Sol-Ark 12K + Enphase IQ8 microinverters

[wheisenburg]

Success!!!

I finally got it working! I appreciate all the advice from everyone on here as I performed many many tests over that past few months from all your suggestions. It was my more in depth conversation with @DER Solar and also suggestions from @BenW that finally got it working.

Just to reiterate my situation:

On the site there is 24 x Enphase IQ8+ micros AC coupled into the GEN port of the Sol-Ark 12K, no DC coupled solar. 16 x Rolls S6 460AGM battery bank used for partial home backup.

That is fantastic news. So just to be clear, what profile did you use that finally got this working? I am using IEEE 1547 default 2015. I have a large contactor in my system that can be used to disconnect the Solar GT PV. The Schneider does not have software that can control a relay on the Gen port. They do some programmable 12 Volt outputs that can be used to drive relays. I have two inverters that both can be used.

I have it setup now so that power on the gen in wires will turn PV off and connect the Gen in. I may just get DC coupled charger that runs directly from the Gen AC to the DC bus. Then I will rewire the contactor to use the NO poles for the GT PV. I should then be able to set up relay driven logic that will turn on the contactor.

1. On anytime the grid has power.
2. I can set control output on one inverter to turn on when SOC is less than 85% and the other when SOC is less than 75%.
3. When the 75% relay turns on it will use one of the poles as a feed back "Latch" so it stays on. When the 85% latch turns off it will turn everything off. This should allow me to cycle the system between 75% and 85% SOC. Within that range the Schneider may try to do frequency shifting.

 

[wheisenburg]

I thought the SolArk was capable of modulating volts/frequency to trigger curtailment on the Enphase side. Guess not.

It can do this, but there are two problems here us IQ8 AC coupled people are facing. One is finding a "Grid Profile" that allows the inverters to turn on at all. This seems to be a big problem with IQ8s. The second is how to make sure you don't operate the AC Coupled PV when the batteries are at too high a SOC. The frequency shifting is not instantaneous, so the batteries need to be available as a place to push power to while the frequency shifting has time to work.

The more batteries you have, and the lower their SOC, the more power they can accept without unwanted voltage rise. On a sunny day my system can put out in excess of 9000 watts so that has to go somewhere. The frequency shifting can take 2-4 seconds depending on the profile. There are systems that use proprietary communications protocols to adjust the Solar GT PV more rapidly. So 9000 watts is 180 amps at 50 volts. If you have three batteries they will each need to take 60 amps. This is faster than you really want to charge a battery. With 6 batteries you can cut that to 30 amps per battery. That is a much more comfortable rate.

Anyway, the idea that you can just pull these things out of a box, hook them up and they will work great is not true. It seems to take a great deal of experimentation and tinkering to get this working well. That is too bad because I think a lot of people would potentially like to add backup power to their GT Micros.

 

[solar8484]

Firmware is - 521-00005-r06-v02.56.01

The Sol-Ark will close the GEN port relay when the SOC reaches 60% turning on the microinverters. Once the batteries charge up to 85% the relay is opened and the solar turns off. It may work at a higher SOC such as 70% instead of 60% but I wanted to make sure the batteries can accept the current inrush from the solar without any issues.

The fact that SolArk requires a relatively low SOC even with a more than adequate battery bank to successfully AC couple with iQ8's suggests that it may not be properly handling battery voltage spike transients relative to the battery voltage high cutoff setting. Can you share the setting?

 

[DER Solar]

Success!!!

I finally got it working! I appreciate all the advice from everyone on here as I performed many many tests over that past few months from all your suggestions. It was my more in depth conversation with @DER Solar and also suggestions from @BenW that finally got it working.

Thanks so much for your dedication to find a solution, hard work trying different solutions and spending incredible amounts time to figure it all out and your patience when things did not work.

