Hybrid Solar System question

[Diceman83]

Hi everyone,

I've been doing a lot of research into solar systems, and I have determined the following things:
- My roof will support (just barely) 24 panels of solar panels
- 16 panels will be south-west facing, 8 will be south-east facing
- Parts of the roof will be shaded in the morning and evening due to trees on neighboring properties
- I currently use about 500 kWh of power per month (16.67 kWh / day on average)
- I want to be grid-tied, but maximize the in-home use of the power coming from the panels and sell extra power back to the grid (I don't trust the POCO to keep their current rates/policies, and who knows what future prices are going to be like)

The system I have designed so far:
- 24x SolarWorld SW 345 XL Mono panels (in 6 groups of 4 panels in series)
- 1x Outback GS8084A Inverter (with Mate3 controller)
- 6x Morningstar TS-MPPT-30 charge controllers (with networked meter and temp probe)
- 48v 400 Ah NiFe battery

From what I've read for my area (Connecticut), these panels should produce at minimum 500 kWh / month.

For the batteries, 400 Ah at 48v and 85% inverter efficiency and a 80% depth of draw should be 13 kWh of power, which should be enough for most of an average day even without solar adding to it. According to the specs for this battery, it should be charged at C/4 (100 amps), though higher charging rates are OK as long as the electrolyte temp doesn't go above 115 F. The minimum charging rate is C/20 (20 amps), but has much lower efficiency. The 8000 Watt inverter can draw a maximum of 78.4 amps from a DC source; 8000 Watts / 120 volts / 0.85 efficiency is my calculation there. Therefore, if the charge controllers are all producing their max current (180 amps total), the batteries can charge at 101.6 amps and the inverter can be producing its maximum power. That part seems fairly balanced to me, anyway.

If you're wondering why I'm looking into 6 small charge controllers, rather than 2 larger ones: I've done quite a bit of research on the SolarEdge system, and I love how each panel gets its own MPPT unit to optimize power production. Unfortunately, their inverters have almost no capacity for working with 3rd party charge controllers or battery systems, so I'm attempting to emulate the strategy with other components. The 30 amp charge controllers have about the same price / amp as larger charge controllers once I add in all the combiners and whatnot, and the Morningstars I selected are extremely robust (no moving parts, solid state components) so they should last a long time, and replacing a failed one shouldn't be too painful.

I have considered other battery types like FLA, Lithium, even a Vanadium flow battery, and even with the high initial cost I really would prefer to go with the NiFe. The price (including racks, water treatment equipment and shipping) for the battery is around $18k. If the battery will typically produce 13 kWh of power (see above), then that works out to $1.38 / watt. Due to IRS policies on rebates for solar systems with batteries (essentially I can't charge it from the grid, which I'm not), I will be getting a 30% federal tax credit, as well as a 30% state credit (if I can get this installed before the end of the year). So the battery price will actually be $7200, which comes out to a much more reasonable $0.55 / watt of usable power, for the entire life of the system (25+ years). The other options had similar price points over that time period, but usually involved a lot more maintenance and effort.

So, due to my circumstances with the panel placement, I doubt I'll ever get the full 180 amps possible between the 6 charge controllers. According to the charging parameters above, I'd want 100 amps going to the battery, and any extra going toward powering the house or being sold to the grid. I'd also want to track power draw at night and only charge the batteries as much as they were discharged, which is something I'd expect a charge controller to be able to track. If the panels stopped producing enough power to efficiently charge the batteries, I'd want it to switch to converting it all to AC for house/grid use. If more power than the panels can produce are needed, it should then draw from the battery until an 80% DoD point is reached, at which point it can pull from the grid. I haven't done the math, since I don't know enough about the charging efficiency or the power buyback rates, but there has got to be a balance point where utilizing grid power vs charging the battery for night-time use has to make sense. I know I'm not saving money by getting a battery, but I am buying some insurance against a future where electricity prices spike and the POCO has lobbied for pricing policies that make fully grid-tied solar a bad choice. Eventually I'd like to get an electric car and install a ground source heat pump, at which point I'd need a lot more solar.

What I'd like to know is this: can the GS8084A track the amperage going to and from the battery and utilize any extra power for house use and/or selling to the grid? I haven't really seen any standard wiring diagrams for this kind of situation, so I'm not entirely sure which parts would be connected in such a way to gather the necessary data. Given the claims in the product details, it seems like what I want to do should be possible, but I'd like more feedback before I commit to buying anything.

Thank you!

