Safety consequence of missing ground conductor on 480V feeder? 4

[bentov]

Got a phone call describing the following system: 480/277 4W meter main, 100a fusible disconnect feeding 1 mile run as follows - 3x350MCM Al 1/2mi to 7.5hp 480V combination panel for 3-phase submersible pump in 16ft deep drainage collector; 3x2/0 Al next 1/4mi to identical setup; 3x#4 Al final 1/4mi to 3rd identical setup - the entire run is 3 wires only in PVC underground conduit. They drove ground rods at each pump location, connected to the panel bonding lugs & pump cable ground conductor. Sounds like it's a fairly new install by someone unfamiliar with NEC Article 250, either passed or didn't get an inspection.

Their question - exactly how unsafe is this, and how can it be made right? The obvious answer is to re-pull everything with the 4th conductor. If & until that gets done though, it's like the old ungrounded days I guess - 1st ground is "free", 2nd creates the line-to-line fault that blows a fuse. Meantime I'm trying to visualize the shock risk to personnel in contact with the panel, piping & water, how that relates to the significant distance back to the earth referenced utility transformer, whether those ground rods should be disconnected or not, etc. Can this be made safe with GFCIs only?

 

[jghrist]

First ground is not "free". There will be ground current flowing back to the grounded source, but it will flow in the earth. To determine safety, you would have to analyze it like you would a substation with IEEE Std 80 to determine touch- and step-voltages. You could get as much as a 277 volt shock between your hand and feet if you are touching "grounded" equipment and standing on earth. There may not be enough fault current to trip the overcurrent protection, so this situation could continue, just like in an ungrounded system, but with the possibility of electrocution. I wouldn't call that "free".

 

[waross]

I would want to see more than one ground rod, or a ground grid at each pump location.
The grounding serves two purposes.
1> To Cause enough current to flow to ensure that the protection operates for a ground fault.
2> To provide life safety by avoiding touch and step potentials.
Local grounding may actually be better than running an equipment grounding conductor from the source.
Some faults often occur when the installation is disturbed by a workman.
In the event of a line to grounded equipment fault, when the grounded equipment is grounded only by a conductor back to the source, the line conductor and the grounding circuit conductor will form a voltage divider. The surface potential of the supposedly grounded equipment may rise to 277 Volts to ground or a little more.
This could be fatal to a worker standing on the ground surface and touching the equipment frame.
Good quality local grounding will maintain the surface of the equipment at the same potential as the local soil.
On long lines, comparing only an equipment grounding conductor with local grounding, the local grounding may be safer.
Ground fault currents may be detected and cleared by monitoring ground current at the source of supply. The setting should be high enough to avoid tripping on charging currents.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[dpc]

Wait - a mile-long 480 V circuit? Problematic in itself. With a long circuit like this, an equipment grounding conductor in the conduit with the phase conductors is a must. Ground rods won't totally fix it. Too much resistance in the earth for the return current.

What jghrist said.

The service is a grounded system so operating without a grounding conductor on a long circuit like this is unsafe and potentially lethal. Normal 480 V ground fault relaying will not prevent electrocution. You will need a 480 V 3-phase GFCI. I've seen these up to 60 A - not for 100 A, but I haven't looked very hard. Not sure if these are UL-listed or not. Google is your friend. You might get by with a 60 A unit based on your load.

I would get this fixed ASAP. At the least it is an NEC violation, if anyone cares about that.

 

[bentov]

I was thinking of the old 480V ungrounded delta systems we still see at remote pump stations, but as you say jghrist, nothing "free" here with the neutral grounded back at the service entrance.

Interesting about the voltage divider, waross. If we had done this run we'd just have the 4th conductor, upsized per NEC at the same ratio as phase conductor voltage drop driven calcs (to provide adequate fault current for SCP clearing). I see in more reading a recommendation that for shock safety it should instead be the same size as phase conductors, to evenly split the 277 thus stay below 150V for anyone in contact with bonded surfaces - makes sense.

So the scenario here - one conductor ground faults, all local bonded parts (panel/deck/pump/water in pipe, etc.) now energized at 277V relative to that earthed neutral back at the service entrance. The (presumably very high) resistance of the 1/2 mile long earth "conductor" will determine current value, very unlikely to blow a fuse (until the 2nd local fault makes it a local phase to phase short). The shock risk for a person standing on the ground there, touching the panel, hinges on his relative resistance - he's a circuit in parallel with the huge concrete sump embedded in the earth (that the panel & pump are bolted to), seems likely not much current will flow through his body. If that is made even more certain by additional local grounding per waross, then maybe this setup can be relatively safe (though unreliable and hard to troubleshoot once it starts "leaking" low levels of fault current).

dpc, per your suggestion Google took me to UL943C Class C protection specs, seems all the info is from Littlefuse (

http://www.littelfuse.com/~/media/f...se_industrialshockblock_sb6000_datasheet.pdf)

- maybe they're the only supplier at the moment, I bet it's expensive. Guess I'll run this by them, see what they suggest. It does appear we have new options for real protection, good to know as there are many unsafe installations out there. Doing field troubleshooting and repair, I see stuff every day that makes me shudder . . . just try to fix what I can, hands off the rest when possible, point out but don't scold when things have been "not right" for decades - what can you say, really, besides stripping out and starting over, cost be damned?

I just skimmed (and now need to study) this 1999 Erico book (

https://www.erico.com/catalog/literature/G157LT99.pdf),

seems like most systems I see (and some I install, to current code) are pretty inadequate in the grounding area, oh well!

Thanks for the replies, lot to think about here . . .

 

[dpc]

I bet it's expensive.

Probably cheaper than re-pulling with a ground wire Google Coleman Cable Inc. for another supplier.

seems likely not much current will flow through his body

But only takes milliamps through the heart to be lethal.

Cheers,

Dave

 

[LionelHutz]

Could you install a neutral grounding resistor with ground fault detection as well as extra grounding at each site and the main to ensure that the required low level ground fault current will flow?

 

[MatthewDB]

Whoever did this created a "TT" type grounding system, quite popular in Europe and in the UK in particular. They always include ground fault detection in at least the main breaker with a trip setting low enough to guarantee tripping on any ground fault.

You can be safe if you follow IEC rules, but you'll never be legal in North America unless you pull the proper ground.