Turbo-AC-coupling - will this work???

[fafrd]

I’m trying to plan for an off-grid system that offers the freedom to move my 3.5kW grid-tied system off-grid (if ever needed).

I’m looking at the Conext SW 4048 which supports AC-coupled solar power connected to it’s output port but is limited to 45A or ~2.25kW of charge power.

My 3.5kW Microinverter-based grid-tied array is overpowered by 4kW of solar panels and will often max out at 3.5kW of output during peak hours of the day.

So the SW 4048 cannot absorb full peak ac-coupled solar power by charging a battery.

The battery charger is 92% efficient and if I assume a lowest-possible battery voltage of 50VDC, is can absorb at most 2446W of AC power.

Hence the need for some ‘turbo’ solution to bump max battery charging power from 2400W up over 3500W…

I have a 240VAC to 50VDC rectifier that can be programmed to consume 2kW (like a Chargenectifier) so my ‘turbo’ idea is to use a current switch to turn on the rectifier to divert 2kW into charging the battery whenever AC-coupled current from the Microinverter string exceeds ~8A.

So the idea in phases would be as follows:

-Overnight: SW 4048 depleting battery by offsetting loads of ~300W average.

-Early morning - AC-coupled solar power < 300W: loads partially offset by AC-coupled solar power but battery continuing to be depleted.

-Mid-morning - 300W < AC-coupled solar power < 2000W (< 8A): SW 4048 first offsets loads and then consumes excess AC-coupled solar power by charging battery (with < 2000W of input power or < 37A of DC charge current).

-Mid-day / Peak-Production - >8A of AC-coupled solar power turns on current switch which turns on rectifier / charger. 2000W of AC-coupled power consumed by rectifier to charge battery @ ~38ADC while remaining AC-coupled power of < 1500W consumed by SW 4048 by first offsetting loads and then charging battery with excess of < 1500W with < 30ADC.

-Afternoon - 300W < AC-coupled Solar Power < 2000W (<8A): current switch switches off turning off rectifier so full AC-coupled solar power of < 2000W goes into SW 4048 to offset load and charge battery with < 2000W of input power or < 37A of DC charge current.

-Evening - AC-coupled Solar Power < 300W: loads partially offset by AC-coupled solar power but battery begins to be discharged to offset remaining load.

-Overnight - battery discharged to fully offset load.

In terms of clouds or other occurrences that temporarily reduce AC-coupled solar output, all of these ‘phases’ can transition up or down and can also cycle through multiple cycles without issue, so from what I can see, so I don’t see any issue there.

The idea would be to have a large-enough battery to power easily absorb a full day’s production but the system needs to also work when the battery starts out near full.

Once the SW 4048 enters float, the rectifier will be programmed to also be in CV and so the AC-coupled solar current will drop under 8A so all AC-coupled solar power of under 2000W will again enter the SW 4048 where it can properly throttle solar output to offset load and CV charge power.

So I’m not seeing any reasons this idea would not work to add a ‘turbo’ charging capability to the SW 4048’s modest AC-charge ing power and would appreciate any insight into what I may be missing…

 

[zanydroid]

Feels like this will work fine for on grid but in an off grid situation you would likely be back to being limited by the built in charger. And you would need to program the external charger to turn off when off grid, or have turning it off in your off grid startup instructions.

 

[fafrd]

Feels like this will work fine for on grid but in an off grid situation you would likely be back to being limited by the built in charger.

Apologies for not being clear - this is true off-grid (there is no grid connection to the AC input.

And you would need to program the external charger to turn off when off grid, or have turning it off in your off grid startup instructions.

Appreciste your rereading through the sequence I laid out in the lead post and let me know where you see a step that won’t work off-grid.

 

[zanydroid]

No problem. It’s also not clear to me that a randomly selected AC charger will present the right power factor to those grid tie inverters. To stay within the typical inverter specs the combined load would need to be corrected to 0.8 or better.

If it was on-grid then the grid will compensate (and I suspect the utility will not count it against the customer), but off grid this would be extra stress on the inverter. I’m not actually sure how the inverter would react.

 

[fafrd]

No problem. It’s also not clear to me that a randomly selected AC charger will present the right power factor to those grid tie inverters. To stay within the typical inverter specs the combined load would need to be corrected to 0.8 or better.

If it was on-grid then the grid will compensate (and I suspect the utility will not count it against the customer), but off grid this would be extra stress on the inverter. I’m not actually sure how the inverter would react.

