Off Grid Home AC or DC Coupled? Help me decide.

[raenrfm]

I keep waffling between choosing an AC coupled vs a DC coupled system. I'd like to use the schnieder XW inverter for it's generator support, but I also like what I hear about SMA, but I'm not sure they can sync with a generator like the XW can. I also understand that an AC coupled system is not self starting if your batteries go south, but I like the cost savings in wire with an AC coupled system and AC is just better understood by electricians etc, also with microinverters, expansion is easy.

Aside from the issue of not being self starting (which could be addressed with proper use of the genset to make sure your batteries never get exhausted), what are the pro's and con's of each system? The two manufacturers I like are Schneider and SMA, I have the design tools for both, and it seems for some reason the SMA wants me to be using 2 Sunny Islands, which really bumps up the cost of that system vs the Schneider. So could anyone comment on either manufacturer in terms of their reliability/warranty/service etc...

[BackwoodsEE]

I've been going through this decision process as well, except for a grid-tie hybrid system. For me, the decision would seem to be even more weighted toward AC coupling, since I already have the grid right here. Start out with some SMA Sunny Boys and then later add Sunny Islands (two of them, for split-phase 120/240V) and a battery to get the off-grid apocalypse functionality that my paranoid little heart desires.

But a couple (no pun intended) of things keep bringing me back to the charge controllers + Outback Radian setup. The SMA stuff seems overly complex, fragile even, and there is a lot of end-user lockout that makes me feel like I'd always be on the outside of the system trying to defeat various barriers to get in and make it do what I want. You have to apply for a password (presumably with some sort of credentials shown, is a BSEE enough?) to make significant changes to the SI's grid-tie settings after ten hours of operation. There are just four battery types (including the never-used NiCad for some reason) supported, and the Lithium capability is restricted to a list of approved batteries that communicate with the SI via its digital bus (proprietary?).

I don't know anything about Xantrex, but have read enough comments about Schneider's lack of customer support to have not bothered looking into it further.

The Morningstar TS-MPPT charge controllers are simple and solid. I have one, a TS-MPPT-60, for my 3-panel experimental system and have checked its performance out thoroughly with a custom IV tracing/monitoring device I designed and built. It pulls out the power that's available and needed from the PV array in fine fashion, even under adverse conditions. I like it a lot. Also, I was impressed with a recent response I got from Morningstar tech support when I inquired about having a LiFePO4 BMS communicate with the charge controller.

There is an appealing aspect of AC coupling: Everything is done at high voltage, never having to be converted down to the 48V of the battery and back up again unless power is actually being stored or withdrawn from storage. On the other hand, storage is done with DC, and you have to go through inversion and then rectification to get DC from the PV array to the battery. But these are really philosophical and aesthetic considerations that I'm probably letting distract me from the real issues: What is the efficiency from PV array to grid or battery in each system, and how well does the charging work?

As far as charging goes, it seems that DC coupling is a bit better. Both the DC charge controller and Outback's Radian charger seem more finely tuned to the battery's needs than back feeding an inverter H-bridge and using it as a rectifier to pump current directly into the battery. The SI has to tell the grid-tie inverter(s) to produce less power by tweaking the frequency of the AC, because a back-fed inverter H-bridge can't regulate its output. Other AC coupling solutions use cruder regulation methods, like Outback's use of a remote-trip breaker to simply open-circuit the connection between Radian inverter and grid-tie inverter. Or, as one prominent YouTube PV installer memorably put it, "mad scientist water-heater dump loads."

One more aspect is NEC compliance. I'm determined to avoid the 690.12 Rapid Shutdown nonsense at all costs, as well as 690.11 Arc Fault. The cleanest way to do that with my planned ground-mounted array is to keep the PV circuits out of my garage entirely. So, either the Sunny Boys or, more likely, the Morningstar CCs, will be mounted outside the garage in their own weatherproof enclosure with either AC or 48V DC going into the garage. There, in a relatively temperature-controlled environment, will be my Sunny Islands (or Outback Radian) and battery bank (either AGM or CALB LiFePO4 cells). The Sunny Boys are probably slightly more suited for sitting outside, and 240V AC is a lot better for making the run from outside to inside the garage than 48V DC. It will be less than 15 feet, but that's not nothing when you are talking about two 60A charge controllers.

