Yet another GT workaround question

leeelson
leeelson Registered Users Posts: 18
I have a SB 3800 grid tie inverter, 18 Sharp 208 panels and a 6 kW Guardian genset. Is it possible to switch my local grid off of the utility when the utility is down and run an off grid inverter instead of my GT? When the utility is down, my plan would be to use the genset in place of batteries (when the panels aren't producing), with the option of turning it off if I don't need any power. My house uses about 18 kwh/day and I could live without power at night. Is this feasible or are batteries required in off grid systems? If required, please explain why. How can I minimize them?
«1

Comments

  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »
    I have a SB 3800 grid tie inverter, 18 Sharp 208 panels and a 6 kW Guardian genset. Is it possible to switch my local grid off of the utility when the utility is down and run an off grid inverter instead of my GT? When the utility is down, my plan would be to use the genset in place of batteries (when the panels aren't producing), with the option of turning it off if I don't need any power. My house uses about 18 kwh/day and I could live without power at night. Is this feasible or are batteries required in off grid systems? If required, please explain why. How can I minimize them?

    Not possible, sorry. Batteries are required to run off-grid. The main reason for this is that demand timing patterns do not coincide with PV production curves. To supply power on demand you must have a stable reserve to pull from, and PV production is all over the map. Solar production is your income, batteries are your bank account.

    Off grid systems are really battery systems; the only reason for the PV is to charge the batteries. PV is optional but batteries are not for grid backup. It would be great if it were the other way round, but that's the way it is.

    The size battery bank you need depends on your usage. You can minimize the size battery bank by powering a small protected loads panel with only a few things connected to it, but your bank has to be big enough so that you will never deplete it past 50% capacity, else you will be replacing your batteries frequently.

    There ain't no free lunch.
  • System2
    System2 Posts: 6,290 admin
    Re: Yet another GT workaround question

    For whatever it's worth it, here is a youtube video of a guy designing GT and OG system:
    http://www.youtube.com/watch?v=Tz6W_He64pA&feature=mfu_in_order&list=UL
    I has several parts and a schematic so make sure you watch them as well. Possibly, it could be adapted to your needs.
    Good luck!
  • leeelson
    leeelson Registered Users Posts: 18
    Re: Yet another GT workaround question
    ggunn wrote: »
    Not possible, sorry. Batteries are required to run off-grid. .


    Thanks for that. Based on your reply, here are my real questions. I plan to add some panels to my 18 Sharp 208's, probably another 6 or so (~1200 W). My SB 3800 won't be sufficient, so I'll need to modify/add to my inverter setup. In addition, in order of importance, I want to:

    1) Keep my grid tie (net metering) capability for all panels.
    2) Make use of my PV panels when the utility is down (just like every other beginner)
    3) Keep the total cost (acquisition + battery replacement) as low as practical.

    From what I've been able to learn, I think I have several choices:

    1) Buy a another SB grid tie inverter on the used market, allowing all 24 panels to be GT. Buy a used off grid inverter and charge controller that would be set up in parallel and activated automatically when the grid goes down.

    2) Buy another SB grid tie inverter on the used market, allowing all 24 panels to be GT. Buy a Sunny Island (which is just a switching box, not an inverter as I understand it) that would then allow all 24 panels to supply the house when the utility is down.

    3) Add a Xantrex XW4024 and charge controller to handle the 6 new panels. The question here is, when the utility goes down, will I have access to all 24 panels, or just the 6? Also, can the XW4024 interface with my existing SB 3800 in grid-tie mode?

    4) Replace the SB 3800 with a Xantrex XW6048 and charge controller. If this would work, can the charge controller handle the increased voltage from my 2 parallel PV strings (12 panels in series on each string) or would I need to re-wire them? When the grid is down, does the XW6048 allow the panels to feed the house, or just charge the batteries from which I can power the house?

    Please comment on which of these won't work and why. Any other possibilities?
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »
    Thanks for that. Based on your reply, here are my real questions. I plan to add some panels to my 18 Sharp 208's, probably another 6 or so (~1200 W). My SB 3800 won't be sufficient, so I'll need to modify/add to my inverter setup. In addition, in order of importance, I want to:

    1) Keep my grid tie (net metering) capability for all panels.
    2) Make use of my PV panels when the utility is down (just like every other beginner)
    3) Keep the total cost (acquisition + battery replacement) as low as practical.

    From what I've been able to learn, I think I have several choices:

    1) Buy a another SB grid tie inverter on the used market, allowing all 24 panels to be GT. Buy a used off grid inverter and charge controller that would be set up in parallel and activated automatically when the grid goes down.