Exactly the sort of input I was seeking when I started this thread

So if the charger has a power factor of 0.8 or higher you’d have no concerns with this ‘turbo’ idea.

But if the charger has a power factor under 0.8, that could be a problem.

The power factor of my rectifier appears to be 0.99: https://www.vertiv.com/globalassets...ems/netsure-731-a412/r48-3000e3-datasheet.pdf

 

[GXMnow]

I am running an AC coupled setup with a Schneider XW-Pro 6848 which says it can handle 6,800 watts of AC coupled solar. I have 4,800 DC watts of solar panel going through 16 Enphase iQ7 inverters which can output nearly 3,900 watts when all 16 are clipping. Most of the time, I am on grid, but my system does go off grid when there is a power failure. I have had 4 grid failures since the system was installed. One time, it was a bit cloudy at 5 pm, so the sun was already going down. Before the grid failed, my Enphase system was only producing about 1,100 watts. My house uses 600 to 1,000 watts most of the time, so it should have just been using a little to no battery to keep up, or even charging a little. NOPE!!

The Enphase inverters are constantly checking if the grid is good. I am not 100% sure what they are doing, but it appears to be a grid impedance check where it looks at voltage change due to current pulses. As stiff as the XW-Pro is, it's not quite good enough. 5 of the iQ7's went offline and never stayed reconnected. A few more would cycle off for a bit and try again. At best, I did see 14 of my 16 working for a bit as the sun was setting and the Enphase system was down to making just 600 watts total. This was with the iQ7's set to the California Rule 21 grid code which should support voltage sag and frequency shift to actually help stabilize the power. Another time, when I went off grid around 11 am, the Enphase was making a lot more power, and it worked about the same. With about 8 (half) of the inverters cycled off an any given time. It was able to provide a little more power than my house was using, so the battery was charging a little. I have updated firmware since on both the XW-Pro and the whole Enphase system, so I am hoping it will work a little better, but my point here is, I had stability issue with the much larger and stiffer XW-Pro inverter. With about the same AC couple solar coming in, I highly doubt the SW inverter will have any chance at keep it stable. Your local micro grid has to be very stiff for the AC coupling to work well. The XW-Pro is a beast of an inverter, and even that is not quite solid enough for Enphase microinverters. Maybe other grid tie inverters might not be as picky, but I think you will have trouble. Every AC coupling guide I have seen says to never exceed the battery inverter power rating with AC coupled PV input. Mine is just about 60%, and it is not always happy.

My recommendation would be to try AC coupling only half of your microinverter array and see if that will stay working in off grid mode. I want to try that with my system, but they combined my 2 x 8 panel groups in the attic and it is only feeding a single 20 amp breaker in the Enphase combiner. That is in spec, and they did run #10 awg wire, so it has low voltage rise. One day, I may still split it to 2 runs for trouble shooting.

 

[fafrd]

I am running an AC coupled setup with a Schneider XW-Pro 6848 which says it can handle 6,800 watts of AC coupled solar. I have 4,800 DC watts of solar panel going through 16 Enphase iQ7 inverters which can output nearly 3,900 watts when all 16 are clipping. Most of the time, I am on grid, but my system does go off grid when there is a power failure. I have had 4 grid failures since the system was installed. One time, it was a bit cloudy at 5 pm, so the sun was already going down. Before the grid failed, my Enphase system was only producing about 1,100 watts. My house uses 600 to 1,000 watts most of the time, so it should have just been using a little to no battery to keep up, or even charging a little. NOPE!!

The Enphase inverters are constantly checking if the grid is good. I am not 100% sure what they are doing, but it appears to be a grid impedance check where it looks at voltage change due to current pulses. As stiff as the XW-Pro is, it's not quite good enough. 5 of the iQ7's went offline and never stayed reconnected. A few more would cycle off for a bit and try again. At best, I did see 14 of my 16 working for a bit as the sun was setting and the Enphase system was down to making just 600 watts total. This was with the iQ7's set to the California Rule 21 grid code which should support voltage sag and frequency shift to actually help stabilize the power. Another time, when I went off grid around 11 am, the Enphase was making a lot more power, and it worked about the same. With about 8 (half) of the inverters cycled off an any given time. It was able to provide a little more power than my house was using, so the battery was charging a little. I have updated firmware since on both the XW-Pro and the whole Enphase system, so I am hoping it will work a little better, but my point here is, I had stability issue with the much larger and stiffer XW-Pro inverter. With about the same AC couple solar coming in, I highly doubt the SW inverter will have any chance at keep it stable. Your local micro grid has to be very stiff for the AC coupling to work well. The XW-Pro is a beast of an inverter, and even that is not quite solid enough for Enphase microinverters. Maybe other grid tie inverters might not be as picky, but I think you will have trouble. Every AC coupling guide I have seen says to never exceed the battery inverter power rating with AC coupled PV input. Mine is just about 60%, and it is not always happy.

