Diagnosing GFCI Nuisance Tripping FROM
Sewage-Pump Ground-Leakage Current

GREG LOWITZ, M.Sc.Eng.
BUILDERA

SUMMARY

After 10 years of reliable operation, a California homeowner experienced nuisance (ghost) tripping of a basement Ground-Fault Circuit-Interrupter (“GFCI”) whenever activating the sewage pump. Upon replacing the GFCI with a new one, the power-tripping problem persisted. As the sewage pump worked reliably for years without tripping the original GFCI—and recognizing that replacing the GFCI did not remedy the new tripping problem—BUILDERA measured suspected leakage current to determine whether a possible pump electrical short was the GFCI-tripping root cause. If left unresolved, abnormal leakage current presented a serious safety hazard that could result in electrical shock or electrocution [1]. The full case study, measurement methodology, and results follow below.

BACKGROUND

Sewage-ejector and dewatering sump pumps are common in basements and below-grade drainage basins (Figure 1). When sewage or other wastewater fills an empty sump tank, after a few minutes a level-sensing float-switch activates the pump, which remains active until the fluid level discharges below a preset threshold. Depending on the pump model, the float switch may be single-pole or double-pole. The former switches only the hot conductor, whereas the later switches both the hot and neutral conductors to the pump—providing an extra measure of protection.

Discharge Hysteresis and Recommended Backup Pump

Hysteresis is the difference between the initial switch-on wastewater level and the subsequent switch-off level—similar in principle to an HVAC thermostat regulating temperature when cycling a furnace every 10 to 20 minutes. Proper hysteresis allows sufficient time to empty the effluent basin without short-cycling the sump pump. In the event of a primary-pump mechanical or electrical failure, Figure 2 illustrates a recommended back-up pump to prevent the basin from overflowing into the basement or crawlspace. For best reliability—and to avoid tripping the subpanel circuit breaker when both pumps are active—the backup pump should also connect to a dedicated GFCI-protected branch circuit (typically 120 VAC, 15 A or 20 A). Optionally, the backup pump may also be a battery-operated DC or AC/DC pump powered, for example, by a 12-volt wet-cell battery (marine battery) and trickle charger (Figure 3). Selection of the backup-battery chemistry and construction—such as the newer sealed Absorbent Glass Mat (“AGM”), gel-cell, and Lithium-Ion (“Li-ion”) technologies—are beyond the scope of this article, with each offering tradeoffs of cost, ease-of-maintenance, discharge characteristics, and longevity.

Pump Seal Keeps Moisture Away From Electrical Components

During normal pump operation, the hot and neutral conductors are electrically isolated from the sewage-pump housing, which is typically constructed from a combination of heavy-duty cast iron or cast bronze and/or engineered plastics [3]. A neoprene [4], nitrile rubber (NBR), or similar watertight gasket prevents moisture from entering the motor cavity or coming in contact with electrical wires.

In the case of an insulating-plastic (ungrounded) sump basin as shown above in Figure 1, the only safety-ground path is from the pump ground wire (via the plug-in receptacle) to the circuit-breaker panel where the neutral and ground conductors bond together at a single point. The ground wire should never carry any return-load current, except during an emergency ground-fault or short-circuit. In addition, the hot and neutral currents should always be identical in magnitude. Any differential current indicates a leakage current, which is typically a shunt to ground—either along the ground conductor itself—or through another grounded object in contact with the energized surface.

Ground Leakage and Differential Current from Seal Failure

Over time, the sewage-ejector pump or switch seal can deteriorate and fail, allowing moisture and corrosive wastewater to penetrate the float-switch or pump cavity. This breach sets up a hazardous condition wherein the 120V hot conductor is in direct or intermittent contact with accumulated wastewater. The wastewater, which is composed of contaminating organic material, salts and other chemicals, conducts electricity to the ground wire via the GFCI and back to the circuit-breaker panel.

Ground-Fault Circuit-Interrupter—Essential for Safety

Notably, after a decade of reliable service, the GFCI began to trip every time a few gallons of wastewater filled the sewage tank. Upon the float switch activating the pump, the GFCI immediately kicked in, shutting off power to the pump. This created an untenable problem because continued flushing, showering, or other water usage in the basement bathroom could quickly fill the basin and overflow into the finished basement through the lowest egress point—especially without a secondary pump on a battery back-up or wired to a separate, dedicated circuit.

