Inverter location, ground loops, and corrosion [RE-wrenches]

[Jim Easton]

<x-flowed>

Here in San Diego we have had some quite costly galvanic corrosion problems 
associated with grounding and DC currents.

DC is more corrosive than AC because one electrode is continuously attacked 
as opposed to the effect of AC where, both grounded electrodes share the 
burden and, to some extent, each half-cycle reverses the effect of the 
preceeding half-cycle.

When an AC neutral is bonded to separate grounding electrodes, there may 
very well be a DC current flowing because of differences in galvanic 
potential, however there should not be DC forced by the load potential drop 
across the neutral conductor joining the two grounding electrodes.

The lightning chokes I have seen are field fabricated using intentional 
slack in the normal current carrying conductors.

Because of the potential for destruction of equipment safety grounds, I am 
very wary of ground loops associated with DC power.  This equipment is 
supposed to last decades with minimal maintenance, so it does not take rapid 
deterioration to ultimately create a serious hazard.  And in this case, 
there seems to be a quite reasonable alternative.

With best regards,

Jim Easton, PE
 (E 11974)
4364 Bonita Road, No. 166
Bonita, CA 91902-1421
Tel: 858-527-0240
Fax: 619-470-8616
----- Original Message ----- 
From: "Bill Brooks" <bill at brooksolar.com>
To: "'Jim Easton'" <jeaston at ucsd.edu>; <RE-wrenches at topica.com>
Cc: <ipl at sover.net>
Sent: Tuesday, July 17, 2007 2:14 PM
Subject: RE: Inverter location, ground loops, and corrosion


Jim,

You bring up some interesting food for thought. I'm trying to see how
galvanic corrosion could possibly be an issue with grounding systems
separated by 850 feet. I also don't see how ac or dc could make much of a
difference in the galvanic corrosion. The difference in potential at each
end of ac neutral with current flow is a dc voltage is it not? It may have
an ac component, but the driver is the dc voltage difference--correct?

You make a good point about lightning protection, and I think that lightning
protection is a matter of a subjective nature rather than being something to
stake a reputation on. You suggest surge arrestors at the array and no
connection to ground at the array. I concur for distances of a 100 feet or
so or less, but if we are talking about 850 feet, I want to ground
everything at the array, install surge arrestors at the grounding system at
the array, and run nothing but current carrying conductors from the array to
the inverter. Chokes might help, but I'm not familiar with suppliers or
listed devices (any info on this would be interesting--probably could make
your own by wrapping wire around a non-metallic cylinder).

Once the current-carrying conductors get to the inverter, these conductors
should have a second set of surge arrestors to dissipate any voltage that is
induced on the wiring from the array to the inverters. The SMA inverters are
internally grounded in the inverter for the GFP function (essentially
disabled in this case for lack of a real current path). This reduces the
copper requirements and conductive paths. Since the distance is so great,
the GFP (not GFCI--just to keep from confusing people about ac devices) will
have no current flow because the ground resistance is too great and voltage
difference too low to get current to move.

Bill Brooks, PE
Brooks Engineering
873 Kells Circle
Vacaville, CA 95688
office and mobile phone: 707-332-0761
fax:707-451-7739
email: billbrooks7 at yahoo.com
web: www.brooksolar.com

Ecclesiastes 11:7 Light is sweet, and it pleases the eyes to see the sun.


