutrals and grounds in the house are tied together?   Why can we not have it like the transformers on the poles where neutral is seperate from ground?   Actually, I probably could, couldn't I?   Outside, there's a grounding wire that's buried in the earth that hooks to the panel.   If I separated the ground and neutral buss-bars in the breaker panel, and made sure my ground buss-bar went to the grounding wire outside, and the neutral went back to the pole, I'd essentially have the same setup as the pole.   Someone in this thread mentioned in their house they'd never have the ground tied to the neutral.   But in every home I've seen, this is how it's always setup.   Is there any dangers to doing what I just described?

Neutrals and grounds are tied together for safety.  I think we have already discussed this.  In many cases water lines go through the ground, houses are built on slabs of concrete that when damp are pretty good conductors in terms of life safety.  If the neutral were not grounded, and wire does have some resistance as do wiring connections, there can be potential on the neutral conductor.  Grounding the neutral mitigates the possible severity of a person coming into contact with a neutral while grounded. 
The transformer on the pole for residential services in the USA almost always has the secondary neutral bonded to earth ground as near to the transformer as possible.  It is usually a fairly small (#8) bare wire that goes down the pole to the bottom.  The bare wire on the 3 wire drop from the pole to your house is the neutral conductor.
WHY on earth would you want to separate the ground form neutral in your house?  It would be a clear violation of the code.  It could put you and your family at a greater risk of shock.  The National Electric Code is revised every 3 years, is based on a lot of history of mishaps and contains the best reasonable ways to mitigate problems.  Look up ebay item 401538337928 it is a 2008 copy of the official NEC handbook for $18.  Just buy it, read through it and keep it for future reference.  The hardcover handbook has a lot of the code explained in real person English
Grounding and specifically proper grounding is a huge field in itself; there are power engineers that have made a career out of nothing but grounding.

[JustMeHere]

There are 240 sockets in the US with a wider prong.  They are for running large window AC units.   

I'd just splice together an adapter wire with the correct plug on each end.

[Spork Schivago]

Ah shit.   We might actually have to go for the transformer still.   The P9S16A's are actually cheaper than the P9S13A's.   We would actually be spending more money purchasing just one, let alone two.   Cost effectively, I think it'd be a hell of a lot cheaper to just wire up the transformer.

[james_s]

I do like the UK term "earth" for referring to what we typically refer to as "ground".

You could ignore the fact that it's AC all together and think of this in terms of DC. If you take a pair of D batteries and connect them end to end you'll have 3V from one end to another. Now pretend that they're concealed and you can't see the + and - markings, that's fine because that information is irrelevant since your hypothetical load doesn't care about polarity. Now take a wire connected to the joint in the middle between the two batteries and attach that to a stake in the ground, now you can call that point ground or earth of you like. Now between the two remaining wires you'll still have 3V, but if you measure between either one and ground you'll see 1.5V. If you connect either one to ground you'll have a short. In DC you'd call it +1.5V and -1.5V but what really matters is the relative potential between two points and since the polarity is constantly changing with AC it's simpler to refer to it as phase.

I don't know how it can get much simpler than this.

[james_s]

Ah shit.   We might actually have to go for the transformer still.   The P9S16A's are actually cheaper than the P9S13A's.   We would actually be spending more money purchasing just one, let alone two.   Cost effectively, I think it'd be a hell of a lot cheaper to just wire up the transformer.

Remember that a transformer that size will be expensive and it will burn up a non-trivial amount of power just sitting there.

I'm not familiar with the specific PDU's you're dealing with but I would be tempted to get a bunch of rack mount power strips and plug them into a row of standard 120V receptacles. You could easily install several dedicated 20A 120V circuits.

[IanB]

Ah shit.   We might actually have to go for the transformer still.   The P9S16A's are actually cheaper than the P9S13A's.   We would actually be spending more money purchasing just one, let alone two.   Cost effectively, I think it'd be a hell of a lot cheaper to just wire up the transformer.

Remember that the transformer suggested earlier by Richard Crowley is 7.5 kVA. If that were to be fully loaded that would draw 7500 VA / 240 V = 31.25 A.

Multiply that by 1.25 to get a continuous load rating (your server farm is a continuous load) and you get 39 A.

