Sol-Ark 15K dropping PV (resolved)

[Oldphile]

I've been trying to resolve this since the system was installed 5 months ago. PV production goes to zero for approximately 5 minutes multiple times a day. Initially I thought it was an arc fault problem until I turned off arc fault and the problem continued. I submitted a support ticket to TIGO and after multiple back and forth they concluded that my RSD initiating circuit was turning off the CCA. I spent a month or more trying to find the cause or prove that isn't the problem. If the CCA was powering off, that would be a RSD activation. The TIGO monitoring only shows no PV power. The Sol-Ark monitoring shows no PV current, but PV voltage is there, So, it's not RSD activated. I should have noted this sooner, but I'm still learning the monitoring tools. Now I've submitted a support ticket to Sol-Ark. Anybody have any thoughts? This was a full sun day without a cloud in the sky.

 

[BentleyJ]

Are you Selling Back to the utility with a net metering agreement?

 

[Oldphile]

Are you Selling Back to the utility with a net metering agreement?

Yes.

 

[BentleyJ]

UL1741 has a provision for a 5 min delay before reconnecting if utility power was disqualified and sell back stopped. Not sure this is relevant, just noticed you mentioned the 5 min interval. Throwing it out there for consideration.

 

[Oldphile]

UL1741 has a provision for a 5 min delay before reconnecting if utility power was disqualified and sell back stopped. Not sure this is relevant, just noticed you mentioned the 5 min interval. Throwing it out there for consideration.

Good thought but, the Sol-Ark monitoring doesn't show any issue with grid voltage or frequency. Also my neighbor has the same system without this issue. This is a 2 unit condominium. Each Sol-Ark has about 6' of wire to an AC disconnect and meter where they connect to the grid on common wires.

 

[GregTR]

So I have noticed a similar thing on my Sol-Ark 15K in October, 2022. Note that I use Solar Assistant so I have much more granular data than you do on mysol-ark.com BTW I just learned that PowerView only stores daily data for a year as I could not look up the matching data over there.

For me the trigger for these dropouts is simple:
1) Be on solar but use more power than what solar can provide (I'm running two AC units
2) Be on battery for the surplus consumption and not draw from the grid
3) suddenly lose the big load (AC turns off)

When this happens Sol-Ark can't divert the excess solar quick enough to the grid and since the battery is still almost full, it can't dump into the battery at the full charge rate so the solar "production" drops to meet my in-house demand but it will not export the excess, not for a solid 5 minutes.

Here is the whole story with load, battery, SOC, and grid power data as well:


So what you're losing is exportable energy in those small windows to the grid which is not "that" much of a loss. I did bring it to Sol-Ark's attention but no promise to fix, nor an acknowledgement that this is a problem was ever given, we kind of went back-and-forth in a few emails and that was that.

 

[Oldphile]

@GregTR I've had a few back and forth with Sol-Ark, but got interrupted by a snow storm and 2 day + power outage. Sol-Ark is telling me that I need a battery so the Sol-Ark doesn't need to curtail PV production. But as you noted that doesn't work when the battery is full. Are you seeing PV current dropping to zero?

I got my batteries Wednesday, the day before the storm. I'm hoping I can better see what's going on now that I have batteries.

 

[GregTR]

It does not drop to zero, it drops to whatever my load is in the house. I'm essentially islanding during the day and only use the grid at night since I get free nights of electricity. This is why I am often at 100% on my battery during the day as I seldom drop below 90% between 7AM (free electricity ends) and before the sun comes up.

So I'm still on full solar during these periods I'm just not using all of it so some of it goes nowhere as it can't go in the battery (full) and it can't go on the grid (this is the bug I think that Sol-Ark wasn't going to fix).

 

[Oldphile]

Batteries solved the problem or at least made it better. I haven't had enough decent production days to say it's 100%.

 

[Oldphile]

Better maybe, but not gone. I did an overnight discharge and let the batteries recharge with PV (Sol-Ark set to batteries first}. After reaching 100% SOC and 56.7 V, the battery voltage continued to rise plateauing at 57.2 and then settled to 57.1 At this time I-PV of both MPPT channels went to zero. This lasted for 45 minutes. At 12:55 for some unknown reason, the battery discharged 2.8 A, battery voltage fell to 56.9 and PV production returned.

The batteries are new and have only been in service for 2 weeks. Could that explain the battery overvoltage shoot? That, of course doesn't excuse the Sol-Ark behavior.

 

[Quattrohead]

That is working exactly as expected now.

 

[Oldphile]

That is working exactly as expected now.

The battery or the Sol-Ark?

 

[Oldphile]

It just got worse. After no PV for 20 minutes I tried PV disconnect for 15 seconds. That resulted in about 300 watts PV. Then I tried AC disconnect. The data snap shot is with AC disconnected. After reconnecting to the grid, I set batteries to discharge at 12 noon. Bang, PV returned at noon. Sky is cloudless, no shade.

 

[DIYrich]

What are your export (grid sell) settings?
What are your Time of Use settings?

Post pics of the screens.

Seems like your batteries are full, and it doesn't want to export excess production to grid.