For code-safety conformance, a GFCI was originally installed in accordance with National Electric Code® (“NEC”) regulations to help protect against a possible shock hazard [5]. Since that time, updated NEC editions have continued to expand where GFCIs are mandated [6],[7]. While the NEC does not explicitly call out GFCIs for sewage, sump or effluent pumps, GFCIs typically reside in areas such as portions of unfinished basements and crawl spaces that mandate GFCIs regardless.

Permanently bypassing or eliminating the GFCI to avoid nuisance trips—as some freely suggest on numerous home-inspection and plumbing forums [8]—creates a vitally more serious condition: notably, a fatal shock potential from energized wastewater. Consider a scenario wherein a failed pump seal energizes accumulated wastewater that backs up into the waste drain while an occupant showers. With the occupant’s feet immersed in electrically-charged water, immediately upon touching a metal shower handle or other grounded surface, lethal currents could flow from the occupant’s feet up through the hands. Just tens of milliamps is enough to paralyze the body at the so-called let-go current, causing severe muscle contractions. At slightly higher levels in the 100 mA to 200 mA range, sympathetic heart fibrillation could cause cardiac arrest and probable death [9]. And while the pump ground wire should carry much of the leakage current in the event of a short, a wet body has very low resistance and could succumb to a burning or fatal shock. Moreover, with a pluggable (non-hardwired) power connection, the safety ground could become compromised.

In accordance with evolving UL 943 GFCI standards [10],[11], a typical residential Class A GFCI trips between 4 milliamps (“mA”) and 6 mA of differential leakage current. Trip time may vary between manufacturers and load; however, typical trip times may be as fast as 20 milliseconds (“ms”) to 25 ms, depending on the short-circuit load.

CHALLENGE

In order to assess whether a short circuit was present at the sewage pump, BUILDERA measured both the load current and ground-leakage current to determine whether the latter exceeded the UL-mandated 4 mA - 6 mA GFCI trip threshold. While measuring high operating current flow above 100 mA is easy and safe using an electrically isolated AC-current clamp meter, measuring ground leakage is more challenging because leakage currents may be as low as three orders of magnitude smaller (1,000 times less) than the load currents measured on the hot and neutral conductors flowing to/from an operating pump.

Moreover, since in this case the GFCI would trip every time the pump activated, measuring steady-state line current or ground-leakage current proved challenging.

SOLUTION

After careful consideration of the environment and risks—and for troubleshooting purposes only—one practical measurement solution was to temporarily bypass the GFCI-protected outlet and connect the sump pump to a nearby unprotected outlet via a short, heavy-duty extension cord. Note that bypassing a GFCI is never advisable in a permanent installation due to the risk of shock. Thus, BUILDERA performed this test with utmost caution and environmental awareness, and is not recommended for homeowners.

Using a FLUKE® 902 FC HVAC Clamp Meter, the pump supply-line AC RMS current measured 8.4 amps (“A”) @ 120 VAC (Figure 4). During normal operation, the ground-leakage current should be near zero milliamps, but in no case any greater than 0.50 mA RMS (60 Hz) for portable, grounded cord-pluggable equipment (ANSI/UL 101 Table 4.1). The Sump and Sewage Pump Manufacturers Association (“SSPMA”) also sets a pump-leakage standard of 0.5 mA RMS (60 Hz), which member manufacturers agree to uphold.

Such low-current measurements called for more specialized equipment—a leakage-current clamp meter—engineered specifically to help diagnose small leakage currents on the grounded conductor, as well as small imbalances between the hot and neutral conductors [12]. This second specialized clamp meter (FLUKE® 368 FC) measured and datalogged ground-leakage current before, during, and after the pump turned on and off. Note that the FLUKE® 368 FC also measures current in the Amps range, but BUILDERA opted to use two adjacent clamp meters in order to measure both load and ground-leakage current simultaneously.

Clamp Meter and Linesplitter Facilitate Safe Current Measurements

This more sophisticated leakage clamp meter has superior noise immunity and low current sensitivity in the microamp range. One microamp is one-thousandth of a milliamp, or a millionth of an amp. In order to avoid clamping directly to wires inside the outlet junction box, BUILDERA employed the HT Instruments LINESPLITTER of Figure 5. Compared to other basic line splitters, this improved device includes an extra loop for ground-leakage testing wherein the ground wire is isolated physically from the hot and neutral. Thus, any current on the ground path via the clamp meter indicates leakage current that could trip a GFCI.

