Generator GFCI trouble

[sparkydave]

I intend to use this generator for ham radio and emergencies, but maybe some RVers can shed some light since I'm scratching my head.

New Predator 4000 generator. Has two GFCI outlets on it. Plug in halogen work lights, works great. Plug in the laptop computer, works great. Plug in a surge protected outlet strip, works great. Plug the laptop into the surge protector while it's plugged into the generator, and it will trip the GFCI (either outlet). Plug the same surge protector and laptop combination into the GFCIs in the house, and there's no problem. Only on the generator, and only when the laptop and surge protector are combined. A 3-light GFCI outlet tester shows no problem with the outlets, and pressing the test button trips them.

Generator is not tied to earth ground (yet), could that be the problem? GFCIs don't need a ground to work, and I've not heard of them not working properly without one.

[westom]

This problem is quite obvious once the nature of so many generators is learned. A 120 volt generator can make 200 volt spikes. That is ideal power for 120 volts electronics. And problematic for near zero joule power strip protectors.

That 200 volt spike can mean current outgoing on the generators hot is connecting to safety ground via lightly triggered MOVs.

Furthermore, those power strips must never be used where 120 volts might create tens of thousands of tiny (less than 330 volt) spikes all day. MOVs are only designed for one or a few large surges. Quickly degraded to ineffective when confronted by constant noise spikes.

Degradation cannot be reported by the "protector good" light.

Anything that power strip protector might do is routinely done better inside electronics. View numbers. How many joules does that power strip claim to absorb? Hundreds? Electronics routinely convert a hundreds or thousand joule surge into rock stable, low DC voltages. Surges are safely consumed to power semiconductors. Same surge can destroy that near zero joule (with massive profit) protector strip.

GFCI trips if the current outgoing on the hot prong is not exactly same as current incoming on the neutral prong. If that current leaks elsewhere via power strip MOVs, then GFCI must trip.

Best power strip has no protector parts and always has a 15 amp circuit breaker.

Other reasons inside a power strip can explain that trip.

Meanwhile build a diagnostic tool: a simple GFCI tester. A 20K adjustable resistor in series with some fixed 1/2 watt resistors connects from hot to safety ground. Resistor is adjusted so that 3 milliamps (hot to ground) cause no trips and 6 milliamps (or more) causes a trip. Now a tester can report what current actually causes a trip. And even if the GFCI is working properly. Those store bought GFCI testers do not provide the same adjustable test and perspective.

If that tester with power strip causes a GFCI trip (when resistor is only at a 3 milliamp setting), then you know how much current is leaking through that power strip. Do same test only with a laptop and that tester. Apparently power strip leaks much current - but not enough to trip a GFCI by itself.

Finally, some GFCIs are also more likely to trip (more sensitive) when AC noise exists. Also get a line filter (that is normally found inside all electronics) from electronics supply house (ie Digikey, Mouser, Allied, Jameco, etc). Build a little box with that filter located sit between the generator and appliance. Does that filter stop GFCI trips? Then know its GFCI is also noise sensitive. And know which appliances are creating too much AC noise. That line filter is also a powerful diagnostic tool.

[road-runner]

Very few of the many web sites that explain gfci operation mention the downstream bonding detection. This results in people who due responsible due diligence feeling very confident of the incorrect information. I was recently a similar fool trying to understand workman's comp insurance. I finally found out there are 2 sets of rules, and the set featured on most web sites is the one "nobody" uses.

[wnjj]

Maybe he meant it won’t trip a downstream GFCI on a bonded generator? That is the case for generators that do not have a built in GFCI. The OP’s generator does and it will indeed trip if you try to bond it (plus he says it’s already bonded inside).

[Bobbo]

ScottG wrote:

DrewE wrote:

enblethen wrote:
Doesn't need earth ground. You need to lie to the genset by installing a dummy plug into it's receptacle with a jumper between ground and the neutral.

I don't think that would help in this case. In fact, it should trip the GFCI on the generator to tie the neutral and ground together by plugging in a bonding plug since that would be a neutral/ground fault after the GFCI (and modern GFCIs have circuitry to detect that kind of fault even in the absence of significant fault current).

It seems that there's something about the combination that is fooling the GFCI. I'd probably lose the surge suppressor and just use a plain outlet strip or multi-tap if needed. There shouldn't be much chance for big surges from a generator if you aren't powering any big inductive loads or having a large power distribution network.

I'm sorry but that is incorrect; It will NOT trip a GFCI to bond the neutral and ground.

