Kohler 12RES Generator tripping GFCI outlets - solved at last!

[Keithvcollins]

@pacifico & @clivusmul are either of you located on a large body of water, on sandy soil, or near a large radar installation? A dock power? I have zigbee, zwave, and wifi devices managed by Smartthings hub and Alexa devices. Most of my zigbee cans are Commercial Electric and will glow if there is >.6v of ghost voltage on that circuit. The other cans are Juno zigbee or Cree Wi-Fi LED floods. Most of the dimmer switches are ZooZ zwave mofset 3way dimmers. 3 ModernForms Wi-Fi DC fans. 2 high efficiency Carrier heatpumps with variable speed fans (VFDs). An elevator with a small APC battery backup, an automated gate and 3 automated doors (all are small dc motors). 15 gfi outlets (6 on isolated circuits with no load) and 2 gfi breakers. The breakers rarely trip. No issues with any of the ARC breakers.
The problem only occurs under low load (<5amps) with high THD > 16%. The gfi’s do not start tripping until about 10 minutes into a test. If I turn on linear loads like the dryer or heat strips for a test no problems will occur. Looking for similarities in environmental conditions, especially because the GFI’s that trip are random and inconsistent across tests. Most of the times with when the gfi outlets trip they continue to blink and cannot be reset until I return to carrier power. Sometimes the relay in gfi outlet will click multiple times before finally tripping. This occurs regardless of brand.

I have run 5 loaded test with the capacitors in place with no issues and one test with the capacitors removed had issues. So even though it looks like I may have a solution I would still like to know exactly what the issue is. My guess is that the Kohler software for the voltage regulator is responding to aggressively to the harmonics under low load conditions and that the capacitors filter out enough of the higher harmonics to keep the regulator from reacting.

[clivusmul]

Keith,

I am located in central AZ at 6100ft, Ponderosa pine forest, on a former pioneer homestead... Hard ground with tons of surface rocks. Water table is about 20ft below the surface at most horizontal pasture levels below my house although my home is uphill from the pasture by about 50 ft... No radar anywhere nearby.

I'm pretty sure the problem is mostly harmonic distortion generated by the Kohler (or other synchronous machines). Nothing is 100% repeatable with respect to diagnosis... Some times with light loads and running for less than 15 minutes, there are no nuisance GFI trips.. Other times with higher AC loads, it seems the GFIs will trip inside of 10 minutes. Long run times with small loads can also see nuisance tripping. There could be a ambient temperature correlation..??

My exterior tripping breakers are all Leviton sited in waterproof enclosures and all have NO AC loads connected. The interior GFIs (kitchen and bathrooms) likewise have nothing connected to their AC receptacles.

My just installed EMI filter seems to help but clearly does not solve all the issues experienced. I know that one problem is the location of my AC noise filters (there are 2ea installed with each filtering one leg of the 240V AC phases to Neutral). Each filter has a Line in and a common in and a Line out and a common out plus a Ground terminal which is designed to take the filtered currents to ground and not allow the noise to pass thru to the outputs of the filters. Anyway, my installation of these filters is close to a sub-panel in my hill barn. This sub-panel is about 35 ft distant from the generator - not ideal as it should be co-located at the generator or located close to the Inverter which in my case is another 75 feet distant from the Hill barn inside my cabin's basement. I may try moving the filter either next to the generator or close to the inverter but as is, it's not ideal for high frequency noise filtering...

My generator is old with about 3900 running hours. I'm sure the rotor slip-rings are arcing more now than when new and that will produce unwanted noise. Some slip-ring brush maintainence couldn't hurt. The generator throttle/RPM control is governed by a electronic gear tooth counter working with a stepper motor so the frequency is very good; usually within 0.1 hz of 60 cps. I've never seen any abnormal voltage excursions and the Schneider does a good job of "qualifying" the AC both prior to switching ON and during all running events. It looks at voltage and frequency but has no means to determine ThD coming in from the generator.

There does appear to be some interactions with the inverter. and generator in combo... I ran the generator for 30 minutes this AM with the inverter totally bypassed from the house load and there were no GFI false trippings noticed during this test... One issue I did observe was that my Schneider Inverter, even in AC Bypass mode will continue to charge the batteries unless you open the AC input breaker at the PDP panel or open the 250 Amp DC breaker going to the battery bank or, shut it totally off using its control panel. The inverter is not internally bypassing AC via its internal Xfer switch because the AC out breaker is opened but the AC in remains applied and the unit seeing AC generator voltage moves to Charger mode and will go into battery charge mode all the while, in parallel, the generator is going around the inverter directly to the loads.

