240VAC + ISO Tranny + GROUND?? NEC vs. ??

[WBear2GCR]

Ok, I know we've batted this around before.
But I keep having trouble with it.

NEC sez run the ground back to the main box and ground.

Now, I'm missing something because that distance back to the main box in my case could be something on the order of 100ft.
That 100ft of run has a non-zero impedance and non-zero resistance.
So I am thinking that something like a lightning hit will generate a substantial amount of current along that so-called ground.

I think what we are talking about here is the "safety ground" not the neutral. (the 4th wire)

I just got some very large 240-240 Isolation Transformers (7kva CCS).
I keep mulling this over in my mind and it keeps seeming to me that it would be better to run a 3 wire 240vac line out 100 feet,
then drop a very good ground rod at that location and make my 120-0-120 + safety ground on the secondary of the isolation iron (it has the requisite taps).

I realize that in this case I'd be counting on the galvanic isolation (breakdown) voltage of the transformer to keep bad things from one side of the thing getting to the other... but that is the idea actually.

So, why is this a bad idea?

                                         _-_-bear

[WA1GFZ]

A M.E. buddy had a problem with a wave solder machine. Every lightning storm would blow out the program so he would have to reload the software. He added an isolation transformer. He still had the problem once in a while. He gave up and called me. I suggested he reestablish a chassis ground near the machine. He was on a ground floor so he just sent a rod through the floor. Result never failed again.
I bet NEC would want the rod bonded to the panel also. Not sure if he did that.

[W3SLK]

Whatever the AHJ says, goes!

[k4kyv]

The NEC is concerned only with safety grounds to prevent shock and electrocution, not with RFI, ground loop hum, poor RF grounds, etc.

If you have a good isolation transformer with excellent regulation, I would see nothing wrong or dangerous with isolating the  shack's a.c. power source from the mains, and using a local ground.  The neutral is not the same thing as the ground. Treat the neutral exactly as you treat the hot wires; it is a  current carrying conductor, and should never be strapped to  ground anywhere except at the entrance panel right at the meter. No need to strap the isolated neutral to ground at the transformer.

There probably is something in the NEC about isolation transformers.  Maybe I'll get out my most recent NEC as well as my 1940 edition and see what they say about it.

Furthermore, you can reduce ground loop and other hum pickup problems in the shack by running balanced 120 v.a.c. power.  It requires a midtapped 120 volt isolation transformer; unfortunately they seem to be made of unobtanium.  You could use a small 120/240-120/240 volt isolation xfmr, wired as for 240v, but  run the whole thing on 120 volts to give 60/60 volts, and connect the midtap to ground.  This way, you have two hots and a ground, but no neutral.  The maximum voltage with respect to ground is 60 volts instead of 120, which improves the safety issue.  Treat the balanced 120v balance line just as you would treat a balanced 600 ohm audio line.  Of course, this means the balanced 120v line is completely isolated from your 240 volt service. The main drawback to this is that you downgrade the transformer's power rating by 50%, since it will be running at half voltage at its maximum current rating, and there may be some degradation in voltage regulation with small transformers.

I have been told that the Navy uses or has used balanced 120 volts a.c. on ships for safety reasons.

[WA1GFZ]

Don,
Navy doesn't like hull currents.

[w3jn]

I would suspect that if you install something non-standard and not in accordance with the NEC, your homeowners insurance will be null and void in the case of an electrical fire.

It doesn't appear you can do this (have two separate grounds) per the NEC.

http://forums.mikeholt.com/archive/index.php/t-122054.html

From a discussion on UPS transfer switches:  http://ecmweb.com/mag/electric_ups_stands_firm/

Grounding systems are vital. Using multiple power sources imposes special requirements on grounding design. First, you must determine if sources are separately derived or service supplied. The NEC defines a separately derived system as

"A premises wiring system whose power is derived from a battery, a solar photovoltaic system, or from a generator, transformer, or converter windings, and that has no direct electrical connection, including a solidly connected grounded circuit conductor, to supply conductors originating in another system."

On the other hand, it defines a service point as the point of connection between the facilities of the serving utility and the premises wiring.

The NEC requires bonding of the neutral to ground at a service point and where a separately derived system exists. The two one-line diagrams (Figs. 3a and 3b of the original article), involve identical equipment, except the transformer in Fig. 3a is utility-owned, and the one in Fig. 3b is user-owned. The grounding requirements of each vary based on who owns the transformer. A utility-owned transformer requires classification of the power entrance into the building as a service point. As shown in Fig. 3a, the required grounding locations of the neutral are at the transformer's secondary windings and at the service equipment. The user-owned transformer in Fig. 3b requires bonding of the neutral to ground at a single location, either at the transformer's secondary windings or at the equipment that provides disconnecting means and overcurrent protection.