My recommendation would be to try AC coupling only half of your microinverter array and see if that will stay working in off grid mode. I want to try that with my system, but they combined my 2 x 8 panel groups in the attic and it is only feeding a single 20 amp breaker in the Enphase combiner. That is in spec, and they did run #10 awg wire, so it has low voltage rise. One day, I may still split it to 2 runs for trouble shooting.

Appreciate this input.

My takeaway fyo your experience with an even larger / stiffer inverter is they DC-coupled solar is still much more ready for prime time than AC-coupling when off-grid.

Let’s hope the rules for post-NEM1/2 make staying on grid a viable option…

 

[zanydroid]

Every AC coupling guide I have seen says to never exceed the battery inverter power rating with AC coupled PV input.

The SW/XW guide extends this to say min of max bulk charge power and battery inverter output power. Which in the case of SW is asymmetric, and with some batteries the recommendation is to set charge lower than discharge.

A picture of the turbo setup would help.

On one of the wiring diagrams I've seen for XW/SW, the microinverters are supposed to be landed in a critical loads panel, and then that panel connected to the inverter. In this case you would put your current switch on the feeder to the inverter, and it may actually need to be a power meter capable of sensing the direction since power can be pushed to the CLP from the SW.

Thinking about how to set the current threshold. I think the microinverters will keep boosting voltage in an attempt to use all DC power available until they hit the cutoff voltage or are curtailed. I think you need the charger to turn on before the SW is tempted to curtail, which I guess is why you picked 8A.

I think the charger would have to be hooked in such that the power meter would not get confused by the power that it draws. For instance if you put it in the CLP then the power going to the inverter will drop.
What happens if the charger stops for some reason while the microinverters are pushing out a lot of juice? That would cause the voltage to spike. What is the ride-through voltage, 255V or something? Hopefully the SW charger & equipment on the CLP can tolerate that voltage.

I'm not sure what to make of the amount of power that the grid tie inverters can surge past the AC charger size (1.5kW).

 

[SamDeleted]

Hence the need for some ‘turbo’ solution to bump max battery charging power from 2400W up over 3500W…

@fafrd , very interesting , I was talking to an old Scottish gentleman that lives off grid on a little croft on a west coast island (rough weather)


He runs his system similar to how your describing , He uses grid tie wind and solar inverters , AC couples to an off grid inverter . but I don't think his is as complicated as yours, I think his inverter can handle the full charging current




Sorry if you've already detailed this in the post, but what is your plan if/when the batteries are full ? Are you going to divert excess power to some kind of dump load ?

 

[fafrd]

The SW/XW guide extends this to say min of max bulk charge power and battery inverter output power. Which in the case of SW is asymmetric, and with some batteries the recommendation is to set charge lower than discharge.

A picture of the turbo setup would help.

GXMNow’s warning about repeated ‘grid impedance check’ has me spooked, so this is all theoretical at this stage and unlikely to be anything I actually implement.

On one of the wiring diagrams I've seen for XW/SW, the microinverters are supposed to be landed in a critical loads panel, and then that panel connected to the inverter. In this case you would put your current switch on the feeder to the inverter, and it may actually need to be a power meter capable of sensing the direction since power can be pushed to the CLP from the SW.

The current switch would be located on the home run wire from the roof into the CLP. No direction needed since there can only be solar power passing through that wire from Microinverters into CLP.

Thinking about how to set the current threshold. I think the microinverters will keep boosting voltage in an attempt to use all DC power available until they hit the cutoff voltage or are curtailed. I think you need the charger to turn on before the SW is tempted to curtail, which I guess is why you picked 8A.

Precisely. Without getting hung up on the rectifier for the moment, the idea is to activate a dump load that will consume 2kW / 8.3A of 240VAC power / current whenever solar currebt exceeds 8A.

I think the charger would have to be hooked in such that the power meter would not get confused by the power that it draws.

The current switch only sees incoming solar power and does not need to know where it is going. Clouds can cause it to thrash (cycle ON and OFF) but CLP loads turning ON and OFF cannot.

For instance if you put it in the CLP then the power going to the inverter will drop.

Yes.