So you would need to provide a 40 amp circuit using the recommended 8-gauge wire to a 40 amp breaker in the panel.

You would not be able to run the transformer off an L6-30R receptacle, you would have to wire it in permanently.

Then you would have to wire up the secondary side of the transformer to your PDU. You probably don't want to just put a fly-lead on the secondary with an IEC receptacle on it, so you will need to source an appropriate commercial use breaker panel that supports 240 V line to neutral with single pole breakers, and then install wall mounted IEC receptacle(s) to plug your PDU into. And then you have to get it to pass inspection.

If you cost all that out, I think it might be cheaper just to buy the two new PDUs.

[james_s]

Especially when you can get used PDUs for peanuts on ebay.

[Spork Schivago]

We gotta back up the train here a bit.   I understand the purpose of the neutral and ground in the house.   If we have a short, we want that current and voltage to go to the path of least resistance, hopefully back to the breaker, to trip it

OK, let’s try to go back to some basics.
I assume you know how a transformer works; I suggest you grab some paper and draw the circuits out for yourself as you follow along; visual reinforcement is a great learning aid and along the way you might have the “Ah Ha” moment where this becomes clear. 
Vin / Vout = Turns in / Turns out and the current capacity is more or less determined by winding size
If you have a 120VAC primary transformer with a 24VAC secondary, you get 24VAC output
Suppose you put this transformer on a metal chassis like old radios etc. and you plug it into the wall with a standard grounded 3 pin plug so that L1 is fused and connected to either wire of the primary and the neutral is connected to the other lead of the primary with the ground connected to the chassis.  All is well with the circuit here; the primary is a completed circuit via the neutral and L1.   

If one of the primary windings anywhere along the coils faults to the chassis, the ground that previously carried zero current suddenly carries the fault current back to the house breaker panel where it completes the circuit via the ground to neutral tie I showed in the previous post and the fuse should blow.   

At this point of the discussion, the secondary winding is completely isolated from the primary.  Let’s call them X1 and X2.  You could connect either lead to the chassis (which is grounded) and nothing would happen; no current would flow.  The other lead of the secondary would then be 24VAC to the chassis, often called chassis ground in this case which is also referenced to earth ground via the 3 prong plug. 

Let’s assume neither X1 nor X2 secondary winding are connected to the chassis for this paragraph.  If you take two 12 volt lamps of the same wattage in series and connect them to X1 and X2, they will both light equally.  They have the same resistance.  If you connect the midpoint between these two lamps to the chassis, nothing bad will happen, it is still isolated and, current will not flow to the chassis.  Now, connect a 5 watt and 10 watt both 12volt lamps in series and connect them to X1 and X2.  The 5 watt lamp gets very bright in a hurry and goes poof; the 10 watt lamp barely glows until the circuit goes open with the failure of the 5 watt lamp; this happens the same if the midpoint is connected to the chassis because it is isolated.

Now let’s change things a bit and connect a wire to the center of the secondary windings, a true center tap I’ll call C. Nothing on the secondary is connected to the chassis at this point.
X1 to X2 is still 24VAC
X1 to C is 12VAC
X2 to C is 12VAC 
Now, connect a 5 watt and 10 watt both 12volt lamps in series and connect them to X1 and X2 BUT with the midpoint connected to C.
They both light with their respective brightness and wattage; the 5 watt one doesn’t go poof!  The unbalanced extra 5 watts used by the 10 watt lamp is carried back to the transformer center tap C.  The 5 watt lamp and half of the 10 watt lamp’s current are the same as two 5 watt lamps in series without the midpoint connected to the transformer; the imbalanced 5 watts for the 10 watt lamp returns back to the transformer via the center tap.

Now if we connect the center tap C to the chassis (which is referenced to earth ground as previously stated) there is no harm, no foul and no current flow between C and the chassis.
And when we repeat the experiment, the results are the same.  The wire from the midpoint of the lamp back to C still carries the 5 watts; chassis ground is just a reference.

The transformer at the pole though....you said, "For instance, if you had 30 amps load on the 120volt side from L1 to neutral and 20 amps load on the other side from L2 to neutral, the neutral conductor would carry the 10 amp difference back to the transformer."