[k4kyv]

I would suspect that if you install something non-standard and not in accordance with the NEC, your homeowners insurance will be null and void in the case of an electrical fire.

It doesn't appear you can do this (have two separate grounds) per the NEC.

http://forums.mikeholt.com/archive/index.php/t-122054.html

From a discussion on UPS transfer switches:  http://ecmweb.com/mag/electric_ups_stands_firm/

Grounding systems are vital. Using multiple power sources imposes special requirements on grounding design. First, you must determine if sources are separately derived or service supplied. The NEC defines a separately derived system as

"A premises wiring system whose power is derived from a battery, a solar photovoltaic system, or from a generator, transformer, or converter windings, and that has no direct electrical connection, including a solidly connected grounded circuit conductor, to supply conductors originating in another system."

On the other hand, it defines a service point as the point of connection between the facilities of the serving utility and the premises wiring.

The NEC requires bonding of the neutral to ground at a service point and where a separately derived system exists. The two one-line diagrams (Figs. 3a and 3b of the original article), involve identical equipment, except the transformer in Fig. 3a is utility-owned, and the one in Fig. 3b is user-owned. The grounding requirements of each vary based on who owns the transformer. A utility-owned transformer requires classification of the power entrance into the building as a service point. As shown in Fig. 3a, the required grounding locations of the neutral are at the transformer's secondary windings and at the service equipment. The user-owned transformer in Fig. 3b requires bonding of the neutral to ground at a single location, either at the transformer's secondary windings or at the equipment that provides disconnecting means and overcurrent protection.

One of the commenters offers a loophole.  The code prescribes what is hard wired into the house wiring, but has no jurisdiction over what is plugged into a power outlet.  Simply use a heavy duty mains outlet like what is used for an electric kitchen range.  Connect the isolation transformer via a flexible pigtail cord like heavy duty appliances use.

[AB2EZ]

It would be helpful if the NEC included an appendix which described the various "rainy day scenarios" that it is designed to protect against. [Maybe there is such a thing]

If the main building is hit by lightning that flows to ground through the electrical wiring in the main building... then all of the electrical conductors in the main building (whether hot or neutral or ground) will rise to a high potential v. the ground in the remote building (assuming the remote building ground is not bonded to the main building ground). I.e. potential is always defined as between two points... and all of the wires in the main building are tied together by various connected loads, having moderate resistance values.

If the voltage difference between the main building ground and the remote building ground is    N x 1000 volts, then the isolation transformer will have N x 1000 volts between its windings.

If the windings arc over...

Stu

[N4LTA]

Actually you can wire the transformer as a separately derived system.

See the rules in the code for grounding a separately derived system. You do not continue the grounding conductor past the transformer. You derive a separate ground. You must have a transformer secondary over-current device. There are some exceptions if the transformer is less than 1 KVA.


Pat
N4LTA

[k4kyv]

I use a separate sub-panel in my shack.  The box in the shack is grounded but it is not bonded to neutral at this point.  I just treat the neutral as I would in any other 240v outlet somewhere else in the house.  There is a measurable a.c. voltage between neutral and ground at the shack sub-panel, indicating an unbalance in the loads on the two 120v legs into the shack.  This is normal since each leg has a different load.  I shorted the neutral to ground just to see  what would happen, and I could see tiny but visible sparks, and with the neutral shorted to ground, a.c. hum suddenly appeared in the audio of almost every piece of equipment in the shack that had been hum-free before.

There is a solid # 6 copper wire bonding the shack ground to the main house ground.  I  run it separately from the power line, which has only 3 conductors, outside the conduit, buried like an antenna ground radial.  The midtap on my balanced isolation transformer is connected to this ground. These grounds are also bonded, through a coax shield, to the radial system at the tower.

[w3jn]

I had a great book that explained the NEC and its intent, and provided discussion and examples of the various items.  Unfortunately I left it in my office when I left my post in Havana.  The discussion on grounding was probably 1/4 of the book.

I don't think the insurance company would consider a 7 KVA transformer a "plug in appliance".  I don't see what the benefit might possibly be of installing it outside the NEC anyway.  Nor do I see a benefit in an additional transformer. 

Bear, run your power normally.  Install your station ground as close to the station as possible.   Then bond your station ground to the service ground per the NEC.  Done.  Fewer failure points.  No problems with your insurance company.  No power loss in an uneeded transformer.

