Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Jamin Lynch]

Disclaimer: I have no plans to do this....information purposes only.

There are several old dancehalls in my area where power was always a problem. In the past some guys would tie into the nearest breaker panel. Often times there was no ground bar. They would attach the ground wire to the neutral bar.

Bad idea?
What would be an easy and safe solution?

[Stephen Kirby]

By nearest breaker panel, you mean some sub panel and not the main service? 

[Jamin Lynch]

By nearest breaker panel, you mean some sub panel and not the main service?

Usually a sub panel. They were old panels probably installed prior to ground bars being required. I guess they were never updated.

The main service was never close enough to the stage

[Jonathan Johnson]

By bonding the EGC of a distro to the neutral bar of a subpanel, you increase the risk of ground current noise if you have interconnected devices that are fed from separate panels. If all interconnected devices are powered from the same panel, it shouldn't be the source of ground current noise.

As for personnel safety, I'll defer to others.

Grounded (via the neutral of a subpanel) equipment in contact with a stage or other metallic building fixture that's bonded to another location may also result in ground current noise. If that circuit (between the grounded equipment and the grounded fixture) is completed by a human in contact with both, that could also be a source of noise. I don't know if the voltage and current potential would be great enough to present a shock hazard. (If it is a shock hazard, then it will definitely be a source of noise... perhaps a 60Hz moan from the performer's vocal chords?  )

[Stephen Kirby]

As I understand it, the safety issue is a failure of the neutral somewhere between the service entrance where it's tied to ground and the end use point.  A real ground provides a path for the hot voltage to return on that doesn't involve a human in the circuit.  In the posited case, if the neutral fails between the main and sub panel, the return path will look for the lowest resistance to a sink or lower potential.  If the resistance is high enough, it will just hold anything it's in contact with at that potential.  But if the resistance gets low enough, like through a waterbag, current will flow.  With the ground connected to a now open stub, it doesn't do anything.

[Jonathan Johnson]

Usually a sub panel. They were old panels probably installed prior to ground bars being required. I guess they were never updated.

The main service was never close enough to the stage

Does the subpanel have a separate grounding electrode? Is there a grounding conductor leading from the panel to a ground rod or water pipe (that is in turn properly grounded)? If so, this can reduce the shock risk to personnel, but it still can result in ground current noise.

In the United States, the National Electrical Code prohibits ground-neutral bonds EXCEPT in the "service entrance" of a building, where the bond is required. However, since a single utility transformer may supply power to multiple buildings, there may be multiple ground-neutral bonds throughout a campus or neighborhood power distribution system -- each building supplied by the transformer will have its own ground-neutral bond, in addition to the ground-neutral bond at the secondary side of the utility transformer. Some power companies also bond the ground/neutral of the transformer secondary with a utility ground/neutral on the primary side. There usually is no EGC between the service entrance and the utility transformer. (Note that for RVs, mobile homes, and other portable buildings, the "service entrance" is an external power pedestal, NOT the breaker panel inside the building. The ground/neutral bond is at the pedestal, and there is an EGC between the pedestal and the breaker panel in the building, where there is no bond.)

This means that some "neutral" current could pass on the building's GEC (grounding electrode conductor) to the ground rod, through the ground rod to another building's or transformer's ground rod, to the secondary of the transformer. Due to the difference in resistance between the neutral conductor and the earth-as-a-conductor, the voltage differential and current flow should be minimal. But it does exist.

If there should be a failure of the neutral in this scenario, the voltage differential and current flow could be quite high, perhaps sufficiently high to cause shocks to personnel. However, if this "circuit" is not protected by a breaker or GFI, the hazard is not automatically removed during a fault condition.

[David Buckley]

What would be an easy and safe solution?

An isolating transformer.  Easy, and safe.  Also moderately heavy, space occupying and expensive.

[Stephen Kirby]

Doesn't the transformer require it's own ground?  My understanding of Jamin's question is with venues where there isn't a legit ground available.

