Author Topic: Fixing a 10-year old photovoltaic grid-tied inverter  (Read 7031 times)

Nords

  • Magnum Stache
  • ******
  • Posts: 3427
  • Age: 63
  • Location: Oahu
    • Military Retirement & Financial Independence blog
Fixing a 10-year old photovoltaic grid-tied inverter
« on: April 19, 2015, 01:57:13 PM »
We're back from Spain and I'll start a separate thread about that later, but first I have a more pressing question.

While we were on travel, our 10-year-old Xantrex 3.0GT photovoltaic grid-tied inverter quit.  I don't know what led up to the problem or what happened when it tripped, and it doesn't record its status.  All I have is the error message "AC voltage fault 3.6v" on the LCD display.  The owner's manual says that the inverter thinks that HECO is only supplying 3.6 Vac to the grid, so it won't start up until it senses roughly 240 Vac. 

Have any of you PV owners ever seen a similar issue?

I'll also spend time with Xantrex's support staff and our local solar supply business.

The other question is whether it's worth repairing a 10-year-old inverter.  I wouldn't mind an excuse to upgrade to a 6-7 KW model.

Our system's been running constantly since 2005 (over 27,000 KWhrs) and I've never seen any error messages before.  From the kilowatt-hour reading it probably tripped sometime in late February or early March.  I think this is a problem with the inverter, not with any of the PV panels or the DC wiring.  (There's over 400 V DC and 300 mA going into the inverter.)   The breaker panel that connects the inverter to the grid shows 240 Vac.  The inverter's AC output lugs show 240 Vac (when it's turned on).  The inverter is not sensing that voltage.   I've shut it off and turned it back on several times, but it shows the same error message every time.

The only serviceable part inside the inverter seems to be the 600v DC ground-fault fuse and a small AC fuse.  After sundown I'll shut off the inverter's circuit breaker to the house and check the AC fuse, but I'm pessimistic that it's the problem. 

Anything else I should check?  Any other ideas?

Nords

  • Magnum Stache
  • ******
  • Posts: 3427
  • Age: 63
  • Location: Oahu
    • Military Retirement & Financial Independence blog
Re: Fixing a 10-year old photovoltaic grid-tied inverter
« Reply #1 on: April 22, 2015, 03:51:41 PM »
This inverter is simplifying the troubleshooting.

Xantrex responded:
"The GT3.0 came with a standard 5 yr warranty in 2004-2007. There is an AC fuse in the upper portion that may be blown, but that should be checked by a qualified electrician. Beyond that, the GT is not repairable. We only offer battery-based solutions for residential anymore."
And then they mentioned that I should buy one of their many fine off-the-grid inverters.  But I just want a new grid-tied inverter.

I pulled the inverter's AC fuse and checked its continuity:  blown.  It's a 10-year-old fuse and I never even knew it was there until I searched for it. 

I replaced the AC fuse last night (20A, 250V, same model used in microwave ovens) and the inverter went through its entire five-minute startup cycle. That was great.

Then about a minute after it started up, it tried to do something that promptly blew the fuse with a muffled "POP!".  It went right back to the old AC voltage fault error message.

My next stop will be Inter-Island Solar Supply:
http://www.solarsupply.com/products/SOLARPV/INVERTERS.aspx 

MikeBear

  • Bristles
  • ***
  • Posts: 390
  • Age: 65
  • Location: Michigan
Re: Fixing a 10-year old photovoltaic grid-tied inverter
« Reply #2 on: April 22, 2015, 07:00:48 PM »
Did it blow that fuse even when there was no devices (load) applied to the inverter? Perhaps something downstream is bad, and the inverter is protecting itself the hard way.

The fact that Xantrex no longer makes them is suspicious, and would make me think they had some sort of inherent fault they didn't want to admit too.