IQ8's are know to be very finicky and difficult to AC-couple with FW, VW, VV but they work fine in a pure Enphase backup environment. They are very sensitive and very fast responding to FW from 60 Hz to 61Hz and shut down at 61+ Hz. The key to your success is not to use any power curtailment, just straight AC coupling with constant frequency and constant voltage. AC coupling via the generator port does that and the generator relay performs the hard switching PV on and off.

My best guess is that the IQ8's can do 60 Hz impedance checking to see if they are connected to the utility transformer (low impedance), Enphase NFT (medium impedance) or TL inverters (high impedance).

If someone has a SolArk with an additional external NFT for handling max load imbalance, it would be interesting to test if the IQ8's will work with the latest UL 1741 SA or SB firmware.

 

[DER Solar]

Feel free to check with PoCo's yourself but in general, PUC/PoCo's interconnection rules don't apply to things behind the PCC. But, just for clarification, there is nothing preventing you from having a grid profile that comply with the CA Rule 21 ride-through requirements but with anti-islanding disabled. Remember, the purpose of anti-islanding is to protect against "unintentional" islanding while connected to the grid for safety reasons. Your micros behind a interconnect compliant (i.e. active anti-islanding on the grid side) hybrid inverter will disconnect (via a switch) your micros from the grid side in grid outages so there is no risk of backfeeding the grid.

When Enphase IQ8s or IQ7's are connected to the IQ system controller and the MID relay is closed, they are directly connected to the grid. They need to be compliant with the grid connect rules IEEE 1547-2018 and IEEE 1547.1-2020 in these states:

• New York, Pennsylvania (PPL), New England states (MA, CN, RI, NH, VT, ME) as of Jan 1, 2023; Hawaii as of Feb 1, 2023.
  
  
• California and New Mexico are starting on April 1, 2023, may have delayed.

UL1741SA is the testing standard for IEEE 1547-2018 and IEEE 1547.1-2020 compliance.

In states that have these requirements, you cannot use a grid profile not compliant with these standards when connected to the grid.
But you can use another profile when off-grid, but I don't know any micro-inverter that allows 2 different profiles: one for grid connectivity and one for off-grid. Enphase may do this internally inside their IQ backup solution.

The major reason for stricter grid connectivity rules is the stabilization of the grid, like LVRT and HVRT aka as fault ride-through for low voltage (brownout) and high-voltage events.

Grid connected systems with backup have MID disconnect relays to guarantee anti-islanding as noted by solar8484. AC coupling on the load side could allow us to remove the anti-islanding control from the profile and still be compliant with grid connect requirements.

Several have reported instability of the IQ8's when off-grid under changing backup loads. When backup loads are switching off, the backup micro-grid impedance increases and could fool/trigger the IQ8 micro-inverter anti-islanding handler to switch offline and stay offline.

I am hopeful we are getting closer to a solution for successful IQ8 AC coupling with a variety of hybrid backup inverters.

 

[Ampster]

When Enphase IQ8s or IQ7's are connected to the IQ system controller and the MID relay is closed, they are directly connected to the grid.

The big unknown is what controls that MID relay? I have heard suggestions that since the SolArk controls that relay and that control will be consistent with CA Rule 21 and/or UL1741SA/B. Furthermore the recommended connection is through the gnerator port which also has a relay. Therefor my assumption is that the micros connected through the SolArk will be disconnected consistent with the requirements of those protocols through the opening of either of those relays. What I do know from my tests is that when I disconnect the grid the gen port relay also opens. What I do not know is if there is a grid event where the grid voltage or frequency goes high what will happen for sure. I can only assume that both relays would open. Previously I had been reluctant to change the grid profile on my IQ7s until I heard the above theory. I have not changed it yet but would like to test that assumption.

