Trace Inverters and GFI circuits - Info and question

[rv_safetyman]

This subject came up about 6 months ago, but I wanted to start a new thread.

Trace full sinewave inverters (both 12V and 24V) often trip GFI circuits.  This could also be the case with other inverters as well.  Most campgrounds will take one look at a converted bus and say that we wired it wrong!!  Several folks have had this problem and Gumpy had the solution.  More about that later.

At our Converted Coach pre-rally this past week several folks from Magnum Inverters gave a great presentation.  Recall that Magnum has many of the old Trace folks.  One of the people was a really nice young engineer.  They fully acknowledged that the inverter will often trip a GFI because of a filter in the system that causes a slight "leakage" to ground (I hope the electronic folks will take pity on me for my simplistic description - Sean are you listening? ).  Sean will appreciate the previous statement, but for those of you who don't know the inside joke, go to the Sean and Louise's blog (http://ourodyssey.blogspot.com/) and look for Louise's 2/20 post

The engineer's solution to the problem was the same as Gumpy's (he called me with the solution, but may not have posted it).  The solution is to simply disconnect the ground in the inverter for a few seconds while plugging the unit in.  Supposedly, the manual describes a special relay to do the same thing.

After the presentation, several folks talked about a "simple workaround".  They made a special short 120 V extension cord with the ground tip cut off.  They then make a ground circuit with a rod driven into the ground.  That sounds a bit dangerous to me.  However, making a short cord with a momentary push switch in it might make sense.

So the solution is to disconnect the ground, but the question is how?  What is the best way?  If a switch is installed, I would think it should be a momentary switch so that the ground is re-established after the filter is no longer a problem.  Is there any current rating consideration for the switch?

BTW, it sure looks like the ex-Trace folks have done a great job of designing the magnum inverter.  I was very impressed.  They seem to be very customer oriented.  Their full sign wave unit can be fully field stripped and repaired easily if necessary.

Jim

[gumpy]

Best way to handle it is to remove the green wire inside the case, and install a switch somewhere nearby that you then connect the green wire to one side, and the other side back into the case where the green wire was originally attached. Clear as mud?

The switch should be at least 30 amp, better if 50 amp.

It should not be momentary, because your arms are not long enough to hold the switch while you plug the shore cord into the outlet which is 15 feet away, and Pat certainly doesn't want to help you plug your bus in every darn time!    Make it a SPST type switch, shut it off, plug in, and turn it on again. it's just that simple. Once that filter is charged up, you can turn the switch back on with no ill effects to the GFCI, and you're ground protection is restrored.

craig

[belfert]

Has anyone found this to be an issue with Prosine inverters?  I'm suprised the inverter manufactuers haven't done anything about GFCIs since they have been used on outdoor 15 amp outlets since about 1980.  I suppose it doesn't matter to most campers as they are pluuging into 30 or 50 amp receptacles not GFCI protected.

[edroelle]

Jim and Craig,

Unfortunately, disconnecting the ground momentarily did not work on either of my 2 conversions.  On my first conversion, I had a Xantrex 4024 Plus.  My present conversion has a pair of RV3012 inverters.

My solution was to use an isolation transformer, thus eliminating the ground fault interrupt from the coach circuits - about $150 on Ebay.

Ed Roelle
Flint, MI

[gumpy]

Unfortunately, disconnecting the ground momentarily did not work on either of my 2 conversions.  On my first conversion, I had a Xantrex 4024 Plus.  My present conversion has a pair of RV3012 inverters.

Ed,

That's odd. I did this on a Freedom Marine 2500 watt unit and it worked perfectly. I haven't actually installed the switch yet on my SW4024 yet, but I'm pretty sure Sean did it on his. Wonder what's different about the 4024 plus.

Do you think there might have been a ground fault in one of your appliances or somewhere else in the coach that was causing the tripping.

Was that 4024 in your MC8?

craig

[edroelle]

Craig,

Yes, it was my MCI 8.   I don't think there was a ground fault in one of the appliance, but I could be wrong.   I switched all coach breakers off, and still was not able to use Jim's approach.   I did have a complicated circuit with an automatic transfer switch though.    Since I could not identify a problem, or trick the system, I used the isolation transformer successfully.

Your results may differ with a different circuit design.