    2) Buy another SB grid tie inverter on the used market, allowing all 24 panels to be GT. Buy a Sunny Island (which is just a switching box, not an inverter as I understand it) that would then allow all 24 panels to supply the house when the utility is down.

    3) Add a Xantrex XW4024 and charge controller to handle the 6 new panels. The question here is, when the utility goes down, will I have access to all 24 panels, or just the 6? Also, can the XW4024 interface with my existing SB 3800 in grid-tie mode?

    4) Replace the SB 3800 with a Xantrex XW6048 and charge controller. If this would work, can the charge controller handle the increased voltage from my 2 parallel PV strings (12 panels in series on each string) or would I need to re-wire them? When the grid is down, does the XW6048 allow the panels to feed the house, or just charge the batteries from which I can power the house?

    Please comment on which of these won't work and why. Any other possibilities?

    2) The Sunny Island is an inverter that makes AC from battery power. It's also a transfer switch and a fairly sophisticated computer. It can work in tandem with Sunny Boy inverters to power a microgrid when the grid is down, but it's more expensive than a GT inverter, 5000 Watts maximum output, only 120V single phase, and cannot be connected to PV modules.

    The bottom line is that you have to have batteries and enough of them to run whatever loads you want to back up for as long as you expect the grid to be down whether there is sunlight or not without depleting the batteries below 50% SOC. If you are imagining a scenario where you use a tiny battery on a battery inverter to fool a GT inverter into thinking that the grid is up when it's not so that you can run off the GT inverter, it won't work. Even if you could get it running during a grid outage, the first time a cloud passes over your house when you are running loads the whole system will shut down. Or whenever you engage more loads than the output of your GT inverter can instantaneously supply, it will shut down.

    You have to get past the idea that you can run your AC loads during a grid outage directly off a PV GT inverter without a battery inverter with significant battery reserves. Yes, it's expensive and the analysis you have to run to size your system must consider many more variables, but it's the real world.

    If grid outages are infrequent and/or of short duration, a straight grid tied system is virtually always the better choice from an economic standpoint. Yes, you will be sitting in the dark like everyone else when the power goes down, but you will be saving money when the grid is up and the sun is shining, and your system may eventually pay for itself.
  • leeelson
    leeelson Registered Users Posts: 18
    Re: Yet another GT workaround question
    ggunn wrote: »
    2) The Sunny Island is an inverter that makes AC from battery power. It's also a transfer switch and a fairly sophisticated computer. It can work in tandem with Sunny Boy inverters to power a microgrid when the grid is down, but it's more expensive than a GT inverter, 5000 Watts maximum output, only 120V single phase, and cannot be connected to PV modules.

    The bottom line is that you have to have batteries and enough of them to run whatever loads you want to back up for as long as you expect the grid to be down whether there is sunlight or not without depleting the batteries below 50% SOC. If you are imagining a scenario where you use a tiny battery on a battery inverter to fool a GT inverter into thinking that the grid is up when it's not so that you can run off the GT inverter, it won't work. Even if you could get it running during a grid outage, the first time a cloud passes over your house when you are running loads the whole system will shut down. Or whenever you engage more loads than the output of your GT inverter can instantaneously supply, it will shut down.

    You have to get past the idea that you can run your AC loads during a grid outage directly off a PV GT inverter without a battery inverter with significant battery reserves. Yes, it's expensive and the analysis you have to run to size your system must consider many more variables, but it's the real world.

    If grid outages are infrequent and/or of short duration, a straight grid tied system is virtually always the better choice from an economic standpoint. Yes, you will be sitting in the dark like everyone else when the power goes down, but you will be saving money when the grid is up and the sun is shining, and your system may eventually pay for itself.

    I accept a need for batteries. That's why my query didn't mention them. All 4 scenarios would have sufficient batteries, I understand that.

    From a financial point of view, wanting to use my PV array when the grid is down makes no sense since my utility is fairly reliable and I have a genset. Like a few other lunatics out there, I happen to believe that a sudden, prolonged utility outage is a possibility that I'm willing to pay to avoid.

    I suspect the Xantrex works similarly to the Sunny Island, except perhaps for the 120 V output. I also suspect that these inverters are less efficient than GT since they operate at low voltages.

