Adding off grid to grid tie system.

[LGsolar]

My biggest gripe with my grid tie system has been losing my solar in a power outage. Last Summer looking through my spare parts I realized there are enough cables and Y connectors to branch off 1/2 my panels upstream of the micro inverters. Running that branch(2800W) or 2 in my case, 10 panels per branch to a pair of 5048 inverters. I would take the output of those 5048s and run them into the input of an manual generator transfer switch. The 5048s would be off until a power outage then I would connect the DC to them through a switch and power up the units then flip the switches on the circuits of the transfer box. Supplying the 5048s inputs with generator and battery bank. In my mind this would work if a car takes out a pole on a sunny day knocking out power to the house or some other weather related outage. Completely aware that turnkey systems are out there that do all this but I already have a working 12KW micro inverter system and didn’t want to start from scratch. There may be something fundamentally wrong with my logic but that is why I’m here in the Danger zone section.

Jim in PA

 

[kernel]

My biggest gripe with my grid tie system has been losing my solar in a power outage. Last Summer looking through my spare parts I realized there are enough cables and Y connectors to branch off 1/2 my panels upstream of the micro inverters. Running that branch(2800W) or 2 in my case, 10 panels per branch to a pair of 5048 inverters. I would take the output of those 5048s and run them into the input of an manual generator transfer switch. The 5048s would be off until a power outage then I would connect the DC to them through a switch and power up the units then flip the switches on the circuits of the transfer box. Supplying the 5048s inputs with generator and battery bank. In my mind this would work if a car takes out a pole on a sunny day knocking out power to the house or some other weather related outage. Completely aware that turnkey systems are out there that do all this but I already have a working 12KW micro inverter system and didn’t want to start from scratch. There may be something fundamentally wrong with my logic but that is why I’m here in the Danger zone section.

Jim in PA

Absolutely this can work. Lots of spaghetti bit it can be done.

 

[Hedges]

"... enough cables and Y connectors to branch off 1/2 my panels upstream of the micro inverters... I would connect the DC to them through a switch..."

You should arrange a "break before make" transfer switch for the DC, so it is never connected to both micro inverters and battery inverters at the same time. One may rely on the the PV wires being floating, and the other may provide an ohmic path from PV wires to earth. Both at the same time could be a problem.

If you have a switch to open-circuit the PV to inverter (or charger) connection, you could then manually transfer MC4 plug connections.

 

[RCinFLA]

It will cost you some money but there is high voltage input charge controllers to charge a 48v battery bank.

There are several inverter types ending in 5048. Did you check if yours can be output AC coupled? There needs to be some control to roll back Grid Tie output if loading is too small to consume solar production. Usually this requires same manufacturers for battery inverter and grid tie inverter, like SMA Sunny Island inverter. Sunny Island inverter varies its output AC frequency slightly which is detected by SMA Grid Tie inverter telling it to vary its output power push when there is not enough consumption by battery charging and/or AC loads. Xantrex does similar thing. I haven't checked lately but it would be nice if there was standardized control method but that would probably be too much to expect as they all want to promote just their own equipment line.

 

[RCinFLA]

As long as battery based grid tie inverter has AC coupling capability and PV grid tie system will not produce more then about half to three quarters of battery inverter power capability there is likely good chance of working fine.

In this situation the PV GT inverter will just shutdown when AC coupled battery grid tied inverter varies the output AC freq. Again, this would only happen when battery based inverter cannot consume via battery charging what might be being back fed into it beyond your AC load consumption.

Any battery based grid tied battery based inverter with AC coupling should have an accurate/stable enough AC frequency output to satisfy PV GT inverter connect criteria. PV GT inverter will periodically try to reconnect but will quickly shutdown again if its power output is not totally being consumed.

This assumes you do not have a cheapo GT inverter that does not pass 1741 specs for grid disconnect timing and +/- 0.1 Hz accuracy of grid. Normal grid active operation will back flow through battery grid-tie inverter's transfer switch.