[ButchDeal]

The inverter can easily handle grid tie bimodal. It will maintain the batteries and primarily act as a grid tie inverter till the grid goes down...

[Diceman83]

Huh, I figured I'd get a little criticism on some of my choices, especially since I'm doing this for the first time and still learning. Maybe the thread just got lost.

It seems I didn't interpret the calculations for the open circuit voltage (Voc) correctly. I was using the the max power point voltage (Vmpp) to match up the string voltage to the charge controllers, and not compensating for temperature. Four panels with 48.7 Voc at 25 C would have 48.7 x 4 x 1.25 = 243.5 Voc if it ever got to the coldest temperatures. Even 3 panels in series would only be usable down to 59 degrees F, which wouldn't be very helpful. So max 2 panels in series for me, if I want to work with 150 volt charge controllers. I've also heard from a system designer that multiple charge controllers working in parallel will end up fighting with each other if the master charge controller thinks the battery doesn't need more amps, even if the inverter can utilize any extra amps for AC loads (or sale back to the grid). Is this true of the Morningstar charge controllers?

I'm also considering being NEC compliant, which means interrupting combiner boxes connected to the Birdhouse. This kind of blows my lots-of-small-charge-controllers idea away, since the combiner boxes are quite expensive. If I do go this way, I think my best bet would be 2 beefy charge controllers like the Midnite Classic 150, which can handle 96 amps to a 48v battery. My panels should be able to produce (at maximum) 345 W x 24 panels / 48 v = 172.5 amps, though I seriously doubt I'd ever see all of that at once. Better to plan for that though, so a pair of Outback 80 amp charge controllers would be undersized. Three Outback 60 amp charge controllers would work, but I don't know if the GS8048 can manage more than two. And it would require more combiner boxes.

So, how does one decide if the higher system optimization (for shading in the morning and evening hours) have priority over a higher component cost? I've estimated about a $450 difference between the two options (two 96 amp charge controllers versus three 60 amp charge controllers). Will the slightly higher output in mornings and evenings offset this difference? Another factor in favor of the three 60 amp charge controllers is possible integration with the inverter (same brand), which I would hope would encourage maximum utilization of the panels, regardless of battery charge state.

Thanks for any and all feedback!

[ButchDeal]

I'm also considering being NEC compliant, which means interrupting combiner boxes connected to the Birdhouse. This kind of blows my lots-of-small-charge-controllers idea away, since the combiner boxes are quite expensive. If I do go this way, I think my best bet would be 2 beefy charge controllers like the Midnite Classic 150, which can handle 96 amps to a 48v battery. My panels should be able to produce (at maximum) 345 W x 24 panels / 48 v = 172.5 amps, though I seriously doubt I'd ever see all of that at once. Better to plan for that though, so a pair of Outback 80 amp charge controllers would be undersized. Three Outback 60 amp charge controllers would work, but I don't know if the GS8048 can manage more than two. And it would require more combiner boxes.

Outback has an NEC rapid shutdown combiner which is a bit cheaper than the birdhouse and a bit more integrated with the system.
I would stick with the OutBack CC which are better monitored by the Mate3, instead of the midnight CC.
you could also look at the Flexpower version of the radian which has most of the components needed integrated already.

[Mike90250]

Multiple, un-sycn'd charge controllers don't 'fight'. As each one thinks the battery is full, it folds back and the other's take over till they think the battery is full. No fighting, charging just slows a bit, depending on how well the voltmeters in each controller agree with each other

[Diceman83]

I'm having trouble finding rapid shutdown components from Outback. I've read through the manual for the GS Load Center, and I didn't see references to a rapid shutdown option. Are there specific components that make this possible, or an example system with part numbers I could reference? It seems like Outback doesn't use the same terminology as Midnite for component names.

As far as the multiple charge controllers discussion goes, is it possible to have non-Outback charge controllers produce their maximum power at all times, so that any power not needed to charge the battery can be converted to AC for house load or grid sale? I still feel like this is an unanswered question, despite a lot of research trying to answer it. Is the only way to extract maximum power from the panels to keep everything Outback so the FLEXnet integration can manage the battery, or is it possible with custom configurations of non-Outback charge controllers? Is that the purpose of the OutBack FN-DC (System Monitor)?

Thanks for the responses so far!

[ButchDeal]

I'm having trouble finding rapid shutdown components from Outback. I've read through the manual for the GS Load Center, and I didn't see references to a rapid shutdown option. Are there specific components that make this possible, or an example system with part numbers I could reference? It seems like Outback doesn't use the same terminology as Midnite for component names.