    That leaves option 1). Is it feasible to have a parallel off-grid inverter/charge controller feeding batteries such that when the utility goes down and the grid tie inverter(s) go down, the OG system takes over? It seems that this should be possible with a manual transfer. What about an automatic transfer?
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »

    From a financial point of view, wanting to use my PV array when the grid is down makes no sense since my utility is fairly reliable and I have a genset. Like a few other lunatics out there, I happen to believe that a sudden, prolonged utility outage is a possibility that I'm willing to pay to avoid.
    What sort of situation would that be? Are you in a hurricane or earthquake zone, or are you worried about terrorists, EMP bombs, asteroid strikes, infrastructure collapse, etc.?
    leeelson wrote: »
    That leaves option 1). Is it feasible to have a parallel off-grid inverter/charge controller feeding batteries such that when the utility goes down and the grid tie inverter(s) go down, the OG system takes over? It seems that this should be possible with a manual transfer. What about an automatic transfer?
    The Sunny Island has an ATS (automatic transfer switch) built in as do the Xantrex hybrid inverters. SMA gear is AC coupled while Xantrex is DC coupled. With either it is possible to configure the system such that the PV continues to operate when the grid goes down.

    With the SMA system your protected loads are powered both from the SI inverter and the SB inverter, taking all the available power from the SB inverter and supplementing it as needed from the batteries. If your PV is producing more than your loads are demanding, the SI uses this excess to charge your batteries.

    With the Xantrex system, the batteries are always being charged by the PV through a charge controller and the inverter takes its power from the batteries and the output of the charge controller (which are on a single DC bus).

    With either system you'll need to decide which loads you want to keep running when the grid goes down and put them in a separate load center (protected loads panel) from the rest of the house. Once you quantify those loads you can size your battery bank accordingly.
  • leeelson
    leeelson Registered Users Posts: 18
    Re: Yet another GT workaround question
    ggunn wrote: »

    The Sunny Island has an ATS (automatic transfer switch) built in as do the Xantrex hybrid inverters. SMA gear is AC coupled while Xantrex is DC coupled. With either it is possible to configure the system such that the PV continues to operate when the grid goes down.

    With the SMA system your protected loads are powered both from the SI inverter and the SB inverter, taking all the available power from the SB inverter and supplementing it as needed from the batteries. If your PV is producing more than your loads are demanding, the SI uses this excess to charge your batteries.

    With the Xantrex system, the batteries are always being charged by the PV through a charge controller and the inverter takes its power from the batteries and the output of the charge controller (which are on a single DC bus).

    With either system you'll need to decide which loads you want to keep running when the grid goes down and put them in a separate load center (protected loads panel) from the rest of the house. Once you quantify those loads you can size your battery bank accordingly.


    OK. Let me re-phrase the question:

    I have a SB 3800 grid tie inverter, 18 Sharp 208 panels and a 6 kW Guardian genset. I plan to add some panels to my 18 Sharp 208's, probably another 6 or so (~1200 W). My SB 3800 won't be sufficient, so I'll need to modify/add to my inverter setup. In addition, in order of importance, I want to:

    1) Keep my grid tie (net metering) capability for all panels.
    2) Make use of all my PV panels when the utility is down.


    I think the most effective way to do this is to buy a another SB grid tie inverter on the used market, allowing all 24 panels to be GT. I would also buy a used off grid inverter, batteries and charge controller that would be set up in parallel and activated automatically when the grid goes down. Is this feasible?

    Note: I've taken Xantrex and Sunny Island out of the discussion.
  • mike95490
    mike95490 Solar Expert Posts: 9,583 ✭✭✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »
    OK. Let me re-phrase the question:

    I have a SB 3800 grid tie inverter, 18 Sharp 208 panels and a 6 kW Guardian genset. I plan to add some panels to my 18 Sharp 208's, probably another 6 or so (~1200 W). My SB 3800 won't be sufficient, so I'll need to modify/add to my inverter setup. In addition, in order of importance, I want to:

    1) Keep my grid tie (net metering) capability for all panels.
    2) Make use of all my PV panels when the utility is down.


    I think the most effective way to do this is to buy a another SB grid tie inverter on the used market, allowing all 24 panels to be GT. I would also buy a used off grid inverter, batteries and charge controller that would be set up in parallel and activated automatically when the grid goes down. Is this feasible?

    Note: I've taken Xantrex and Sunny Island out of the discussion.