Other advantage is you can put a small generator into battery based grid-tied inverter Gen input and excess power needs beyond generator's capability can be supplied via batteries during night time or no sun periods. You setup max Gen AC current draw on inverter AC input and it will automatically manage what is given to charging batteries based on AC loads getting first priority on generator's available power. By setting limit to 50-80% of generator's capability you also get maximum fuel efficiency from generator. Inverter Gen AC input will not allow PV power backfeed push into generator.

 

[Hedges]

A micro inverter is transformerless, and probably has an electrical path between AC lines and PV. Connecting them also to an SCC could be a problem, biases them to some voltage.

Rather than a "Y" connector branching to your off-grid system, I would suggest using MC cables to extend the PV panel and micro inverter to an accessible location. I would feed the PV circuit through a DC rated disconnect, and inside the disconnect I would have wires for micro inverter and SCC.

Turn the switch off, which unlocks the cover so you can open it. Swap MC cables from micro inverter to SCC. Close cover and turn switch on.
This interlock prevents you from disconnecting the MC connectors under load.

Something like this - has 3 poles, so could turn off 3 panels:

SQUARE D DISCONNECT TYPE 3R HU361RB 30AMP 600VOLT 3POLE NON-FUSIBLE | eBay

Find many great new & used options and get the best deals for SQUARE D DISCONNECT TYPE 3R HU361RB 30AMP 600VOLT 3POLE NON-FUSIBLE at the best online prices at eBay! Free shipping for many products!

www.ebay.com

 

[fafrd]

I figure this is as good of a thread as any to ask my question.

I’m also considering automatically ‘switching’ some panels between a quad-input Microinverter and an SCC.

For the Microinverter, the multiple single-panel strings need to be isolated, while runnng into the SCC, they need to be combined.

So I’m thinking about a solid-state relay to activate the connection to the SCC: https://www.amazon.com/gp/aw/d/B079BGGVYX/ref=ox_sc_saved_image_4?smid=A1LH3TFU4S09BS&psc=1

40A is overkill for the peak current of under 14A flowing out of my panels, but a good amount of headroom (and a heatsink) seems like a prudent investment.

I’ll use a normally-closed relay controlled by the grid signal to assure those SSRs are open / high-impedance while the quad-Microinverter is producing so that when the grid goes down, the deactivated Microinverter goes high-impedance and the SSRs connect all of the strings in parallel to the SCC to use available PV energy to charge my 14kWh backup battery…

Is there a problem with this plan that I am missing?

Obviously, to make this work, the quad Microinverter will be located in the basement near the SCC rather than under the panels (and I will run a seperate 1S1P string from roof to basement for each of my DC-coupled ‘backup-capable’ panels…).

 

[Hedges]

I’m also considering automatically ‘switching’ some panels between a quad-input Microinverter and an SCC.

For the Microinverter, the multiple single-panel strings need to be isolated, while runnng into the SCC, they need to be combined.

Do you know if the quad-input microinverter is four independent inverters? Or a single inverter, either with four PV inputs connected in parallel, or in series with RSD built in?

I'm going to guess they are four independent MPPT (allowing multiple angles) onto a single DC rail. Would it work or be unhappy if the four were in parallel? (Not sure if RSD these days includes wattage limit, or just voltage.)

Maybe Y cable and combine through diodes (both positive and negative leads) into your SCC. When SCC not active, microinverter gets to pull each down.
One remaining problem: SCC probably biases negative side to ground.

So ... transformer isolate all the microinverters you're messing with. With no reference to ground on the AC side, you could arbitrarily bias the DC side.

For the price of 2 diode drops as loss (when using SCC), and transformer loss (using microinverters), this might work. If you inadvertently activate both, I don't think there is harm but could confuse MPPT.

 

[fafrd]

Do you know if the quad-input microinverter is four independent inverters? Or a single inverter, either with four PV inputs connected in parallel, or in series with RSD built in?

I assume it is four independent input stages (MPPTs) optimizing input power to a single 240V 60Hz inverter.