[WBear2GCR]

Johnny,

My problem is with the large inductance and Idrop between the local station ground and the run back to the main panel.
That just seems wrong to me. Very very large currents could flow between the two, as far as I can see...

Pat, I like the "separately derived system" term - I'll have to look into it some more. Is that part of the NEC?

Don, seems to me that maybe the hum that suddenly appeared is an example of just that "ground difference" and the Idrop between the local coax (RF/station) ground and the two AC "grounds" you have running back to the house? That's my thinking on this, and why I want one single local very good ground on the end of a local isolation transformer.

As far a plugging stuff in, as long as you have the requisite plug and socket combo (Hubbel Twistlocs anyone?) + breaker  ur golden, afaik.

                   _-_-bear

[k4kyv]

Bear, the difference in potential between the ground and neutral was because I was pulling more current through one 120v leg than I was the other.  That means the neutral wire was carrying current.  Current through a wire causes voltage drop. So unless both 120v legs are carrying equal current or you are pulling a 240v load, there will be a difference in potential at the outlet between neutral and ground.  

That is why, per NEC, the only place where neutral and ground are strapped together should be at the entrance to the building, near the meter, and everywhere else the neutral should  be treated with the same  respect as the hot.  When I shorted neutral to ground at the shack, the ground lead was sharing some of the unbalanced return current, which it is not supposed to do.  That means there is a voltage gradient all along the ground wire.  Unless every single thing in the building is grounded to exactly the same point, this gradient will result in a potential difference between every piece of equipment.  Any interconnecting cables will also carry a current due to this potential difference, and  thus induce hum into the signal line.

The inconsistency in the NEC is that they say the breaker box in the shack is a sub-panel and should be treated just like an entrance panel, and that a safety ground should be strapped to neutral where it enters the building. This ground should go to a ground rod, which should be bonded via a ground wire to the main ground rod.  Instead, I just treat it as I would have if the shack were simply another room in the house and not a separate building, and keep the neutral separate from ground.  I don't see any conceivable way that having the wooden structure of the shack physically detached from the house could have anything to do with the safety of the electrical system. I do have the ground rod at the shack that is bonded to the main ground rod, but I don't strap neutral to it at the service entrance to the shack.

The reason I use an isolation transformer for the 120v equipment in my station is to get rid of the induced a.c. hum caused by the unbalanced a.c. line that has the neutral grounded at the main entrance panel.  I treat the 120 volt power line exactly as I would treat a balanced audio line.  The unbalanced a.c. line with "hot" and a neutral very close, but not exactly, at ground potential, creates hum in exactly the same manner as grounding one side of a filament line to chassis does in an an audio amplifier. If I strapped the "neutral" from the  secondary of my isolation transformer to the house neutral, I would  no longer have a balanced system, and the isolation transformer would serve no useful purpose. Instead, the ground wire is connected to the midtap of the secondary of my isolation transformer. There is no neutral, just two 60v "hots". I like the idea someone mentioned of using a variac across the secondary of an untapped isolation transformer, and adjusting it to null out the hum.

Ideally, all household appliances would run off 240v and there would be no such thing as a 120v outlet. What we call "neutral" would be the safety ground, and would never carry load current.  That way the electrical system would be inherently balanced and 90% of  this ground loop crap would not exist. But our present 120-120/240v system is like the QWERTY keyboard; it was establish in the very early days of electrical distribution and we are still using existing equipment going back more than a century. It is so embedded that we can't get rid of it and instead have to work around it.

Most of the NEC rules are based on sound engineering and common sense, but their one-size-fits-all approach sometimes leads to inconsistencies and contradictions.

[N4LTA]

Yes,

The NEC allows that a generator or transformer have it's own grounding system. It is called a separately derived system.

I can't remember chapter and verse - but it may be article 230.3 - in the grounding chapter - i am at home and not near a code book.

You will need to have essentially a new service ground with a grounding electrode, grounding electrode conductor, and over-current protected disconnecting means. That allows you to have a ground close by (as good of one as needed) and not to have to rely on the service ground.

I am a distribution engineer and I do this for a living and will make up a little drawing for you if you need it.  just would need the actual diagram of the transformer.

The best NEC code book is the one with the code and an explanation of most articles incorporated together. It is called the NEC Code Handbook. Beware of any book not published by the NFPA as interpretations in these are just opinions and are often dated.

Most jurisdictions are at least one and sometimes two editions behind the most recent code.

Pat
N4LTA

[WBear2GCR]

Pat,

The transformer is made by Signal Transformer. It's in their standard catalog. It's not an autoformer. The primary has taps for voltage adjust but it is basically 0-120 & 0-120 primary and the same for the secondary (of course rather than just one single 120 tap there are various taps). I'd be happy to have a look at a diagram in case there is something I don't know about or miss!