    So to do this, you need a High Voltage DC controller - only one is the XW 600-80
    and inverter of your choice ?
    And on a sunny afternoon, when the batteries are full, where does the power go, and how does it get steered that way ?
    Powerfab top of pole PV mount | Listeroid 6/1 w/st5 gen head | XW6048 inverter/chgr | Iota 48V/15A charger | Morningstar 60A MPPT | 48V, 800A NiFe Battery (in series)| 15, Evergreen 205w "12V" PV array on pole | Midnight ePanel | Grundfos 10 SO5-9 with 3 wire Franklin Electric motor (1/2hp 240V 1ph ) on a timer for 3 hr noontime run - Runs off PV ||
    || Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
    || VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A

    solar: http://tinyurl.com/LMR-Solar
    gen: http://tinyurl.com/LMR-Lister ,

  • BB.
    BB. Super Moderators, Administrators Posts: 33,433 admin
    Re: Yet another GT workaround question

    In theory, get a True Sine Wave Off Grid Inverter with Internal transfer switch. Connect the AC Line to Inverter AC input (or one pole of your transfer switch), and connect the output of the inverter to your load + GT inverters (GT inverter output must be less than the maximum power rating of your off-grid inverter).

    When AC power is good, the TSW OG Inverter is in pass-through mode. And the loads/GT inverters are fully functional.

    When the AC fails, the TSW OG Inverter disconnects from the Mains and supplies energy to the loads and line sync to the GT inverter (note, TSW inverter probably needs to be have a fairly accurate time base--less than ~1% +/- Hz.

    The next issue--say you have light loads and good GT output back driving your AC Coupled OG inverter... Some inverters (like XW and SunnyIsland) will adjust the frequency up as the batteries reach "full charge"... Eventually, frequency will go high enough and start dropping the GT inverters (with 5 minute timeout for reconnect).

    What will the "generic" TSW OG Inverter do? You might have to call the vendor and find out (Magnum might be a good place to call... So might Exeltech).

    The option is to install a dump controller+resistive (or other) loads on your battery bank (just like setting up a Wind Turbine).

    I would try Magnum first... They have 120/240 VAC split phase inverters (120/240) which, I assume is your first choice (saves having to play games trying to match 120v OG inverter to 240 only GT inverter and powering your 120/240 VAC home.

    I don't know enough about inverters to say that all TSW inverters can be AC coupled/back driven to charge the battery bank... And certainly a much smaller set has been designed and tested for this sort of operation.

    From what I understand, probably no MSW inverter will make the GT Inverters work/sync up correctly (and you would get into AC/battery grounding issues).

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »
    OK. Let me re-phrase the question:

    I have a SB 3800 grid tie inverter, 18 Sharp 208 panels and a 6 kW Guardian genset. I plan to add some panels to my 18 Sharp 208's, probably another 6 or so (~1200 W). My SB 3800 won't be sufficient, so I'll need to modify/add to my inverter setup. In addition, in order of importance, I want to:

    1) Keep my grid tie (net metering) capability for all panels.
    2) Make use of all my PV panels when the utility is down.


    I think the most effective way to do this is to buy a another SB grid tie inverter on the used market, allowing all 24 panels to be GT. I would also buy a used off grid inverter, batteries and charge controller that would be set up in parallel and activated automatically when the grid goes down. Is this feasible?

    Note: I've taken Xantrex and Sunny Island out of the discussion.
    In my opinion, it could work - conditionally - but some conditions could take it down. One such is the case where it's cloudy (or dark) and your loads exceed the capability of your battery inverter. Another is when the sun is shining, your loads are less than the output of your PV, and the batteries are fully charged, though in a Sunny Island/Sunny Boy system the Sunny Island will munge the frequency of the microgrid when this happens, which will dial down the output of the Sunny Boy(s).

    Another thing to consider is that battery inverters which are grid interactive have a limit on their passthrough current capability which is usually somewhere near what their power delivery capability is. If you put a small one between a large PV system and/or large protected loads panel and the grid connection, it will be a bottleneck for current flowing either direction. In some systems, a bypass switch can be installed to route current around the battery inverter when the grid is up, but if your battery inverter also has PV on it, that PV is then off line. That's another case where the Sunny Island may be preferable because it does not support PV.

    In any case (and I don't mean to overstate the obvious), you must separate your loads into protected and non-protected (grid connected) panels, and IMO a good design would dictate that the sum of your protected loads not exceed the capability of your battery inverter.

    To cut to the chase, a main consideration is that your battery inverter be capable of delivering enough power so that your loads are covered as the output of your PV bounces around. I don't mean to talk down to you but a common misconception is for folks to look at the rated output current of a PV inverter and think that means that the inverter puts out that quantity of current steady state as long as it is running. Nothing could be further from the truth. It puts out the current equal to the power generated by the PV array divided by the line voltage, subject to its efficiency and limited by its maximum - no more, no less. You cannot control it.