What happens if the charger stops for some reason while the microinverters are pushing out a lot of juice?

I assume you are asking about the rectifier (or dump load) failing (hardware failure). This would cause power into the SW to surge past its ability to absorb it so frequency should shift to the level that Microinverters turn off.

There would be no hysteria in this circumstance so I would want to avoid endless thrashing and would probably want to rig up some kind of safety to keep the AC-coupled string OFF in the case of a hardware failure of the rectifier.

But additional protections in the case of hardware failure is not my priority for now as much as understanding whether the basic idea is sound or flawed. (Repeated hard-to-pass grid quality checks being the biggest potential flaw uncovered so far).

That would cause the voltage to spike. What is the ride-through voltage, 255V or something? Hopefully the SW charger & equipment on the CLP can tolerate that voltage.

Good point.

I'm not sure what to make of the amount of power that the grid tie inverters can surge past the AC charger size (1.5kW).

Oh, as long as the AC power reaching the SW is less than the ~24kW the SW can absorb, it should be used to charge the battery (that is one of it’s central features).

That should continue until the SW enters CV mode as the battery nears full charge at which point it should try to decrease incoming power by increasing frequency.

With Microinverters supporting UL1741SA/B ffequency-shift curtailment, that might or might not work smoothly but since my older Micro-inverters are only UL1731 and do not support frequency-shift curtailment, only frequency-shift ON/OFF control, the SW should increase frequency until the Microinverters shut down.

That’s another situation in which there is no hysteria (other than the battery SOC) and anytime the battery gets charged enough to put the SW into CV mode, I’ll have all the solar charge I need for the day and so would want to keep the Microinverter string off even once the battery voltage drops low enough to put the SW back in CC mode.

But that’s another design detail that can be resolved once the basics are proven to be sound.

I’m assuming that if my AC-coupled solar string was only half as big as it is and cold never exceed 7.4A (absolute maximum ratings) the SW could handle that amount of AC-coupled power without issue (though the issue of thrashing once the battery nears full SOC would still need to be solved).

GXMNow’s warning about how his Enphase Microinverters cause thrashing with his Conext XW just due to repeated grid-quality checks has me spooked.

I though AC-coupling with an off-grid inverter was more mature than it apparently is…

 

[fafrd]

@fafrd , very interesting , I was talking to an old Scottish gentleman that lives off grid on a little croft on a west coast island (rough weather)


He runs his system similar to how your describing , He uses grid tie wind and solar inverters , AC couples to an off grid inverter . but I don't think his is as complicated as yours, I think his inverter can handle the full charging current

If you have a chance to inquire, would love to know what off-grid and Microinverters he is using (and whether he’s ever had issues getting his mucroinverters to qualify the ‘grid’ as presented by his off-grid inverter).

Sorry if you've already detailed this in the post, but what is your plan if/when the batteries are full ? Are you going to divert excess power to some kind of dump load ?

Once the battery nears full, I’ll have all the solar power I can use from that day, so my plan was to just shut down the string until the next morning…

 

[GXMnow]

Once the battery nears full, I’ll have all the solar power I can use from that day, so my plan was to just shut down the string until the next morning…

While this may be true, my battery can reach full as early as 1 pm. So I still like to run on solar until the sun goes down instead of running the battery for 5 hours. That might be tough to do off grid though. Since I am on grid, I just revert back to letting the Enphase system export the extra to grid again and back up those credits for the few times I do have to buy power.

One of the common dump loads is an electric water heater. I wonder if anyone has tried to use something like a programable high power light dimmer to adjust the current into a water heater to match the extra solar production?? Use all the power the solar panels are producing so the battery state stays flat at your full charge. Use more load in the house, or a cloud goes by, then it lowers the power to the water heater. Obviously, you still need another shut off if the water heater get's too hot, but that can take a very long time.

 

[SamDeleted]

If you have a chance to inquire, would love to know what off-grid and Microinverters he is using (and whether he’s ever had issues getting his mucroinverters to qualify the ‘grid’ as presented by his off-grid inverter).

Once the battery nears full, I’ll have all the solar power I can use from that day, so my plan was to just shut down the string until the next morning…

I don't think he's on micro inverters . He told me to try set your grid tie inverter to 'off grid' or 'ac couple' mode if at all possible



What happens if you forget to turn the array off ?

 

[solar8484]

Since it's pure off grid you can disable anti-islanding in the grid-tied inverters to avoid the problems others faced. Also, what is your battery bank size? Is it rated for 3500W charging?