Apologize in advance for the lengthy reporting...
bob

[Keithvcollins]

The run capacitors are wired in parallel directly to each breaker for each leg and to the common inside the generator housing. They are
Dayton
2MEC7D
20uF ‡5%
370 VAC (VCA) 60/50H2
Rated to 85 °C Max Temp
Protected 10,000AFC
NON PCB

[clivusmul]

Keith, it's a bit hard to see the photo exactly but it looks like your generator is "hard-bonded" Neutral to earth ground per the large White wire and the green lead and I see two caps on the right side (sorta background) that look to have one of two leads each to the L2 or L1 breaker node and the other one going to common Blue buss bar at the top...
Understand this is inside the generator breaker box too... correct?

bob

[Keithvcollins]

I had the same response the first time I saw the white wire to ground as well. When we first started trouble shooting the issue the installer decided to ground the 110 from the house that charges the battery in the generator. He didn’t have a green wire so he used a piece of white wire. The generator ground is bonded only to the buss bar in the transfer panel. The only earth ground is from the meter box which is bonded to the two transfer panels.

You are correct that L1 connects to one capacitor and L2 connects to the other capacitor. Each capacitor then connects to the blue common buss bar. This connection is directly at the breaker inside the generator. Basically as close to the source as we could get.

[clivusmul]

Got it.. Thanks!

[Pacifico]

@pacifico & @clivusmul are either of you located on a large body of water, on sandy soil, or near a large radar installation? A dock power?

Well, I'm sort of "close" to a large body of water known here on the west coast as the Great Pacific Firebreak. Being 3 miles distant, I don't think it's a factor.

The soil is rocky clay, clayey gravel and variably decomposed argillite atop bedrock of the 160 million year old Late Jurassic Otter Point Formation according to my geological report. There're no radar installations in the area; we're pretty much out in the boonies and off-grid.

I'm looking forward to going over your solution involving these run capacitors with my solar installer who has been out of town until yesterday. None of these issues occurred when I used my Honda EU6500i generator; it was only after the Kohler 12RES was installed that these issued popped up and then, only when the Kohler generator is used. When I switch back to the Honda, the problems disappear.

[clivusmul]

Keith, I'm curious why you/your electrician choose to install 20uFd non-polar caps across your Kohler main breaker? I clearly understand/get-it that we want to suppress the slip-ring and/or brush noise that is being generated by the Kohler but why exactly was 20uFd chosen? Would 15uFd or 10uFd have worked just as well? or, perhaps the electrician had two spare 20's on his truck? I'm going to go up next weekend and install a pair of 15uFd caps at the same location on my Kohler. I have already installed a pair of series pass EMI filters on each generator phase but as prior mentioned, they fail to resolve all the nuisance tripping experienced... I'm thinking the caps at the generator breaker in addition to the EMI filters should do the job... If my 15uFD selection fails to work, I can always make them bigger...

I'm not sure my generator's breaker box has room inside for the two large caps but have bought some hardware tape and grommets to allow the caps to be mounted external to the box and have the connecting wires enter via the grommet'd holes I will drill (if interior box room is the challenge I think it will be...).

In the RF world (ham radio, etc), large caps can look inductive to higher frequencies due to lead length and general construction realities... We would often shunt a larger electrolytic cap with a smaller ceramic capacitor to handle the really high frequencies...

Based on all the discussion, I remain convinced the nuisance tripping GFI breaker issue is being caused by noise on the AC mains stemming from the generator...

[Keithvcollins]

I’ve tried to get a response from the installer why/how he settled on 20uF. I know he ordered them specifically. I’ve asked multiple times if Kohler participated in the sizing and all I get are crickets now that the issue appears to be resolved.

I am happy to report that with this last hurricane we lost power for 4 hours and had zero issues. I still have the fluke35b so I hooked it up and just let the scope record. I’ll attach a picture. It includes the switch back to carrier power at the end. Because my load is low <5a on the leg measured you can see the impact of the 1.8a load created by the 20uF. One take away is that any consistent load might also resolve the issue. In fact one of the guys at Outback said he had wired incandescent light bulbs on each leg to solve odd problems before.

[Keithvcollins]

[Keithvcollins]

The response from my inquiry about the sizing: “The generator is in its stock form no adjustment have been made to it the capacitors don’t change anything about the generator they stop the noise coming back from the house this proves that the light switches and other devises are what causing the problem Kohler suggested a permanent solution is to call the manufacturer of all the lights and switches and appliances that can cause this problem and buy filter kits from them”

Hmm…. Not sure Kohler is understanding the scope of the issue they face as more customers add inverters and home automation.

[clivusmul]

Keith, the response above from Kohler is ridiculous...