As far as the multiple charge controllers discussion goes, is it possible to have non-Outback charge controllers produce their maximum power at all times, so that any power not needed to charge the battery can be converted to AC for house load or grid sale? I still feel like this is an unanswered question, despite a lot of research trying to answer it. Is the only way to extract maximum power from the panels to keep everything Outback so the FLEXnet integration can manage the battery, or is it possible with custom configurations of non-Outback charge controllers? Is that the purpose of the OutBack FN-DC (System Monitor)?

Thanks for the responses so far!

http://outbackpower.com/outback-products/manage-the-system/flexware-combiner-boxes/item/flexware-ics-plus

http://outbackpower.com/outback-products/make-the-power/category/flexpower-radian-series

in a typical outback bimodal install the inverter will send all the extra power to the grid.

[Diceman83]

Wow, those components are difficult to find. Thank you for the links!

The prices I've found seem to indicate it would be about $1062 for my situation, but I could have the wrong pieces.

I'm hoping I don't even need a rapid shutdown system, if my interpretations of this article are valid:
http://www.homepower.com/articles/so...art-1/page/0/1 The article makes these points:

Rapid shutdown requires controlling specific PV system circuits. These circuits include:

• PV source and output
• Inverter input and output
• Charge controller input and output
• Batteries and/or inverter input

Circuits are not subject to rapid shutdown (i.e., are not required to be controlled) if they are:

• On a building and less than 10 feet from the PV array
• Inside a building, with a length less than 5 feet

I'm hoping I won't need controls, since:

• The roof is right next to where the combiner boxes will be located (hopefully within 10 feet)
• The charge controllers will be within 5 feet of where the PV wires will enter the basement
• The charge controllers will be within 5 feet of the battery bank
• The battery bank will be within 5 feet of the inverter
• The inverter will be within 5 feet of the main breaker panel

Any thoughts on these conclusions? I'm sure it would be ideal if every solar installation has a rapid shutdown system, but is it worth the extra expense when the inverter controls most of it anyway? Could I rig up an AC disconnect that would accomplish the same thing?

[ButchDeal]

Well if you need rapid shutdown it is determined by your local code
you also could need the same parts for AFCI though.

[Diceman83]

Would a MNDC-GFP63 (http://midnitesolar.com/productPhoto...tOrder=2&act=p) count for the AFCI requirement? I'm not sure if GFP is equivalent to AFCI, though they seem very similar.

[ButchDeal]

No it does not

[Sunking]

There is so much wrong with your plan, no one wants to touch it.

Your first big error is ROI. There is no such thing as ROI with batteries involved. You must be a democrat as they are the only fools who think if it cost me $1 dollar to make a Kwh, it should sell it 10-cents. You can make up for the losses by selling more.

Time to wake up. Any thing you take off grid is going to cost you 5 to 10 times more than you can buy the power for. Forget the batteries. FWIW FLA is th ebest buy. NiFe and lithium will not cut it.

[Diceman83]

Sigh... I was hoping you wouldn't join this discussion, Sunking. I haven't found your arguments against batteries in other threads to be persuasive, and you seem to make the assumption that everything will remain the same. I've already stated I wasn't expecting an ROI; I'm banking against a dismal future.

Let's do a little math.

I currently pay about $90 / month for about 500 kWh of power. Over 30 years, if nothing changed, I'd be paying $32,400 (in today's dollars). We all know rates are only going to go up, and that figure is doubtful, but it's a starting point.

If I get a fully grid tied system that produces at minimum 500 kWh, it would cost around $16,000 up front, and I'll probably get $7000 of that back (between tax credits and loan costs), so final cost is around $9000 for me over 30 years. I'll be producing more than I use, but most of what I need will come from the grid (say 400 kWh per month). If nothing changes, I'll probably make enough extra each month to cover grid fees and maybe even make some profit back from surplus power. Sounds really good, and this is where your argument usually ends. However, considering how US politics and lawmaking works, I'm very doubtful of this outcome, since the utilities must be guaranteed a profit on their networks so they can maintain them. I seriously doubt an even sell/buyback rate will persist anywhere in the US for much longer; Hawaii is already having issues with excess solar power on the grids. Therefore, I would expect that regardless of my usage and production, I'll still end up paying $25+ per month average over 30 years. That leaves the fully grid-tied system at $18,000 between the initial cost and grid usage fees. It also has no backup if the grid goes down, which is something I'd like to protect against.

