AC Coupling Design

[YellowBuddy]

A picture is worth 1000 words...

Would it work?

I should note that I already have the Sunny Islands, just trying to figure out if I can make it work with the existing system.

 

[Hedges]

I think in your "proposed", you have Solar Edge (at least)sometimes on output of SI-5048, also going through ATS to grid?
1) Does Solar Edge support frequency shift?
2) Grid should never connect to output of SI, only input. You could put a transfer switch on input of SI, selecting either generator or grid.
I presume in existing system where I see arrows cross, you do not connect grid to SI output? Not sure what Manual Disco does there, unless just disconnect from critical loads panel.
3) Sunny Island and/or Sunny Boy can be damaged if they are connected to an AC source, and then that source is snapped to another source out of phase. The inverter was happily shoving current into the grid, and suddenly generator is connected and happens to be 180 degrees out of phase. (or, inverter was supplementing generator, no backfeed allowed, and suddenly it sees the grid out of phase.) Something blows up unless you get lucky and fuses/breakers protect transistors.

Manual says transfer switch must be "off" for 5 seconds before switching. I'm using interlocked manual breaker for that reason.

From the other thread where you mentioned batteries draining, separate PV panel and critical loads panel, with load-shed relay, lets SI keep grid up for Sunny Boy until the sun returns. I've started thinking "sleep mode" could extend that, since with my four SI6048 I would drain 1200 Wh during a 12 hour night.

Another recovery option is DC coupling, which will charge battery back up. If you plan to run loads until SI shuts off you should consider a different low-battery cutout voltage to reduce wear on battery.

I'd like to be able to AC couple a generator through Sunny Boy (Windy Boy? Hydro Boy?) as an alternative to switching AC side. But I don't want to just switch rectified AC into the input capacitors, too much current.

 

[Hedges]

I assume you plan to have the four SI 5048 as 2S2P. Manual says to match wire lengths. I still got a 3:1 difference in current. Problem was QO270 breakers on output having different resistance. The 63A DIN mount Schneider breakers I used on the input were fine. Depending on how things work for you it might be necessary to use DIN breakers or fuses rather than QO (or whatever your brand/model) plug-on breakers.

 

[YellowBuddy]

I think in your "proposed", you have Solar Edge (at least)sometimes on output of SI-5048, also going through ATS to grid?

Correct. The thinking was that the Solar Edge when grid active will power the Priority Loads panel, but also continue to grid feed excess energy when the grid is up. When the grid is down the ATS functions as an automatic disconnect switch to prevent back feed to the grid and allows the SI to take over.

1) Does Solar Edge support frequency shift?

Yes in two ways as far as I can tell.

Basic is the search for max/min frequencies (59.3-60.5Hz). Outside of the range and the inverter will reboot and search for grid parameters. Not an optimal solution. So this is preferred

Then they have "Active Power Control" with a "Power Frequency" configuration.
"P(f) – Power Frequency: This is used when frequency-based power reduction is required. This defines a linear graph set by two points. The inverter de-rates power according to the defined graph, until the frequency reaches the trip value and the inverter disconnects (the trip point is preset per country therefore does not need to be defined as one of the two points)."

So in theory it should play together. But I haven't found someone who's mixed SE and SI together yet.

2) Grid should never connect to output of SI, only input. You could put a transfer switch on input of SI, selecting either generator or grid.

This is helpful. I was wondering about that, I played with a version that had the grid and generator on AC2 with an ATS as well. The one concern I had was sending the entire SolarEdge system through the Sunny Islands. The (4) SI-5048 will has a continuous load rating of ~165A on the output side. I wasn't sure if sending the SolarEdge system through the SI's and into the grid would end up overloading either the Sunny Islands or the panel bus. The SolarEdge is coming in at (peak) 150A/240V.

I presume in existing system where I see arrows cross, you do not connect grid to SI output? Not sure what Manual Disco does there, unless just disconnect from critical loads panel.

I see the confusion. I should have added the junction box to the diagram or at least offset the lines. The SE isn't going to the critical loads panel in the current. The SE is going to the grid via a service line tap. (Black up and down arrow to main panel). The manual disconnect is for the SolarEdge system only and acts as a manual shut down/disconnect for code.