Yes, that is true as in the lamp scenario above; but to carry it further:  It would be very helpful for you to draw this on paper.
Suppose you have another transformer connected at the primary exactly as above but with two 12VAC secondary windings, completely isolated from each other.  So as not to confuse this with the above the first winding leads will be labeled X3 and X4; the other will be X5 and X6.
If you connect your 5 watt 12V lamp to X3 and X4 it will light; circuit completed.  If you then connect the 10 watt lamp to X5 and X6 it too will light; a completely separate circuit.  With both lamps on it you have 4 wires to the transformer secondary 
With the above 4 wire scenario in mind, if you connect X4 to X5 there will be 24VAC between X3 and X6; and the X4&X5 connection is the center tap and behaves EXACTLY the same as the single winding with the center tap mentioned above.  Only three wires are needed for your two lamps and the center tap only carries the imbalanced current.

On a side note, not relevant to this discussion, if you connect X3&X5 as well as X4&X6 both making two parallel windings (they have to be wound the same direction or phasing if you prefer) you will get 12VAC at double the current capacity. 

Where are these loads coming from?    Some place in the transformer, or in the house?   Why would there ever be an excess of current?   A 60 watt bulb will always draw 60 watt, even if you have it hooked to a 200-amp breaker.   So how could you ever have an excess of current?   Even if you were to directly wire hot to neutral or hot to ground, if neutral and ground are at 0V, wouldn't there be an instant volt drop where they connected and a lot of heat generated?

The loads mentioned are hypothetical.  Every time you turn on a light or 120 volt appliance the situation changes.  Think of the transformer as the source of your power however you want to think about it; volts, amps, watts etc. If you have a 60 watt bulb connected from L1 to neutral and you turn on your 1200 watt 120V toaster connected from L2 to neutral, there will be an imbalance.  The 200 amp breaker doesn’t care unless it sees an overload on one or both legs L1 and/or L2.  The breaker doesn’t see the neutral, nor does it have to as the neutral amps can never be greater than either L1 or L2 in normal circumstances. 
If you directly wire L1 or L2 to neutral or to a properly connected ground it will go boom.  If you can stand them, Electroboom’s YouTube videos are happy to demonstrate this so you don’t have to.
And to answer your question, there would be heat generated but it would be governed by the available voltage and current as well as the time before something (like the breaker or wire vaporizing) breaks the circuit.  If you turn off the breaker and connect line to neutral, the breaker should pop back off before any real damage is done and you won’t experience any real heating.
You would be better not thinking of ground and neutral as zero volts but as a reference point.  Zero volts is like absolute zero temperature, unattainable in real life but a good reference point.  Calling it ground or neutral is a lot more specific, purposeful and doesn’t mess with your head as much.

I had an old camaro and the ignition went bad.   I had to hot wire it.   I'd use the fuse panel.   I'd hook to a 12V source, then use a wire to tie into the fuse for the Ignition.   Then I'd hook into the 12V source and tie into the starter, but just long enough to get running.   One day, I was a little careless and grounded that 12V source.   The wire evaporated in my hand and left a white indent.   At first it didn't hurt, then it hurt real bad.   But there wasn't no excess current that just traveled safely back to the cells in the battery.   Wouldn't an excess current be called a short?

I guess you found out the hard way why it is called “hot wiring” a vehicle.  FYI a “short” is basically any unwanted return to source.  In your case the wire you were holding made a circuit from the +12VDC source (basically the battery or unfused path to the battery + terminal) to the engine block or chassis which are very well bonded to the – terminal of the battery.  A good car battery is easily capable of producing over 1000 amps at about 10 volts for a short time.  That is 10 kilowatts available but as your wire that is the weak part of the circuit heats up, its resistance goes up significantly which probably saved you from being burned a lot worse and / or welding something you didn’t want welded.  In the Camaro case the current wasn’t excess, it was just available from the source.  Your wire was the same as a filament in a headlamp as far as the battery was concerned; it was connected between the battery poles and the current took the path of least resistance in the newly made circuit.  The wire you used did have resistance, all wire does.  If you had used a really big wire, you could have blown up the battery. 