In addition, the clamp meter can also measure the combination of the hot and neutral conductors (without the ground). Normally these currents should always cancel, resulting in a net current reading of 0 mA. However, any imbalance between the hot (supply) and neutral (return) indicates a lossy condition, wherein another load (or person) is shunting current to ground.

In order to capture transient data vs. time, the FLUKE® 368 FC also includes a bluetooth-enabled link to a mobile-phone app, in this case FLUKE CONNECT®. Other test and measurement manufacturers offer similar logging functionality and may be acceptable alternatives. In this case, the app wirelessly logs at a sustained refresh rate of approximately four (4) samples per second—based on the refresh rate of the clamp meter itself. As the internal memory-based data logger is limited to a minimum of one-second intervals, BUILDERA used the faster logging method via direct connection to the app. While an even faster logging rate would more precisely capture rapid transients, such as inrush current, this was adequate for our purposes. The measurement process was as follows:

MEASUREMENT PROCESS

The following procedure should only be performed by a qualified electrician or electrical engineer.

Empty Existing Wastewater

1. Using protective gear, plug the sump pump into a fully grounded (3-wire), non-GFCI 120 VAC receptacle. This assumes the pump is rated for a 15 A circuit. Do not use an ungrounded outlet. If a grounded outlet is not available, stop immediately to avoid a serious shock hazard.

2. If an extension cable is required to temporarily reach a separate outlet, use a cord rated for at least 15 A. Keep the cord length short (ideally less than 25 to 50 feet). Longer cables have higher capacitance, resulting in additive leakage current that may interfere with tests.

3. If the pump doesn’t start immediately, fill enough wastewater into the sump pit to activate the pump. Usually a few gallons is sufficient to raise the level high enough to activate the pump. If the pump starts but appears to be straining, unplug the pump immediately to avoid possible damage. The pump may be clogged with debris and require servicing.

4. Otherwise, allow the pump to run until the sump basin drains and the pump reverts to its normally-off state.

5. With the sump pit confirmed to be empty, measure the hot current and the leakage current using a line splitter with appropriate clamp meters. Both clamp meters should read near zero since no current should be flowing while the pump is off.

6. With both clamp meters turned on, set the line meter to Amps (“A”) range. Using the MAX-HOLD feature may also be helpful as data logging is not necessary for the line current. Connect the line clamp meter to the x1 loop on the line splitter.

7. Connect the leakage clamp meter to the isolated “ground” path on the line splitter. Configure the app to log data to the phone (in this case an iPhone, but it could also be Android).

8. With the data logger running, turn on water at a nearby tap or flush the toilet multiple times until enough water begins to fill the sump tank. After a minute or two, the water level in the sump tank should rise sufficiently to trigger the float switch, turning on the pump. This test may be repeated across the L-N loop, which should yield similar results to the ground-leakage test. Note that GFCIs do not trip on ground leakage per-se, but rather, on differential current between line and neutral, which is typically due to a ground short.

9. During this initial transition as the pump turns on, the clamp meter may over-range momentarily while the measurement stabilizes until the pump turns off—usually just after 5 to 15 seconds. Any leakage-current value sustained above 4 mA to 6 mA would cause a real GFCI to trip, thus shutting off the pump to prevent an electrical shock.

10. Note that the response time and sampling rate of the leakage clamp meter is too slow to capture the initial motor-current inrush. However, within a few samples it will measure any leakage detected on the ground conductor. Although the FLUKE® 368 FC meter has a mA range (with uA resolution), setting it to the “A” range may help to prevent transient overload upon initial pump activation, but at the expense of measurement resolution. BUILDERA has written to FLUKE that a user-fixable range of 300 mA would provide the best balance of range and resolution for this type of GFCI test.

RESULTS

Figure 7 shows that the leakage current reached an overload peak in the vicinity of 30 mA to 40 mA RMS and decayed to 20 mA RMS over several seconds until the pump completed its emptying cycle and shut off. The peak leakage current was at least six to eight times higher than the nominal 5 mA GFCI trip current that UL stipulates. This represented approximately a 3,000 ohm instantaneous shunt to ground (120 V / 40 mA), which increased to 6,000 ohms (120V / 20 mA) as the wastewater level declined.

Figure 8 shows the results after repeating the measurement, but this time with the GFCI inline with the sump pump. In this test, the leakage current surpassed at least 30 mA causing the meter to overrange. The leakage current decayed back to 0 mA after tripping tripping the GFCI. An audible click ensued as the GFCI circuitry disabled power to the pump.