The dotted line from the service line to the critical loads panel is a typically non-active line. There's a manual interlock in place that prevents the line from going into the SI. It's there as a backup; in case the SI fails, the panel can switch to receive power from the grid while I order a replacement.

Another recovery option is DC coupling, which will charge battery back up. If you plan to run loads until SI shuts off you should consider a different low-battery cutout voltage to reduce wear on battery.

This was certainly something I'm considering as well. The plan isn't actually to run it until it shuts off. My battery bank and recharge rate just isn't high enough right now to maintain a high SOC. So the goal here is to get the SolarEdge into my microgrid; and along with a backup generator, and larger bank, to prevent my batteries from draining down.

From the other thread where you mentioned batteries draining, separate PV panel and critical loads panel, with load-shed relay, lets SI keep grid up for Sunny Boy until the sun returns. I've started thinking "sleep mode" could extend that, since with my four SI6048 I would drain 1200 Wh during a 12 hour night.

As of right now, my batteries operate in float/standby. They can last for 6 hours without sun from Spring-Fall. In the winter, I have to deal with whether I want heat or not. Without heat, they'll last the same amount of time - but I might not. With heat, they last 30 minutes. So from a design perspective the goal is to get battery + generator to be able to operate for 12 hours in the winter with heat. The sleep mode would work for the summer to shed almost everything, but it would be tough to meet the heat requirements.

 

[Ampster]

1) Does Solar Edge support frequency shift?

Yes Solaredge is a 1741 compliant inverter. I believe the model number including the H means it is the latest HD Wave technology and can modulate, not just turn on or turn off. I think that means it is 1741 SA compliant.
I have a question about the design of the @YellowBuddy system that your knowledge of Sunny Islands might be able to answer.
The question I have is about the ratio of Sunny Islands and their batteries to the total 30 kW capacity of the Solaredge inverters. I know my 5kW Outback Skybox is only cabable of AC coupling up to a 7.6 kW GT inverter. I don't completely understand the limitation but believe it has to do with how fast frequency shifting can ramp down a GT inverter and it may relate to the battery headroom of the hybrid inverter. Do you know the answer about that ratio?

Thus may relate to your other thread on Frequency shifting?

 

[YellowBuddy]

I assume you plan to have the four SI 5048 as 2S2P. Manual says to match wire lengths. I still got a 3:1 difference in current. Problem was QO270 breakers on output having different resistance. The 63A DIN mount Schneider breakers I used on the input were fine. Depending on how things work for you it might be necessary to use DIN breakers or fuses rather than QO (or whatever your brand/model) plug-on breakers.

This is very helpful, it's a QO panel so I'll have to remember that.

The question I have is about the ratio of Sunny Islands and their batteries to the total 30 kW capacity of the Solaredge inverters. I know my 5kW Outback Skybox is only cabable of AC coupling up to a 7.6 kW GT inverter. I don't completely understand the limitation but believe it has to do with how fast frequency shifting can ramp down a GT inverter and it may relate to the battery headroom of the hybrid inverter. Do you know the answer about that ratio?

I might be able to solve that with the ATS and the Active Power settings. On the SE inverters there's a Wakeup Grad configuration:
Wakeup Grad – Wakeup Gradient: enables gradual power production when it begins operation after a fault or an inverter reset. For gradual power production during normal operation, use the Ramp Rate option.

So in the case the grid faults, this would kick in and allow the SI time to perform frequency shift so it doesn't get slammed with the full 30kW.

@Hedges 2nd attempt:



Here's some of my logic..
1. DC Generator, bypasses complexity of potentially backfeed into the generator, and dealing with an additional ATS.
2. SI-5048 now grid fed via a tap in the junction box, which ultimately goes to the meter. This avoids potentially overloading the main panel or the sunny islands and allows the Solar Edge to directly feed the line as it currently does.
3. In the event of an outage, ATS (2) kicks in and disconnects the SolarEdge from the grid and reconnects them to the Priority Loads panel. These will now feed into AC1 of the Sunny Island - I believe it's exactly how my Sunny Boys do on the Critical Loads Panel. On a sunny day that will in turn go back to the SI and be able to charge the batteries.
4. ATS (2) would use an open circuit switchover to prevent feeding the service line into AC1 of the Sunny Island. The disconnect would also trigger a grid fault in the Solar Edge system and allow the ramp up described above.