    To run successfully and sustainably off grid you must have enough battery capacity to cover the deficiency when the output of your PV inverter falls below the demand of your loads, as well as to absorb the excess when the output of your PV inverter exceeds your loads. One or the other of those two conditions exists virtually all the time.
  • leeelson
    leeelson Registered Users Posts: 18
    Re: Yet another GT workaround question
    mike90045 wrote: »
    And on a sunny afternoon, when the batteries are full, where does the power go, and how does it get steered that way ?

    How does this happen in most off grid configurations?
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »
    How does this happen in most off grid configurations?

    Off-grid systems have their PV output regulated by the charge controller. Once the batteries are charged the PV is disconnected from the battery. Battery Voltage drops; it reconnects.

    GT inverters do not function this way. They take whatever DC is available from the panels and convert it to AC and push it to the grid by virtue of their output being higher Voltage than the grid (Voltage difference becomes current flow). They do not need to regulate output other than by frequency and not exceeding Vmax: the grid can take everything they put out due to its massive electrical inertia. A stand-alone generator is not meant to be back fed current at all, and an off-grid inverter is not designed to but can accept a limited amount of back-feeding. It turns this into DC current flowing to the batteries, but has no system to regulate that back-fed current once the batteries are full.

    The Xantrex XW and SMA Sunny Island are two battery-based inverters actually designed to take back-fed AC and regulate it. If this is what you need/want to do use one of those. Do not expect a device to function properly (or at all) under circumstances it was never designed to handle.
  • leeelson
    leeelson Registered Users Posts: 18
    Re: Yet another GT workaround question
    Off-grid systems have their PV output regulated by the charge controller. Once the batteries are charged the PV is disconnected from the battery. Battery Voltage drops; it reconnects.

    GT inverters do not function this way. They take whatever DC is available from the panels and convert it to AC and push it to the grid by virtue of their output being higher Voltage than the grid (Voltage difference becomes current flow). They do not need to regulate output other than by frequency and not exceeding Vmax: the grid can take everything they put out due to its massive electrical inertia. A stand-alone generator is not meant to be back fed current at all, and an off-grid inverter is not designed to but can accept a limited amount of back-feeding. It turns this into DC current flowing to the batteries, but has no system to regulate that back-fed current once the batteries are full.

    The Xantrex XW and SMA Sunny Island are two battery-based inverters actually designed to take back-fed AC and regulate it. If this is what you need/want to do use one of those. Do not expect a device to function properly (or at all) under circumstances it was never designed to handle.
    I guess I was pretty vague in a previous post. I'm trying to determine whether I can have 2 parallel systems: a GT system that is in effect when the utility is up. When the utility goes down, a separate system (inverter, charge controller, batteries would take over. That off-grid system would then function as any off grid system, hence my question to mike90045.

    I gather that the problem here is that its not obvious whether the OG inverter would be smart enough to switch automatically when the utility goes away. Another problem (a deal killer, perhaps) is how to find a charge controller that can handle the high voltage and power (5 kw) configuration of my PV panels that need to be configured for GT operation.
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »
    I guess I was pretty vague in a previous post. I'm trying to determine whether I can have 2 parallel systems: a GT system that is in effect when the utility is up. When the utility goes down, a separate system (inverter, charge controller, batteries would take over. That off-grid system would then function as any off grid system, hence my question to mike90045.

    I gather that the problem here is that its not obvious whether the OG inverter would be smart enough to switch automatically when the utility goes away. Another problem (a deal killer, perhaps) is how to find a charge controller that can handle the high voltage and power (5 kw) configuration of my PV panels that need to be configured for GT operation.

    Actually that's not at all difficult. Having an off-grid inverter handle emergency loads when the grid goes down is easy: all inverter-chargers have built-in transfer switches which connect loads to AC IN when AC is present. When it isn't they switch to inverter mode and pick up the loads seamlessly.

    To recharge the batteries from an existing GTI's array requires either some custom array re-wiring or the very expensive Xantrex XW 80MPPT-600 (the only charge controller I know of capable of the up to 600 Volt input normally used by central GT inverters). Depending on the actual configuration of the existing GT array, it is possible to switch out enough panels to supply recharging power for the batteries (it is not likely you'd need the full 5kW array for recharging limited back-up power reserves).
  • leeelson
    leeelson Registered Users Posts: 18
    Re: Yet another GT workaround question

    To recharge the batteries from an existing GTI's array requires either some custom array re-wiring or the very expensive Xantrex XW 80MPPT-600 (the only charge controller I know of capable of the up to 600 Volt input normally used by central GT inverters). Depending on the actual configuration of the existing GT array, it is possible to switch out enough panels to supply recharging power for the batteries (it is not likely you'd need the full 5kW array for recharging limited back-up power reserves).