 

[DER Solar]

The grid-to-load AC pass-through relays in the SolArk, the MID relay in an Enphase IQ controller, the FranklinWH aGate MID relay, the similar relay in a Tesla main panel and the Schneider XW Pro AC disconnect relays on the AC board are all required primary MID disconnect devices and are under the control of either the primary hybrid inverter with the main AC MID or pass-through relay or the MID panel with an MID logic controller. The inverter or MID controller needs to be listed for anti-islanding.

Connecting micro-inverters to the grid-side requires anti-islanding enabled, but connecting them to the backup side behind the existing MID switch may not require anti-islanding. The SolArk generator disconnect relay could actually stay on during a grid-to-backup failover if the SOC is not too high. But it is safer to switch them off to avoid overcharge on a high SOC battery.

In testing IQ7's, IQ8's and Hoymiles micro-inverters with a Schneider XW, they all disconnected during grid-to-backup failover with the default IQ grid profile "IEEE 1547 default 2015" and the latest Hoymiles "UL 1741 SB" profile. Will do more IQ inverter testing with a UL 1741 SA profile and see how they handle LVRT events during fail-over.

Also noticed different grid-fault/switch-over behavior on 2 Enphase IQ gateway versions: the older IQ gateway (800-00555-r03, 2017) reboots, the newer (800-00655-r09 , 2021) gateway stays on.

It would be great if hybrid inverters like SolArk, LuxPower or EG4-18KVP had an additional relay controlled AC coupling port so we could have both: relay-controlled AC coupling and generator backup.

 

[Ampster]

It would be great if hybrid inverters like SolArk, LuxPower or EG4-18KVP had an additional relay controlled AC coupling port so we could have both: relay-controlled AC coupling and generator backup.

I agree not because I want to plug in a generator but because I want to use an auxiliary port for load control of loads like an electric dryer or electric oven. My former Outback hybrid had a very good AC coupling algorithm but did not have the AC coupling capacity which my SolArk has. I have always assumed it was because the SolArk could open the relay if the lag on modulating the GT inverter{s} was too slow. I have few power outages and most of them are short enough that I don't have to worry about utilizing the AC coupled solar since I have 2kW of DC coupled solar and am going to add another 1.2kW DC coupled solar.

 

[DER Solar]

I agree not because I want to plug in a generator but because I was to use an auxiliary port for load control of loads like an electric dryer or electric oven. My former Outback hybrid had a very good AC coupling algorithm but did not have the AC coupling capacity which my SolArk has.

An auxiliary relay-controlled AC circuit for external loads would be great! Many hybrid inverters do not have enough flexible i/o control ports. You have to add too much external logic controls and a few external logic controllers to do load shedding, load diversion and to drive an external MID relay or contactor.

FranklinWH aGate MID panel has 2 smart circuits for load control: load shedding and future load diversion. 80A targeted for an EV charger and 50A targeted for a general purpose load shedding device like an oven or AC unit or a subpanel.

Schneider XW Pro is unique too: it has i/o control ports for driving an external MID device (BCS 2200) or for driving an external contactor, 200A or more. Separating MID from the stacked inverters facilitates the relocation of the inverters away from the main service to a location suitable for NFPA regulations, like a fire-proof room for LIF batteries.

Load diversion is often under-utilized during a grid outage, you can divert excess PV energy into DHW or an EV charger instead of shutting PV off.
The next generation Tesla Powerwall 3 will have support for both load shedding and load diversion on the EV charger.

 

[solar8484]

When Enphase IQ8s or IQ7's are connected to the IQ system controller and the MID relay is closed, they are directly connected to the grid. They need to be compliant with the grid connect rules IEEE 1547-2018 and IEEE 1547.1-2020 in these states:

• New York, Pennsylvania (PPL), New England states (MA, CN, RI, NH, VT, ME) as of Jan 1, 2023; Hawaii as of Feb 1, 2023.
  
  
• California and New Mexico are starting on April 1, 2023, may have delayed.

UL1741SA is the testing standard for IEEE 1547-2018 and IEEE 1547.1-2020 compliance.