Thoughts? I appreciate all of the feedback btw!

 

[Ampster]

That sounds good. I think the DC generator is an elegant solution since you already have plenty of AC capability.

 

[Hedges]

Yes Solaredge is a 1741 compliant inverter. I believe the model number including the H means it is the latest HD Wave technology and can modulate, not just turn on or turn off. I think that means it is 1741 SA compliant.
I have a question about the design of the @YellowBuddy system that your knowledge of Sunny Islands might be able to answer.
The question I have is about the ratio of Sunny Islands and their batteries to the total 30 kW capacity of the Solaredge inverters. I know my 5kW Outback Skybox is only cabable of AC coupling up to a 7.6 kW GT inverter. I don't completely understand the limitation but believe it has to do with how fast frequency shifting can ramp down a GT inverter and it may relate to the battery headroom of the hybrid inverter. Do you know the answer about that ratio?

Thus may relate to your other thread on Frequency shifting?

Official instructions from Sunny Island manual is 6.7 kW of AC-coupled sources (e.g. Sunny Boy) per Sunny Island for grid-tied systems, because 56A programmed limit for relay x 120V = 6720. The relay is actually 3 pole 30A, but they're obviously being careful not to assume even distribution.

Off grid, the manual says AC coupled PV can be 2x the inverter's rating (so 12 kW), but wind or hydro only 1x (6 kW).

As for battery, minimum 100 Ah at 48V per Sunny Island, and 100 Ah per 1 kW of PV.
When you quick-start configuration of battery, a default max charge rate of 0.55C is set. If you do have 100 Ah per 1 kW, that limits it to 0.2C

As you indicate, a large battery may let it suck down excess power while it slowly ramps of frequency. But does it do that on a full battery? Maximum charge current is 140A, at 48V that's the same 6.7 kW, so doesn't explain handling 12 kW of PV.

That 30 kW of Solar Edge would be 12% above 4x6.7 kW, which could cause the inverters to disconnect. Even with the total within spec, any imbalance in current between parallel branches could do the same.

Off-grid, 24 kW should be OK on two Sunny Island, and 48 kW on four.

It was those two points that had me considering a transfer switch, back when I had 12.5 kW of Sunny Island and a single SI 5048US. I would keep 5 kW always on Sunny Island, and transfer the other 7.5 kW during off-grid operation. But, they are not to be snapped between two different AC sources, need a 5 second dead time.

As for batteries, I presently have 405 Ah, 4x SI, about 15 kW of PV, so my battery bank is 1/4 of recommendation. Testing a couple times cycling loads while off-grid it seemed to be OK as it charged batteries and loads were switched. But I was only trying about 1500W of load with 24 kW of SI. I haven't tried heavy loads.

 

[Hedges]

This is very helpful, it's a QO panel so I'll have to remember that.



I might be able to solve that with the ATS and the Active Power settings. On the SE inverters there's a Wakeup Grad configuration:
Wakeup Grad – Wakeup Gradient: enables gradual power production when it begins operation after a fault or an inverter reset. For gradual power production during normal operation, use the Ramp Rate option.

So in the case the grid faults, this would kick in and allow the SI time to perform frequency shift so it doesn't get slammed with the full 30kW.

@Hedges 2nd attempt:

View attachment 19509

Here's some of my logic..
1. DC Generator, bypasses complexity of potentially backfeed into the generator, and dealing with an additional ATS.
2. SI-5048 now grid fed via a tap in the junction box, which ultimately goes to the meter. This avoids potentially overloading the main panel or the sunny islands and allows the Solar Edge to directly feed the line as it currently does.
3. In the event of an outage, ATS (2) kicks in and disconnects the SolarEdge from the grid and reconnects them to the Priority Loads panel. These will now feed into AC1 of the Sunny Island - I believe it's exactly how my Sunny Boys do on the Critical Loads Panel. On a sunny day that will in turn go back to the SI and be able to charge the batteries.
4. ATS (2) would use an open circuit switchover to prevent feeding the service line into AC1 of the Sunny Island. The disconnect would also trigger a grid fault in the Solar Edge system and allow the ramp up described above.