This test confirmed that the GFCI in the original installation was working as expected by tripping quickly after detecting a current fault to ground that exceeded approximately 5 mA. This test also corroborated the original hypothesis that the pump switch or seal was likely compromised with moisture intrusion, although post-mortem analysis of the pump was not performed. Ground-leakage current was present only when the wastewater level was sufficiently high, and it returned to zero shortly after the GFCI tripped—before the pump could finish its emptying cycle. This test showed why bypassing a GFCI that started to trip is a bad idea. In fact, the tripping source in this case was due to a failing pump that presented a safety hazard. The only remedy was to repair or replace the pump.

Figure 9 shows a typical pump alarm that flashes and provides a loud audible noise were the wastewater level to reach a high level, indicating that, for whatever reason, the sewage pump had failed to empty the sump tank. In this particular installation, the alarm was also plugged into the same GFCI-protected outlet which posed an obvious concern: if the pump tripped the GFCI, it would concurrently disable the alarm, rendering the alarm useless. In practice, the pump should be on its own branch circuit, and the alarm on a separate circuit. Moreover, this installation did not include a secondary back-up pump, which as noted earlier, is advisable to guard against possible wastewater from flooding a finished basement.

CONCLUSION

This case study demonstrates the importance of using a GFCI-protected outlet for basement sewage and sump pumps. Nuisance or ghost tripping where none existed previously indicates either a failed GFCI (which should be replaced periodically) or a failing pump seal or insulation. In this example, the GFCI had not failed, but the the pump was beginning to show leakage between the hot line and ground. The GFCI did its job.

Ultimately, this was not a “nuisance” or “ghost” trip, but rather, a true ground short and indicator of potential shock hazard in the basement wastewater system. In the event of a sewage backup into a shower or onto the basement floor, the shock potential was very real requiring mitigation by inspecting and replacing the pump, as well as rechecking the leakage current during normal operation. After replacement, the leakage current should measure well below the 4 mA to 6 mA GFCI trip threshold, namely, no greater than 0.50 mA RMS @ 60 Hz for portable 3-wire equipment. In addition, given the age of the original GFCI, it should be permanently replaced with the latest model that includes the new UL self-test mode and end-of-life indicator. Regardless, GFCIs (even with the new automatic self-test mode) should be manually tested monthly for proper operation.

ACKNOWLEDGMENTS

BUILDERA is grateful to M. Klein for his cooperation in this case study and granting access to the premises for investigation and analysis of the sewage pump and GFCI tripping.

REFERENCES

All references below include active website links (URLs) that were tested and active on the date of original access. Links can move or change over time and may not be reflected in the following.

[1] L. Leyton. “Ground Fault Circuit Interrupters (E321).” PDH Online Course Notes. 2003. https://pdhonline.com/courses/e321/e321content.pdf (accessed Jan. 18, 2020).

[2 ] T. Scherer and K. Hellevang. “Electric Backup Sump Pumps for Houses - AE1771.” North Dakota State University Extension Services. Aug., 2015. https://www.ag.ndsu.edu/publications/home-farm/electric-backup-sump-pumps-for-houses/ae1771.pdf (accessed Jan. 18, 2020).

[3] T. Scherer. “Individual Home Sewage Treatment Systems - AE892.” North Dakota State University Extension Service. Oct., 2015. https://www.ag.ndsu.edu/publications/home-farm/individual-home-sewage-treatment-systems/ae892.pdf (accessed Jan. 18, 2020).

[4] Zoeller Company. “Model 266 and 267 Submersible Sewage, Effluent or Dewatering Pumps.” FM3121 Sales Slick. 2017. https://www.zoellerpumps.com/content/literature/FM3121-Sales-Slick-266267.pdf (accessed Jan. 18, 2020).

[5] National Fire Protection Association (NFPA). “NFPA 70® National Electrical Code® (2017 Edition).” https://www.NFPA.org (accessed Jan. 18, 2020).

[6] L. Park and A. Park. “New GFCI Code Changes in the 2017 National Electric Code (NEC).” Home Inspection Professionals, Inc. http://www.homeinspectionpro.com/news/new-gfci-code-changes-in-the-2017-national-electric-code-nec/ (accessed Jan. 18, 2020).

[7] T. Domitrovich. “For Safety’s Sake: NEC 2020 Increases GFCI Protection.” EATON Blog Series. May 29, 2019. https://www.eaton.com/us/en-us/company/news-insights/for-safetys-sake-blog/NEC-2020-increases-GFCI-protection.html (accessed Jan. 18, 2020).