    The current array configuration has 2 sets of 9 panels (each of the nine in series) in parallel with each other. Each panel's open circuit voltage is 36 V. I want to add 6 more panels. I want all 24 panels to be grid tied when the utility is up. Since I will have to add a GTI to accommodate the 6 new panels, I suppose I can put the 6 on a separate string. When the utility is down, perhaps I can use 1 or 2 of the 3 strings.
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Yet another GT workaround question
    Off-grid systems have their PV output regulated by the charge controller. Once the batteries are charged the PV is disconnected from the battery. Battery Voltage drops; it reconnects.
    Not to contradict you, but a Sunny Island/Sunny Boy system does this a bit differently, since it is AC coupled and does not have a charge controller. The SI alters the AC frequency in the microgrid when the batteries approach 100% SOC, and you can set the Sunny Boy(s) to interpret this as a signal to dial down their output. When the batteries reach 100% SOC, the frequency is shifted to the cutoff point for the SB's, and they shut down completely.
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »
    I guess I was pretty vague in a previous post. I'm trying to determine whether I can have 2 parallel systems: a GT system that is in effect when the utility is up. When the utility goes down, a separate system (inverter, charge controller, batteries would take over. That off-grid system would then function as any off grid system, hence my question to mike90045.

    I gather that the problem here is that its not obvious whether the OG inverter would be smart enough to switch automatically when the utility goes away. Another problem (a deal killer, perhaps) is how to find a charge controller that can handle the high voltage and power (5 kw) configuration of my PV panels that need to be configured for GT operation.
    Why have two PV systems? You could put a Sunny Island (with batteries) between your GT system and the grid with a protected loads panel (microgrid) between them. When the grid is up, the SI is in bypass mode and is transparent. When the grid goes down, the SI disconnects the microgrid from the grid and arbitrates the power to the microgrid between the PV and the batteries.

    The SI would handle not overcharging the batteries (see my previous post in this thread). Ideally, your GT inverter(s) would be Sunny Boy(s) which would taper off their output as the batteries approach 100% SOC, but I believe other inverters would work, too, although they would shut off ungracefully when the frequency hit their tolerance limit.

    You would have some details to work out; the SI is a 120VAC inverter, so if your GT inverter is 240VAC, you'd have to include an autotransformer (if you can live within the 5000W limit of the passthrough mode of the SI) or a second SI (if you can't). Also, you'll have to have enough battery capacity to handle your loads when the PV is not producing enough (or at all).

    I want to be gentle about this, but it seems to me that you are struggling to design a rather complex PV/backup system with limited knowledge of the alternatives that are out there and of how all this stuff works. That is not a slam, just an observation.
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Yet another GT workaround question
    ggunn wrote: »
    Not to contradict you, but a Sunny Island/Sunny Boy system does this a bit differently, since it is AC coupled and does not have a charge controller. The SI alters the AC frequency in the microgrid when the batteries approach 100% SOC, and you can set the Sunny Boy(s) to interpret this as a signal to dial down their output. When the batteries reach 100% SOC, the frequency is shifted to the cutoff point for the SB's, and they shut down completely.

    I wasn't talking about the Sunny Island or the Xantrex XW because he had "taken them out of the discussion" and inquired about "most off-grid systems".

    There's quite a difference between trying to run the whole GT system connected to a battery-based OG inverter of any type and having two separate systems. The XW or the SMA SI work fine for the first scenario, as they were designed for it. For the second any OG inverter will do (some a bit easier to apply than others).
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »
    The current array configuration has 2 sets of 9 panels (each of the nine in series) in parallel with each other. Each panel's open circuit voltage is 36 V. I want to add 6 more panels. I want all 24 panels to be grid tied when the utility is up. Since I will have to add a GTI to accommodate the 6 new panels, I suppose I can put the 6 on a separate string. When the utility is down, perhaps I can use 1 or 2 of the 3 strings.

    That's two strings of nine panels in series.

    You probably will not be able to use a whole string as its Voltage will be too high for any charge controller other than the XW MPPT80-600.

    The next step is to determine what you back-up power system needs to be in terms of inverter size and battery bank capacity. Then you need to figure out how much panel is needed to recharge those batteries.

    There are controllers from Midnite Solar which can take up to 250 Volts max input. This would allow you to use 2/3 of a string of panels: 36 Voc * 6 = 216 Volts and 1248 Watts. For this you would have the strings' negative common to the input of the controller and install some method of switching out the panels between #6 and #7 in the string to feed positive to the controller when the grid is down. You may need more than one string depending on what you go for in back-up power.