Thoughts? I appreciate all of the feedback btw!

How fast does ATS2 switch? When does it switch? Probably SE rides through a brief switchover. If SE is on SI and grid returns, SI will wait 5 minutes before synchronizing and connecting. If ATS2 gets impatient it could yank SE off SI and slam it onto grid out of phase.

Just noticed you have SI 6048 and SI 4548. Probably the same pass-through current.
Your 3x10 kW SE and 2x5 kW SB is very close to pass-through capability of 6xSI, 40 kW to 40.x kW
If you rearranged as 2x SE + 1x SB on 4xSI, that's 25 kW vs. 26.8 kW pass-through. 1x SE + 1x SB on 2xS, 15 kW vs. 13.4 kW pass-through.
Are all your arrays oriented identically? If you have some at different angles that reduces peak. Maybe you could parallel two different strings on one GT inverter and program it for reduced limit.
You might be able to have all GT inverters behind SI, so no ATS just SI handing disconnect from grid.
Then one ATS to put main panel on 4xSI, with a load-shed relay. But if automatic you'd want to disable heavy loads like heating, or enable them only if battery 80% full and/or sun shining.

What current is the grid capable of delivering? I think you have at least 200A + 200A, possibly +100A if that can be switched in. If overload is possible you could add a class T fuse right after the meter. I was lucky to have a 200A breaker right after the meter, convenient as a disconnect when I work on other things like the separate main panel.

They make it difficult not letting us backfeed 200A from the far end of the panel. Your Square-D if 225A with 200A main allows 70A PV. If 150A main then 70A & 50A. Eventually I'll have a fused disconnect of 100A or more tapped after the meter, before my panel. I suppose this is why you have SI on multiple panels.

 

[YellowBuddy]

How fast does ATS2 switch? When does it switch? Probably SE rides through a brief switchover.

I haven't bought the ATS yet, so the switch specifics are TBD. But if memory serves me correctly, there's digital ones that have programmable time delays, so I could customize it. If I build in a 5 minute time delay, it should avoid the ride through right?

If SE is on SI and grid returns, SI will wait 5 minutes before synchronizing and connecting. If ATS2 gets impatient it could yank SE off SI and slam it onto grid out of phase.

Since the SE is on AC1 and the grid is on AC2, if SE gets pulled by ATS2; won't the SI still maintain the microgrid off the batteries? If so, wouldn't it still try to sync with the grid?

As I type that I guess you could mean slamming the SE out of phase back onto the grid. But if there's a time delay on ATS2, then it should force a reboot of the SE which would re-initialize according to the grid phase right?

Just noticed you have SI 6048 and SI 4548. Probably the same pass-through current.
Your 3x10 kW SE and 2x5 kW SB is very close to pass-through capability of 6xSI, 40 kW to 40.x kW
If you rearranged as 2x SE + 1x SB on 4xSI, that's 25 kW vs. 26.8 kW pass-through. 1x SE + 1x SB on 2xS, 15 kW vs. 13.4 kW pass-through.
Are all your arrays oriented identically? If you have some at different angles that reduces peak. Maybe you could parallel two different strings on one GT inverter and program it for reduced limit.

Unfortunately I don't think re-arranging is an option. That would take quite a bit of trenching, rewiring, and permit/inspection work as the systems were installed by different companies and considered two different "plants" by my utility company. Everything from their respective meters back would have to stay the same if I didn't want to get re-approval from the state and utility company. Anything after the meters is just the local construction office and much faster and easier.

What current is the grid capable of delivering? I think you have at least 200A + 200A, possibly +100A if that can be switched in. If overload is possible you could add a class T fuse right after the meter. I was lucky to have a 200A breaker right after the meter, convenient as a disconnect when I work on other things like the separate main panel.

They make it difficult not letting us backfeed 200A from the far end of the panel. Your Square-D if 225A with 200A main allows 70A PV. If 150A main then 70A & 50A. Eventually I'll have a fused disconnect of 100A or more tapped after the meter, before my panel. I suppose this is why you have SI on multiple panels.

From the meter, there's a line tap that uses 4/0 (200A) wiring to my main panel and a separate 4/0 (200A) wiring to the manual disconnect/SE. The grid line itself is capable of supporting a 400A panel (320A continuous) but I haven't looked at what wiring it uses.