[raenrfm]

Do you know if the Sunny Islands will sync with a generator? I really like that idea, using the generator/inverter combo to assist with transient heavy loads.

The other thing I like about AC coupling though is that power from the array will go directly to loads and any access will go into the batteries, which seems like a better strategy in terms of reducing the cycling on the batteries and also while you will be converting DC-AC-DC to charge the batteries, is that potential loss in efficiency greater than the increase efficiency of using power right off the array before you store the access?

I still can't decide. Also what you say about the SMA programming being firewalled is a little concerning. If I buy a unit I want full control of it. Also if for example I'm away and my wife needs to look after something with the system, I don't want it overly complicated, and really want it to be as autonomous as possible.

[ButchDeal]

The other thing I like about AC coupling though is that power from the array will go directly to loads and any access will go into the batteries, which seems like a better strategy in terms of reducing the cycling on the batteries and also while you will be converting DC-AC-DC to charge the batteries, is that potential loss in efficiency greater than the increase efficiency of using power right off the array before you store the exccess?

D.C. Coupling has similar logic but more importantly because of the AC coupling control methods, AC coupling will have more battery cycling generally.

[B-17]

If it's off grid and smaller than about 16kW, then go DC coupled. If it's grid interactive then AC coupling can be worth considering if it's at least 7kW in size. AC coupling can have it's place but DC coupling often has more advantages

[raenrfm]

D.C. Coupling has similar logic but more importantly because of the AC coupling control methods, AC coupling will have more battery cycling generally.

See I would have thought the opposite would be true because your loads during the day would first use power from the array without even involving the batteries, and only if your loads exceed the arrays output would it start to draw from storage. I guess it really depends on the size of the array and the loads you're drawing.

[ButchDeal]

See I would have thought the opposite would be true because your loads during the day would first use power from the array without even involving the batteries, and only if your loads exceed the arrays output would it start to draw from storage. I guess it really depends on the size of the array and the loads you're drawing.

On a D.C. Coupled system you can use power from the array without involving the batteries. It is impossible to charge and discharge batteries at the same time you know.

[jflorey2]

I keep waffling between choosing an AC coupled vs a DC coupled system. I'd like to use the schnieder XW inverter for it's generator support, but I also like what I hear about SMA, but I'm not sure they can sync with a generator like the XW can. I also understand that an AC coupled system is not self starting if your batteries go south, but I like the cost savings in wire with an AC coupled system and AC is just better understood by electricians etc,

I don't know of many electricians who understand AC coupling. It's not just wiring an inverter/battery bank to a solar array.

also with microinverters, expansion is easy.

With strings that is true as well.

Aside from the issue of not being self starting (which could be addressed with proper use of the genset to make sure your batteries never get exhausted), what are the pro's and con's of each system?

Pros for hybrid system:
Lower wire cost (500VDC from an array requires less copper than 240VAC from array, and the high current 48V wiring is short)
Systems are already designed for doing this and all the hardware is readily available

Pros for AC coupled system:
Can use microinverters

The other thing I like about AC coupling though is that power from the array will go directly to loads and any access will go into the batteries, which seems like a better strategy in terms of reducing the cycling on the batteries and also while you will be converting DC-AC-DC to charge the batteries, is that potential loss in efficiency greater than the increase efficiency of using power right off the array before you store the access?

In a DC coupled system you do not cycle the batteries when solar power exceeds loads. The power goes directly from the solar array, through the charge controller, through your inverter and to your AC loads.