[Mike Sokol]

What I used to do in this exact circumstance 40 years ago was to use a pair of vice-grips with a big lug on the side to run a piece of #8 wire to the copper water pipes in the basement of the club. I connected the other end to my distro ground. It worked well because ALL my grounds were tied to this one grounding point.  That's not a good idea nowadays (and certainly a code violation) since water pipes are plastic. However, the idea still has merit. If you can get an approved ground lug connected to the service panel, then run a heavy piece of EGC wire to your distro ground, I think it would accomplish the same thing.

I'm seeing the same sort of related issue in a few fire halls where they've offered to let me plug my 50-amp/240-volt distro into a 3-wire "stove" receptacle. I've refused for the reasons mentioned above, but have suggested they install a proper 4-wire 50-amp/240-volt receptacle for future shows. They have the breakers, so it just needs 10 feet of the proper wire and a NEMA 14-5 receptacle. Of course, the EGC of this receptacle will be bonded to the ground bus bar in the sub-panel, which is exactly how it's supposed to be done.

I figured out a long time ago that I could make a lot of strange wiring work, but if it was in violation of code and something happened (fire, electrocution, etc...) I would be on the hook for the damages and lawsuits. So I won't wire up something that's in violation just to save a few bucks. Far better to do it right and legal.

[Stephen Kirby]

So Mike, in Jamin's situation where he's tying to a sub panel which has no ground bar connected to the service entrance ground with a discrete conductor, where do you go?  Back out of the gig unless they run a proper EGC to the sub-panel?  I think it was a hypothetical question based on past experience but looking forward to the possibility in the future.

[David Buckley]

Doesn't the transformer require it's own ground?  My understanding of Jamin's question is with venues where there isn't a legit ground available.

Nope, that's the beauty of it. The transformer isolates you - both ways - from the venue power, and provides (through a neutral/ground bond on the output) a local, "real" ground for everything plugged into it.  Same as using a generator with neutral/ground bond on it.

So from the perspective of the venue, the transformer is just a load between hot and neutral.  From the perspective of the load on the other side of the transformer, it looks and works just like a modern, fully code compliant feed.

[Jamin Lynch]

Some of these old dancehalls were built years ago and still do not have up to date electrical. They are way out in the country and I guess overlooked by inspectors.

Most of the sub panels had no ground bar at all.

I've seen some guys "attach" the ground wire to the hard conduit that comes into the sub panel.

We don't have basements anywhere in my area. 

[Stephen Swaffer]

Nope, that's the beauty of it. The transformer isolates you - both ways - from the venue power, and provides (through a neutral/ground bond on the output) a local, "real" ground for everything plugged into it.  Same as using a generator with neutral/ground bond on it.

So from the perspective of the venue, the transformer is just a load between hot and neutral.  From the perspective of the load on the other side of the transformer, it looks and works just like a modern, fully code compliant feed.

Actually, the transformer provides a local, real neutral or grounded conductor.  Code still requires the transformer derived neutral to be bonded to ground.  The problem in a sub panel fed by only three wires is that typically load current is conducted by the neutral wire raising it to some voltage above "ground" which gives rise to audio noise.  This will still happen if you install a transformer and ground to the neutral in the panel.  Code specifically prohibits current other than fault current from flowing on an Equipment grounding conductor, so a new transformer will have to have a new wire run to meet code.  If you are going to run a new wire, why not just run a ground for the panel?

From a safety standpoint, you could run everything on GFCI's-code allows this. This may or may not be acceptable from an audio standpoint.

If every load on the panel is 240 volt, then the third wire could be used as ground.  In that case (perhaps a heating sub panel ) a transformer could work-but that would be an uncommon situation.

[David Buckley]

Code still requires the transformer derived neutral to be bonded to ground.

Yes, code does mandate (250.30) the derived neutral (ie the transformer secondary) to be bonded to the supply ground with a separately derived system as the NEC calls it.  But this is one of those occasions where the NEC is just plain wrong.  It gets even more dangerous when an isolating transformer is used in a marina or on a boat, where it gets really deadly.  The requirement for the bond defeats the very purpose of an isolating transformer.