My proposed plan would be around $34,000 up front, with maybe $15,000 back (between credits and interest), so final cost is $19,000. I won't be producing nearly as much extra power, but I'm hoping that at least 80% of the power I generate will be used by me, meaning I would only need 100 kWh total from the grid monthly. This greatly reduces probable future costs and fees related to grid utilization, to say an average of $5 / month over 30 years. The total cost would therefore be around $21,000. I also have about a day's worth of backup power if the grid goes down and there's no sun (more if I reduce usage).

Considering how I'm planning to use the batteries, a FLA battery bank of equivalent power availability will be about $10,000 and last 10 years. If you found a 640 Ah 48v battery for less, please let me know, but it would still need to be less than $6000 to compete with the NiFe over a period of 30 years. Initially, I can get some tax credits to reduce this cost, but by 2022 the federal solar tax credits will completely expire, meaning the batteries I purchase at that point will be their entire up-front cost (not sure how much of a credit you'd get from recycling the old batteries). In any case, the repeated large purchase, disposal and installation effort, and other ongoing expenses related to FLA are factors against this option.

I'm sure other battery technologies will emerge that last longer and charge more efficiently than NiFe within the next 10 years, and maybe they will make more sense to buy over FLA and NiFe batteries. Maybe fuel cells will finally make sense, and solar energy will be used just for electrolysis to produce hydrogen (doubtful, since the basic process is so inefficient). In any case, it will likely involve another large un-subsidized expense and some rework of the system.

The whole "$1 to make, $0.10 sell" argument you make is insulting, and is not how a home owner thinks of a solar system's value. Sure, if I wanted to try to make money with solar, with your figure of $0.10 / kWh I would start turning a profit in 35 years, which is more than the expected life of the system and I may as well have just stayed on the grid. You're absolutely right, it makes no sense if all you look at is the current situation.

I believe that the current situation is unsustainable however, and we will see major changes in the next 10 years or less. The US government thought we'd be in a crisis by 2015, and that was only held off by fracking and other environmentally disastrous mining methods. The primary reason for this belief is our complete dependence on cheap oil for everything; transportation and power first and foremost. We had a taste of $100 / barrel a few years ago, and I'm pretty sure it was a factor in the burst of the housing bubble. When the global oil supply is outpaced by demand, we're really going to miss the $50 / barrel we're enjoying now, and the price of everything is going to go up significantly. If the government is smart, it will heavily tax gasoline and subsidize diesel for the trucking/bus industry, in order to keep prices for most goods (food, etc.) stable. People will find other ways to get to work if the price of gas is high enough. The alternative is a breakdown of basic social norms, as prices for everything skyrocket and wages remain the same. Trucking will still eventually fail, and our only alternatives are going to be short range smaller electric vehicles and electric trains powered by renewable sources. I'm not politically affiliated, but I am a realist. If everything stays the same, oh well, I paid a little too much for a solar system. If it all goes to hell, I'll be more prepared than most to weather the fallout.

In a few years, I'll probably be installing more PV to support home heating (ground source heat pump) and power for electric vehicles. I'm guessing I'll need another system of the same size as I am currently planning, but we'll see.

[inetdog]

Considering how I'm planning to use the batteries, a FLA battery bank of equivalent power availability will be about $10,000 and last 10 years.

Not bloody likely. And this is the pivot from which your calculations diverge the most from Sunking's.
As long as you are going to be cycling your FLA battery bank rather than keeping it on hand as a reserve for power failures, you will not get 10 years of life from them. Batteries designed for cyclic use (based on chemical composition and mechanical design) will deteriorate just sitting there fully charged.

[Diceman83]

Heh, I was just giving it the benefit of the warrentee time, as if the manufacturer would support a heavy use application. NiFe specs give it a 11,000 cycle (80% DoD) lifetime, which is literally once a day for 30 years. I'm hoping not to do daily 80% DoD, but I'm sure it'll happen fairly often with weather. I'm nowhere near what I would need for a fully off-grid system, but I'm aiming to balance grid usage with price. A generator would certainly reduce my dependence on the grid, probably entirely if I got it large enough, but that adds an oil dependency to the system and my long term goal is to eliminate oil usage.

[jflorey2]

The system I have designed so far:
- 24x SolarWorld SW 345 XL Mono panels (in 6 groups of 4 panels in series)
- 1x Outback GS8084A Inverter (with Mate3 controller)
- 6x Morningstar TS-MPPT-30 charge controllers (with networked meter and temp probe)
- 48v 400 Ah NiFe battery

Outback recommends at least 400ah of lead-acid batteries, partly because to supply 8000 watts (17,000 watts surge) you need up to 350 amps at 48 volts - which means you need a VERY low ESR in your battery pack. NiFe has a fairly high ESR compared to lead-acid, so you will likely need a larger battery system to achieve a similar surge capacity to lead-acid.