[karrak]

But a couple (no pun intended) of things keep bringing me back to the charge controllers + Outback Radian setup. The SMA stuff seems overly complex, fragile even, and there is a lot of end-user lockout that makes me feel like I'd always be on the outside of the system trying to defeat various barriers to get in and make it do what I want. You have to apply for a password (presumably with some sort of credentials shown, is a BSEE enough?) to make significant changes to the SI's grid-tie settings after ten hours of operation. There are just four battery types (including the never-used NiCad for some reason) supported, and the Lithium capability is restricted to a list of approved batteries that communicate with the SI via its digital bus (proprietary?).

I don't know anything about Xantrex, but have read enough comments about Schneider's lack of customer support to have not bothered looking into it further.

The Morningstar TS-MPPT charge controllers are simple and solid. I have one, a TS-MPPT-60, for my 3-panel experimental system and have checked its performance out thoroughly with a custom IV tracing/monitoring device I designed and built. It pulls out the power that's available and needed from the PV array in fine fashion, even under adverse conditions. I like it a lot. Also, I was impressed with a recent response I got from Morningstar tech support when I inquired about having a LiFePO4 BMS communicate with the charge controller.

I agree with you about the overcomplexity and vendors trying to lock you in to only using their equipment with LiFePo4 (LFP) batteries. For high end applications like EVs which require high charge and discharge rates which at times run the batteries close to their operating limits I can see the need for complex communications between the charger, load and battery BMS.

For standard off-grid systems where the battery will be operating at less than half their rated charge and discharge rates I think the only communication that is necessary is a signal from the BMS to the charger and loads to say STOP which could be as simple as a relay or solid state switch. Older charge controllers and inverters had no such facility so the usual approach was to have the BMS operate large power relays to disconnect the solar array from the charge controller or disconnect the battery from the inverter. A large number of newer inverters and charge controllers now have the facility to be switched off remotely via a simple switch or relay. If you have a charge controller with a remote battery temperature sensor that has no option to remotely stop charging and have an electronics background it would not be to hard to knock up a little circuit connected to the BMS with a relay and a couple of variable resistors that would fool the battery temperature compensation circuit to think the battery temperature was too high and stop charging the battery.

Simon

Off grid 24V system, 6x190W Solar Panels, 32x90ah Winston LiFeYPO4 batteries installed April 2013
BMS - Homemade Battery logger github.com/simat/BatteryMonitor
Latronics 4kW Inverter, homemade MPPT controller

[BackwoodsEE]

If you have a charge controller with a remote battery temperature sensor that has no option to remotely stop charging and have an electronics background it would not be to hard to knock up a little circuit connected to the BMS with a relay and a couple of variable resistors that would fool the battery temperature compensation circuit to think the battery temperature was too high and stop charging the battery.

That remote temp sensor trick is exactly what Morningstar suggests doing for their TS-MPPT charge controllers. They recommend leaving a resistor in place at all times and then having the BMS switch another resistor in parallel with it when the CC is to be stopped. Seems a little gimmicky to me, given that the consequences for the resistor not switching in properly are very expensive. But I suppose it might be all right when used as just a backup for the CC's own battery voltage sensing.

[Sunking]

For high end applications like EVs which require high charge and discharge rates which at times run the batteries close to their operating limits I can see the need for complex communications between the charger, load and battery BMS.

That is complete nonsense. EV's charge rate is less than an Off-Grid Solar Battery System at Level 1 rates using 120 VAC. Slightly less than or equal to Off-Grid at Level 2 208-240 VAC. Only at level 3 and 4 do they charge as fast or faster than Off-Grid. Verry few ever use it in the USA because of the expense and large amount of power required. 95% of EV customers in the USA use Level 1 Charge Stations at their home. L1 charge rate varies depending on which EV battery size you have. Typically C/30 to C/15. Off-Grid Solar if sized properly will charge at as fast as C/2, or as slow as C/8. You keep exposing how big of a fraud you really are Karrak.