This is a situation where there is a choice: safety or compliance.  I'll choose safety every time, thanks.

Here is the NEC way from ecmweb, note the solid green ground from the supply on the left all the way through, and also note that the ground fault current represented by the mostly left facing arrows on the solid green line end at the transformer secondary centre tap, and do not go all the way to the supply



This is the rest-of-world way of doing it; the fault current path is identical, there is no compromise on safety, and the isolated supply has nothing at all to do with the the primary supply.

[Stephen Swaffer]

Yes, code does mandate (250.30) the derived neutral (ie the transformer secondary) to be bonded to the supply ground with a separately derived system as the NEC calls it.  But this is one of those occasions where the NEC is just plain wrong.

IMO, this is a dangerous mindset.  Even if you use an iso transformer, you need a ground. You claimed the transformer creates a local ground-which it cannot do.  Even the "rest of the world way" has a ground. Otherwise, you have a situation where someone lays a 58 on a stage and creates one "ground" connection unintentionally and another fault becomes dangerous. Metal objects that are not intended to be energized need to be grounded, and the right way to mitigate hazards is with GFCI protection. Code compliant, safe, defensible if something goes really bad. Y

Another issue with the isolated, no bond setup is what happens when (and sooner or later it will happen)' there is a fault to ground. With no neutral bond, no fuse or circuit breaker will blow.  There will not be any obvious indication of a fault until the second fault occurs. Whether you realize it or not you no longer have an isolated system.

In a sub panel with no neutral bar, I would look for a box bonding screw (I assume if you are in a panel, you are qualified and if you are qualified finding that or verifying a absence should be doable). If it exists, clamping to conduit is the same as attaching to the neutral bar.  If not, conduit is an acceptable EGC, if it is installed correctly for that purpose-which you probably can't really verify. At some point you have to hope the installer got it right.  From a practical standpoint, I would either go to conduit, or if bonding screw exists go to the neutral bar and be very vigilant in my testing and verification.  The main hazard would be the neutral voltage being raised by a load and that raising the "ground" voltage which becomes on issue if you are on a damp concrete floor or near building steel, etc.

[Stephen Kirby]

Thanks Stephen,  I've not dealt with iso transformers but I've had lots of step up/down transformers in industrial applications and they all had a ground rod at them with a neutral bond.

I am thinking that the old buildings Jamin is talking about probably have full metal conduit.  Although some of the joints may be a bit corroded by now.  And the neutral bar bolted to the box which is effectively the same as the conduit.  Probably have all iron piping as well.  Although someone may have spliced plastic in some above ground repair.  Would an iron pipe that goes into the concrete be more trustworthy?

[Stephen Swaffer]

Would an iron pipe that goes into the concrete be more trustworthy?

The forums favorite answer-it depends  Actually that would make a good ground in most cases-somewhat depending on how much of the pipe is in contact with the concrete.  However, the biggest issue here is really properly called bonding.  Bonding is intentionally creating a metallic conductive path-and is required by code for everything on a service that is likely to or can become energized.  Mike posted a link to a video not too long ago regarding the difference between bonding and grounding.  "Grounding" to the metal pipe would likely not ever trip an overcorrect protective device, if that pipe is bonded to the service an overcorrect protective device would trip.  If that pipe is a good enough ground to be considered a grounding electrode, code still requires all grounding electrodes on a premises to be bonded to each other.

In short, you really need a properly sized metallic path back to the ground neutral bond to really have a safe setup-and that path cannot intentionally carry current, which rules out using the neutral.

[Mac Kerr]

Thanks Stephen,  I've not dealt with iso transformers but I've had lots of step up/down transformers in industrial applications and they all had a ground rod at them with a neutral bond.

I am thinking that the old buildings Jamin is talking about probably have full metal conduit.  Although some of the joints may be a bit corroded by now.  And the neutral bar bolted to the box which is effectively the same as the conduit.  Probably have all iron piping as well.  Although someone may have spliced plastic in some above ground repair.  Would an iron pipe that goes into the concrete be more trustworthy?