You're exactly right. I considered just feeding the main panel since I don't need the extra circuits. But then I got concerned with overloading and surging the main panel if I fed the entirety of the SE into it. The SB/SI system isn't a lot and as you stated is using a 70A backfeed. I'm not sure the SE on 150A would pass inspection if I fed it to the bottom of the panel.

You might be able to have all GT inverters behind SI, so no ATS just SI handing disconnect from grid.
Then one ATS to put main panel on 4xSI, with a load-shed relay. But if automatic you'd want to disable heavy loads like heating, or enable them only if battery 80% full and/or sun shining.

Do you mean like this?


It could technically work. According to the SMA manual I can't mix the SI-6048, SI-4548, and SI-5048 in a double split phase configuration. So I would have to put the entire system on SI-5048; which 41.7A of output which is 166.8A continuous; and I'd be able to hold 216A for 30 minutes.

I'd have to pay the utility to do a disconnect to install the ATS, but I could sell the SI-6048 and SI-4058 to recoup the cost.

Just two potential concerns.
1) I'd still be feeding all of solar edge through the main panel in this case and still have overloading concerns. I'd probably need to upgrade the main panel to a 320A panel.
2) Doesn't this assume the SB and SE communicate with the same scale back frequencies and react the same way? I'm not sure if they do or don't so it might not be an actual concern.

 

[Hedges]

I haven't bought the ATS yet, so the switch specifics are TBD. But if memory serves me correctly, there's digital ones that have programmable time delays, so I could customize it. If I build in a 5 minute time delay, it should avoid the ride through right?



Since the SE is on AC1 and the grid is on AC2, if SE gets pulled by ATS2; won't the SI still maintain the microgrid off the batteries? If so, wouldn't it still try to sync with the grid?

As I type that I guess you could mean slamming the SE out of phase back onto the grid. But if there's a time delay on ATS2, then it should force a reboot of the SE which would re-initialize according to the grid phase right?



Unfortunately I don't think re-arranging is an option. That would take quite a bit of trenching, rewiring, and permit/inspection work as the systems were installed by different companies and considered two different "plants" by my utility company. Everything from their respective meters back would have to stay the same if I didn't want to get re-approval from the state and utility company. Anything after the meters is just the local construction office and much faster and easier.



From the meter, there's a line tap that uses 4/0 (200A) wiring to my main panel and a separate 4/0 (200A) wiring to the manual disconnect/SE. The grid line itself is capable of supporting a 400A panel (320A continuous) but I haven't looked at what wiring it uses.

You're exactly right. I considered just feeding the main panel since I don't need the extra circuits. But then I got concerned with overloading and surging the main panel if I fed the entirety of the SE into it. The SB/SI system isn't a lot and as you stated is using a 70A backfeed. I'm not sure the SE on 150A would pass inspection if I fed it to the bottom of the panel.



Do you mean like this?
View attachment 19519

It could technically work. According to the SMA manual I can't mix the SI-6048, SI-4548, and SI-5048 in a double split phase configuration. So I would have to put the entire system on SI-5048; which 41.7A of output which is 166.8A continuous; and I'd be able to hold 216A for 30 minutes.

I'd have to pay the utility to do a disconnect to install the ATS, but I could sell the SI-6048 and SI-4058 to recoup the cost.

Just two potential concerns.
1) I'd still be feeding all of solar edge through the main panel in this case and still have overloading concerns. I'd probably need to upgrade the main panel to a 320A panel.
2) Doesn't this assume the SB and SE communicate with the same scale back frequencies and react the same way? I'm not sure if they do or don't so it might not be an actual concern.

Being disconnected 5 seconds should be enough to have the GT inverter disconnect, after which it waits 5 minutes of monitoring before reconnecting. I think that is the UL1741 requirement. Disconnecting for 5 minutes would be more than plenty.

Yes, SI would maintain the microgrid. If grid returned SI would monitor for 5 minutes, raise frequency to 65 Hz to knock all GT inverters off microgrid, then synchronize to grid. The problem would be if ATS recognized grid before those 5 minutes were up and snapped SE onto grid. But as you say an off-time before putting SE on grid would fix that.

