Using 2 inverters for reliability

I am a beginner at DIY solar and am planning my first installation. All is going well but having trouble finding out one aspect of the installation. I was using an inverter 30 years ago on a live aboard boat and found it to be very unreliable. I know inverters have improved vastly since then but I still want increased reliability. So with that background ... here's the question.

Can I parallel the output of 2 inverters if inverter one is connected to half the panels and inverter two is connected to the other half. Panels will all be identical and neither of the two strings subject to shading? Right now the inverters I'm considering are the Growatt 6kw grid tie inverters (MIN 6000TL-X).

I realize in the event one inverter go's out I will only have half my rated capacity (but half is better than nothing ... right ?) Are there any serious cons to doing it this way ?

Thanks in advance

All grid-tied inverters are paralleled on the outputs.
So the answer, is yes.
But they require a grid connection. Hopefully, this is not going on your boat.

schuh8 said:

Can I parallel the output of 2 inverters if inverter one is connected to half the panels and inverter two is connected to the other half.

Click to expand...

I think its worth mentioning that you can only parallel the same brand, and model of inverters and most importantly they need to communicate with each other to mach phasing of the AC output. To put it more clearly, If you took a renogy inverter and paralleled the out put with an amazon special inverter, I would think that it would probably result in at the very least an unhappy inverter, but probably you would let the magic smoke out of both of them.

I am new at this and I hope I am wrong... But I currently dont think so.

ScrotusGobbleBottom said:

I think its worth mentioning that you can only parallel the same brand, and model of inverters and most importantly they need to communicate with each other to mach phasing of the AC output. To put it more clearly, If you took a renogy inverter and paralleled the out put with an amazon special inverter, I would think that it would probably result in at the very least an unhappy inverter, but probably you would let the magic smoke out of both of them.

I am new at this and I hope I am wrong... But I currently dont think so.

Click to expand...

That only applies to off grid inverters.

Edit for clarity:
Also applies to hybrids, because of their off grid capability.

Sorry, should have mentioned this will be a grid tied system. But given that info ... If one of my two matching inverters goes out in 10 years and the Growatts are no longer available, I am assuming I could replace it with a similar 6KW grid tied inverter by another manufacturer ?

You can have any brand grid tied inverters, mix-n-match. Because they all interact through the bottomless grid.
If one fails, just need another that accepts the voltage, current, and number of PV strings in some manner. (preferable not to have to rewire panels.)

But you will likely want battery backup for some loads, so consider hybrid inverters, or grid-tie inverters that integrate nicely with battery inverters. Know what is available and at what cost before buying.

Having a battery can also let you deal with time of use rates more cost effectively.

schuh8 said:

Sorry, should have mentioned this will be a grid tied system. But given that info ... If one of my two matching inverters goes out in 10 years and the Growatts are no longer available, I am assuming I could replace it with a similar 6KW grid tied inverter by another manufacturer ?

Click to expand...

Correct
Any size, any manufacturer.

I run multiple (redundant capable) inverters but not in the way you're thinking. Each inverter must power separate circuits - the 2 x inverters cannot be synchronized or AC output hooked to each other.

Each 12,000w 240v/120v inverter goes to it's own distribution panel. However, the distribution panels can power each other with an interlock set of breakers/wiring. If an inverter dies, I can power both distribution panels from the other inverter.


Of course I no longer have 24,000w but only 12,000w... but this can be managed by serializing use of household items (i.e. don't run the dryer at the same time as the cooktop) until the situation can be corrected.

And in fact, one of my AIMS died at 5yrs / 25,000hrs of operation and I used the top AIMS to cover the gap (power both distirbution boxes) till I got the new 12,000w SGP to replace it. I was able to get the AIMs repaired after a few months - so now I have 3 x inverters with one as a backup.