Just like most old electronics, the first thing I'd check is for bad or bulging electrolytic capacitors. be VERY careful if you poke around in one of those, as charge caps of the size they use can seriously hurt or kill you!

zolotiyeruki

  • Walrus Stache
  • *******
  • Posts: 5658
  • Location: State: Denial
Re: Fixing a 10-year old photovoltaic grid-tied inverter
« Reply #3 on: April 23, 2015, 03:19:07 PM »

Just like most old electronics, the first thing I'd check is for bad or bulging electrolytic capacitors. be VERY careful if you poke around in one of those, as charge caps of the size they use can seriously hurt or kill you!
Agreed with this.  10 years ago was kind of the peak of "stuff breaking because of bad capacitors".  Crack it open and start looking.  The bad caps are easy to spot, cheap to source, and easy to replace.  All you've got to lose is a couple hours of time :)

Nords

  • Magnum Stache
  • ******
  • Posts: 3427
  • Age: 63
  • Location: Oahu
    • Military Retirement & Financial Independence blog
Re: Fixing a 10-year old photovoltaic grid-tied inverter
« Reply #4 on: April 23, 2015, 04:35:59 PM »
Did it blow that fuse even when there was no devices (load) applied to the inverter? Perhaps something downstream is bad, and the inverter is protecting itself the hard way.

The fact that Xantrex no longer makes them is suspicious, and would make me think they had some sort of inherent fault they didn't want to admit too.

Just like most old electronics, the first thing I'd check is for bad or bulging electrolytic capacitors. be VERY careful if you poke around in one of those, as charge caps of the size they use can seriously hurt or kill you!
I prefer to do this type of troubleshooting at dawn or dusk just in case I make a stupid mistake.  Before I put in the new fuse, the PV panel strings were generating about 200 VDC and HECO's bus voltage was 240 Vac.  The inverter went through its five-minute startup cycle and then displayed "Insufficient solar energy" since there wasn't enough sunshine to convert DC to AC.  ("Not enough sunshine" is rated at 195 volts DC or less from the PV panels.)  After that it displayed its typical night-time "Inverter offline" message.

I didn't time the next event, but it was suspiciously close to a minute later that the replacement fuse blew.  The inverter probably tried to check something (DC voltage?  HECO's AC voltage?) and that let through electrons higher than 250 Vac or 20 amps.  The circuit breaker between the inverter and HECO's grid is also rated at 20 amps and that never tripped.  I think the problem is inside the inverter's AC circuitry, perhaps exacerbated by whatever power is present on the DC side and getting across to the AC side.

I'm not sure whether "downstream" refers to the PV panel DC voltage or HECO's AC bus voltage.  But some circuit component seems to have failed and is letting through way too much power. 

Xantrex made a pretty good grid-tied inverter 10 years ago, but today the field is very crowded.  It's hard to tell whether they've abandoned their gear or decided to move into different markets. 

Just like most old electronics, the first thing I'd check is for bad or bulging electrolytic capacitors. be VERY careful if you poke around in one of those, as charge caps of the size they use can seriously hurt or kill you!
Agreed with this.  10 years ago was kind of the peak of "stuff breaking because of bad capacitors".  Crack it open and start looking.  The bad caps are easy to spot, cheap to source, and easy to replace.  All you've got to lose is a couple hours of time :)
I can only see about half of the components without taking off the cooling fins, but all the capacitors that I can see look good. 

This inverter has spent 12 hours/day of the last 10 years converting 600 volts DC (and several amps of current) into 240 volts AC.  I'd hope that if a capacitor or some other component went bad then I'd see large stains or scorch marks.  After I swap it out I'll have to take it apart for a full postmortem and see whether I can fix it.

I'm leaning toward a SMA Sunny Boy in the 4000-6000 watt range.  They're a German company but their gear has been around the island for at least the last five years.  (I remember snickering about the name the first time I saw it.)  They include two MPPT processors to extract the most out of each of the two strings of PV panels, which is probably worth doing for these aging cells.  Unlike most grid-tied inverters, it even has a little backup receptacle for limited house power if HECO goes down.  10-year warranty.  Better yet, Inter-Island Solar says it's in stock. 
http://files.sma.de/dl/18726/SB5000TL-US-DUS143847.PDF (that link downloads the PDF spec sheet)
« Last Edit: April 23, 2015, 04:41:15 PM by Nords »

Syonyk

  • Magnum Stache
  • ******
  • Posts: 4610
    • Syonyk's Project Blog
Re: Fixing a 10-year old photovoltaic grid-tied inverter
« Reply #5 on: April 23, 2015, 05:08:40 PM »
You're looking for bulging, venting out the top, brown residue around the base, anything "weird" with the capacitors.  That was definitely capacitor plague time.

https://en.wikipedia.org/wiki/Capacitor_plague

Otherwise, without good wiring diagrams, a multimeter, and some decent troubleshooting skills, it's probably just easier to replace it.  High voltage electronics are not something to casually poke around.