I think my utility feed is 2/0 which is good for 200A with 90 degree insulation. As someone pointed out elsewhere, terminals of a breaker are rated more like 75C which may mean wire ampacity needs to be based on that temperature instead. Also wires underground get wet so the insulation has a lower temperature rating in that application (but I would guess it is kept cooler by the environment.) 4/0 with 260A at 90 degree sounds like good margin.

"320A continuous" but 2x 200A breakers on it sounds like living dangerously. Maybe they're allowed to get away with that?

All three 10 kW SE on 4xSI would be too much AC to pass through to grid. Adding two 5 kW SB on top would be even worse grid-tie but still within what SI could manage off-grid. 4x SI will never backfeed more than 112A. Don't know if you would be allowed 125A backfeed breaker + 150A main = 275A on a 225A panel allowed 270A main + backfeed. If I go with 100A backfeed breaker + 150A main, then I should stay below 80A backfeed for 25% larger breaker to avoid nuisance trips. I would like 150A fused disconnect branching to two 2 pole 63A breakers for SI.

Maybe keep 2x SE on SI all the time, and use transfer switch to move third SE between grid and SI. Need to coordinate ATS settings so third SE is never on SI when SI is on grid. Or just leave 3rd SE on grid all the time and get by with 2x SE for off-grid backup situation.

2x SE would be 83A AC, ok for 4x SI and 100A backfeed breaker on 225A panel with 150A main. You could put 1x SE always on SI with 70A backfeed breaker and switch one or two SE between grid and SI with ATS. If only one switched it won't exceed SI limit and won't trip a 70A breaker in a short time, if some disagreement between SI and ATS on what is in-spec grid. Sustained on-grid time with two SE would trip. Some unbalanced current could be tolerated without one of the SI exceeding relay current. If a relay output from SI tells ATS when to switch that should work well.

The new grid-support GT inverters are OK to be on-grid, will drop off when necessary. A Sunny Boy in backup or off-grid mode is not, so it can't be switched between grid and SI. Since an RS-485 signal from SI tells SB to go from grid parameters to backup parameters, I thought switching that too would ensure SB was in grid-tie UL1741 mode when switched on-grid. But I've never been able to grab a scope image of that signal. If it is a keep-alive signal sent every few seconds that would work, but if a one-time switchover signal it would not.

Relay function: "GdOn Utility grid Relay switching when utility grid is available and connected."

That might have a firmware delay but should be good enough for switching something that is a moderate overload for breaker but not for relays.

You might as well keep it two systems with two main panels, since it is > 200A and needs 6x SI to handle all the PV.

You can also consider DC coupled PV if that provides any benefit. Usually that would only get used in an off-grid situation when battery voltage drops, but since your 2x SI can put a bit more backfeed through its relays than 2x SB, a charge controller set to higher voltage and a setting in SI would make it a grid-interactive inverter. That may be "PvFeedTmStr", not sure.

 

[YellowBuddy]

"320A continuous" but 2x 200A breakers on it sounds like living dangerously. Maybe they're allowed to get away with that?

From what I understand when the electrician was out, it's how it's designed. A 200amp panel is actually rated for 160A continuous and when people ask for 400A service, it's 2x200A panels but actually 320A continuous. I didn't do any additional research on it though and just took him at his word.

Maybe keep 2x SE on SI all the time, and use transfer switch to move third SE between grid and SI. Need to coordinate ATS settings so third SE is never on SI when SI is on grid. Or just leave 3rd SE on grid all the time and get by with 2x SE for off-grid backup situation.

All of the SE's combine prior to the meter. So I'm not sure I'd be able to do this without going through the interconnect application process again. But what may be possible is the active power settings. I haven't explored it in depth but I may be able to program SEs so the 3rd one just shuts off and doesn't ramp at all at a certain frequency. But that would only be helpful in proposal 2, I don't think it helps in proposal 3 as it would have to be permanently derated.

You can also consider DC coupled PV if that provides any benefit. Usually that would only get used in an off-grid situation when battery voltage drops, but since your 2x SI can put a bit more backfeed through its relays than 2x SB, a charge controller set to higher voltage and a setting in SI would make it a grid-interactive inverter. That may be "PvFeedTmStr", not sure.