Ed

[gumpy]

Well, I don't know anything about the 4024 plus, so maybe they changed something in it.

I can't figure out why they haven't figured out how to fix the problem, though. They've known about it for several years.

[H3Jim]

Not to mention Trace not providing an easy solution to the grounding problem.

[Sean]

Since my name's been bandied about, I'll chime in.

In the Xantrex/Trace SW series, the culprit is a ground wire for a filter.  Trace has conveniently made this connection with a spade lug, for easy removal.  Note that this is NOT the ground for the entire inverter -- the chassis ground is still intact if this wire is removed.  Also note that the wire is, IIRC, 14-gauge -- it's not intended to carry as much as 30 amps, as it is strictly to control EMI.

I used to have a switch that made or broke this connection as needed for connecting to GFCI power sources.  (I also had to take one more step -- more on that in a moment.)  A couple months ago, I had to replace the inverter (for an unrelated issue), and on the replacement, I did away with the switch, and simply yanked this wire, insulated the spade, and called it done.

The simple reason was that I found there to be no significant increase in EMI emissions with the switch in the "off" position.  As it was not a safety issue (the grounds for all the output power are continuous back to the main input, and the chassis is still grounded), I decided that messing with the switch was more trouble than it was worth.

[Side note:  If you are in the business of opening up inverters, then there is a potential safety issue:  this ground connection allows the filter capacitors to drain off when the inverter is powered down.  With the wire disconnected, if you remove power from the unit, then quickly open up the case and touch the filter capacitors, you could get a shock.  The simple solution to this is to reconnect that wire and drain the caps before performing any internal service to the unit.]

The second step I had to take to avoid tripping GFI's has to do with my ground-to-neutral bonding relay, and I suspect most people using Trace SW series units have the same issue.  The act of connecting the shore power lifts the bonding relay.  With shore power off, the relay is closed, and ground and neutral are connected to each other in the coach.  The legal requirement for GFCI receptacles requires them to trip in a period of time that is much shorter than any electromechanical relay can break the ground-neutral connection, and, as a result, even if the inverter has no nuisance tripping problem, the bonding relay will cause the circuit to trip 100% of the time.

Once I cured the Trace issue, I spent some time where every time we needed to use a GFCI, I would open up my ATS can, and slip a dollar bill under the contacts of the bonding relay, breaking the bond until I could get plugged in.  Once the relay opened electrically, the dollar bill would slip out (even if I forgot to remove it), and the protection would be intact should shore power be removed.

I spent some time noodling on a more automated way to fix this.  Some modern inverters do it automatically by controlling their relays electronically, and timing the circuit to avoid the problem: both the hot and neutral from the shore are switched through an initial connection relay, and the inverter does not close this until several milliseconds after it lifts the ground-neutral internal bond.

(Unfortunately, if you have an inverter with an internal bonding relay, but without this sophisticated timing sequence, you are SOL:  the inverter will trip a GFCI every time.  But read on, because the way I solved my bonding problem will work for this type of inverter.)

I thought about what it would take to retrofit my custom ATS to implement this kind of timing, and rejected it as too complicated (and dangerous if the sequence failed at the wrong instant).  I also thought about Jim's suggestion of a momentary pushbutton that could be held down to disconnect the ground input during the process of plugging in.  I rejected that because it is clearly a code violation and dangerous to boot -- if the switch fails in the open position, then there is no ground protection whatsoever for the entire coach.  And this switch would be in the circuit 100% of the time, not just when using a GFCI.

What I ended up doing was making an adapter that I use only when connecting to 20-amp GFI-equipped receptacles.  It consists of a standard, heavy-duty, right-angle NEMA 5-15 plug wherein I have cut down the ground tang with my Dremel tool until is is just a hair shorter than the hot and neutral tangs.  (Ground tangs are always made longer than the current carrying conductors, to ensure that the ground connection is the first one made and the last one broken.)  While this, too, is something of a code no-no, I consider it the safest alternative short of simply electing not to connect to GFI outlets.