> -----Original Message-----
> From: Jim Easton [mailto:jeaston at ucsd.edu]
> Sent: Tuesday, July 17, 2007 7:27 AM
> To: RE-wrenches at topica.com
> Cc: bill at brooksolar.com; ipl at sover.net; boillbrooks7 at yahoo.com
> Subject: Re: Inverter location, ground loops, and corrosion
>
> A caveat about multiple grounds (Ground Loops) in DC systems.
>
> If a DC neutral is grounded at multiple locations, even if the ground
> current is too small to trip tha GFCI (Assuming one is installed), it will
> still cause corrosion.  The amount will depend on the total neutral
> current
> and the ratio of resistances between the neutral conductor and the
> ground-to-ground path.
>
> DC is much more aggressive than AC in causing electrolytic corrosion. This
> may, over time, cause a failure of the equipment ground with serious
> safety
> implications.
>
> When wishing to provide lightning protection for a remotely located PV
> array, rather than bonding the neutral to the array ground, I would
> suggest
> instead installing a lightning arrestor at the array, and possibly putting
> in lightning chokes (Coils having negligible DC resistance but impedance
> at
> high frequencies) between the arrestor and the wire run from the array to
> the inverter(s).
>
> Unfortunately, I do not know of any UL/NRTL Listed arrestors for this
> purpose, however our inspectors here are usually sympathetic to using UL
> Recognized arrestors for this purpose.
>
> Jim Easton, PE
>  (E 11974)
> 4364 Bonita Road, No. 166
> Bonita, CA 91902-1421
> Tel: 858-527-0240
> Fax: 619-470-8616
>
> ----- Original Message -----
> From: "Bill Brooks" <bill at brooksolar.com>
> To: <RE-wrenches at topica.com>
> Sent: Monday, July 16, 2007 1:02 PM
> Subject: RE: Inverter location [RE-wrenches]
>
>
> >
> > David,
> >
> > Sorry for taking so long to respond. I just took a closer look at your
> > design and I'm really wondering why you are only putting 12 modules in
> > series. You cannot use 310Volts as the operating voltage for that
> circuit.
> > It will be more like 250Vdc on a nice day and much lower on a really hot
> > day
> > (225Vdc before degradation and actual output lower than manufacturer
> > promised). That is way too low a voltage for that inverter on a hot day.
> > Since you have such a long run, you should run the array at highest
> > allowable dc voltage. This will reduce voltage drop and wire size. The
> > slight penalty in inverter efficiency is absolutely worth it.
> >
> > If this is a California system, you should be looking at a different
> > number
> > of modules in series. I would recommend 15 in series, a total of 75
> > modules,
> > with 3 sets on one 7kW unit and 2 sets on one 5kW unit. I'm not sure
> where
> > this system is located, but you really need more modules in series. Even
> > in
> > the coldest part of the country, you can get 14 modules in series on
> that
> > inverter. Given the long dc run, I would combine at the array and run
> back
> > to the inverter at the house.
> >
> > Since these inverters have ground fault protection, you have to be very
> > careful about connecting dc negative to ground. With a very long
> distance,
> > you can ground the negative at the array and then the manufacturer does
> it
> > again in the inverter (not at the dc disconnect since that will blow the
> > GFP
> > fuse). The equipment grounds on the array would have to be connected to
> > the
> > same ground rod as the dc negative at the array, and the string
> combiners
> > would be the easy place to make all the ground connections (Outback
> > combiner
> > split in two halves). As you pointed out, doing it this way necessitates
> > non-metallic conduit and no ground wire in the conduit from the combiner
> > box
> > back to the inverter (850 feet).
> >
> > I hope that helps.
> >
> > Bill.
> >
> > Bill Brooks, PE
> > Brooks Engineering
> > 873 Kells Circle
> > Vacaville, CA 95688
> > office and mobile phone: 707-332-0761
> > fax:707-451-7739
> > email: billbrooks7 at yahoo.com
> > web: www.brooksolar.com
> >
> > Ecclesiastes 11:7 Light is sweet, and it pleases the eyes to see the
> sun.
> >
> >
> >> -----Original Message-----
> >> From: David Palumbo, Independent Power & Light [mailto:ipl at sover.net]