Also, how come our neutrals and grounds in the house are tied together?   Why can we not have it like the transformers on the poles where neutral is seperate from ground?   Actually, I probably could, couldn't I?   Outside, there's a grounding wire that's buried in the earth that hooks to the panel.   If I separated the ground and neutral buss-bars in the breaker panel, and made sure my ground buss-bar went to the grounding wire outside, and the neutral went back to the pole, I'd essentially have the same setup as the pole.   Someone in this thread mentioned in their house they'd never have the ground tied to the neutral.   But in every home I've seen, this is how it's always setup.   Is there any dangers to doing what I just described?

Neutrals and grounds are tied together for safety.  I think we have already discussed this.  In many cases water lines go through the ground, houses are built on slabs of concrete that when damp are pretty good conductors in terms of life safety.  If the neutral were not grounded, and wire does have some resistance as do wiring connections, there can be potential on the neutral conductor.  Grounding the neutral mitigates the possible severity of a person coming into contact with a neutral while grounded. 
The transformer on the pole for residential services in the USA almost always has the secondary neutral bonded to earth ground as near to the transformer as possible.  It is usually a fairly small (#8) bare wire that goes down the pole to the bottom.  The bare wire on the 3 wire drop from the pole to your house is the neutral conductor.
WHY on earth would you want to separate the ground form neutral in your house?  It would be a clear violation of the code.  It could put you and your family at a greater risk of shock.  The National Electric Code is revised every 3 years, is based on a lot of history of mishaps and contains the best reasonable ways to mitigate problems.  Look up ebay item 401538337928 it is a 2008 copy of the official NEC handbook for $18.  Just buy it, read through it and keep it for future reference.  The hardcover handbook has a lot of the code explained in real person English
Grounding and specifically proper grounding is a huge field in itself; there are power engineers that have made a career out of nothing but grounding.

Thank you for your explanation but in all honesty, I was so tired from lack of sleep last night, I don't even remember writing the post.   I had some stuff to finish and I think when I called it a night, it was over 36 hours of me being awake, it was around 0300 but after getting into bed, I wasn't able to actually fall asleep until 0600.   For the Official NEC, you don't need to waste your money on a copy of the book.   You can do what I do, just go to https://www.nfpa.org/NEC    and click on Free online access to the NEC® and other electrical standards.   Then you're not getting the outdated standards, it also lists the changes from the previous edition to the latest (the latest is 2017 I believe), and best of all, it's totally free.   You just need to sign up for an account.   There are options to download a digital copy (for money) or purchase an actual physical copy (for money), but you don't need to do that to get access.  You can browse the entire 2017 NEC for free using that site.

When I question something (for example, what wire type do I need in the basement, can I drill holes through the first floor floor joists and run the wire or should I staple it, do I need to use the plastic staples, etc), I just head there, login, and check it up.   It's extremely handy.

After reading what I wrote, I think maybe what I was thinking was having an isolated ground from neutral.   I dunno why in earth I would want that.   I do have a need for an isolation transformer for a couple DUTs, so maybe that's where it came from?   I dunno.   Usually, I try to be in bed by 2100 ~ 2200, but sometimes, there's deadlines I have to meet, and I run behind so I have to stay up longer than I like.    I know I could have slept a lot longer today but I had my wife wake me up at 1100 so I could get some work done.

Anyway, thank you, and everyone else, for helping me understand everything properly now!   I've sent the email to my Account Executive to see where we stand, but because of the price tag actually being more for the North American version (with providing almost half of the version I currently have), I might just still buy the transformer.

As IanB said though, that might throw off the inspector.   So I'll first see what my Account Executive says and on Monday, while I'm waiting for a response, I'll call the local code enforcer.   I got his cell number but I don't want to disturb him on the weekend.   We're not buddy buddy kinda thing, you know?   I'll see if he'd recommend against the 240:240 transformer.

Little bit a sleep and it suddenly makes perfect sense how to wire it up and everything.   I don't know why I couldn't grasp it the other day.   Thanks though.

[Spork Schivago]

I do like the UK term "earth" for referring to what we typically refer to as "ground".