[8] International Association of Certified Home Inspectors. “Sump Pump Dedicated Outlet GFCI.” https://forum.nachi.org/t/sump-pump-dedicated-outlet-gfci/55736 (accessed Jan. 18, 2020).

[9] C. Ferguson. “Electric Shock and the Human Body. JMK Engineering.” Jun. 13, 2016. https://jmkengineering.com/electric-shock-human-body/ (accessed Jan. 18, 2020).

[10] Underwriter Laboratories. “UL Standard for Safety for Ground-Fault Circuit Interrupters, UL 943.” Fourth Edition. Feb. 1, 2006. https://standardscatalog.ul.com/standards/en/standard_943_5 (accessed Jan. 18, 2020). This has been superseded by an updated UL specification, however, it provides useful historical context.

[11] N. El-Sherif. “UL’s New GFCI Classes - Consulting - Specifying Engineer.” Consulting Specifying Engineer Magazine. Jan. 6, 2014. https://www.csemag.com/articles/uls-new-gfci-classes/ (accessed Jan. 18, 2020).

[12] Fluke Corporation. “Chasing Ghost Trips in GFCI Protected Circuits.” FLUKE Blog. Jan. 1, 2019. https://www.fluke.com/en-us/learn/blog/grounding/chasing-ghost-trips-in-gfci-protected-circuits (accessed Jan. 18, 2020).

[13] HT Instruments. “LINESPLITTER Mains Plug L-N-PE Splitter Conductors Accessory.” Datasheet. May 30, 2017. https://www.ht-instruments.us/en-us/products/linesplitter/download/datasheet/ (accessed Jan. 18, 2020).

Other Helpful Reading

D. Chandler. “Understanding GFCI Nuisances: Why is my GFCI Tripping?” Henderson Headlines. Jan. 24, 2019. https://www.hendersonengineers.com/insight_article/understanding-gfci-nuisances/ (accessed Jan. 18, 2020).

Energy Education Council. “Ground Fault Circuit Interrupters (GFCIs).” https://safeelectricity.org/ground-fault-circuit-interrupters-gfcis/ (accessed Jan. 18, 2020).

Western Automation. “Ground Fault Circuit Interrupters for AC & DC Systems.” Article. Mar., 2016. https://www.westernautomation.com/wp-content/uploads/2016/03/GFCIs_For_AC_DC_Systems.pdf (accessed Jan. 18, 2020).

IAEI Magazine. “UL Question Corner.” https://iaeimagazine.org/magazine/columns/ul-question-corner/ul-question-corner-leakage-current/ (accessed Jan. 20, 2020)

SSPMA. “Recommended Standards for Sump, Effluent, and Sewage Pumps.” 2013. http://www.sspma.org/uploads/8/3/9/2/8392851/recommended_standards_2013.pdf (accessed Jan. 20, 2020).

Underwriters Laboratories Inc. (Dec. 30, 2019). “ANSI/UL 101-2019 Standard For Safety For Leakage Current For Utilization Equipment And Scope To Reflect That The Impact And Application Of The Standard Includes Other Products As Well As Appliances.” https://www.shopulstandards.com/ProductDetail.aspx?UniqueKey=33159 (accessed Jan. 22, 2020).

Underwriters Laboratories Inc. (Jul 07, 2016). “UL 778 Standard for Motor-Operated Water Pumps.” Edition 6. https://www.shopulstandards.com/ProductDetail.aspx?UniqueKey=31321 (accessed Jan. 22, 2020).

R.C. Worst & Company, Inc. “How to Test Wire Using a Megohmmeter.” https://www.youtube.com/watch?v=YC2jptaU0qc (accessed Feb. 05, 2020).

ATTRIBUTION

G. Lowitz. “Diagnosing GFCI Nuisance Tripping from Sewage-Pump Ground-Leakage Current.” BUILDERA Case Study CS2020-1. https://www.buildera.com/case-study-gfci-tripping-sewage-pump-ground-leakage-current

SEARCH TAGS

GFCI, ground-fault circuit interrupter, RCD, nuisance tripping, ghost tripping, sewage ejector pump, effluent pump, sump pump, lift pump, building code, electrical code, clamp meter, 5 mA, shock hazard, electrocution, UL 943.

COPYRIGHT

©2020 BUILDERA. All rights reserved. Please contact BUILDERA for licensing options, custom derivative works, and attribution to reproduce electronically or in print.