The SB would be pretty easy to convert to DC PV based on the wiring and location. The SE would not be. Is it worth the complexity if that whole system is working well right now?

All three 10 kW SE on 4xSI would be too much AC to pass through to grid. Adding two 5 kW SB on top would be even worse grid-tie but still within what SI could manage off-grid. 4x SI will never backfeed more than 112A. Don't know if you would be allowed 125A backfeed breaker + 150A main = 275A on a 225A panel allowed 270A main + backfeed.

Does it work like that? Because on my current main panel it's a 200A main breaker and a 70A backfeed breaker from the SI/SB - professional installer not my work. It doesn't add up to 270A, otherwise wouldn't the main breaker would trip? Or am I misunderstanding the statement?

You might as well keep it two systems with two main panels, since it is > 200A and needs 6x SI to handle all the PV.

Ok so back to proposal 2. I'm trying to interpret if all of what you just stated applies to this proposal as well. I don't think it will because it'll utilize the existing backfeed when on grid. When off grid, the SI would be able to adjust the output, and there would be no backfeed. When the grid is up, it would only supply priority panel 2. Is that a correct understanding?

 

[Hedges]

From what I understand when the electrician was out, it's how it's designed. A 200amp panel is actually rated for 160A continuous and when people ask for 400A service, it's 2x200A panels but actually 320A continuous. I didn't do any additional research on it though and just took him at his word.



All of the SE's combine prior to the meter. So I'm not sure I'd be able to do this without going through the interconnect application process again. But what may be possible is the active power settings. I haven't explored it in depth but I may be able to program SEs so the 3rd one just shuts off and doesn't ramp at all at a certain frequency. But that would only be helpful in proposal 2, I don't think it helps in proposal 3 as it would have to be permanently derated.




The SB would be pretty easy to convert to DC PV based on the wiring and location. The SE would not be. Is it worth the complexity if that whole system is working well right now?



Does it work like that? Because on my current main panel it's a 200A main breaker and a 70A backfeed breaker from the SI/SB - professional installer not my work. It doesn't add up to 270A, otherwise wouldn't the main breaker would trip? Or am I misunderstanding the statement?



Ok so back to proposal 2. I'm trying to interpret if all of what you just stated applies to this proposal as well. I don't think it will because it'll utilize the existing backfeed when on grid. When off grid, the SI would be able to adjust the output, and there would be no backfeed. When the grid is up, it would only supply priority panel 2. Is that a correct understanding?

Each 200A breaker should only have 160A continuous load so it doesn't nuisance trip. A few minutes at 200A is OK.
If a 200A continuous load was applied it might trip, or it might not.
If two 200A breakers each fed a 200A continuous load and didn't trip, they would draw 400A continuous from the wires with 320A ampacity and overheat them. For that reason, breaker rating is supposed to be no more than wire ampacity, possibly 320A in this case.

Utility feed may be allowed to exceed that. Similar concern for wiring with skinnier neutral that assumes some imbalance, but if a customer had 10x 20A circuits fully loaded and not tripping, drawing 200A though a hot wire sized for 200A and returning all 200A through an undersize neutral, that would be a problem.

With 200A main and 70A backfeed, the main will at most carry 200A from grid or 70A (coming from PV) and push it back into the grid. The busbar is only rated 225A. If 200A and 70A are at opposite ends, the most busbar carries is 200A (feeding all loads, zero coming from PV.) If main carries 200A and backfeed carries 70A, each point on busbar carries 200A or less. Loads total 270A, but current is flowing in opposite directions and subtracts doesn't add.

If 200A main and 70A backfeed are both at one end of busbar, then busbar carries 270A to loads. In excess of its rating but probably doesn't burn up. The code lets us have main + backfeed exceed busbar by 20% to allow PV, knowing that even if someone rearranging breaker locations later gets it wrong nothing will actually burn up.

If we put 200A main on one end of 225A busbar, and a 125A backfeed plus a 100A backfeed on the other end for 225A, that would still work fine. Busbar would never carry > 225A. But code doesn't allow it. There seems to be an exception for a panel which is just PV aggregator without any loads, though.