> >> Sent: Wednesday, July 11, 2007 6:33 PM
> >> To: RE-wrenches at topica.com
> >> Subject: RE: Inverter location [RE-wrenches]
> >>
> >>
> >> Bill,
> >>
> >> I apologize for the false wrench-mail start earlier.
> >>
> >> This is a timely reminder for me to hear right now. I am working on the
> >> design for a 13kw system with a wire run of 850 feet!
> >> An expensive wire run any way we do it.
> >>
> >> AC looks simpler at first. Inverters (2 SMA SB 7000's) at the arrays
> >> pushing
> >> electrons thru 750 MCM AL conductors in 4" conduit to the service
> >> entrance.
> >> But as Bill points out we do not want the inverter(s) to see higher
> >> voltages
> >> on it's terminals due to voltage drops. Best to play that safe since
> any
> >> resistance in that wire run, and connections, not accounted for in your
> >> calculations can shut down the inverters if the utility voltage is on
> the
> >> high side.
> >>
> >> DC wire run involves more conductors but is more of a sure thing.
> >> We may use 72 Evergreen 180's. They are 25.9 vmp and 6.95 amps. With
> >> series
> >> strings of 12 modules 310 volts 6.95 amps. I calculate 6 pairs of #4
> Cu,
> >> or
> >> #2 AL.
> >>
> >> I have a question pertaining to paralleling series strings together (3
> >> strings in this case) before the long wire run to the inverter. Is this
> >> an
> >> acceptable/workable practice? That would mean that we are still
> >> transmitting
> >> 310 volts mp but at 20.85 amps (26.06 amps cold temp.) into the SMA
> 7000.
> >> If
> >> we could do this we would only need 2 pairs of DC conductors. I
> calculate
> >> 1/0 CU, or 3/0 AL in this case.
> >>
> >> Back to the grounding lesson of a couple of weeks ago Bill. Let me run
> my
> >> understanding of that by you for this case. This assumes inverters at
> >> service entrance and an 850 foot PV wire run. With a separately derived
> >> system I can ground the pv arrays and bond negative to ground at the dc
> >> combiner box(es). No ground wire in pvc conduit and no ground wire in
> >> trench. At the DC disconnect(s), just before inverter(S), I will bond
> DC
> >> negative to ground. This is my understanding of separately derived
> >> systems.
> >> Your posting on June 22nd is quite clear and the best concise
> description
> >> that I have seen on pv grounding issues that I have read. I normally do
> >> not
> >> bond the negative to equipment ground at the array combiner box. I
> >> normally
> >> do run #6 copper in the trench, along side the conduit, to connect the
> >> grounding electrode system of the pv array(s) together with the
> grounding
> >> electrode system at the house. I have done this with pv arrays that are
> >> up
> >> to 250 feet away from the house. I have had very low incidence of
> >> lightning
> >> damage over my 20 plus years of installing. Other systems that I have
> >> serviced, that do not have as much copper in the ground, seem to suffer
> >> more
> >> damage from lightning induced electrical surges.
> >>
> >> Dave
> >
> >
> > - - - -
> > Hosted by Home Power magazine
> >
> > To send a message: RE-wrenches at topica.com
> >
> > Archive of previous messages:
> > http://lists.topica.com/lists/RE-wrenches/read
> >
> > List rules & how to change your email address:
> > www.mrsharkey.com/wrenches/etiquette.php
> >
> > Check out participant bios: www.mrsharkey.com/wrenches/
> >
> > Moderator: michael.welch at homepower.com
> >
> >
> >
> >


- - - -
Hosted by Home Power magazine

To send a message: RE-wrenches at topica.com

Archive of previous messages: http://lists.topica.com/lists/RE-wrenches/read

List rules & how to change your email address: www.mrsharkey.com/wrenches/etiquette.php

Check out participant bios: www.mrsharkey.com/wrenches/

Moderator: michael.welch at homepower.com
--^----------------------------------------------------------------
This email was sent to: michael_welch at sbcglobal.net

EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Qcs.bz9JC9.bWljaGFl
Or send an email to: RE-wrenches-unsubscribe at topica.com

For Topica's complete suite of email marketing solutions visit:
http://www.topica.com/?p=TEXFOOTER
--^----------------------------------------------------------------





</x-flowed>