Our Kohler generators have never been the recipient "fail point" of a noise problem stemming from some home appliance, that in my case, have never been activated. My nuisance GFI breaker tripping occurs, over sufficient generator running time, with no appliance being active in the house. Your "inverter mode" Honda generator used as a source of power never causes a GFI nuisance trip. To date, only the Kohler units are a part of the problem definition.

I doubt there is any high frequency EMI filtering designed into the Kohler RPM/Vout control regulator or its pre-breaker connections. My unit has a gear tooth RPM counter that controls a throttle stepper-motor. This unit maintains constant RPM under varying loads... It is inherently slow to regulate RPM which is a good thing as it remains mechanically "overdamped" in terms of negative RPM feedback; it can not act too quickly... The voltage regulator controls the unit's Field current to maintain constant Vout under varying loads... The two individually or together need some form of closed-loop compensation to avoid RPM 'hunting' or undamped output voltage excursions but all synchronous machines need these controls... Kohler's design does both things well but it ignores high freq. noise and brush hash that is unavoidable in these types of machines (especially as they age...) unless a EMI filter is installed at the unit's breaker or immediately upstream from this breaker... I have the schematic for my 10RY unit and it shows no internal EMI filtering at the primary unit output...

My guess is 20uFD is a good compromise between a effective lowpass filter that isn't too "loading" yet provides sufficient filtering to resolve your/my noise induced nuisance tripping of unloaded breakers... A 10kWh to 24kWh AC standby generator has no issues driving a extra 1 amp parasitic load on each of the output phase legs; my unit as is includes 3ea, 50 amp breakers... Larger values are no more effective and result in extra and unnecessary parasitic power loading and smaller capacitor values may lack sufficient filtering to be effective. My guess is any cap between 5uFd and 30uFd will do the job based on individual unit's output noise signature... I think really large caps will just get hot and not filter any more effectively than your installed 20ufd's...

[clivusmul]

Keith -

I just performed my first test run following installation of 2ea, 15uFd (460V AC rated) non-polar AC motor starting capacitors.

I used spade connectors to the caps and ring terminals on the other end of 4ea, 8 to 10 inch long #10 stranded copper wire/s.. This wire is overkill in terms of diameter but it was handy and well supports the two capacitors hanging under the breaker box... The Capacitor "stack" lies directly under the Kohler breaker box and feeds thru slots cut into the bottom of the breaker box to increase air circulation. The other end of the cap wires goes directly to the output lugs of the Kohler 50 amp primary output breaker.

I did notice one "interesting" item that I believe has zero to due with nuisance tripping of home GFI breakers... The Kohler dealer installed a 115V AC weatherproof power outlet on the right side of the breaker box and connected it to 1-phase of the un-breakered generator output lugs... So, it's entirely possible to open the breaker and still power this 115V receptacle. I would have connected this ancillary receptacle to the breakered side of the unit so when the breaker is OPEN, all power is off to the external loads... Anyway, nothing is plugged into this outlet and has rarely ever been!

First test run result was stable and there were no nuisance trips of my many GFI home breakers. It's too soon to claim victory! I failed to check for the normal ceiling fixture light initial ON flicker which I always see prior to today... I did not see any flicker but may have failed to view sufficiently quickly after Kohler activation... I'll do another few test runs later this afternoon and report. My batteries were nearly full at generator activation so I reached full charge/float in about 10 minutes and then shut down the Kohler...

Bob

[Keithvcollins]

@pacifico and @clivusmul any more updates?

[clivusmul]

With the 2ea, 15 ufd caps installed, I have run the Kohler twice without any nuisance GFCi tripping of unloaded outlets. I plan to incrementally raise battery charging current from 80 to 120 amps & note performance & will report results.
Bobj

[Pacifico]

@pacifico and @clivusmul any more updates?

I have an appointment with my system installer on Monday to go over the possible solutions presented here. He has been very responsive about trying to solve this problem but ran out of ideas until you came in, reporting on your experience and solution.

It was easy to put this project on the back burner as summer time generator demand is zero. I've been preoccupied with other projects and my system installer has been busy as well. With winter coming, We're both motivated to move priority of this project up to the top of the list.

I will definitely be reporting back on progress as it occurs. I appreciate both of you taking time to detail your experience. It gives me hope!

[Keithvcollins]

@pacifico any luck with the capacitors yet?
@clivusmul any update on additional test runs?
Would you two mind sharing info on your installer/kohler dealer you are working with. I am working with R&T Power out of Morehead NC. Once we have all our results I’d like to see us summarize together to get it back to Kohler. I think it would really help if Kohler dealers had this as part of their standard protocol if they encounter nuisance tripping.