There are some things to consider:
When paralleling two inverters they must communicate with each other via a communications cable between inverters. One acts as the master and the other a slave. This ensures the proper phase relationship in the AC output of both. It is true that for some inverters they sync with the grid and if two are syncing with the grid they more than likely will be in phase. But what happens when the grid is down. The clock in the Master is used by the slave for phase syncing since both are now running solely on PV/Battery.
Equally important if you have two 120VAC inverters paralleled to create 240 VAC, the "slave" must phase shift it's output 180 degrees from the Master, thus the communications between inverters. You can't be certain that inverters from different manufactures will communicate that same data across the communications link.

schuh8 said:

Right now the inverters I'm considering are the Growatt 6kw grid tie inverters (MIN 6000TL-X).

Click to expand...

You appear to be in the US, I don’t think you will be able to grid tie the Growatt MIN 6000TL-X inverter because it’s not intended for US split phase.
If I’m understanding you properly, the MIN 6000TL-XH-US would be the version to use in US. These inverters do not have grid pass thru, in some cases it’s best to select a brand that does. To grid tie two of them at the same time you would need connect them electrically to separate panels and have a power meter for each ( or use what Growatt call a SYN200) such as to a sub panel and main panel or in some cases people have two main panels in a large home as we do and some others on this forum. Alternatively, if you want spare, buy two bigger identical ones and install in a manner that they are easy to swap. Test both and put one back in box or leave both side by side with one disconnected. If you are wanting consider other Growatt equipment types look close at what timselectric is doing. It’s not for me but would do well for lots of folks.

Don't look at anything I'm doing for information on grid-tied. I'm strictly off grid. I don't want any part of the grid-tied nonsense.

Since you have reliable power generation and storage, have you considered providing it to neighbors, for a fee?
Run wires, tie them into it ...

You could call it Tim's Gas and Electric

timselectric said:

Don't look at anything I'm doing for information on grid-tied. I'm strictly off grid. I don't want any part of the grid-tied nonsense.

Click to expand...

I was referring to what I think/thought you were doing, part off grid with some loads still fed from grid separately?

MajicDiver said:

I was referring to what I think/thought you were doing, part off grid with some loads still fed from grid separately?

Click to expand...

That is currently the case, because my system isn't complete yet.
But, there is not and will never be any grid-tied equipment.

(At least not connected to the actual grid) lol

OffGridInTheCity said:

I run multiple (redundant capable) inverters but not in the way you're thinking. Each inverter must power separate circuits - the 2 x inverters cannot be synchronized or AC output hooked to each other.

Each 12,000w 240v/120v inverter goes to it's own distribution panel. However, the distribution panels can power each other with an interlock set of breakers/wiring. If an inverter dies, I can power both distribution panels from the other inverter.
View attachment 198929

Of course I no longer have 24,000w but only 12,000w... but this can be managed by serializing use of household items (i.e. don't run the dryer at the same time as the cooktop) until the situation can be corrected.

And in fact, one of my AIMS died at 5yrs / 25,000hrs of operation and I used the top AIMS to cover the gap (power both distirbution boxes) till I got the new 12,000w SGP to replace it. I was able to get the AIMs repaired after a few months - so now I have 3 x inverters with one as a backup.

Click to expand...

So to be clear, you can have 2 inverters(not all in ones) connected to the 1 battery if they each power their own circuit? For instance a 1500 watt aims inverter to run large loads with a dedicated circuit and a small Victron 375 watt inverter to run small things like lights, etc ?

james1 said:

So to be clear, you can have 2 inverters(not all in ones) connected to the 1 battery if they each power their own circuit? For instance a 1500 watt aims inverter to run large loads with a dedicated circuit and a small Victron 375 watt inverter to run small things like lights, etc ?

Click to expand...

Yes. Each inverter powers it's own distribution panel (with regular breakers) out to circuits. So each inverter powers a different set of circuits. However, the panels are cross connected so that by flipping breakers either panel can power the other. So a single inverter can power both distribution panels - and all the circuits - if need be. Of course if all loads are on at the same time it will overload a single inverter (that's why I have 2 inverters) but that can be managed by serializing use or disabling some of the circuits (turning breakers off) during the emergency period.