POCO's enforce interconnection requirements at the PCC (or POI) which is not just any electrically connected point. As long as a distributed generation system can meet the requirements at the PCC then it doesn't matter what generators/inverters do behind the PCC. So, if you have a grid interface controller that connects directly to the PCC and meets all the interconnection requirements then the generators/inverters behind the controller don't need to. Tesla Powerwall/Gateway is probably the most widely deployed example of such grid interface controller.

 

[solar8484]

Also noticed different grid-fault/switch-over behavior on 2 Enphase IQ gateway versions: the older IQ gateway (800-00555-r03, 2017) reboots, the newer (800-00655-r09 , 2021) gateway stays on.

The newer iQ Envoy is UL1741SB compliant so it must stay up during ride through events to mediate any power control communications from the POCO's during the events.

 

[DER Solar]

The newer iQ Envoy is UL1741SB compliant so it must stay up during ride through events to mediate any power control communications from the POCO's during the events.

Correct, that's exactly what's happening.

 

[DER Solar]

POCO's enforce interconnection requirements at the PCC (or POI) which is not just any electrically connected point. As long as a distributed generation system can meet the requirements at the PCC then it doesn't matter what generators/inverters do behind the PCC. So, if you have a grid interface controller that connects directly to the PCC and meets all the interconnection requirements then the generators/inverters behind the controller don't need to. Tesla Powerwall/Gateway is probably the most widely deployed example of such grid interface controller.

Not sure what you are implying.

Existing AC coupled inverters behind the MID relay are grand-fathered in and can stay and don't need to be compliant with new rules. New AC coupled inverters installed at the time of the MID device installation or afterwards need to comply with the POCO interconnection requirements because the load side is directly connected to the grid side via the closed MID relay.

But here is an interesting case where I don't know the answer: if there is a PV disconnect relay, like the one on the SolArk generator port, the MID controller could disconnect non-compliant AC coupled inverters from the grid and reconnect after the backup is operational. Don't know if a POCO will allow installation of non-compliant inverters that don't directly support LVRT and HVRT, most likely not.

 

[solar8484]

Not sure what you are implying.

Existing AC coupled inverters behind the MID relay are grand-fathered in and can stay and don't need to be compliant with new rules. New AC coupled inverters installed at the time of the MID device installation or afterwards need to comply with the POCO interconnection requirements because the load side is directly connected to the grid side via the closed MID relay.

But here is an interesting case where I don't know the answer: if there is a PV disconnect relay, like the one on the SolArk generator port, the MID controller could disconnect non-compliant AC coupled inverters from the grid and reconnect after the backup is operational. Don't know if a POCO will allow installation of non-compliant inverters that don't directly support LVRT and HVRT, most likely not.

It just means there are multiple ways to meet interconnection requirements at the PCC. A grid interface controller (MID may be a component of it) can meet interconnection requirements potentially with or without assistance from the generators/inverters behind it. So, a grid interface controller may need to be upgraded to meet the current interconnection requirements when adding inverters behind it. For example, if you add new inverters behind an existing Tesla Powerwall/Gateway system then the Tesla Powerwall/Gateway may need to be upgraded to meet the current interconnection requirements if the new inverters don't do it (even if they are capable) for some reason.

 

[DER Solar]

It just means there are multiple ways to meet interconnection requirements at the PCC. A grid interface controller (MID may be a component of it) can meet interconnection requirements potentially with or without assistance from the generators/inverters behind it. So, a grid interface controller may need to be upgraded to meet the current interconnection requirements when adding inverters behind it. For example, if you add new inverters behind an existing Tesla Powerwall/Gateway system then the Tesla Powerwall/Gateway may need to be upgraded to meet the current interconnection requirements if the new inverters don't do it (even if they are capable) for some reason.

All of this is very interesting and challenging. Let's define the semantics on the terms we are using, I'll take a first pass on the terms I am familiar with.