You could ignore the fact that it's AC all together and think of this in terms of DC. If you take a pair of D batteries and connect them end to end you'll have 3V from one end to another. Now pretend that they're concealed and you can't see the + and - markings, that's fine because that information is irrelevant since your hypothetical load doesn't care about polarity. Now take a wire connected to the joint in the middle between the two batteries and attach that to a stake in the ground, now you can call that point ground or earth of you like. Now between the two remaining wires you'll still have 3V, but if you measure between either one and ground you'll see 1.5V. If you connect either one to ground you'll have a short. In DC you'd call it +1.5V and -1.5V but what really matters is the relative potential between two points and since the polarity is constantly changing with AC it's simpler to refer to it as phase.

I don't know how it can get much simpler than this.

Yeah, I don't know how it could get simpler either or why your posting that.   I definitely understand it now.   Last night, I just didn't have any sleep, so maybe that was why I had a hard time grasping it?   I know my wife has mentioned some problems to the docs about my memory.   She goes to the neurologist with me and a few other appointments.   She says there's times that people that I know real well seem to disappear from my memory.   It's just not people, but time as well.   Sometimes, I guess I lose a year or two or can't remember my birthday.   But what's odd, I can't ever remember not being able to remember stuff.   So, it's like some memories are isolated from others and when I cannot remember my best friend, for example, I might remember something that I can't remember now, but when I can remember my best friend, that something is gone and cannot be reached!

They're having me do these exercises, brain ones, that's supposed to help my brain find new pathways.   In a way, the brain is much like the circuit boards we work on I guess, and what do you do when you have a burned out trace?   You can run a jumper wire or essentially, find another way for the electrons to flow.   I guess the brain works similarly.   If someone has a stroke, they have to sometimes "reteach" their mind how do things that it used to know, like walk.   My issue isn't so much with walking.   The damage I sustained affects the prefrontal cortex and for some reason, there are holes there, or legions.   What is bad, they're spreading and I have to go to some two week hospital thing where they monitor me and try to stress me out to try and figure out why they're spreading.    I have the global volume loss of T2 white matter (something along those lines), where she says I'm losing brain function.   Anyway, exercising the brain is supposed to help and my wife is supposed to "quiz" me every day.   What did you eat for breakfast?  What day is it?   What TV show were we just watching?   What day is it?   What did you eat for breakfast?   What did you eat for lunch?   Stuff like that.

[Spork Schivago]

Ah shit.   We might actually have to go for the transformer still.   The P9S16A's are actually cheaper than the P9S13A's.   We would actually be spending more money purchasing just one, let alone two.   Cost effectively, I think it'd be a hell of a lot cheaper to just wire up the transformer.

Remember that a transformer that size will be expensive and it will burn up a non-trivial amount of power just sitting there.

I'm not familiar with the specific PDU's you're dealing with but I would be tempted to get a bunch of rack mount power strips and plug them into a row of standard 120V receptacles. You could easily install several dedicated 20A 120V circuits.

Transformer is actually cheaper than the PDU I believe, but I haven't gotten the quote, but you do bring up a very important part that I keep on thinking about.   How efficient will this transformer be?   Someone I think referred to it as a hot piece of metal hanging on the wall.   To me, that means it won't be very efficient at all.   In the end, maybe the separate North American P9S13A's would be financially cheaper than the transformers power loss that shows up on the electric bill...In the long run, the P9S13A's are probably the less expensive way to go.   And the best part, I'd get the redundancy which is something we need eventually.   If I can get that now, that's one less thing to worry about later on.

[IanB]

So you would need to provide a 40 amp circuit using the recommended 8-gauge wire to a 40 amp breaker in the panel.

And you would be running your whole system off one 40 amp circuit, with no redundancy. If that breaker tripped for any reason then everything would shut down.

[Spork Schivago]

Ah shit.   We might actually have to go for the transformer still.   The P9S16A's are actually cheaper than the P9S13A's.   We would actually be spending more money purchasing just one, let alone two.   Cost effectively, I think it'd be a hell of a lot cheaper to just wire up the transformer.

Remember that the transformer suggested earlier by Richard Crowley is 7.5 kVA. If that were to be fully loaded that would draw 7500 VA / 240 V = 31.25 A.

Multiply that by 1.25 to get a continuous load rating (your server farm is a continuous load) and you get 39 A.

So you would need to provide a 40 amp circuit using the recommended 8-gauge wire to a 40 amp breaker in the panel.