EV's built-in Charger and BMS do communicate using several different proprietary topologies. In the rare event balancing is required, the BMS will signal the Charger to reduce charge current to the amount the Bypass Circuit can shunt around the charge cell. If it did not do that would still leave current flowing through a charged cell. Then once all cells are at the same SOC the BMS will turn off the charger. At no time will any commercial EV BMS ever allow the customer to fully charge the battery. Under normal operating conditions where no Balance is required is a straight forward CC/CV charger, and the BMS will turn off the charger when current tapers to a set value determined by the manufacturer.

There are no off the shelf systems like that for Solar Charge Controllers using LiFeP04 cells. At best just a signal to the charge controller to turn off the controller which is a horrible design for two reasons.

1. Uses Vampire Boards that cannot bypass all the charge current, thus overcharging cells, It takes cells to 100% plus SOC which you should never fully charge a lithium battery.
2. When the Charge Controller turns off, you are on battery power despite it may be high noon with 6 to 9 hours of daylight left leaving your solar panels doing nothing the rest of the day while you use battery power.

With a Low Voltage 16S or less LFP battery no BMS is required. You just use a simple CC/CV set to 3.3 to 3,45 vpc to take the cells to roughly 90% SOC and let them float on solar power until the sun sets before you stat using battery power. That is something Karrak has not wrapped his thick skull around yet.

[raenrfm]

Hey sunking I appreciate your input but I wonder how that's going to help me decide on the design of my off grid home? I don't have an EV note do I plan on getting one. I want to know people's experiences with AC vs DC coupled systems as I'm wrestling with the pros and cons of each.

[jflorey2]

For standard off-grid systems where the battery will be operating at less than half their rated charge and discharge rates I think the only communication that is necessary is a signal from the BMS to the charger and loads to say STOP which could be as simple as a relay or solid state switch.

I would also note that Midnite has a remotely operated DC circuit breaker that can open the circuit when a cell goes under or over voltage. Since pretty much all battery installations have a main breaker, this doesn't require any new switching in the DC circuit (the highest current circuit in the system.)

[brucet9]

That is complete nonsense. EV's charge rate is less than an Off-Grid Solar Battery System at Level 1 rates using 120 VAC. Slightly less than or equal to Off-Grid at Level 2 208-240 VAC. Only at level 3 and 4 do they charge as fast or faster than Off-Grid. Verry few ever use it in the USA because of the expense and large amount of power required. 95% of EV customers in the USA use Level 1 Charge Stations at their home. L1 charge rate varies depending on which EV battery size you have.

At no time will any commercial EV BMS ever allow the customer to fully charge the battery. Under normal operating conditions where no Balance is required is a straight forward CC/CV charger, and the BMS will turn off the charger when current tapers to a set value determined by the manufacturer.

So much out-dated information!

There is no way that 95% of EV customers in USA charge at Level 1 at home anymore since battery sizes and driving ranges of real EVs have gotten bigger.

Small battery hybrids may charge at 120V AC, but longer range BEVs charge L2 at home at 240V and more than 6kW. 7,500 Chevy Bolt EVs can charge at 240V, 7.6kW. Teslas (more than 130,000 in the US) charge at 240V 9.6kW or 17.2kW. Nissan Leafs (about 100,000 in US) charge at 240V 6.6kW. BMW i3 (28,000 in US) at 240V 7.4kW. Every BEV coming out in the future will be at least 6.6kW charge rate, most of them considerably higher.

All of the large battery EVs enable charging to 100% of rated charge.

I haven't seen any home storage battery rated for more than 5kw and none with storage capacity above 13kWh; enough for only about 45 miles per day.

[jflorey2]

Small battery hybrids may charge at 120V AC, but longer range BEVs charge L2 at home at 240V and more than 6kW.

Agreed. Of the BEV/PHEV owners at my company (260 the last time we polled them) 78% had level 2 chargers. Generally the people who had only level 1 (or no) chargers were the PHEV crowd.