    If you are considering this take note: It would be necessary to make sure the wiring/switching involved followed NEC regs and would be able to handle the current without much Voltage drop. Even so, probably neither the NEC nor an inspector would not approve of such an arrangement although I can't think why).

    Another reason why it is probably much simpler to just buy a Sunny Island and utilize the full array and GT inverters already in place.
  • leeelson
    leeelson Registered Users Posts: 18
    Re: Yet another GT workaround question
    ggunn wrote: »
    Why have two PV systems? You could put a Sunny Island (with batteries) between your GT system and the grid with a protected loads panel (microgrid) between them. When the grid is up, the SI is in bypass mode and is transparent. When the grid goes down, the SI disconnects the microgrid from the grid and arbitrates the power to the microgrid between the PV and the batteries.


    The SI would handle not overcharging the batteries (see my previous post in this thread). Ideally, your GT inverter(s) would be Sunny Boy(s) which would taper off their output as the batteries approach 100% SOC, but I believe other inverters would work, too, although they would shut off ungracefully when the frequency hit their tolerance limit.

    You would have some details to work out; the SI is a 120VAC inverter, so if your GT inverter is 240VAC, you'd have to include an autotransformer (if you can live within the 5000W limit of the passthrough mode of the SI) or a second SI (if you can't). Also, you'll have to have enough battery capacity to handle your loads when the PV is not producing enough (or at all).

    I want to be gentle about this, but it seems to me that you are struggling to design a rather complex PV/backup system with limited knowledge of the alternatives that are out there and of how all this stuff works. That is not a slam, just an observation.

    Again, I've been too vague. I'm not proposing 2 PV systems. If you look at earlier posts, I'm trying to be GT when the utility is up and OG when its down, using 1 PV system. And yes, I do have "limited knowledge of the alternatives". That's why I'm posting on the "beginners" forum: to increase my knowledge :D.

    I'm trying to get a general idea of which is the most cost effective approach to take, not design a system in detail. From what I can glean, I have 3 general options. Let's examine 2 of them: using the Sunny Island and using an off-grid parallel system. Both options include 24 Sharp panels, sufficient batteries and GT inverters.

    The Sunny Island (at about $4K) would probably also require an autotransformer since I need 240 split phase. I don't understand the "passthrough" comment. Where does the 5 kw limit come from? Is this just a battery charging limit or does it affect the grid interaction? Anything else required? Does the SI require a separate switching device? Cost?

    The parallel system would require an off grid inverter (3-5 kw) and charge controller. It looks like I could get a 4 kw OG inverter new for around $2.5 K and a charge controller for around $1.2K but I suspect I could buy these used for about half that (at least the inverter). Anything else needed?

    If the above is correct, it appears that the parallel system has cost advantages, although it provides less power and is more complex to set up.

    Make sense?
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Yet another GT workaround question
    leeelson wrote: »

    The Sunny Island (at about $4K) would probably also require an autotransformer since I need 240 split phase. I don't understand the "passthrough" comment. Where does the 5 kw limit come from? Is this just a battery charging limit or does it affect the grid interaction? Anything else required? Does the SI require a separate switching device? Cost?

    When the grid is up, the SI is in passthrough mode, i.e., the AC in (grid) is connected to the AC out (protected loads, aka microgrid) through its internal transfer switch, the inverter section is asleep, and the battery charger is pulling whatever it needs from the AC bus. Power flows freely and bidirectionally through the SI, except that it is limited to (I believe; it's from memory) 52A (check that number). So no, it's not related to the battery charging limit and yes, it affects the grid interaction in that it limits the current in and out of the protected loads panel (microgrid) from the grid side.

    If the current limit is a problem, then you may stack two SI's with one on one leg of the 240VAC and one on the other to double the limit and obviate the need for an autotransformer to connect with your GT inverter. Actually you can combine as many as you like up to some crazy limit.

    When the grid goes down, the SI opens the internal transfer switch (disconnecting the protected loads from the grid) and turns on its inverter section, which feeds the AC out on a demand basis. If the PV inverter on the microgrid is providing less than the demand, the SI makes up the difference from the batteries. If the PV inverter is oversupplying the demand, the SI pumps the excess into the batteries. If the batteries are full, the SI dials down the output of the PV inverter (if it's a Sunny Boy) to fit the load.

    I'm not sure how nice the SI will play with non-SMA GT inverters.