 

[Oldphile]

Yes batteries are full when this happens. It was also happening without batteries, but would recover in 5 minutes. Now I have to slightly discharge the batteries. All my TOU settings are just test values.

 

[DIYrich]

I prefer looking at the screen on the sol-ark.
Turn off TOU if you are not using it. Just confuses things.

Why are you limiting export to 5500w?

 

[Oldphile]

I prefer looking at the screen on the sol-ark.
Turn off TOU if you are not using it. Just confuses things.

Why are you limiting export to 5500w?

Sol-Ark tech support seemed to be concerned about rising grid voltage. He asked if it had risen to 126 volts. There's 3 Sol-Ark 15K on the property with a total of 22.9KW PV. I've seen as much as 127 volts on one line, so I restricted export limit to a total of 18.5KW. I'm just trying to eliminate a red herring.

I'm using TOU to discharge the batteries. I'm also using it to clear this fault. If I disconnected from the grid for the night and then reconnected in the morning, the batteries would charge until late morning and then the Sol-Ark would shed the PV. For how long, I don't know. The longest it has gone before I intervened is 80 minutes. I'd rather not lose all my afternoon PV production. By intervene, I mean use TOU to discharge the batteries after Sol-Ark sheds the PV.

 

[Brucey]

Can you set it to only charge up to say 95% SOC? So then even if it's "full", the battery can still take the charge/excess as the AC load drops and the inverter ramps down.

 

[Oldphile]

Can you set it to only charge up to say 95% SOC? So then even if it's "full", the battery can still take the charge/excess as the AC load drops and the inverter ramps down.

There's no SOC charge limit setting for PV. The best I can do is set loads first, not batteries first. That just extends how long it takes for the batteries to be fully charged.

 

[Cmiller]

There's no SOC charge limit setting for PV. The best I can do is set loads first, not batteries first. That just extends how long it takes for the batteries to be fully charged.

If you set your TOU setpoints to 95% as suggested by @Brucey then pv will only charge to that before pushing the rest to grid. Think of the TOU setpoint as the "discharge to...." setpoint. The Sol-Ark will even "sell" out of the batteries down to the TOU setpoint.

In other words, "battery first" will come into play until the batteries get up to the TOU setpoint, then PV gets pushed to AC until TOU calls for a higher SOC%.

 

[Oldphile]

If you set your TOU setpoints to 95% as suggested by @Brucey then pv will only charge to that before pushing the rest to grid. Think of the TOU setpoint as the "discharge to...." setpoint. The Sol-Ark will even "sell" out of the batteries down to the TOU setpoint.

In other words, "battery first" will come into play until the batteries get up to the TOU setpoint, then PV gets pushed to AC until TOU calls for a higher

@Carlos_Sol-Ark has set it to 99% to try that.

 

[DIYrich]

Turn off grid charging. Let the solar fill batteries as much as it can. Let your load drain batteries as much as it can. That is the most efficient way to use batteries, unless you have time of use rates. Solar first, battery, last is grid.

 

[Cmiller]

Turn off grid charging. Let the solar fill batteries as much as it can. Let your load drain batteries as much as it can. That is the most efficient way to use batteries, unless you have time of use rates. Solar first, battery, last is grid.

Unfortunately there isn't one "cookie cutter" setup that "is the most efficient way" to run every system! There are multiple things that factor into what makes the most sense and is the most efficient. On HV DC bus all-in-one inverters like the Sol-Ark, Eg4, etc. you have a lower efficiency when pushing power to and from the batteries, than when you are keeping everything running between PV and AC. So if a person has a net metering agreement that is a 1:1/kWh:kWh within the billing cycle, and at the end of the billing cycle only the difference is billed or credited, it usually makes the most sense to actually not even cycle the batteries and use the grid as a battery with the physical batteries there for backup when grid goes down! The efficiency when keeping the batteries full for backup and using grid as your battery is low to mid 90s % (probably 92-95%) and when you push most of your produced power into the batteries then draw it back out overnight your efficiency drops to more like 80-85%!

A few things that factor into what "is the most efficient" setup:

• Net metering/ net billing agreement details
• price per kWh for buying/selling from/to the PoCo
• how much battery storage one actually has vs. daily kWh usage
• how much PV one has vs. daily kWh usage
• what kind of system equipment one has (eg. AIO, inverter with charge controller separate, AC coupled, etc.)

 

[Oldphile]

a net metering agreement that is a 1:1/kWh:kWh within the billing cycle, and at the end of the billing cycle only the difference is billed or credited

This is what I have. I'm only cycling the batteries to see how the batteries and system perform. I got batteries solely for backup. I would join Connected Solutions, but the only battery Eversource accepts into their program is Enphase.

 

[Oldphile]

I found normal 100% SOC voltage for these batteries to be 56.7
If I turn off TOU, battery voltage will range from 56.8 to 57.1 when PV is producing.
The Sol-Ark wants to utilize the batteries like a shock absorber. With TOU turned off, this doesn't happen and PV production is curtailed. The only way to bring it back is to turn on TOU and set SOC to 99%. The battery then alternates charging and discharging with battery voltage varying between 56.3 and 56.8

The Sol-Ark should be able to function without utilizing TOU.
My other Sol-Ark 15K doesn't have this problem and doesn't even have a battery.