I'm going to guess they are four independent MPPT (allowing multiple angles) onto a single DC rail. Would it work or be unhappy if the four were in parallel? (Not sure if RSD these days includes wattage limit, or just voltage.)

I’ve read mixed things about having 2 parallel MPPTs active on a single string. I want the strings separate during the shady part of my morning to optimize output on a single-panel level (when grid is active) but in rarely-used backup mode (no grid) don’t care about losing some efficiency from parallel connection…

Maybe Y cable and combine through diodes (both positive and negative leads) into your SCC. When SCC not active, microinverter gets to pull each down.
One remaining problem: SCC probably biases negative side to ground.

Interesting. So you’re saying a diode on the path to a common SCC bus will give preference to the Microinverter as long as it’s MPPT is active.

My SCC does bias negative to ground but do Microinverters typically require floating ground?

When the SCC is starved, it will drive string voltage down to Vbat + 2V occassionally, so I don’t see any way yo avoid a problem without cutting off the current flow…

So ... transformer isolate all the microinverters you're messing with. With no reference to ground on the AC side, you could arbitrarily bias the DC side.

Not quite understanding but would like to.

Using a common-ground between Microinverter strings and SCC, I can use an SCR on each string hot leg to connect to the common-string + into the SCC when grid is down - do you see anything easier / cheaper?

For the price of 2 diode drops as loss (when using SCC), and transformer loss (using microinverters), this might work. If you inadvertently activate both, I don't think there is harm but could confuse MPPT.

Still not quite understanding. I understand the use is a single diode so that string voltage to a common plus bus into SCC is one diode drop higher than the string voltage straight into the active Microinverter (so Microinverter will grab most string power), but when the SCC occasionally drops string voltage close to 0V, it will still streal all current from the Microinverter…

The SCR will drop 1V / ~2.5% which is not the end of the world and will guarantee isolation of the string(s) from the SCC until the Microinverter.

If there is an alternative that is cheaper and only loses 0.7V / 1.75% of potential power, I’m all for it, but I need to assure that the starved SCC cannot steal power / current from the Microinverter…

 

[Hedges]

If SCC is off, microinverter(s) each get to pull down each PV panel as desired.
If microinverters are off, SCC gets to pull down all PV panels together, through the diodes.

If both inadvertently on, might act funny.

Probably, the transformerless design of microinverters applies some bias voltage to the PV panel. It will be a boost converter to achieve the higher voltage needed, but not isolated. Likely negative PV terminal follows the negative most phase (at any time, switching with H-bridge.) Therefore, an SCC that pulls negative PV terminal to ground would be a problem.

An isolation transformer on AC side should fix that. Negative most side of AC lines can be held no lower than ground without problem. It will only present a capacitive load on the line (primary to secondary or to core coupling capacitance.)

I'd at least start out with diodes on both positive and negative leads. That way microinverters don't talk to each other on the negative side either.

 

[fafrd]

If SCC is off, microinverter(s) each get to pull down each PV panel as desired.

The only way to turn the SCC off is to disconnect it from the battery…

If microinverters are off, SCC gets to pull down all PV panels together, through the diodes.

Yes, once the Microinverter is off, the SCC can clearly control all strings through diodes.

My issue is that while the Microinverter is still on, the SCC can still ‘steal’ some/all PV power by dropping resistance/voltage low enough…

If both inadvertently on, might act funny.

Exactly.

Probably, the transformerless design of microinverters applies some bias voltage to the PV panel. It will be a boost converter to achieve the higher voltage needed, but not isolated. Likely negative PV terminal follows the negative most phase (at any time, switching with H-bridge.) Therefore, an SCC that pulls negative PV terminal to ground would be a problem.

An isolation transformer on AC side should fix that. Negative most side of AC lines can be held no lower than ground without problem. It will only present a capacitive load on the line (primary to secondary or to core coupling capacitance.)

I'd at least start out with diodes on both positive and negative leads. That way microinverters don't talk to each other on the negative side either.