OffGridInTheCity said:

Yes. Each inverter powers it's own distribution panel (with regular breakers) out to circuits. So each inverter powers a different set of circuits. However, the panels are cross connected so that by flipping breakers either panel can power the other. So a single inverter can power both distribution panels - and all the circuits - if need be. Of course if all loads are on at the same time it will overload a single inverter (that's why I have 2 inverters) but that can be managed by serializing use or disabling some of the circuits (turning breakers off) during the emergency period.
View attachment 204309

Click to expand...

Thanks. Sounds good.

Redundancy is overrated. I'd rather have one good system and make it less likely to fail.

Hypocritically though, I built my entire system in parallel to standard grid feeders, so my backup is always just switch back to grid.

2 is 1 and 1 is none.

timselectric said:

2 is 1 and 1 is none.

Click to expand...

I love redundancy but in my case with my current single inverter of 1500 watt I have to keep draw slightly below 1500. By having a second small inverter in the system I can continuously run the large load and keep the lights and demand water pump and TV etc running simultaneously. In other words I will get the equivalent of a 1900 watt inverter by just adding a small 400 watt inverter to the system.

OffGridInTheCity said:

Yes. Each inverter powers it's own distribution panel (with regular breakers) out to circuits. So each inverter powers a different set of circuits. However, the panels are cross connected so that by flipping breakers either panel can power the other. So a single inverter can power both distribution panels - and all the circuits - if need be. Of course if all loads are on at the same time it will overload a single inverter (that's why I have 2 inverters) but that can be managed by serializing use or disabling some of the circuits (turning breakers off) during the emergency period.
View attachment 204309

Click to expand...

Do you have any pictures from when you assembled your batteries? Particularly how you handled creating the main negative and positive rails on each battery?

I am building a 13S10P 48V battery using BAK 21700's, 5A continuous, 15A max. Basically a 2.6kWh when at 4V (52V x 50A). One of the things I plan on doing is limiting discharge current to 40A, and charging current to about 25A (Its expensive building a battery), and doing a 75-25% Cycle.

I have a battery charger that fits up to 21700's, and on a 1A charge, they take about 3.5 hours to charge. On a 1.5A charge, they get a little warm, but take like 2.5-3 hours to charge.

ShirBlackspots said:

Do you have any pictures from when you assembled your batteries? Particularly how you handled creating the main negative and positive rails on each battery?

Click to expand...

The base physical unit is 4 x 19 slots. Stack/populate/zip-tie 2 of these to get an even 260ah (5lbs) per unit. You can see 1 x 260ah unit here using flattened copper pipe with hand-drilled holes and bolts to connect the 2 x 6awg buses on each base unit.


These 260ah units are put in series using the same flattened copper pipe segments and bolts


14s = 48v for me. 4/0 wire with lugs and bolts to parallel and do the main + and - at opposite ends of the overall paralleled stack of batteries.

**And yes, the negative take-off is about 20ft of 4/0 wire vs the plus is 6ft (to the main buss) but at the overall low C it doesn't make any difference I can tell. I have Batrium so I can see each of the 126 units to confirm there's no noticeable stress. Low C make potential current issues toothless - one of the benefits of a larger powerwall.

ShirBlackspots said:

I am building a 13S10P 48V battery using BAK 21700's, 5A continuous, 15A max. Basically a 2.6kWh when at 4V (52V x 50A). One of the things I plan on doing is limiting discharge current to 40A, and charging current to about 25A (Its expensive building a battery), and doing a 75-25% Cycle.

Click to expand...

I don't limit individual batteries - but rather monitor overall powerwall load which gives me load per cell. I'm at ~200a peak load which is 22.2a/battery or 222ma/cell or 0.1C'ish - low stress and nothing to worry about. Yearly, overall average is 52a. Plus I have a 400a shunt-trip so the max possible overall load is 444ma/cell and all cells were tested at 500ma.