Don, the "balanced" AC is very popular among some audiophiles these days. I wonder about it though... do you find that it reduces hum quite a lot?

                            _-_-bear

[KB2WIG]

FHIW,

The Code is avaiable at

http://www.nfpa.org/aboutthecodes/AboutTheCodes.asp?DocNum=70&EditionID=238&cookie%5Ftest=1

and ... Its FREE.....


klc   

[k4kyv]

Don, the "balanced" AC is very popular among some audiophiles these days. I wonder about it though... do you find that it reduces hum quite a lot? 

It seems to noticeably reduce it in some pieces of equipment, and I can see a difference on the scope on my xmtr, but nothing drastic.  In my receiver it changes the tonal quality of the hum but doesn't actually reduce the level. I think it is significant that it changes anything at all.  But I badly need to re-do my grounding, since stuff has been added peacemeal over the years, and I have a few daisy-chain grounds.

[w3jn]

I stand by my comments.

THe voltage drop in 100 feet of #8 to bond your service entrance and station grounds together pale in comparison to the impedance of the earth.  You should *never* depend on the earth between two grounding points.  Because there is considerable (~ 10 ohms or more) resistance it follows if there's current flow between them there's gonna be a voltage potential between them.

Your station ground is therefore NOT at zero potential compared with the electrical service.  The leakage in your isolation transformer will create a potentially dangerous situation, and current will flow from your isolation transformer thru your equipment to your station ground and that's gonna cause you all sorts of problems.

You're trying to solve a non-existent problem which will cause more trouble.  Or so it seems to me.

[WBear2GCR]

I guess the "earth" ground is higher Z than a #8 copper wire of equivalent length...

But I am unclear on where this "leakage" voltage is coming from?
 
The whole idea of the isolation transformer is to create galvanic isolation between the two areas.
It would seem to me that the only ground that needs to work is the one on the secondary of the iso tranny, if you are on the secondary side of the affair. There is no clear need that I can see for the ground potentials on both sides of the iso tranny be identical.

I don't see how this ground is any better or worse than the ground rod dropped next to the incoming service?
We're not relying upon the "ground wire" on the pole to conduct all the way back to the generating station or substation, yes?

Of course I may be missing something from this equation...

                      _-_-bear

[WA1GFZ]

Just for the heck of it put a 100 watt bulb between Hot and neutral. Measure the voltage across the bulb. then put the bulb between hot and safety ground and measure the voltage across the bulb. Maybe your problem is a high resistance in the neutral leg so current is diverted through the safety ground lead.??
That is one nice transformer you have.

[w3jn]

First, I'm not sure what problem you're having that you need an isolation transformer to resolve it.  I can't think of a reason to have one that big.  A small one to work on hot chassis sets, etc. yeah...

The reason for the ground rod at the service entrance is to provide *one* ground reference for the structure.  If you have two that aren't bonded together, there's gonna be a voltage differential between them.

If everything's working right there probably would be no safety hazard.  However the NEC standards are intended to ensure fire and personnel safety in the case of a variety of faults.  Example - what happens if the primary of your transformer develops a fault to its chassis?  In your arrangement, the fault will not be cleared by tripping the circuit breaker on the primary, because you're re-deriving the ground at the secondary of the transformer and the fault will travel to your station ground, thru the earth, and up from your service entrance ground to the neutral bus of the service entrance.  Could cause a fire, or at least a very hazardous situation since all of your equipment is now hot with juice from your main service.

At work we used to have similar arrangements for communications equipment, the idea being to isolate noise from disrupting the circuit.  A separate ground was always provided, not bonded to the service entrance.  I can't tell you how many shocks I've gotten from touching equipment connected to the isolation transformer's secondary while also touching conduit, etc., in the room.

Finally someone recognized enough was enough and the separate ground strategy was abandoned.  The ground loop problems that guys used to tear their hair out trying to solve, and the shocks, all disappeared.

[WD5JKO]

I too have a unused ~ 7 kva tapped primary 230v:120v stepdown transformer. I live in a ranch style home with the shack over 100' from the power service entrance. The shack power is a really droopy GFCI circuit which is shared with kitchen and bathroom stuff. I really need a nice dedicated source of both 230v and 115v in the shack. The breaker box however is a long way away, and all positions are full.

So my idea after reading this thread is to plug that transformer into the kitchen electric stove outlet 50A 230v (3 wire) which is thru 1 wall from the shack. (I have a gas stove). Install a ~ 15amp 2 pole breaker local in a small distribution box, feed the transformer, and then I have a stiff 230v source, and after the transformer I have a very stiff 120v circuit with another 1 pole breaker set to 20 amps for the 120v service.

