SPS Secure Power Supply in SMA Sunny Boy inverters

[phx]

SMA Sunny Boy inverters SB7.7-1SP-US40 and SB6.0-1SP-US41 have a "Secure Power Supply" (SPS) component, which can be wired to a separate SPS outlet. It can provide up to 2kW of 120V AC power in SPS mode in case of a grid failure. In such a case, it must be manually activated via an SPS switch.

Both "Neutral" and "Hot" wires in SPS are separate from the main "Hot" and "Neutral" wires in grid-tied circuits.

Does anybody know if the SPS "Neutral" is bonded with the main "Neutral" inside the inverter? Is it safe to make such bonding outside of it?

The idea is to use a common house "Neutral" for both SPS and grid-tied lines and a relay as an "Automatic transfer switch." It should automatically transfer power for the most critical 120V circuit in the house to the SPS in case of a grid failure.

It should work this way:

• If the grid is active, the solenoid in the relay is energized and holds "Normally closed" contacts in the "Open" status. Those contacts are connected to the inverter as an "SPS activation switch." Thus, when the "Grid is on," the SPS is NOT activated.
• If the grid fails, the relay is not energized, and those "Normally closed" contacts will close and activate an SPS.
• Another contact group in the relay will transfer the "Hot" wire in the critical house circuit (server, telework computer, network equipment, etc.) from the main house panel (with no power at this time) to the SPS output in the inverter.
• Computers and network equipment all have their own UPS blocks. So, short power hiccups of up to 5-10 minutes each are not critical as long as the SPS can provide power for them within that timeframe.

 

[SamDeleted]

Interesting question

@Hedges might know

 

[Hedges]

I don't know yet. I acquired a -41 but haven't tried SPS.

Sunny Boy's output is isolated single-phase not split phase, 240V and 208V L-L for these models. (Some other models have supported 277V L-N)
They have neutral connection, not to carry current or balance load, just to monitor that each voltage is within spec, and to auto-detect split-phase grid or two legs of 3-phase.

For SPS, the L1 and L2 outputs of inverter supply L and N of SPS outlet. Of course both poles would be isolated from grid L1 & L2.

Manual page 8:

"No galvanic isolation
...
• The neutral conductor of the AC output is not bonded to ground within the product.
• The neutral conductor of the AC output for secure power supply operation is bonded to
ground within the product."

That says N-G bond for SPS is in fact provided by Sunny Boy. So feeding it into your house wiring would result in two N-G bonds, and "objectionable" return current in ground.

You say "most critical 120V circuit" in house. If you can put both L and N through a transfer switch, that should take care of it.

It is tempting to use DPDT relay as transfer switch. My concern is how small a gap between contacts. It doesn't meet the spacing requirements for an "isolator" separating a circuit from the grid which I think is 1" between contacts. Safety switches and UL listed circuit breakers do, while "supplemental protectors" (look just like circuit breakers) do not. Perhaps isolator is only required for human safety when touching wires, and not for electrical transfer? I'm not sure. Larger gap would reduce the chances that a multi-kilovolt transient on power lines arcs across and causes damage. A true transfer switch would be preferred.

But, inside SMA battery inverters, it is just a relay that separates grid AC from island AC, so maybe not as big a worry as I'm thinking. I'd say at least disconnect grid AC with a circuit breaker before handling SPS output or critical circuit (especially since power to them are controlled by electronics.)

The idea is to use a common house "Neutral" for both SPS and grid-tied lines and a relay as an "Automatic transfer switch." It should automatically transfer power for the most critical 120V circuit in the house to the SPS in case of a grid failure.

It should work this way:

• If the grid fails, the relay is not energized, and those "Normally closed" contacts will close and activate an SPS.

I think someone tried a jumper rather than a switch to activate SPS, while neither grid nor battery inverter was connected. It seemed like after PV DC was removed and reapplied, SPS did NOT reactivate. Needed to have enable signal toggled.

 

[phx]

Thank you, Hedges. That completely answered my question. I looked into the "Manual" and found those statements you mentioned. I wish I saw them earlier.

• The neutral conductor of the AC output is not bonded to the ground within the product.
• The neutral conductor of the AC output for secure power supply operation is bonded to
ground within the product."

So, bonding the SPS Neutral to the main house Neutral would be a code violation and a fire hazard as such bonding has to occur only in one place at the utility service entrance.

However, transferring BOTH L and N wires via a relay for the critical circuit as you suggested, might be the best option. That circuit just needs to be rewired a little to isolate its N from the main panel and connect it there through the "Normally open" relay contacts. In such a case when the grid is on and the relay is energized, both N and L will be connected to the main panel (grid) via "Normally open" relay contact groups.

When the grid is off, those contact groups will switch both L and N to the SPS circuit.

 

[Hedges]

However, transferring BOTH L and N wires via a relay for the critical circuit

Transferring L1 & N may not be as bad, but I did try DPDT 30A relay to transfer L1 and L2, not between two sources but rather to connect a resistive heater with two elements in series vs. parallel. It either stuck and twisted, or arced, and shorted L1 to L2. That didn't trip the local 50A breaker but it tripped an upstream 70A breaker.

I would not rely on a double-throw relay to actually do "Break before Make", based on that experience.

Two single-throw relays with interlocks and dead time between them, or a double-throw knife switch, might provide better isolation. I've been contemplating a dead-man transfer switch to connect my loads (refrigerator, etc.) to grid if inverter fails. I envision something like spring or weight operation in one direction, air or vacuum in the other.

You should be able to find a transfer mechanism meeting your tolerance for reliability and protection against damage.