 

[fafrd]

Since it's pure off grid you can disable anti-islanding in the grid-tied inverters to avoid the problems others faced. Also, what is your battery bank size? Is it rated for 3500W charging?

Haven’t built this yet - just trying to plan ahead.

I have 3.5kW of on-grid Microinverters currently but want to add an off-grid system for CL backup and to offset the continuously-running loads (fridges & freezers).

I don’t think I have any way to disable anti-islanding but appreciate the idea and will check with the Microinverter manufacturer…

Battery bank is 14.3kWh @ 100% / 11.5kWh @ 80% and my daily consumption is ~12kWh.

The grid-tied array puts out more than twice what I need in the summer months, so if things reach the point that I want to stop exporting for (increasingly paltry) credit, I’d just shut down the ac-coupled array once the battery gets fully-charged and not worry about the wasted / uncaptured solar energy.

 

[fafrd]

I don't think he's on micro inverters . He told me to try set your grid tie inverter to 'off grid' or 'ac couple' mode if at all possible

I’ll check with the manufacturer if that is possible but end-users / customers have no ability to modify any settings.

What happens if you forget to turn the array off ?

The SW will increase frequency to the point that the Microinverters shut down, so I think there is little risk of anything bowing up.

Biggest issue is to avoid thrashing (ideally full-ON until battery nears full charge then have the SW increase to shut it off until it turns it back on overnight to begin producing again the next morning…).

 

[fafrd]

While this may be true, my battery can reach full as early as 1 pm. So I still like to run on solar until the sun goes down instead of running the battery for 5 hours. That might be tough to do off grid though.

Since I am on grid, I just revert back to letting the Enphase system export the extra to grid again and back up those credits for the few times I do have to buy power.

Which is what I would do until my NEM 1 grandfathering ends…

I just want to plan the system so I have the option to take the grid-tied array off-grid in the case that post-NEM1 terms make it prohibitive to keep grid-tied solar without utility-authorized ESS…

One of the common dump loads is an electric water heater. I wonder if anyone has tried to use something like a programable high power light dimmer to adjust the current into a water heater to match the extra solar production??

My rectifier can be programmed to control input power through CAN…

I’m concerned the reaction time would be too slow to keep up with my constantly-changing Microinverters, however.

And in any case, this does not address the central concern you raised with true off-grid (no connection to grid) because of frequent grid qualification by the Microinverters…

Use all the power the solar panels are producing so the battery state stays flat at your full charge. Use more load in the house, or a cloud goes by, then it lowers the power to the water heater. Obviously, you still need another shut off if the water heater get's too hot, but that can take a very long time.

If I can still export (after my NEM 1.0 expires), this is all a non-issue.

If I can’t (because I decide it’s not worth it under the terms in place at that time), I’ll have more power than I can use on a daily basis, so no real incentive to optimize further once the battery is full.

Hopefully I’ll never need to do this, I just want to select a hybrid inverter that gives me an ‘escape hatch’ to go true off-grid and convert my grid-tied array to AC-coupled off-grid (decommissioned from the utilities perspective).

 

[JBertok]

I was reading what you said about your NEM 1.0 Permission to Operate agreement expiring. Even if you end up with eventually nothing back from your agreement, the Grid could still be a no-cost diversion load, couldn't it?

 

[fafrd]

I was reading what you said about your NEM 1.0 Permission to Operate agreement expiring. Even if you end up with eventually nothing back from your agreement, the Grid could still be a no-cost diversion load, couldn't it?

Absolutely, but it may be prohibitively expensive.

Even if credit drops to $0.00/kWh, that ‘solar tax’ they approved in the Preliminary Decision (since eliminated from the Final Decision last December) may make ability to export any energy to the grid at all too expensive (cheaper to be a plain ol’ utility customer with an off-grid system who only pays the utility for occasionally-needed ‘backup’ power… (AC charger to keep battery from fully depleting).

 

[zanydroid]

I was reading what you said about your NEM 1.0 Permission to Operate agreement expiring. Even if you end up with eventually nothing back from your agreement, the Grid could still be a no-cost diversion load, couldn't it?

Depends on if we ever get consistent negative spot prices during summer days. It's theoretically possible if you have non-dispatchable power that can't be ramped down operating at the same time as the solar glut, but 1741SA is supposed to give utility knobs to turn down solar. And the vast majority of solar at that point in the future should be responsive.

Unless 1741SA was mis-designed. Then it would be a shitshow. But 1741SB's explicit control systems would also have to be mis designed.