Your proposal 2, with 3x SE for 30 kW on-grid is OK, what you have today. If switched by (slow) ATS to SI during grid failure, that should work. You may be able to have SI command the switching. SI will knock SE off-line before synchronizing and connecting, so a delayed signal to put SE back on grid is OK. The key seems to be 5 seconds or longer dead time in ATS so SE is shut down before connecting to an out-of-phase source, and SI telling ATS to switch when SI connects/disconnects from grid so you never have 30 kW going through SI's internal relays. You also will never have to worry about imbalance of PV backfeed current through paralleled SI. Only imbalance of loads fed from grid. Connect a couple space heaters and go read the SI master and three slaves. Each slave reports its own power feed-through, and master reports sum of all four.

 

[YellowBuddy]

Your proposal 2, with 3x SE for 30 kW on-grid is OK, what you have today. If switched by (slow) ATS to SI during grid failure, that should work. You may be able to have SI command the switching. SI will knock SE off-line before synchronizing and connecting, so a delayed signal to put SE back on grid is OK. The key seems to be 5 seconds or longer dead time in ATS so SE is shut down before connecting to an out-of-phase source, and SI telling ATS to switch when SI connects/disconnects from grid so you never have 30 kW going through SI's internal relays. You also will never have to worry about imbalance of PV backfeed current through paralleled SI. Only imbalance of loads fed from grid. Connect a couple space heaters and go read the SI master and three slaves. Each slave reports its own power feed-through, and master reports sum of all four.

That's a great suggestion to have SI control it. I'll have to see if there's compatible ones with a two wire external control or if I'd have to make one. I know the SI only requires 5 seconds, but I'd personally like it to be longer. We do get brown outs and occasional interrupts from our grid. So this would also prevent rapid switching to just go a little longer. The reason I would also want to go 5 minutes (maybe 2 or 3 will do) is because we are 100% electric and want to be careful not to fast cycle all the motors (well pump, heat pumps) as that would cause rapid wear.

 

[YellowBuddy]

I just found some new/unused surplus TMobile DC generators for sale. They're 7.5kw, 150A air cooled generators. Built in surge protection and 200A breakers. Runs higher and louder than I'd like (3600rpm / 65db @ 23ft) but since they're 48V direct, I shouldn't need an additional charger. They accept a 2 wire remote start, so I could have the SI kick them on and off still. They seem perfect, I'll have to start a post in the main forum to make sure my understanding of DC generators is correct. I'm more familiar with AC generators.

 

[abilityonline]

@YellowBuddy I've ac coupled my (2) solaredge 7600 HD wave GT inverters to my (2) sunnyIsland hybrid inverters. They work perfectly. I have California 21 enabled on both solaredges to not over charge the batteries. But once 90% battery charge is achieved, the solaredges cycle offline as the frequency shifts up. It's not ideal but at least they aren't over charging the battery. Did you find any specific settings to avoid the recycling on the solaredges?

 

[Ampster]

But once 90% battery charge is achieved, the solaredges cycle offline as the frequency shifts up. It's not ideal but at least they aren't over charging the battery. Did you find any specific settings to avoid the recycling on the solaredges?

I know this was directed at YellowBuddy but it is a generic question that I will try to answer. Unless you have somewhere else for the energy to go, it has to be turned off. Try putting a load on the inverter like heating water or something that you do often but could time it when batteries are full.

 

[abilityonline]

I know this was directed at YellowBuddy but it is a generic question that I will try to answer. Unless you have somewhere else for the energy to go, it has to be turned off. Try putting a load on the inverter like heating water or something that you do often but could time it when batteries are full.

@Ampster thank you for taking the time to answer. Yes, that's the easy solution. I have an EV that I plan on charging to load shed above 90%. Nevertheless, I believe there are settings in the Mode Q(U)+Q(P) that should throttle down the GT inverter production without it going offline. I was curious if someone had perfected those settings for a SolarEdge / Sunny Island AC couple system.

 

[Ampster]

As far as I know CA Rule 21and UL1742SA are the same and a new SolarEdge should be compliant. The issue might be on the SI. This is DejaVu. have we already had this conversation?

 

[abilityonline]

We have. There are settings inside California 21 that are adjustable within the Mode Q(U)+Q(P) settings