MID:
MID is a MicroGrid Interconnection Device, typically a relay or contactor disconnecting the micro-grid and required and sufficient for anti-islanding, as a relay it is required but not sufficient for supporting FRT fault ride-through. However, it could be a full-blown electronic device - Grid Interface Controller - that connects and disconnect the microgrid from the utility grid including FRT support and more.

Grid Interface Controller:
see above, everything that facilitates grid connect and disconnect, support FRT, but may not necessarily include/imply a grid inverter.

FRT (Fault Ride Through), LVRT (Low Voltage Ride Through), HVRT (High Voltage Ride Through):
LV and HV are events and FRT, LVRT and HVRT are actions or counter measures to stabilize the grid so it does not collapse.
A 1741 SB MID compliant relay needs to stay closed longer than a UL 1741 SA MID relay so the grid-tied inverter can ride-through the fault event to stabilize the grid.

LV (low voltage) or "brown-out" is a grid event where there is too much demand and not enough supply. Voltage drops, lights dim, but as long as the voltage stays above the LV shut down or within a defined 0-V time frame for several milliseconds, the grid-tied inverters needs to stay connected and hopefully inject power into the grid to add supply. At a minimum the inverter should stay on, and if possible increase supply, either from export-limited PV or from battery.

HV (high voltage) is over-supply and not enough consumption, it happens during spring and fall months near noon at high solar irradiation. Too much PV power supply in a residential area because everyone is at work not consuming power at home and not charging the EV at home and AC units are not yet cranking. If HV is below the upper HV cut-off, PV grid inverters need to stay connected, curtail PV power via PoCo induced FW, VW or VVar, increase consumption by charging batteries or by increasing reactive power which essentially puts on the "breaks" by shifting phase on the injected AC current, consumed energy gets converted into heat and the inverter fans turn on.

PCC and PCI:
These are terms you could elaborate on, I have only an intuitive understanding what they mean and what they imply.
Is a "PCC" all the way down via the service entry, through the MID switch, into the backup-panel, into backup sub-panels and into the outlet?
Or does it stop at the MID switch?

Tesla Powerwall/Gateway:
Regarding the Tesla Powerwall/Gateway, it needs to be upgraded unless the new grid-compliant inverters are connected either directly to the grid or to the backup side of a new additional grid-compliant backup system. Question: can a Tesla Powerwall AC battery/inverter be added to a non-compliant Powerwall gateway to increase backup power?

Inverter Stacking:
Here is another interesting case: I have Schneider XW+ inverters that are UL1741 compliant. They can stay connected as long as the setup is not changed. I cannot stack another UL1741 XW+ unit where UL1741 SA would be required. I have to upgrade all inverters to be UL 1741 SA compliant. But I can create a new second UL 1741 SA compliant backup system with new separate backup load panels and leave the old UL 1741 backup system unchanged.

 

[Techy D]

That is fantastic news. So just to be clear, what profile did you use that finally got this working? I am using IEEE 1547 default 2015. I have a large contactor in my system that can be used to disconnect the Solar GT PV. The Schneider does not have software that can control a relay on the Gen port. They do some programmable 12 Volt outputs that can be used to drive relays. I have two inverters that both can be used.

I have it setup now so that power on the gen in wires will turn PV off and connect the Gen in. I may just get DC coupled charger that runs directly from the Gen AC to the DC bus. Then I will rewire the contactor to use the NO poles for the GT PV. I should then be able to set up relay driven logic that will turn on the contactor.

1. On anytime the grid has power.
2. I can set control output on one inverter to turn on when SOC is less than 85% and the other when SOC is less than 75%.
3. When the 75% relay turns on it will use one of the poles as a feed back "Latch" so it stays on. When the 85% latch turns off it will turn everything off. This should allow me to cycle the system between 75% and 85% SOC. Within that range the Schneider may try to do frequency shifting.

The grid profile is IEEE 1547 default 2015 1.0.11.