Charging is similar in that max PV -> Charge Controller (<300a max) and Chargeverter (100a) at the same time and I'm still <400a shunt-trip and <444ma/cell - so no worries.

I agree it's good to put limitations (main shunt-trip in my case) because one forgets details as the years go by (I'm in year 7 now) and let's say I add a 2nd Chargeverter for another 100a and not even think about 500a total potential charging.... but with the 400a max shunt-trip I'm protecting the powerwall in spite of myself *Yes, the shunt-trip is mainly to protect the wire - but it has the secondary affect of protecting the health of the cells in the powerwall.

This is my battery. That's an 8AWG wire there on the main negative side of the battery. I'm using 0.1 x 10mm nickel strips that are double stacked. The interconnection between + and +/- and - are two strips of .1x5mm. I'm using 0.1 strip because my spot welder was having difficulty making consistent spot welds with 0.2 strips. I am completely redoing the strips, pulling off the 0.2 strips (if I can) to put the 0.1 strips on.

My BMS is a 100 Balance (Daly) 60A active balance BMS. It has 30A continuous charging, 60A max. It comes with TWO temperature probes, with an optional port for two more, and a WiFi dongle. It has a port for an optional switch to turn the BMS on or off, and supports RS485/Can, next to the UART port. I has an optional 4" and also an optional 7" screen to display information, with a special order wire to have the WiFi dongle connected at the same time.

My soldering skills (or my soldering iron) aren't that great, but I did order some flux and 60/40 soldering wire from Amazon that should be here tomorrow (so I can solder that 8ga wire to the main negative and positive), along with a Victron 150/45 MPPT and the VE.Direct to USB cable.

On that second picture, on the bottom of the battery, that's the 0.2 x 8mm strips that I am going to remove and do the end terminals correctly with the 0.1 x 10mm strips.

I might build a second one in the future, and connect the two to the MPPT with 6ga wire.

Also -- I assume that are your batteries recycled, based on all the mixed batteries? I see (8x13)-2 = 102 batteries x 14 = 1428 batteries. At $4.99 a battery, that's $7125! Times that by 9 packs, and that's about $64k in batteries.

schuh8 said:

Sorry, should have mentioned this will be a grid tied system.

Click to expand...

If you are grid tied, then why is inverter reliability a critical issue?

If the inverter is down, then you run on grid until you get it fixed or replaced. You lost some production, but no functionality of the site.

Maybe a microinverter solution is for you. That's lots of little inverters, all grid tied, and failure of any one of them takes out only one panel.

If you want power during a grid outage, that changes things considerably, however.

Mike C.

ShirBlackspots said:

Also -- I assume that are your batteries recycled, based on all the mixed batteries? I see (8x13)-2 = 102 batteries x 14 = 1428 batteries. At $4.99 a battery, that's $7125! Times that by 9 packs, and that's about $64k in batteries.

Click to expand...

Most were 2nd hand batteries from BatteryHookup, Battery Clearing House, Jag35, and Power2Spare (no longer active). Average price over the years has been ~$1.10/cell for 95%+ original capacity, 2200-3000mah range, name brand, hi quality cells. Overall including BMS, wiring, solder, shelves, everything... over the last 6+ years I'm running ~$220/kwh and with 30% tax credit that's ~154/kwh effective. Oldest battery nearing 2,000 cycles and going strong.

One cool thing about 18650 (individual, smaller cells) is you can maintain 260ah/unit just by varying the # of cells depending on the type. This let's me build new batteries that are same capacity as older batteries even though 6+ years have passed and the offerings change year by year.

Cell/Battery detail. Over the years... started with 2nd hand ebike packs, then modem packs were everywhere, and then 'never used' medical packs and few misc purchases in-between. Bat #9 was a Frankenstein build of left overs from #1 thru #8.