I can't believe I am saying this again... A "ground rod" stuck in the the dirt has nothing to do with electrical safety bonding. It is for lightning protection, not electrical safety. It will not trip a breaker in a fault condition. It will trip a GFCI because it will create a leakage current path, but it will not pass enough current to trip a breaker.

To trip a breaker you need a return path for fault current that returns to the source. If you have a separately derived system like a generator that is where the safety ground conductor needs to return to, and that is where the neutral/ground bond needs to happen. If the generator neutral is not bonded to the generator ground (chassis/ground lug) there is no return path to the source to trip the breaker.

Mac

[Mike Sokol]

In short, you really need a properly sized metallic path back to the ground neutral bond to really have a safe setup-and that path cannot intentionally carry current, which rules out using the neutral.

Yup, and I use a Ground Impedance Tester to verify that the bond has a low enough impedance to quickly trip any over-current device. All my GIT's have set threshold of 1 ohm. So any bond impedance greater than 1 ohm is a fail, while a bond impedance less than 1 ohm passes. Any bond that measures really low (less than 1/10 of an ohm, I think) is a suspect bootleg ground.

As Mac has stated, ground rods really aren't the same thing as your EGC "Bond" point. In fact, a ground rod can have an impedance of up to 25 ohms to earth and still be within code. A little quick math shows that 120 volts / 25 ohms = 4.8 amps, which isn't enough to trip any over-current device. That's why a ground rod isn't a substitute for a proper ground BOND to the Neutral conductor inside the service panel, and by extension any sub-panel. Ground rods are for lightning protection and keeping the local ground plane of your building close to earth potential. And they do provide enough leakage current to trip a GFCI.

[David Buckley]

IMO, this is a dangerous mindset.  Even if you use an iso transformer, you need a ground. You claimed the transformer creates a local ground-which it cannot do. 

No, I didn't claim anything of the sort.  What I claimed, correctly, was  "...provides (through a neutral/ground bond on the output) a local, "real" ground for everything plugged into it".

That's just fact. 

When I say "everything plugged into it", I'm making the assumption that the "everything" is your disto.

It may not have been clear in the explanation above, so let me make it really clear now.  The isolation transformer with a neutral/ground bond creates a gound point.  About that there is no doubt.  That ground point still needs to be connected to everything on the stage, which, assuming one uses normal three prong plugs with three core cable, will be the natural case assuming one wires one's stage right and uses normal techniques.

When a sound provider puts up a PA for a band, it is really important that every conductive element that the band can touch on the same are at the same potential, to prevent shock.  The band need to be in an equipotential zone.  Every ground wire from everything on the stage needs to end up at a single point that you can trust.  Since Buddy Holly's day, the most hazardous situation that occurs onstage is live guitar, grounded mic. 

The audio world is very familiar with distros, a device intended to distribute power, but which actually has the more important function of providing a common ground point for everything connected to it.  Step down a level, and there is the poor mans distro, a device that provides a common ground to everything plugged into it.

As long as everything is plugged into the distro, then everything has the same ground potential.  So if our guitarist touches the mic, both things are at the same potential, no shock.   

I really do know what I'm on about with this stuff, I'm not making it up as I go along.  If there is something you don't understand, just ask.

[Stephen Swaffer]

What I don't understand is how this is any safer than using the neutral bar in the non compliant panel as your single bonding point for your equipotential plane? Neither case meets code, and as long as you are using the neutral, a bad neutral creates problems either way.  Yes, it could work IF you had complete control-until someone plugs a fan in and sets it on stage or a late musician (won't ever happen I know) plugs his amp into the nearest receptacle instead of the distro, or whatever. To really be safe you still have to be bonded back to the service. Most of the guys on the NEC code committees are pretty smart, too.