Using SSRs just seems easier / more direct. There is a (tiny) bit more power loss but there is no question about whether a panel/string is under the control of the Microinverter MPPT or the SCC MPPT…

 

[Hedges]

A combiner box with breaker pole per PV panel takes care of turning off SCC. Or single breaker, and fuse per panel.
A breaker on AC of microinverters turns them off.
Or are you trying to automate power transfer during grid failure?

What I like about my idea is passive stuff, and tolerant of forgetting to turn one off.

Have a schematic of how you'd wire everything together with SSR?
I have this suspicion that PV negative may be at -170V with microinverter. Don't know if design does boost of both poles, or just one.
I suggest you plug in a Y cable and measure.

 

[MurphyGuy]

I would just do it the right way and install an battery inverter capable of AC Coupling...

Turning your system into a Rube Goldberg Frankenstein monster is probably going to create as many problems as it solves.

 

[fafrd]

A combiner box with breaker pole per PV panel takes care of turning off SCC. Or single breaker, and fuse per panel.
A breaker on AC of microinverters turns them off.
Or are you trying to automate power transfer during grid failure?

Yes, my goal is to make the switchover from grid-tied Microinverter output to DC-coupled SCC output automatic…

What I like about my idea is passive stuff, and tolerant of forgetting to turn one off.

I like that too, but only if it works…

SCRs/relays driven by active grid signal or AC current exceeding a threshold is good enough for me…

Have a schematic of how you'd wire everything together with SSR?

No, but it’s pretty simple: three strings parallel-connected to both an SCC and a quad-input Microinverter. I’m looking for a way to have only one connection or the other active…

I have this suspicion that PV negative may be at -170V with microinverter. Don't know if design does boost of both poles, or just one.
I suggest you plug in a Y cable and measure.

Pretty certain the Microinverter is common-negative (but I’ll check).

 

[fafrd]

I would just do it the right way and install an battery inverter capable of AC Coupling...

Turning your system into a Rube Goldberg Frankenstein monster is probably going to create as many problems as it solves.

That’s certainly a concern and I appreciate the input.

I’ve already invested in the DC-coupled battery-based off-grid microgrid solution. My goal is to see whether I can find a straightforward, safe way to repurpose that backup PV array for some additional grid-tied production for the 99.99% of the time the grid is up…

 

[Hedges]

Pretty certain the Microinverter is common-negative (but I’ll check).

It could be common-negative, but if a non-isolated topology boot converter, then PV negative would follow a negative-going half-wave rectified sine wave, peaking at -170V.

That would be the simplest, cheapest microinverter design I could come up with. Boost converter from PV to 340V, H-bridge to L1, L2 of 240V split-phase. (Same topology for 120V single phase, and same issue.)

If it is an isolated topology as shown in the following link, then this problem doesn't exist.

Fig. 11. Enphase micro-inverter fly-back topology

Download scientific diagram | Enphase micro-inverter fly-back topology from publication: Grid-Connected Solar Electronics | | ResearchGate, the professional network for scientists.

www.researchgate.net

I suggest connecting an MC "Y" cable to each of negative, positive PV leads. Measure AC voltage relative to earth. Connect a high impedance (e.g. low wattage lightbulb like 120V Christmas bulb) between lead and earth, measure again. Knowing whether PV is floating (can be biased to ground in SCC) or driven to line voltage (or some voltage) will tell you if you need all-lead isolation. If isolation required, transformer would do that.

 

[fafrd]

It could be common-negative, but if a non-isolated topology boot converter, then PV negative would follow a negative-going half-wave rectified sine wave, peaking at -170V.

That would be the simplest, cheapest microinverter design I could come up with. Boost converter from PV to 340V, H-bridge to L1, L2 of 240V split-phase. (Same topology for 120V single phase, and same issue.)

If it is an isolated topology as shown in the following link, then this problem doesn't exist.