You would not be able to run the transformer off an L6-30R receptacle, you would have to wire it in permanently.

Then you would have to wire up the secondary side of the transformer to your PDU. You probably don't want to just put a fly-lead on the secondary with an IEC receptacle on it, so you will need to source an appropriate commercial use breaker panel that supports 240 V line to neutral with single pole breakers, and then install wall mounted IEC receptacle(s) to plug your PDU into. And then you have to get it to pass inspection.

If you cost all that out, I think it might be cheaper just to buy the two new PDUs.

Yes, I've actually already thought about all of this and talked to the wife and said if we can send it back, I think we're going to send it back.   I told her we'd need the second panel, the wires for it, the breakers for it, we'd still need the receptacle, we'd still have to pay for inspection (which very well could fail, and then what?  Stuck with the PDU and still gotta buy two of the smaller units!)   Hopefully, they'll accept the return based on that's what HPE said I _needed_ and couldn't power it off my current panel, although, I was stupid, and didn't word it properly like I should have.   I called it 120-0-120 split-phase, which I shouldn't have done.   That's how I've been referring to it the whole time, even when they did the on-line power analysis.    That's my fault.   They probably didn't even know what the hell it was and just said in their minds, say what?   Hell no!

[Gregg]

Now that you say you understand the concept of 120/240 volt US power I am wondering why you want that expensive PDU.  In your basement, you hardly need remote switching and the PDU you chose doesn't seem to be able to meter the power.
If I were wiring 7.5 KW worth of servers in my basement, I would put a sub panel near them and run separate circuits to each major server and maybe separate breakers to each of the power supplies in those that have dual power supplies.  There is a lot of scalability and some redundancy in such a system and it costs a lot less in the long run.  You can then put very inexpensive power strips in the equipment racks and easily change everything as your service grows.  I would also wire in a wrap around bypass panel ahead of the panel to later install a UPS without interrupting any of the servers. 

[vk6zgo]

Now that you say you understand the concept of 120/240 volt US power I am wondering why you want that expensive PDU.  In your basement, you hardly need remote switching and the PDU you chose doesn't seem to be able to meter the power.
If I were wiring 7.5 KW worth of servers in my basement, I would put a sub panel near them and run separate circuits to each major server and maybe separate breakers to each of the power supplies in those that have dual power supplies.  There is a lot of scalability and some redundancy in such a system and it costs a lot less in the long run.  You can then put very inexpensive power strips in the equipment racks and easily change everything as your service grows.  I would also wire in a wrap around bypass panel ahead of the panel to later install a UPS without interrupting any of the servers.

That was my thought,too.

Most of the places I've worked, the power distribution was either designed  & built " in house", or was supplied as part of the equipment we were using, with the requirements specified when the EEs were ordering the system.
In both of the above methods, the power distribution setup could be "tailor made".

In the OP's case, it should be easy to customise the setup with discrete bits.

[IanB]

Now that you say you understand the concept of 120/240 volt US power I am wondering why you want that expensive PDU.  In your basement, you hardly need remote switching and the PDU you chose doesn't seem to be able to meter the power.
If I were wiring 7.5 KW worth of servers in my basement, I would put a sub panel near them and run separate circuits to each major server and maybe separate breakers to each of the power supplies in those that have dual power supplies.  There is a lot of scalability and some redundancy in such a system and it costs a lot less in the long run.  You can then put very inexpensive power strips in the equipment racks and easily change everything as your service grows.  I would also wire in a wrap around bypass panel ahead of the panel to later install a UPS without interrupting any of the servers.

Exactly. All you need is two or three dedicated 30 amp circuits feeding L6-30R receptacles, and then a power strip like one of these for each receptacle to feed your devices with dedicated C13/C14 power cords. You don't need an expensive PDU with features you are not using.

http://www.apc.com/shop/us/en/products/SADE-89TKZQ/P-AP7541

In each server you can put redundant power supplies fed from different circuits.