[Jonathan Johnson]

Yup, and I use a Ground Impedance Tester to verify that the bond has a low enough impedance to quickly trip any over-current device. All my GIT's have set threshold of 1 ohm. So any bond impedance greater than 1 ohm is a fail, while a bond impedance less than 1 ohm passes. Any bond that measures really low (less than 1/10 of an ohm, I think) is a suspect bootleg ground.

As Mac has stated, ground rods really aren't the same thing as your EGC "Bond" point. In fact, a ground rod can have an impedance of up to 25 ohms to earth and still be within code. A little quick math shows that 120 volts / 25 ohms = 4.8 amps, which isn't enough to trip any over-current device. That's why a ground rod isn't a substitute for a proper ground BOND to the Neutral conductor inside the service panel, and by extension any sub-panel. Ground rods are for lightning protection and keeping the local ground plane of your building close to earth potential. And they do provide enough leakage current to trip a GFCI.

Academic point: if you use the neutral busbar as a common bonding point between the distro neutral and ground, AND the chassis of the subpanel is also bonded to this same point, AND you also drive a ground rod into the dirt and connect it to that SAME bonding point, you have achieved the SAME scenario as the service entrance that is fed from the utility transformer without a separate ground wire from the transformer to the service entrance. (Yes, it's common for utilities to supply a service entrance with a neutral conductor but not EGC.) If current on the neutral that's feeding that subpanel causes a voltage rise at the bonding point, the ground rod helps to minimize that voltage rise by attempting to bring the EGC in the distro to the same voltage potential as the dirt.

If there is a hot-to-grounded-chassis fault you may not trip a breaker UPSTREAM of the subpanel, but you should still trip the breaker of the branch circuit the faulty device is connected to. If there is a hot-to-EGC fault in the cord supplying the distro, it should trip the breaker feeding that cord (you did tie into a breaker in the subpanel, didn't you?).

If there is a hot-to-dirt fault, you might not trip a breaker. But even if everything was installed exactly according to code, and there is a hot-to-dirt fault, you might not trip a breaker.

EDIT: If the neutral busbar is NOT bonded to the chassis of the panelboard, and ALL of the existing circuits are in conduit, then you should NOT tie the EGC of your distro to the neutral busbar. (There may be later additions that have installed a branch circuit EGC and connected that to the neutral busbar. This is incorrect practice.) In that case, even though there is no grounding busbar, the conduit system acts as the EGC (Equipment Grounding Conductor). If you attach the EGC of your distro securely to something metal that is bonded to the panel chassis, that would suffice. For example, you might have an electrician bolt a lug to a hole drilled in the cabinet. You might want to remove paint or install a star washer between the lug and the panel. Even then, it's a good idea to do a ground impedance test, as a loose fitting on the conduit supplying the subpanel could render the ground useless.

The best option is to have an electrician install a grounding busbar. This should be bonded to the cabinet. The busbar should also have a jumper to a grounding bushing installed on the conduit supplying the panel. If possible, a separate EGC should be added to the conduit from the main panel to the subpanel.

[David Buckley]

What I don't understand is how this is any safer than using the neutral bar in the non compliant panel as your single bonding point for your equipotential plane?

It's safer because using a random neutral as ground is just that - random, one has no idea of the integrity of the neutral.  If the neutral fails then you have the "ground" wire raised to line potential, and there may well be other things that penetrate the equipotential zone that are grounded to a different potential causing a genuine shock hazard.  if the premise has missing or inadequate ground wiring, then it may well have missing or inadequate bonding too.

With an isolating transformer and its N/G providing a local ground, then you are certain of the integrity of that ground at the point of supply.  If the building neutral fails, then you lose the show, but there is no shock potential.

... until someone plugs a fan in and sets it on stage or a late musician (won't ever happen I know) plugs his amp into the nearest receptacle instead of the distro, or whatever.

OK, lets take that case where the PA and backline is supplied through an isolating transformer, with all this kit running off the distro thus bonded to the locally created ground.  We'll further assume there is a metal encased fan on stage plugged into some local outlet, and we don't know anything about the ground pin of that outlet.  Lets further assume that the vocalist, SM58 in hand touches that fan.  The SM58 is connected to our local ground.