The way this works is that I put the adapter onto my shore cord first.  Then I plug the adapter into the GFCI.  Since the hot and neutral tangs are just a hair longer than the ground tang, current flows to the bonding relay coil, and lifts the bond.  As I continue to insert the plug in the receptacle, the ground makes complete contact, and all is safe and well again.  If the plug manages to work its way loose, there is a small chance than it can get back to a position where hot and neutral are connected, but ground is not (although unlikely, as there would not be enough friction left to hold the plug in the receptacle), and this could be construed to be a potentially dangerous situation.  However, remember that the entire power source, in this situation, is GFCI protected, which tends to minimize any such danger.

I sleep well at night with this arrangement.  It lets me use otherwise unusable GFI outlets when that's all that are available, but in the vast majority of situations (where there are no GFI's involved), I have a completely code-compliant and safe installation, as the special adapter is not used.  And the only non-code-compliant part in the whole arrangement is a 5" long adapter cord, which I will keep to myself so there is no danger that the next owner of Odyssey will have any code issues (unless he or she chooses to make a similar adapter).

Obviously, I can't recommend this solution to anyone else.  But it's what I did to solve an otherwise intractable problem.

-Sean
http://ourodyssey.blogspot.com

[gumpy]

Thanks Sean.

Thanks for pointing out that the green wire is not the unit's chassis ground. I should have realized it was not.

I must say I need to spend more time looking at this issue on my coach. I'm not sure if my transfer switch will have the same problem or not. I have actually only had a problem with this on two GFCI outlets. One was at my father-in-law's place, and I think the other one was in my garage, but has since been removed. There are other GFCI outlets at in-law's that I use all the time with no problems, but I know there's one on his garage that he had a problem with his inverter, which is why we installed the switch on it, so I know where I can test mine.

So I haven't actually installed the switch on the ground wire of my own inverter. I think think this spring I'll play with it some more. I need to build a custom dogbone for connecting to 120V outlets so maybe I can play with this also. 

craig

[rv_safetyman]

Sean, thank for the great reply.  I hope you did not take offense at my attempt at humor.  Indeed, I had hoped you would reply.

OK, so now we have a simple, elegant, safe solution to a 15A 120 volt system.  However, I am told that 30A GFI circuits are becoming more prevalent and supposedly there are some 50A GFI breakers out there.  Would the shortening of the ground tab work for the 30A and/or 50A circuits if we were to run into a GFI breaker protected circuit?

We have been told to turn the shore power breaker (both the one in the bus and the one on the pole) off while plugging in.  Sean's solution would dictate having both breakers in the on position.  Are there any issues there?

Great dialog on a subject that has really bugged me.

Jim

[Sean]

I hope you did not take offense at my attempt at humor.

I don't know, Jim -- I think you're a bit too fond of that cartoon Louise posted...

...  However, I am told that 30A GFI circuits are becoming more prevalent and supposedly there are some 50A GFI breakers out there.  Would the shortening of the ground tab work for the 30A and/or 50A circuits if we were to run into a GFI breaker protected circuit?

Yes, the same technique would work BUT
(1) I would be much more concerned about modifying a 50-amp plug than a 20-amp one
(2) In nearly four years of full-timing, we have never, ever encountered a GFI-protected 50-amp receptacle.  The NEC does not require them, and I am not aware of any state codes that do either.  They are very expensive, so parks are not going to install them unless mandated.

In that same time frame, we have encountered precisely one GFI-protected 30-amp receptacle.  (It was a COE campground on Trinidad Lake, in Colorado: http://ourodyssey.blogspot.com/2005/07/we-are-finally-out-of-oppressive-heat.html)  That's out of hundreds.

One quick-and-dirty solution to the 30-amp issue, should you ever encounter one, is to simply use a standard 30-to-20 dogbone, available at many camping outlets, or make your own from a 30-amp RV pigtail and a receptacle from the hardware store.  Then you plug the funky adapter I discussed earlier onto your shore cord, and the standard dogbone into the pedestal, and then make the final connection between the two.  Do as I did, and use #10 for both adapters, and you can safely run 30 amps through the whole arrangement, as long as you buy high-quality, heavy duty plugs and receptacles.

The solution we settled on, should this ever happen to us again, is to carry a GE and a Square-D 30-amp breaker with us.  There is no such thing as a 30-amp GFI "outlet" (at least, I've never seen one) -- the way 30-amp GFI's are implemented is by using a standard NEMA TT5-30 receptacle powered from a 30-amp, GFCI breaker.  We would simply yank the GFCI breaker and replace with our own for the duration of our stay.