In addition, NiFe requires a greater range of system voltages (both for charge and discharge.) A cell will often need 1.8 volts to charge and will discharge down to .9 volts - so make sure your inverter/charger can handle that.

[jflorey2]

>Where are you getting those numbers?

Math. The GS8048 will surge to 17,000 watts (to start things like AC compressors.) 17,000/48=354 amps. You can't get around that; it's math. If you try to draw that much through a good lead acid pack you will see a few volts of drop as the resistance in the pack causes the voltage to drop. (Note that most circuit breakers have delay curves that allow short-term surges for just this reason - which is why you don't need 400 amp or 2x200 amp breakers for Radian installations.) Since the Radian will work down to 40 volts you are OK; the voltage sags to 43 volts for 100 milliseconds then recovers.

Problem is that nickel-iron has a much higher internal resistance. So in exactly the same case, the battery sags to 38 volts, the inverter trips and shuts down, and your house goes dark.

How do you get around this? Use a larger battery size so that the increased inherent resistance is mitigated by more active plate material in parallel.

[Diceman83]

I still don't understand where the 17,000 watts figure comes from. My main panel is rated for 100 amps at 120 volts AC, which is 12,000 watts. I've never tripped the 100 amp breaker, so I've never had surges as high as 17,000 watts. If there's a common expectation of surges being over double the rated, I guess I've never heard or read anything about it.

Fortunately, it doesn't matter anymore, since I did a much more in depth analysis of the true cost of batteries and decided to go fully grid tied.

[jflorey2]

>I still don't understand where the 17,000 watts figure comes from. My main panel is rated for 100 amps at 120 volts AC, which is
>12,000 watts. I've never tripped the 100 amp breaker, so I've never had surges as high as 17,000 watts.

When you are starting a large compressor (like a central air conditioner) you can see surges from 2 to 8 times the steady state run current. Your breaker doesn't blow because breakers have delays built into them for just that reason. For example, a Square D/Schneider Class 0600 AC breaker rated for 20 amps will pass 16 amps forever, and will generally pass 20 amps for a long time. Above that the trip time varies. At 30 amps it will blow after 80 to 300 seconds. At 40 amps, 20 to 100 seconds. At 100 amps, 4 to 15 seconds. That means that if you have an AC compressor that normally takes 12 amps at 240 volts, and it takes 70 amps for 20 milliseconds to start during a short cycle, then your 20 amp Square D breaker won't blow. (And your Radian inverter will start it, since it can output 70 amps for a short time.) But your batteries WILL see that load, and must support it.

>Fortunately, it doesn't matter anymore, since I did a much more in depth analysis of the true cost of batteries and decided to go fully grid tied.

Sounds like a good decision.

[Mike90250]

As far as the multiple charge controllers discussion goes, is it possible to have non-Outback charge controllers produce their maximum power at all times, so that any power not needed to charge the battery can be converted to AC for house load or grid sale?

Easy, set the charge controller to a higher voltage, and let the inverter siphon power away to feed the grid. At not solar hours, you rest the batteries.
Or just go simple grid tie, forget the hybrid inverter

[Sunking]

Good luck on 10 years.You are only fooling yourself, not us. There is a damn good reason US manufactures quit making NiFe batteries 41 years ago, and every pro answering is teling youu are making a HUGE MISTAKE with MAKE BELIEVE CALCULATIONS based on FAIRY TALES. You will get a nasty $18,000 lesson you will never forget.

FWIW NiFe cannot deliver or take high C-Rates. I am surprised no one has picked up on that yet. Because once you figure out what that means is you are going to need a lot more battery or a huge cut in power. That will be your only choices.

Oh, I do not care if you want to hear from me or not. Not my problem. But if you listen, I will save your arse, and I don;t even like you. Between Mike, JFlorey, inetdog, Butch, and myself have over 100 years of professional experience telling you are full of crap. Perhaps you better wake up and face reality before you do something really stupid.

[inetdog]

The one person active on the forum with extensive off-grid experience with NiFe is Mike.
I think the jury is still out on his system, but there were a bunch of unexpected problems to balance out the tolerance of abuse and potentially long lifetime. Search through Mike's previous posts on NiFe for a comprehensive story.