That sounds like a solid plan of attack.

 

[Techy D]

I thought the SolArk was capable of modulating volts/frequency to trigger curtailment on the Enphase side. Guess not.

The Sol-Ark is capable of doing that, I had it working VERY briefly where it was modulation the frequency without the micros turning off completely. I tried many different grid profiles as well but none seemed to help it.

 

[Techy D]

Thanks so much for your dedication to find a solution, hard work trying different solutions and spending incredible amounts time to figure it all out and your patience when things did not work.

IQ8's are know to be very finicky and difficult to AC-couple with FW, VW, VV but they work fine in a pure Enphase backup environment. They are very sensitive and very fast responding to FW from 60 Hz to 61Hz and shut down at 61+ Hz. The key to your success is not to use any power curtailment, just straight AC coupling with constant frequency and constant voltage. AC coupling via the generator port does that and the generator relay performs the hard switching PV on and off.

My best guess is that the IQ8's can do 60 Hz impedance checking to see if they are connected to the utility transformer (low impedance), Enphase NFT (medium impedance) or TL inverters (high impedance).

If someone has a SolArk with an additional external NFT for handling max load imbalance, it would be interesting to test if the IQ8's will work with the latest UL 1741 SA or SB firmware.

Yes, lots of patience was needed indeed!

These IQ8’s have a brain of their own and are super responsive to very small changes - I’m completely okay with using the GEN relay opening and closing on the Sol-Ark instead of frequency modulation.

 

[wheisenburg]

Yes, lots of patience was needed indeed!

These IQ8’s have a brain of their own and are super responsive to very small changes - I’m completely okay with using the GEN relay opening and closing on the Sol-Ark instead of frequency modulation.

There are disclaimers in some of their "Off Grid Profiles" that they only work with IQ7s.

 

[12kw_2021]

I experimented off grid with IQ7a inverters and the ieee2015 zero export grid profile. AC coupled to the solar 15K genpins.

Goal was to have the enphase zero grid export profile control the iq7a's and ramp them up and down as needed to support battery charging and load.

My findings are if the grid is connected this zero export grid profile works just fine and I ran it for several months before I had permission from my local electric company to export power.

Zero power was exported. Of course the amount of PV generated each day was also reduced as there was no place for any excess power to go but it did support all loads with available PV and charge the battery each day.

However when I disconnect the grid using the same profile that worked perfectly for months on grid the solarc genpins do not supply voltage and frequency that is stable enough and the enphase iq7a's completely turn off and don't appear to turn back on unless I go out to my combiner box and reboot the system by turning off the breaker on the combiner box which causes a reset and a refresh of each micro inverter.

Data from the enphase installer application shows that the solark AC Gen pins frequency shifts as high as 60.6 hz and this causes the iq7a to turn off.

The behavior is very odd running off grid with this profile initially I'll see full PV power generation of 7 or 8 or 9 kilowatt hours and then it gradually decreases to one or two kilowatts and then it shuts itself off when that AC frequency from the Sol-arc shifts .

This is operating with the battery at a very low state of charge around 30%, so the solarc has plenty of places to put 100% of the available PV.

So my conclusion is the solarc gen pins off grid are not stable enough to allow proper operation of the iq7a microinverters using a zero export grid profile.

I'm going to do more experiments with other enphase grid profile and I guess I'll share here but it looks like in my case the solarc can only be used to turn the AC coupled solar on or off depending on state of charge of the battery.

Talking to sol-arc technical support their feedback is graceful control of the enphase micro inverters by varying the frequency it's not possible using AC coupled solar on the Gen pins the sol-arc can just turn the inverters on or off.

My operational plan was to run with the normal IEEE 2015 grid profile standard mentioned above when running on grid and exporting power then switch to the IEEE 2015 0 export grid profile when running off grid and have the enphase microinverters ramp up and down depending on the load but my experiment failed.