(I have to readily admit here, though, that the real reason we carry those breakers, as well as a two-pole 50-amp GE, is because we have encountered, more than once, pedestals that simply would not work due to marginal or bad breakers.  In one case, before we started carrying our own spares, I got our pedestal working by robbing the 50-amp breaker from a neighboring camp site: http://ourodyssey.blogspot.com/2007/08/hotter-than-hades.html.)

We have been told to turn the shore power breaker (both the one in the bus and the one on the pole) off while plugging in.  Sean's solution would dictate having both breakers in the on position.  Are there any issues there?

My transfer switch, bonding relay, and inverter are ahead of my main disconnects in the bus.  There was no code requirement for a breaker upstream of that when I built mine -- I haven't seen a more recent copy of the NEC but I am told there is now such a requirement.  But, in any case, I never, ever turn off any switches or breakers in the bus before connecting to shore power.

I do, normally, turn off the shore receptacle before connecting or disconnecting.  (And I always test shore receptacles before using them).  This is a good practice if, for no other reason, it prevents arcing between the plug and receptacle when connecting or disconnecting under load.

In the special case of 15- or 20-amp GFCI receptacles, I need to have the receptacle hot in order for the trick to work.  However, in this case, I've already dialed my inverter down to 12 or 16 amps (80% of the receptacle/breaker rating) before connecting, the inverter will be the only "load" the circuit sees (other than the ATS and bonding relay coils themselves), and the inverter does not present any real load to teh circuit until it has seen "good" AC power there for 20 seconds, so I am just not worried about inserting this plug while the receptacle is hot.

As always, you should test the receptacle before using it.  A voltmeter is preferable, but even one of the three-light testers is fine for most purposes.  And, while I'm talking about them, my three-light tester, an LED model from Radio Shack, has a GFCI tester built in -- it's a little red button on the tester that, when pushed, should trip any GFCI.  I use this tester on all single-pole receptacles that we use, whether 15-, 20-, or even 30-amp (for the 30, I use the same 30-to-20 adapter I talked about above to connect the tester).  This will alert me if the receptacle may be GFCI-protected even if not so marked (by, for example, a GFI breaker upstream).  On more than one occasion, it has alerted me to the fact that a GFCI receptacle in a pedestal had a non-working GFI (in other words, the outlet was hot and the tester showed two green lights, but the GFI portion simply wouldn't trip).  In those cases, we simply connected to the outlet the same way we would connect to any non-GFI source.  (Clear evidence, by the way, that GFCI's do go bad -- test yours regularly.)

HTH,

-Sean
http://ourodyssey.blogspot.com

[Sam 4106]

Hi Sean,
Yes, GFI's do go bad. But, there is another possible reason your tester won't trip a GFI. If a GFI is wired so that the hot and neutral feed wires are hooked to the load terminals, instead of the line terminals, a GFI will not trip. You probably know that already, but others may not.
Thanks, Sam 4106

[Sean]

... there is another possible reason your tester won't trip a GFI. If a GFI is wired so that the hot and neutral feed wires are hooked to the load terminals, instead of the line terminals, a GFI will not trip. ...

You are quite right, Sam, and it is possible that at least some of the non-working GFI's that I have encountered were simply mis-wired when they were installed.  I certainly was not going to take the time to open up the pedestal ad figure it out -- it was enough for me to know that (a) the hot, neutral, and ground wires were all intact and connected, as indicated by the three LED's on the tester, and (b) the GFI was not tripping, so I did not need to use my magic adapter.

For the uninitiated, when you install a GFCI receptacle, there is a "line" side and a "load" side.  The line side is where you connect the incoming power to the receptacle.  In a "dedicated" application, which would (should) include all 15- and 20-amp campground pedestals, that's the only connection you use.  In a residential application, though, the code permits several receptacles to be protected by one GFCI device.  It is common, for example, for several kitchen outlets to be "ganged" to one GFCI circuit.  In this case, the GFCI receptacle is the first on wired downstream of the panel, and all the other receptacles are then fed from that GFCI device.  That's what the "load" terminals are for.  As Sam mentions, connecting the "line" to the "load" terminals will defeat the GFI protection.

-Sean
http://ourodyssey.blogspot.com