This is just plain wrong. I have a bonding plug for my Honda generator since I have a PI EMS-HW30C. If you want to drive to my house, I will let you plug it into a GFCI outlet and see what happens.

[ktmrfs]

ScottG wrote:

DrewE wrote:

enblethen wrote:
Doesn't need earth ground. You need to lie to the genset by installing a dummy plug into it's receptacle with a jumper between ground and the neutral.

I don't think that would help in this case. In fact, it should trip the GFCI on the generator to tie the neutral and ground together by plugging in a bonding plug since that would be a neutral/ground fault after the GFCI (and modern GFCIs have circuitry to detect that kind of fault even in the absence of significant fault current).

It seems that there's something about the combination that is fooling the GFCI. I'd probably lose the surge suppressor and just use a plain outlet strip or multi-tap if needed. There shouldn't be much chance for big surges from a generator if you aren't powering any big inductive loads or having a large power distribution network.

I'm sorry but that is incorrect; It will NOT trip a GFCI to bond the neutral and ground.

a neutral/ground bond AFTER (downstream) of a GFCI WILL trip the GFCI. they will trip on either a hot/neutral imbalance OR a ground neutral bond.

[time2roll]

If power returns through the ground it will be an imbalance between hot and neutral.
5 mA and GFCI clicks off.

[jkwilson]

ScottG wrote:


I'm sorry but that is incorrect; It will NOT trip a GFCI to bond the neutral and ground.

Modern GFCIs do detect neutral ground faults via a second transformer.

[MrWizard]

If you need the power strip for more gear

You could try a small isolation transformer between the laptop supply and the surge strip

[ScottG]

DrewE wrote:

enblethen wrote:
Doesn't need earth ground. You need to lie to the genset by installing a dummy plug into it's receptacle with a jumper between ground and the neutral.

I don't think that would help in this case. In fact, it should trip the GFCI on the generator to tie the neutral and ground together by plugging in a bonding plug since that would be a neutral/ground fault after the GFCI (and modern GFCIs have circuitry to detect that kind of fault even in the absence of significant fault current).

It seems that there's something about the combination that is fooling the GFCI. I'd probably lose the surge suppressor and just use a plain outlet strip or multi-tap if needed. There shouldn't be much chance for big surges from a generator if you aren't powering any big inductive loads or having a large power distribution network.

I'm sorry but that is incorrect; It will NOT trip a GFCI to bond the neutral and ground.

[sparkydave]

Fair enough, I figured the surge protector probably had MOVs clamped to ground, so between that and the laptop's power supply there may have been enough leakage to ground to trip it.

Prevailing wisdom was to plug ham radio gear and computers into a genny through a surge protector, in the event something goes wrong or it turns out RFI drives the generator's AVR berserk. Guess I'll take my chances, or unplug the laptop before I key up the radio the first time. Other plan is use an adapter with the twist-lock outlet that doesn't have a GFCI.

Thanks!

[road-runner]

I'd guess that the combination of the outlet strip and laptop produces enough leakage to the ground wire to trip the gfci. Why doesn't it trip the home gfci? Easy answer: The generator gfci has a lower trip threshold than the home gfci. More likely answer: Not being an inverter generator, the sine wave output waveform is likely distorted, resulting in the laptop and surge protector's leakage to exceed the gfci trip threshold. Forget the bonding plug. If the neutral wasn't already bonded, the gfci would never trip.

[time2roll]

Don't use the surge protector.

[DrewE]

enblethen wrote:
Doesn't need earth ground. You need to lie to the genset by installing a dummy plug into it's receptacle with a jumper between ground and the neutral.

I don't think that would help in this case. In fact, it should trip the GFCI on the generator to tie the neutral and ground together by plugging in a bonding plug since that would be a neutral/ground fault after the GFCI (and modern GFCIs have circuitry to detect that kind of fault even in the absence of significant fault current).

It seems that there's something about the combination that is fooling the GFCI. I'd probably lose the surge suppressor and just use a plain outlet strip or multi-tap if needed. There shouldn't be much chance for big surges from a generator if you aren't powering any big inductive loads or having a large power distribution network.

[sparkydave]

enblethen wrote:
Doesn't need earth ground. You need to lie to the genset by installing a dummy plug into it's receptacle with a jumper between ground and the neutral.

Actually, the generator already has neutral and ground joined behind the panel. The outlet tester shows this and I confirmed with an ohmmeter.