[Spork Schivago]

Now that you say you understand the concept of 120/240 volt US power I am wondering why you want that expensive PDU.  In your basement, you hardly need remote switching and the PDU you chose doesn't seem to be able to meter the power.
If I were wiring 7.5 KW worth of servers in my basement, I would put a sub panel near them and run separate circuits to each major server and maybe separate breakers to each of the power supplies in those that have dual power supplies.  There is a lot of scalability and some redundancy in such a system and it costs a lot less in the long run.  You can then put very inexpensive power strips in the equipment racks and easily change everything as your service grows.  I would also wire in a wrap around bypass panel ahead of the panel to later install a UPS without interrupting any of the servers.

There are reasons we need this PDU that I'd rather not go into.   Originally, we wanted a switched AND metered one, but HPE said they didn't make one that would work with our power supply.

I'm wondering though, now that we know they make mistakes, do any of you guys see a Switched AND metered horizontal 2U PDU that will work with my 120-0-120 split phase?

I feel comfortable saying it's not just power that we need the PDU to provide.   It's the features that come with it that are important.   And once we had this wired up, we were going to purchase another for redundancy.   Knowing now that we can definitely power those PSUs with our electricity, if we can send this back, we're definitely sending it back and at least purchasing two PDUs.   Hopefully we can split the load equally on both, but I'm really hoping just one can power the entire load.   That way, if one dies, we're notified and it automatically provides power to the broken PDU (depending on how it breaks) so our servers don't go down.

Money is always something to be conscious of, so even if we cannot send the PDU back, I think we're going to look into voiding the warranty and modifying it to work safely with our electricity, replace the European style plug with a nice NEMA L6-30P.   Someone else suggested that early on (at least the plug part).   We just have to wait to see what they say tomorrow.

Can you go a bit more in detail about the wrap-around bypass panel?    Eventually, the entire business will be on a seperate line, totally isolated from the residential stuff.   It'll have a dedicated panel (not a sub-panel), and a dedicated meter.   When we purchase the whole home generator, it'll have an ATS (Automatic Transfer Switch) that will have breakers in it.   We'll run them to both panels (well, have the company do it).   We still need a UPS.   We're slowly getting there though.   At least now, I can power my server again and install the upgrades (assuming the one 1400 watt can handle the entire load).   And then I can work on configuring that (which means another monthly bill for the VDA license).   But we're getting there.

[Spork Schivago]

Now that you say you understand the concept of 120/240 volt US power I am wondering why you want that expensive PDU.  In your basement, you hardly need remote switching and the PDU you chose doesn't seem to be able to meter the power.
If I were wiring 7.5 KW worth of servers in my basement, I would put a sub panel near them and run separate circuits to each major server and maybe separate breakers to each of the power supplies in those that have dual power supplies.  There is a lot of scalability and some redundancy in such a system and it costs a lot less in the long run.  You can then put very inexpensive power strips in the equipment racks and easily change everything as your service grows.  I would also wire in a wrap around bypass panel ahead of the panel to later install a UPS without interrupting any of the servers.

Exactly. All you need is two or three dedicated 30 amp circuits feeding L6-30R receptacles, and then a power strip like one of these for each receptacle to feed your devices with dedicated C13/C14 power cords. You don't need an expensive PDU with features you are not using.

http://www.apc.com/shop/us/en/products/SADE-89TKZQ/P-AP7541

In each server you can put redundant power supplies fed from different circuits.

That PDU would not work at all for us.   We ARE using features of the advanced PDU that we have, hence the reason we purchased it.   But it'd be real nice if we could find an equivalent to the P9S13As that was still a G2 but supported the Switched AND Metered, like we originally wanted, but HPE said no way without three-phase (for horizontal).   We have a 36-unit rack, and actually prefer the horizontal units.    We will buy double of what we need for the load, just to daisy chain them and have the redundancy.   But there's still a lot more expensive stuff we need to work on as well.

We need the UPS, we need the whole home generator, we need the dedicated panel, the dedicated meter, the auto-loading / rotating rack mountable tape backup unit, tapes, still need to finish wiring up the house, still haven't received the proper ends for the shielded ethernet (I don't want to go in on why we need shielded, so please don't ask!), we still have to finish the contract (almost done), finished the other one which gave us the rest of the money, so we're good for a bit on cash, but it's not going to pay for everything, still need to setup the VMs, reenable the VDA license, etc, etc, etc.    Time and money is always something I'm aware of.

There are features these PDUs, like the one we have, we need.   Features that don't come with normal power strips or power distribution units.