    Clear?
  • jagec
    jagec Solar Expert Posts: 157 ✭✭
    Re: Yet another GT workaround question

    Hmm, I guess a real cheapskate way to roll would be to hook your grid-tied array directly to a shunt controller with an appropriate dump load. That would crash your array voltage all the way down to battery voltage, making your nice 3.7 kW array a mere 600-watt array (assuming 48V battery bank), but at least it would still be producing power. Add a 48-to-12V DC-DC converter and a cheapo MSW inverter, and you're in business for a tenth of the price of the Sunny island. Of course, you're throwing away most of your solar harvest when the grid is down, but how often does that really happen?

    Disclaimer: There might be some problems with this idea that I haven't foreseen; I'm sure the experts will jump in and correct me if that's the case.
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Yet another GT workaround question

    What could possibly go wrong with running a several hundred Volt, 3.7 kW PV array at its Isc limit for several hours a day? :roll:
  • jagec
    jagec Solar Expert Posts: 157 ✭✭
    Re: Yet another GT workaround question
    What could possibly go wrong with running a several hundred Volt, 3.7 kW PV array at its Isc limit for several hours a day? :roll:

    I don't know; my 100% legal, grid-tied, professionally installed array runs at 94% of Isc full-time without issue. If there is an operational difference between the two different points on the power curve I genuinely would like to hear it...I'm here to learn!

    shunt.png
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Yet another GT workaround question

    Well for one thing PWM battery charge controllers are meant to have an input Voltage in the neigbourhood of the expected Vmp for the system Voltage. Applying several hundred Volts to one will probably fry it instantly.

    On the bright side, if there are any loose or high resistance connections anywhere in the system you'll find them very quickly. The firemen might even let you run the siren! :p
  • jagec
    jagec Solar Expert Posts: 157 ✭✭
    Re: Yet another GT workaround question
    Well for one thing PWM battery charge controllers are meant to have an input Voltage in the neigbourhood of the expected Vmp for the system Voltage. Applying several hundred Volts to one will probably fry it instantly.

    On the bright side, if there are any loose or high resistance connections anywhere in the system you'll find them very quickly. The firemen might even let you run the siren! :p

    Which is the reason for using a shunt controller (which diverts the input between the batteries and a resistance-matched dump load as needed, thus maintaining the solar panels at a continuous low voltage) rather than a standard PWM (which connects and disconnects the array to the batteries as needed, resulting in the array reaching Voc when the duty cycle is below 100%).

    Loose and high resistance connections would already be well known, as the system operates at the same amperage (but a much higher voltage) in grid-tied mode.

    Again, I would love to hear what the potential issues are with this thought experiment, rather than just dismissing it out of hand. Yes, it is unconventional and inefficient--I certainly don't plan on setting it up on my system--but why wouldn't it work?
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Yet another GT workaround question

    The panels themselves won't suffer for it. The problem will occur in whatever section of the system all that power is choked down to battery Voltage level. If you had a shunt controller designed for such high Voltage it might work. Do you know of any designed to operate at 250+ Volts? I don't. Switching high Voltage DC isn't easy, and any unit capable of it isn't likely to be cheap either.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,433 admin
    Re: Yet another GT workaround question

    The battery should keep the voltage at ~14.5 volts (or whatever the bank voltage is)--The shunt controller will have to bleed off excess power... And, in theory, you should have two independent shunt controllers + diversion loads (backup/redundancy/safety).

    If you get a broken connection somewhere, the higher voltage sure make the chances of a sustained arc higher.

    My two cents.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • jagec
    jagec Solar Expert Posts: 157 ✭✭
    Re: Yet another GT workaround question
    BB. wrote: »
    The battery should keep the voltage at ~14.5 volts (or whatever the bank voltage is)--The shunt controller will have to bleed off excess power... And, in theory, you should have two independent shunt controllers + diversion loads (backup/redundancy/safety).

    If you get a broken connection somewhere, the higher voltage sure make the chances of a sustained arc higher.

    My two cents.

    -Bill

    Very true. Perhaps an upstream overvoltage disconnect would work as well, at a lower cost than full redundancy (I believe they make low-voltage surge suppressors for the telecom industry that might work?). Of course, the necessary added levels of safety make this system somewhat less inexpensive than it appears at first blush.
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Yet another GT workaround question

    jagec you may be on to something here.

    Consider this; since the dump load (undefined) is going to be needed anyway, perhaps putting it (or 'a') in series with the panels to drop the over-all Voltage would work. As you said, the full power of the array would be drawn down to a small portion anyway, so why not make use of it instead?

    Hmm. Serial loading of high Voltage arrays. Got to think about that for a bit. ;)