We are adding essential circuits and panelboards to a facility that previously had none. We are installing the essential circuit panelboards in a different part of the building than the normal (existing) branch circuit panels are located. The normal branch circuit panels supply lighting and receptacles in the same rooms that our newly installed essential circuits are in. I do not see any exceptions to 517.14, and unless I am missing something it looks like we will be required to comply. I was trying to find a workaround as it is not going to be easy to get a bonding conductor to those remote panels.

What am I missing?

The requirement appears to add nothing to the (general) requirement to run a ground wire / EGC with all feeders- and your grounds are required to be connected at transfer switches, transformers, etc. as well.

I could see this requirement as making sense if you had planned on using the conduit as your ground. As I see it, though, you're going to be running an EGC a lot larger than #10 anyway.

Reno, I think the issue is bonding the grounding bus' together with a #10 bonding conductor. That is required.

BigB:

There is NO exceptions to circumvent the #10 bond conductor. I strongly doubt that you could qualify for a variation, IF that option is available in your location.

George, perhaps my ignorance is showing here ....

As I understand the trade, every panel has a nice, big, fat ground wire going from the ground buss of the source panel to the ground buss of the fed panel. Even where you have different systems interconnect - for example where there is a transfer switch - you have those EGC's.

All that remains is the mater of size- and I'm pretty hard pressed to imagine a circumstance where this feeder EGC could ever be sized as small as #10.

Is 517 asking you to run an additional EGC? Or, does this requirement address the idea that some might use the conduit of the feeder as the EGC between panels? (If so, then wouldn't the rest of the NEC have 'caught up' to 517 and made this clause redundant?)

Or, perhaps, is this requirement addressing the use of multiple ground busses in the same panel? Personally, I've never been comfortable with relying on the mounting screws and panel case as the only means of connection. If so, then the #10 wire would be 'upsizing' if all the circuits landing there were 15 or 20-amp circuits.

That last condition can't be the OP's issue, as he's asking about panels well separated from each other.

Let me stretch my imagination a little bit further ... what about different systems? Say, the 120v convenience circuita and the 480v mega-machine? Hmm. Those could be from separate services .... bonds back to the service, service bonded to grounding electrodes ... different grounding electrodes are to be connected .... yup, there's the bond, already present.
----------------

Have I been in error in believing that you cannot rely on the conduit as yout EGC for feeders? Is that where I'm getting confused? Are folks out there still feeding sub-panels without pulling an amp-sized green wire in the pipe?

This is in addition to the EGC going back to the distribution equipment and is just to mitigate any ground shift along the way by bonding all panels in the patient care area together. It is similar to the bonding we do in a pool.

Well Reno— I don't profess to being an expert on hospital wiring or anywhere near what some people are who actually install the systems. I do know that the code panels have wrestled with the bonding/grounding wording for some time and will continue. The 2011 code eliminated the term "grounding conductor" as an example of the confusion. I do believe the goal in 517 is to "Bond" the grounding bus' together to eliminate any difference between them. The analogy that Greg offered is valid and I will have to agree that it seems overkill for the requirement in 517 to have to run a seemingly redundant #10 wire between grounding bus' but that's what it says.

Thank you ... I think that now I understand as well. They don't just want the panels bonded via the distribution; they also want a local connection, direct between the panels.

Not having done any hospital work, it's all news to me.

The above is correct. ALL panels serving the patient care area MUST have the ground termination bars bonded with a min. #10 insulated, continuous copper. This is in addition to the ground conductor with the feeders.

Reasoning is that there is the availability of multiple power sources in patient care areas and the bond assures a zero potential. Greg hit it right on.

To those not familiar with HC facilities, it may seem like overkill; not being aware & getting a 'red tag' may seem crazy, but...it is what it is.

I did one big hospital and spent a week with AHCA, the Florida hospital police. I picked up a little of what they look at but it is still a lot of read and believe.

Greg:

HCF is one that 'the book' has to be available! Yes, there is a lot of reading. The Plan Review for most HCF here (NJ) is done by DCA, which I feel is a blessing for me.

We will be running the #10 insulated bonding conductor in 3/8 flex for physical protection since it has to go over 150 feet through ceilings and walls. We will use 3/8 flex fittings to enter each enclosure. Anyone see anything I might be missing?

Time to show my Ignorance of HCF Design Protocols...

Let's see if I have the concept of Section 517.14 correct:

The Specified Redundant Grounding Conductor is Installed, so as to Bond all the Panelboards in the same Individual Patient Care Vicinity, and therefore create an Equipotential Plane - where the Potential (Voltage) between any Two Panelboards in this Identified Vicinity will be Theoretically Zero.

The Potential (Voltage) between Equipment / ECG, and "Ground" (Ground being anything from the Floor directly beneath the Panelboard, to the Earthen areas of which the Building's Foundations have been cast in) is of no Relevance in this Design constraint; only the Potential Difference between any Panelboards supplying Branch Circuitry to the same Individual Patient Care Vicinity.

If the above is correct, the Specific Bonding Conductor (and Design constraints for it), is very much similar to the concepts of Swimming Pool Bonding...
i.e.:

----------------------------------------------------------
Within the Bonded Equipment Plane (Swimming Pool, Water in the Pool, The Concrete Decking surrounding the Pool, and the Mechanical Equipment serving the Pool), if someone was to drop an Extension Cord in the Pool's Water (Cord is connected to a 120VAC Source), Persons entering or exiting the Pool will not experience a Shock, as the Plane is of one singular Potential (120V).
-----------------------------------------------------------

Not being involved with the Design Protocols of Health Care Facilities, I am very curious to the ways and Technical aspects of these Projects.

Thanks to all who participated in this Thread!

-- Scott

BigB:

JUST A HEADS UP:

You may want to check if the flex connectors you are using are UL for grounding/bonding the flex jacket to the enclosures. Some interpert any grounding/bonding conductors installed in metalic raceway, to require the raceway to be bonded, using listed/labeled connectors.

Thanks, that was one thing I was thinking about. I have the screw in type that screws inside the flex and I have some 90s, but I do not know the listing status. The 90s are Arlington, not sure about the straights.

In such a situation the fear is that the bonding conductor can be 'choked' -- thus leaving a voltage potential if even for only a millisecond -- something that can kill a person who is hooked into the equipotential plane via every manner of conductive gear.

The normal resistance to voltage we have is entirely lost when our saline blood is penetrated by various conductors.