Fig. 11. Enphase micro-inverter fly-back topology

Download scientific diagram | Enphase micro-inverter fly-back topology from publication: Grid-Connected Solar Electronics | | ResearchGate, the professional network for scientists.

www.researchgate.net

I actually haven’t chosen a Microinverter yet, but if PV negative can move around as much as you suggest, seems like yet another reason to go with a double-pole disconnect from the SCC (but you are correct that I need to assure any voltage swings on either PV- or PV+ are within the limits of whatever SSR/Relay I use…).

I suggest connecting an MC "Y" cable to each of negative, positive PV leads. Measure AC voltage relative to earth. Connect a high impedance (e.g. low wattage lightbulb like 120V Christmas bulb) between lead and earth, measure again. Knowing whether PV is floating (can be biased to ground in SCC) or driven to line voltage (or some voltage) will tell you if you need all-lead isolation. If isolation required, transformer would do that.

Good idea to measure some voltages before setting anything up, but it’ll need to wait until I’ve selected an inverter.

Is there a way PV floating versus or biased to ground can be determined from the datasheet / specs?

 

[Hedges]

Is there a way PV floating versus or biased to ground can be determined from the datasheet / specs?

For string inverters, yes. SMA indicates some older inverters are negative grounded, some optionally positive grounded (through a 1A GFCI fuse).

Others are "transformerless", require PV array not exceed some capacitance to ground (at least for 3-phase 208V connection), and documentation indicates voltage bias swings during testing when it powers up. In operation I think they bias balanced +/- on split-phase, and carry an AC signal in 3-phase. That's my 10000TLUS, I assume new models are similar.

Microinverters, they aren't likely to say much, just that PV is not to be grounded.

 

[fafrd]

For string inverters, yes. SMA indicates some older inverters are negative grounded, some optionally positive grounded (through a 1A GFCI fuse).

Others are "transformerless", require PV array not exceed some capacitance to ground (at least for 3-phase 208V connection), and documentation indicates voltage bias swings during testing when it powers up. In operation I think they bias balanced +/- on split-phase, and carry an AC signal in 3-phase. That's my 10000TLUS, I assume new models are similar.

Microinverters, they aren't likely to say much, just that PV is not to be grounded.

Here’s the manual for my existing BDM300x2 Microinverters: https://northernep.com/downloads/manuals/BDM-600-LV-micro-inverter-manual.pdf

Footnote 1 on page 6 states: ‘DC circuits of BDM-300X2 are isolated and insulated from ground. An integrated ground protection circuit is included in the micro inverter.’

I don’t anything in the specifications to indicate whether the PV ‘-‘ can be biased to and/or swing to a very different voltage than 0V (~ground) during operation or not…

If I’m electrically disconnecting both ‘+’ and ‘-‘ PV inputs from the SCC while Microinverter is active, seems like I just need to assure that any voltage swings on either PV input does not go outside of the 0-60V range specified for the SSR.

The Tigo rapid-disconnect switches are essentially doing the same thing, so seems like it shouldn’t be too difficult to solve this problem for the average / typical inverter..,

 

[Hedges]

So they say, but I'm not entirely convinced. But it could be true.
It could be isolated from ground, but if not isolated from AC, the AC circuit could provide galvanic connection to ground.

If you isolated both '+' and '-' from SCC, it would be fine regardless of what voltages they might get driven to (so long as not exceeding standoff voltage of disconnect.)
But your SSR, correct, need to see what it does. An isolated PV panel would get shoved around by capacitive coupling to AC.
When connected to SCC, you're depending on microinverter having PV wires isolated from AC side. The transformerless topology I imagined could have a transistor connecting them.

I think a test with Y cables for voltage (likely some DC, possibly AC voltage present) and then biasing with a high resistance and seeing if you can pull to zero volts will answer the question.

Not sure RSD actually isolates. It seems at least some have a buck circuit that drops voltage. For instance, Solar Edge optimizers deliver 1V per panel while in shutdown. SMA now has a way to enable their Secure Power feature with grid down, which suggests they get power from PV while in shutdown.