My shack ground currently consists of the copper water pipe, and some ground rods. These tie together at a ground point on my metal equipment rack. Of course there is also the power safety grounds in place from 3 prong 120v power cords. With my power distribution idea, at the transformer service feed I would bond the stove 230v power ground wire to to the 120v circuit (xfmr secondary)  ground and neutral, and to the 3 wire 230v ground. I would NOT make a separate ground as was discussed in this thread.

I was thinking about taking that stove 230v feed and doing a #8 trifiliar wound coil around a suitably large ferrite   rod(s), kind of like a filament choke. The idea is to preserve the safety ground, and yet keep the RF from backing up into the rest of the house, and instead from an RF perspective the ground is the copper water pipe, and the ground rods.

So, since I'm wondering out loud here in writing, what's wrong with my idea? How could I pull this off and still be NEC compliant? Adding a bigger whole home service box and two new LONG circuits is NOT going to happen. $$$$

Jim
WD5JKO

[N4LTA]

A couple of things that might be noted:

1. The power company does not supply some great all encompassing  ground. In fact they supply no grounding conductor at all. They supply a grounded conductor (miscalled a neutral). The grounding conductor is developed at the bonding point  - at the service

2. The quality of the service ground is determined by the service ground system - not the power company.

3. A 100 foot length of #8 is a very good low impedance ground for  60 Hz fault current. It is worthless as a high frequency ground - The NEC could care less about high frequency grounding. Any grounding conductor in a conductive conduit is worthless for high frequency. The metal conduit acts as a single turn transformer - just like a sleeve on a coax antenna feeder. You might note that "pole grounds" installed by the power company have a cover on the pole grounds - but it does not encircle the conductor.  If it did - the conductor would make the lightning arrestor installed on the transformer worthless.

4. Lightning is a very high frequency pulse that acts like a traveling wave. The pulse is easily reflected by impedance "bumps" and at these points, the voltage can double. That is why dead- ends on the power system have so many lighting problems.

5. The power system ground in most residences is poor - usually a 5/8" ground rod - possibly tied to a water pipe that may or may not be a good ground.



Pat
N4LTA

[k4kyv]

My stab at a solution was to bond the station ground to the entrance ground using a run of #6 copper.  But rather than bundling it together with the 240v hots and neutral and following the zig-zag path of the wiring between the two buildings, I ran it in a straight line between the ground rods, and buried it about 6" into the soil so that it would also serve as a ground radial for rf and lightning. I have a few radials attached to each one the two ground rods, each radial a little over 20' long.

The line that runs between the buildings was given to me by a friend who worked at the power company.  It is a run of the twisted trio that runs from the pole transformer to the service entrance of a house, and was designed for direct burial. To protect it from frequent digging in the ground here, I enclosed it in PVC conduit.  It has only 3 conductors, so the neutral is not distinct from the ground.  I treat the neutral the same as the two hots, and rely on the separate bonding wire as the safety ground.

Lightning pulses resemble rf more than they do DC or 60~ a.c. A shallow radial system should provide better lightning protection than would a single ground rod.  But another factor is the cross sectional area of the ground conductor.  So, if you are depending solely on a radial system for lightning protection, I would  recommend using enough radials so that  the total  cross sectional area is equal to that of a ground rod, although due to skin effect, most of the pulse current should travel near the surface of the  rod.  That's why copper clad steel rods are acceptable, vs solid copper rods. The radials should be fanned out over as wide an angle as possible.

A friend who supervises maintenance of the state's communications systems referred me to an article that recommended treating lightning as 10 mHz pulses, and protecting vulnerable systems with radials and choices of conductor lengths for maximum effectiveness at that frequency, paralleled with traditional grounding techniques.

My tower and antenna system have been in place for almost 30 years.  The tower has sustained several direct hits.  The radial system consists of 120 radials, each 133'4" long (¼λ for 160m).  I use a ball gap at the base of the tower across the insulator, but also manually ground everything using knife switches whenever I think there is danger of lightning.  None of the antenna tuners or feedlines have ever sustained an iota of lightning damage, even a couple of times when I had failed to manually ground everything before a severe storm hit.

I have heard vague rumours claiming that a direct lightning hit has caused the structural failure of an entire tower.

[WA1GFZ]

Don,
The town I grew up in put up a new tower Rohn 25 for the town communication. It was hit that summer. It was taken down and sat in a field near the Police Dept. I checked it out once thinking I could grab a couple sections and build a tower. Every section was twisted and bent.