[clivusmul]

Keith,
Similar to your situation, our weather has been mostly warm and cloud clear so my solar panels are keeping my battery bank easily charged to 100% daily... I have performed 4 or 5 test runs lasting 10 to 15 minutes with zero nuisance GFI breaker tripping observed. This is a very good sign however, my results won't be confirming until I have run the generator at high output power for 2 hours multiple times. This mandates I start with my battery bank at say, 50% SOC which is difficult to do... I need a few cloudy days for this to be a opportunity. I can easily set the battery charge amps to 100 but this runtime is quite short given the high SOC starting state of the charger. After the battery becomes full, the XW-Pro reverts to Float or "Idle" and I have a lightly loaded generator providing the static load needed by my cabin which is only 250 watts or so...
I do feel confident in the following scenario - I think the Leviton GFIC breakers have a sensitivity to the noise signature emanating from the Kohler (non-brushless) generator. I believe our "consumer grade" GFIs have a frequency "band" of applied noise it tends to rectify on the AC mains (its primary connection) and somehow direct this rectified signal to ground via a internal capacitor or diode that causes the imbalance in leg current which, in turn, trips the breaker. These residential GFI breakers are extremely sensitive to load current imbalance and will trip at a fraction of a milliamp of imbalanced current... I further think the Kohler synchronous machine design is a legacy unit lacking noise filtering at the machine's output which has never been a problem until the last 10 or so years when building codes began to mandate GFI breakers at many locations in new home constructions...
So IMHO, it's a perfect storm of a legacy, i.e., "non-brushless" generator producing noise that has not been a problem for older home installations (those lacking multiple GFI breakers), coupled with new home builds having multiple GFI breakers exposed to this noise produced by the legacy generator. Kohler could likely fix this problem by adding similar bypass caps at the generator's main breaker panel at the generator's output - just like you and I did...
Kohler's problem will be adding large bi-polar filter caps to a generator that's already in some state of high volume production. I would bet their new "brushless" models won't show this noise tripping sensitivity to load GFI breakers so they will likely decline to modify existing inventory of machines and the new models don't need the additional filtering... Some form of field note acknowledging the problem and a suggested field fix would be a good thing to issue however... Bob

[Pacifico]

@pacifico any luck with the capacitors yet?
@clivusmul any update on additional test runs?
Would you two mind sharing info on your installer/kohler dealer you are working with. I am working with R&T Power out of Morehead NC. Once we have all our results I’d like to see us summarize together to get it back to Kohler. I think it would really help if Kohler dealers had this as part of their standard protocol if they encounter nuisance tripping.

My installer is eager as well to try this fix and see what happens. I'm out of town at present but will be coordinating schedules sometime this week. I agree that putting our collective experience together will be good for everyone. I'm very much looking forward to installing the filters and then report back on what happens.

[clivusmul]

A summary - Two of us, using Kohler's legacy standby genetators, have each installed a pair of high-voltage, non-polar, 15 to 20 uFd oil filled motor starting capacitors at the generator's primary output breaker (L1 & L2 to N) and, to date, have experienced no further distributed house receptacle GFCi breaker nuisance tripping.

In all cases before the caps were installed, various GFI Breakers, all without loads, and over various exposure times would trip offline... This appears to now be resolved.
Rgds, Bob (ClivusM)

[Keithvcollins]

Bob, just wanted say my capacitors are run capacitors.

[clivusmul]

Copy Kevin... The distinction is name only.. Induction motors use both Run &/or Start capacitors (based on hp & use) to create winding phase shift to produce both run rotation & starting torque boost... If the motor is designed to run using a 20 uFd "run" capacitor, a 20 uFd "start labeled" capacitor serves equally well.
Often the AC/induction motor needs both (Air conditioning compressor is 1 good example) caps to work & are often both contained in 1ea metal can housing... 2 caps in 1 housing of same or different values... The motor usually uses a mechanical relay to provide the switching engaged or disengaged by rotation.

[Pacifico]

Keith or Bob, any insight on installation location of the run capacitors? The two options are at the generator like Keith did or at the Outback load center where the generator power comes in to the system and where there is a little more room to work. I thought I saw mentioned that closest to the generator was best but after reading through all the posts, didn't find reference to this specific question.

I have two run capacitors on order (Dayton model 2MEC7 that Keith used) and my installer can come about October 23rd. I'm looking forward to getting these installed and reporting back with the results.

[Keithvcollins]

My installation has the capacitors wired directly to the breaker inside the Kohler generator enclosure. From bobs post on the 8th you can see the description of his install. Also wired directly to the breaker at the generator.

[clivusmul]

Correct.. Close as possible to the generator's outout breaker enclosure. Keith's are inside the breaker enclosure; mine are directly underneath & connected using short wires passing through the enclosure's vent slots...
Bj/Clivus