The important thing here is that there is no common reference between the local ground of the SM58, and the supply feeding the fan.  If the socket feeding the fan is as it should be, then there may be a slight voltage difference with a very small current drive capability between the SM58 and the fan, from leakage,  Less than you would see if the the socket was wired with reversed N/G (bootleg), which is a small voltage with significant current drive capability. 

If the fan socket has no ground, then it is also possible that the fan may be faulty, in which case its metalwork is at mains hot potential.  Our vocalist might well get a bit of a tingle when he touches it, due to leakage between our local ground and the supply ground. 

Contrast that with the case of what would happen without the isolating transformer.  The vocalist would get the full mains across him with high current drive capability, and would be seriously shocked.

The bottom line is this: if there is a properly grounded supply available, then one can build a distributed supply that is adequately safe for entertainment purposes by just using the right kit and plugging it together.  When such a supply is not available, then the question then becomes how can one safely put on a show.  Getting a set of mole grips out and attaching a cable to a bit of conduit and calling it ground, or using a neutral bar as a bootleg ground may well get you through the night, but these are not approaches without risk.  Using an isolating transformer with a local ground is a strategy that will always work and deliver a safe outcome.

Most of the guys on the NEC code committees are pretty smart, too.

Yeah, they are.  But they are not perfect.  They fail to see the advantages of isolating transformers.   Every year, Americans die because of electrical incidents around water and marinas, and these folks could be protected were isolating transformers used to break the ground circuit.  GFCI/RCDs are fabulous, one of the best inventions in the world of electricity, ever, and an essential part of any distro, but a GFCI/RCD cannot protect against ground to ground shocks.  No protective device you can (still) buy provides ground to ground shock protection.  This is why its really important to get grounding right.

I think the code making guys are influenced by long standing practices in power distribution in the USA.  For an interesting (but somewhat political!) summary of some of it, have a read of he Hazardous Multigrounded Neutral Distribution System (PDF, 256KB).

[Stephen Swaffer]

Electrically there is no difference between your isolating transformer and the scenario that occurs if someone wires a home (or outbuilding) with a 3 wire service from the POCO and does not install a grounding electrode.  The POCO's transformer is essentially an isolating transformer-yes it has a grounding electrode, but the bottom line is that with your "local ground" your connection to earth-and anything else on the premises is an unknown quantity.  Set an equipment rack or lay a mic cord on a concrete floor and spill a drink and you just earthed the "local ground".  It doesn't take much of a connection to carry 6 mA.

Your "equipotential" zone is exactly what the NEC demands of EVERYTHING on the premises.

What happens if there is an insulation breakdown in your isolating transformer? Undersize the transformer to save money and weight and this becomes a real concern.

The line between "feeling a tingle" and a dead musician is an awful fine one to walk.

Then the elephant in the room.  If something bad happens and you wind up in court it will be your word against the wording of the NEC.  If those on the panel at the NEC cannot understand the advantages of an isolating transformer, do you really expect to successfully explain that to a judge or jury? 

Probably the best plan for the OP's scenario would be to carry enough #6 or #4 wire to run a ground wire back to the main panel, unless a true EGC can be ascertained at the panel in question. In an steel framed building the building frame should be part of the grounding electrode system so it would provide a ground connection point-but those are the only connections I would really be comfortable with. Even better, invest in the equipment necessary to verify the integrity of the connection to the service ground/neutral bond.

Obviously, the problem in the OP's scenario is business.  If you create a stink, there is probably someone else that will just hook up and go.  In my business as an electrical contractor, I finally figured out that I don't need every job out there-if the customer doesn't want to do things right and safe often it is better to just walk away-the customers that are willing to do things right are usually more successful and ultimately a more profitable customer anyway.  It should be worth at least a service call to get an experts opinion-and if there is no ground bar but there is conduit, installing a properly bonded ground bar as shouldn't break the bank.

[ProSoundWeb Community]