Date:         Sat, 11 Apr 2009 07:50:59 -0700
Reply-To:     Rocket J Squirrel <camping.elliott@GMAIL.COM>
Sender:       Vanagon Mailing List <vanagon@gerry.vanagon.com>
From:         Rocket J Squirrel <camping.elliott@GMAIL.COM>
Subject:      Re: gfci outlet buzzes with inverter
In-Reply-To:  <49e02d72.1ac1f10a.04a4.ffffa88a@mx.google.com>
Content-Type: text/plain; charset=UTF-8; format=flowed

The powered AC receptacle is "floating," which means that it doesn't know,
or care, about chassis ground or earth-earth. David's right -- neither the
hot nor the neutral wire is referenced to ground. They don't know about
the chassis of the van, they don't care, they are not interested in going
there.

There may be some eentsy leakage through the inverter, which has one side
of its 12V input tied to chassis, but the resistance of that path will (or
darn should) be of such high impedance that were you to lay your tongue
onto the AC's hot or neutral wire ("or" is an operative word here, laying
one's tongue /across/ the two wires would cause a cosmic experience) while
grasping an unpainted part of the van's chassis, you should be in no risk
of electrocution.

To test this theory: from the inverter's hot wire, connect a 1,500 ohm
resistor to chassis and measure the AV voltage across that resistor.* The
resistor simulates the conductivity of your body and is sufficiently low
so that it will bleed down the trivial leakage (if any) within the
inverter. If the voltage across that resistor is greater than 0.7 volts,
then the power at the receptacle carries the risk of electrocution and
should be treated with the same degree of respect as a household outlet.
Repeat this test on the neutral side of the inverter's outlet.

Note that if you were to do this test on a household outlet, the hot side
would read your 120 VAC household voltage, while the neutral side would
read essentially nothing because in a properly-wired house, even without
the leakdown resistor, the neutral side is meant to be at the same
potential as earth/ground so no volts there. But the output of the
inverter, as David said, is not referenced to ground so you want to check
the neutral side in your van, too.

If neither the hot nor neutral show any significant voltage across the
resistor then one could consider going on to the holding of the guitar
while putting your lips on the microphone test. Me, I'd not be worried
about doing that as long as I knew that the van was not connected to any
shore power, and not even if a wire from the chassis of the van was
dangling into a puddle on the ground. But I'm not advising you to do that
because it would really mess up my day if I got a call from your widow's
attorney referencing this message.

W/r/t GFI buzzing -- you're not feeding it a nice smooth sine wave, but
one with steps. It's gonna buzz. I doubt it would do anything useful with
the floating output of the inverter, because a GFI compares the "send"
current (call it the hot wire) with the "return" current (say, the
neutral) and as long as everything going into the load (your toaster) is
also returning out of it, then it's happy. But when your puppy knocks that
toaster into the bathtub, some of the send current will NOT be returning
where it should but will rather be going through whoever is in the tub to
earth/ground. This imbalance of send and return currents will cause the
GFI to shut off power.

In the van, though, the current really doesn't want to go the
earth/ground/chassis, so you could pitch the toaster into the tub and the
send current would/should return right back to the neutral wire like good
little electrons do. So the GFI would never pop. But and always in matters
of safety, I urge that you proceed carefully before doing a full-scale
test. In other words, use a mother in law before checking it on yourself,
is my advice.

===========
* This is a standard load for testing for leakage current from AC-powered
appliances. You'll also want to use a true-rms reading meter because the
output of an inverter is not a sine wave. More details about this at the
bottom of http://repairfaq.ece.drexel.edu/sam/safety.htm for more information.

--
Mike "Rocket J Squirrel" Elliott
84 Westfalia: Mellow Yellow ("The Electrical Banana")
74 Utility Trailer. Ladybug Trailer, Inc., San Juan Capistrano
Bend, OR
KG6RCR



On 4/10/2009 10:41 PM David Beierl wrote:

> At 08:28 PM 4/10/2009, pickle vanagon wrote:
>> I'm concerned about whether the buzzing is something to be worried about
>> otherwise.  It's faint enough that you don't really hear it unless you
>> get
>> down on all fours right next to the outlet.  But is the vibration
>> creating
>> an unsafe situation?
>
> The maker of the GFCI should know better than we do whether that
> stepped square wave is ok for his device...
>
> ...but we can say that the vibration is not going to hurt anything.
>
> Whether the rest of the installation is safe is a different
> question.  Tentatively, it sounds ok to me, *with the transfer
> switch.*  Keep in mind that normally neutral and ground are tied
> together somewhere nearby, which would probably destroy the inverter
> if it connected to such a situation; but with the outlet isolated I
> think you're all right unless you plug in a device with a hot-ground
> or neutral-ground short in it, which will take down the inverter.
>
> The inverter doesn't really have a neutral -- both terminals are hot
> w/ respect to chassis ground, just like a (US) household 220 vac
> circuit.  Some inverters use a dummy ground terminal on their socket(s).
>
> Another point -- I'm quite sure I've read that modern GFCIs do not
> measure current flowing in the ground leg, but instead algebraically
> sum the hot and neutral legs and trip when the difference exceeds 5
> ma.  Thus they don't care where the missing current is going, and can
> operate fine with a two-wire hookup.
>
>
>  --
> David Beierl - Providence RI USA -- http://pws.prserv.net/synergy/Vanagon/
> '89 Po' White Star "Scamp"
>