Visit MikeHolt.com
Ground Fault Protection and the Multi-wire Branch Circuit - A Troubled Marriage  

 
Image

We’ve been contacted by Randal Andress who, as a recently retired engineer became interested in ESD about a year and a half ago. Last August he participated in the Fire Protection Research Foundation research planning meeting on "Hazardous Voltage/Current in Marinas, Boatyards and Floating Buildings" as a call-in.

Both before and since then Randal had been thinking about the effects of marina basin background currents (sometimes called foreign or stray currents) on the measurement of AC leakage from boats by clamping their shore cord with an ammeter. He found that the effect of background current depends on whether it originates from the same or opposite leg (L1/L2) of the distribution source as the current leaking from the boat. That led to the question of how faults/leaks from different legs of a main or feeder panel would be seen by a 2-pole (L1/L2/N) ground fault protection breaker.

What he found was that the protection provided is attended with subtle if not troublesome and problematic differences from the protection provided by a single pole, GFP breaker on a 2-wire circuit (H/N). The manifestation of these differences in a marine environment is of particular interest/concern.

Randal had not seen this addressed anywhere and so he began writing an article on the subject. Following are excerpts from the article. See the entire article (4th Draft) by clicking here. He is particularly concerned to get the analysis correct and easy to understand. He has made the text very wordy so as to stand fairly well without complex illustrations. And for the moment he has included illustrations that are very busy so as to be able to support an ad-hock briefing without having the text. Randal would appreciate your comments so please post them below.

 

Bottom Line Effects

In a nutshell, then, what we find is that when feeding a MWBC, the 2-pole ground fault protection breaker trips when the imbalance caused by the difference in the L1 vs. L2 faults/leaks exceeds the trip limit. In other words, it trips when:
| (L1 leaks & faults) – (L2 leaks & faults) | > Trip Limit.
...
To put it another way, the common, multi-wire branch circuit cannot be protected against ground faults to the same degree and with the same precision, the same measures of protection, expectation of behavior, etc. as a branch with its own neutral. This is not all that surprising when you consider that by sharing a neutral, the return current of each branch is made anonymous as to its origin (L1 vs. L2).

Here are some of the effects of using a 2-pole ground fault protection on a MWBC:

1. Fault current on one leg in excess of the rated ground fault protection level is required to trip the breaker in the presence of the inevitable leakage/fault current difference (I.e., when L1_leakage - L2_leakage is not = 0) – thus decreasing fault detection sensitivity.

2. The manual test button will not consistently perform a valid test. - i.e., a good breaker will test bad - when L1 fault/leakage current is greater than that of L2 (plus test current margin) or when L2 fault/leakage current is greater than that of L1 (plus margin), depending on whether or not the current for the test button is drawn from L1 or L2.

3. L1 and L2 circuits may be able to sustain leakage/fault currents well in excess of the rated ground fault protection level. I.e., leakages/faults may be arbitrarily high without tripping the breaker so long as their fault/leakage difference (L1 vs. L2) is less than the rated ground fault protection level. E.g.,a faulty/leaky boat which trips a 2-pole GFCI/GFPE when connected to a 120V/30A, L1 circuit could work just fine if moved to an adjacent L2 powered receptacle.

4. The removal of an appliance or device from a circuit could cause a trip. This happens when the fault/leak in the removed device is of such value that, when removed, it increases the L1 vs. L2 difference up to the trip level. The order of appliance connection and disconnection determines the possible, trip/non-trip, circuit states.

5. The likelihood of so-called nuisance trips is increased. A very small added fault or leakage can cause the breaker to trip since lop-sided L1 vs. L2 fault/leak current increases the trip sensitivity on one leg while decreased sensitivity on the other.

What Circuits Are Affected

These effects apply to virtually every 120V multi-wire branch that consists of two ungrounded conductors from opposite legs (L1 and L2) of the 240V source. However, the implications in some cases are decidedly more significant than in others.

On the less troublesome end of the spectrum might be the dishwasher and garbage disposal split outlet connection that is fed by a 2-pole Type A GFCI breaker. In the first place the Class A GFCI trip point is low (~5 mA) … Secondly, the test button dead zone for a test current of 7 mA would be only about 3 mA wide ... Thirdly, there is only one appliance connected to either leg (L1/L2) … Lastly, but perhaps most importantly, both are subject to mandatory standards and codes.

At the other end might be a 2-pole (L1/L2/N), 100 mA GFPE multi-wire circuit at a marina that feeds 120V/240V as well as and both L1 and L2 120V shore power receptacles. … The circuit connections – the boats – vary widely both as to the mix of appliances on board and also with time (here today, gone tonight, in a different slip tomorrow) and are subject to no mandatory codes. In addition the likelihood of lethal exposure is greater since the circuit faults/leaks could far exceed a safe limit and could be flowing into the water.

[--------------- end excerpts ------------------]

Randal Andress is an Electrical Engineering Graduate, retired after 30 years with TRW/Northrop Grumman and is an NFPA member and a published author. Please post your comments below.

randalPandress@gmail.com

 

Click Here to Post a Comment Click Here to View Comments Click Here to Unsubscribe
 
Image
 
Visit: Exam Preparation | Continuing Education | Code Products | Newsletters | and more
Copyright© Mike Holt Enterprises of Leesburg, Inc. All Rights Reserved This article is protected by United States copyright laws and may not be published without prior written permission.
Mike Holt Enterprises of Leesburg, Inc. 3604 Parkway Blvd. Suite 3 Leesburg, FL 34748
"... as for me and my house, we will serve the Lord." [Joshua 24:15]
Comments
  • We have seen this effect in the field doing ESD research. The latest instance was in measuring a voltage gradient in the water near the propeller of a boat. Both L1 and L2 loads were running, each on it's own 30 amp shore cord. Was measuring a gradient of 125mv/ft in the water. When L1 supply shore breaker was turned off, the gradient jumped to over 200mv/ft.

    While these levels are not lethal (2v/ft considered lethal for nominal person), they indicate the cancelling affects we see when there is leakage into the water from both L1 and L2 on the same branch circuit.

    We have also seen this when clamping shore cords. Let's say you are clamping a single 30amp cord with loads running and measuring 100ma on the cord. Then you turn off the shore supply breaker at the source (pedestal). Normally you would expect the clamp reading to go down. Have seen it a number of times where it goes up! This has to be the effect of cancellation of L1 and L2 fault leakage into the water.

    Randall has done some excellent work in this area. Perhaps there is way to prevent this with a simple modiification, similar to the addition of a second CT to a GFCI to prevent leak to ground from traveling in the ground then back up into the neutral such that it is not seen by the GFCI.

      August 20 2016, 7:43 pm EDT
  • Reply from: Randal Andress   August 20 2016, 11:29 pm EDT
    David, thank-you for your kind words and for both these first hand observations of the implications of the mix of L1 and L2 fault/leak currents especially in a marine environment. This is precisely what happens in the 2-pole GFCI. The L1 fault/leak current, in returning to the distribution neutral-ground connection, shares its conductive path with that from L2 currents which are of opposite sign and so the net current is the difference. As our devices seem to tolerate more complexity (The grounded neutral detection you mentioned and the AFCI are examples.) we should not shy away from thinking "outside the box".


Reply to this comment
* Your Name:
   Your name will appear under your comments.

* Your Email:
   Your email address is not displayed.
* Comments:

Email Notification Options:
Notify me when a reply is posted to this comment
Notify me whenever a comment is posted to this newsletter