MikeBear

  • Bristles
  • ***
  • Posts: 390
  • Age: 65
  • Location: Michigan
Re: Fixing a 10-year old photovoltaic grid-tied inverter
« Reply #6 on: April 23, 2015, 05:40:50 PM »
I think swapping it out for a new one is by far your best solution. Putting the bad one on Ebay with good photos and a full description of what it's doing (IE: blowing the fuse, etc) might still get you some good money from a person that knows how to fix it already, and he's willing to take a chance on it...

I'm always amazed what some Ebay buyers will pay for certain things. All it takes is a little research (Completed auctions), a better phrased ad compared to the next guy, GOOD 10-11 photos of everything, being 100% honest in the description, and a fair starting price as bait...

Unless of course the place you are buying the new one from will take this in on a trade-in?
« Last Edit: May 05, 2015, 08:19:50 PM by MikeBear »

Nords

  • Magnum Stache
  • ******
  • Posts: 3427
  • Age: 63
  • Location: Oahu
    • Military Retirement & Financial Independence blog
Re: Fixing a 10-year old photovoltaic grid-tied inverter
« Reply #7 on: May 05, 2015, 06:59:09 PM »
Inverter tech has come a long way in a decade.  It's too early to tell (and I'm still doing the Dilbert Engineer Dance of Joy), but it's possible that this new inverter will extract an extra 20% power from the old array.  I'll check the data every week or two for the next month and compare it to history.

10 years ago I paid $2500 (plus installation) for a 3KW inverter with only one microprocessor.  It included an RS-232 port and its display looked like the output of a nine-pin printer.  Maybe the installer got a bulk-purchase deal and marked me up to retail, but back then a dollar a watt was the going rate.

Last month I paid $2733.29 for a 4KW inverter with two MPPT microprocessors, network-ready, and an animated LCD display that shows about 10x the information in the 4x the screen size.  It even has the capability of feeding a 1500-watt 120v receptacle directly from the inverter.  In other words, this grid-tied inverter is set up to power a house even if the grid is down.  (Well, as long as the sun is up.)  That receptacle is not a substitute for an emergency generator, but it'll keep an aquarium air pump going after the hurricane knocks out HECO's grid.  It'll even power the controller & pump for our solar water heater.  That's an extra hundred gallons that I can easily turn into potable water if the island loses water pressure.

The SMA Sunny Boy has its separate DC disconnect box, which makes it easier to replace a broken component without buying an entire new inverter.  (Murphy's Law says that this system will last at least 20 years just because it's modular.)  I can actually run four strings off the inverter, but I've left it at one per microprocessor so that it can optimize each string without letting one string's performance drive the show.  The display even shows the individual string's voltage and current so that I can compare panel performance.  It's been a while since I've had a 10 KWHr day on the old inverter, but in less than half the day the new one has already racked up 6 KWHr. 

The old Xantrex must have been designed with an internal AC neutral/ground connection, because it only had two hot wires (plus a ground wire) feeding into its AC side from the circuit breaker.  The Sunny Boy has AC connections for L1, L2, and N.  When I tried to run it with just the two wires in L1 and L2 it kept complaining about an AC grid voltage fault.  Once I grounded the N terminal it happily commenced cranking out the watts.

Coincidentally the Xantrex was blowing fuses on an AC grid voltage fault.  I wonder if its internal ground has a problem.  I'll do that post-mortem next.

Life in Hawaii:  when I opened up the main feed panel to the house to check the AC circuit breaker for the inverter, there were three gecko skeletons draped across the meter side of the bus bars.  (Gecko skeletons seem to have a much higher resistance than gecko flesh.)  I've been blaming all of our power hits on HECO but maybe a flash-fried gecko can do that too...

enigmaT120

  • Bristles
  • ***
  • Posts: 389
  • Location: Falls City, OR
Re: Fixing a 10-year old photovoltaic grid-tied inverter
« Reply #8 on: May 06, 2015, 02:21:05 PM »
That's creepy.  They don't have a higher resistance, they conduct power more readily sans flesh, I guess. 

 

Wow, a phone plan for fifteen bucks!