Sol-Ark 12k Problem Voltage spikes and Output Power to Zero

[SolarRich]

There's something going on with either the Sol-Ark or the arrays. Arrays are 9 panels 400 watt Solaria. 2 arrays. One array per dual axis tracker.
The main symptoms I see are PV voltage spiking while at the same time the PV current going to 0 or near 0. It seems to occur to both arrays at the same time however today was an exception. Today PV2 array(blue line) was looking worse than PV1 array(green line). In prior says I turned off 1 array while leaving the other on and vice-versa. Same results from either array when the other array is off.The disconnect switches cut off both positive and negative and are fused.
Here are some charts from PowerView



The other mystery is that PV2(blue) has a high Vpv when there is no sun. I try to measure this voltage with my DMM (AC or DC) and just get a couple volts that varies alot. So I can not confirm this voltage.

I did call Sol-Ark today but that was a couple hours before closing. I left a voice message.
Any ideas of where to start troubleshooting?

This is an older system that now has new panels and inverters. There are 2 more identical trackers/arrays that go to a SunnyBoy grid tie inverter. This half of the system has no problems.

 

[Solar Guppy]

Is the Sol-Ark selling to the grid? if so you are likely have a grid event ( over voltage ) would be my best guess. If this is the case, you need to monitor the grid and see where or what is causing the over-voltage. Since you have a Sunnyboy and the Sol-Ark is new that should help narrow down the issue.

 

[SolarRich]

Yes, selling to the grid. Did not think of it as a grid issue so now I need to look there. Here's a chart of the output voltage. It's hard to tell but the voltage spikes on the PV input do not correlate to high AC volts - at least not directly in the charts. But AC does hit 125VAC from time to time. Perhaps I should try a day without the SunnyBoy on?

 

[SolarRich]

Is the Sol-Ark selling to the grid? if so you are likely have a grid event ( over voltage ) would be my best guess. If this is the case, you need to monitor the grid and see where or what is causing the over-voltage. Since you have a Sunnyboy and the Sol-Ark is new that should help narrow down the issue.

Thanks Solar Guppy!
It's about 2PM here and so far the Sol-Ark has been OK. The SunnyBoy is turned off.
So if this continues to be good, how do I make the Sol-Ark and Sunny Boy exist together?

Both inverters feed a dedicated 200A panel that only has the outputs of the inverters coming into the panel and then an 80A breaker with #2 gauge wire to the PV disconnect switch on the outside of the house (about 5 ft wire each way). After the disconnect the 2 gauge wire is spliced into 4/0 wire that goes out to a 200A breaker and then the power meter.
This was all done many years ago and has worked w/o problems until the Sol-Ark arrived.

Each inverter will push 29amps at 240V to the grid based on the PV panels at STC rating. So lets say a max of 60amps for both inverters.Most sunny days are at about 25A output per inverter.
The SunnyBoy is still working with this setup.

Does the location of the breakers in the 200A panel make a difference?
I've seen 125VAC on each of the 2 legs based on the Sol-Ark Power View chart as posted above. But perhaps there's more to the story. Spikes of voltage the chart is not capturing that cause the Sol-Ark to reduce power for that instant?

The SunnyBoy chart of power output does not show any variations or spikes like the Sol-Ark chart. The SunnyBoy data and chart are on the SMA SmartView website. I also have a revenue grade power meter on the output of the SunnyBoy that I pull(modbus RS485) data from so I can produce my own charts. I may be able to get get line voltage out of this meter also.

Any ideas?

 

[Solar Guppy]

I suspect the data logging may not be fast enough ( sample rate ) to see the grid go over voltage, the reason you see the PV is the inverter likely goes off line for 5 minutes which is a common value, while the grid maybe a single over 254V event measured in seconds. 250 is very high by the way, nominal should be 240V.

I have never had hands on a Sol-Ark, but it maybe possible to increase the voltage threshold in the Sol-Ark in the inverter menu, sorry I have no idea where, but likely under something called grid profile as a guess.

Depending on the age of the SMA, it likely is using 264V as the trip voltage, which was allowed in older UL1741 versions.

So if you can change the profile for the Sol-Ark, that's one option, the other is to use larger wire to lower the resistance ( inverse of drop as inverters are the power source ). You could also measure the voltage at each step ( inverter to meter ) to the grid to see where the largest increase in voltage is coming from, it could be a weak breaker, lose lug ect.

Finally, it maybe the connection from the local utility transformer to the meter, if so, you would need to discuss with the power company about options on that. Did the utility approve 15kw? that's a fairly large solar system and the utility would need to check their end to see if they can handle selling of that amount of energy.

 

[Clint-L]

There's something going on with either the Sol-Ark or the arrays. Arrays are 9 panels 400 watt Solaria. 2 arrays. One array per dual axis tracker.
The main symptoms I see are PV voltage spiking while at the same time the PV current going to 0 or near 0. It seems to occur to both arrays at the same time however today was an exception. Today PV2 array(blue line) was looking worse than PV1 array(green line). In prior says I turned off 1 array while leaving the other on and vice-versa. Same results from either array when the other array is off.The disconnect switches cut off both positive and negative and are fused.
Here are some charts from PowerView
View attachment 116402
View attachment 116404

The other mystery is that PV2(blue) has a high Vpv when there is no sun. I try to measure this voltage with my DMM (AC or DC) and just get a couple volts that varies alot. So I can not confirm this voltage.

I did call Sol-Ark today but that was a couple hours before closing. I left a voice message.
Any ideas of where to start troubleshooting?

This is an older system that now has new panels and inverters. There are 2 more identical trackers/arrays that go to a SunnyBoy grid tie inverter. This half of the system has no problems.

I’ve been dealing with same issue. I had a 12k hooked on same array and it was fine. They have bugs in the 15k. With everyone seeming to be going crazy buying Solarks with the Russia thing going on, support has become pretty non existent lately.

I had voltage spikes going over 600v but at the same time the PV would drop to zero. It’s like the MPPT tracking is all jacked up too.

They put a new software version on my 15k just last week 10/16/22. They admitted to having some issues of the 15k reading incorrect voltage on arrays as well as ramping up the grid voltage too high. My unit was actually disconnecting from grid as well.

I’ve been highly disappointed in the 15k. I believe they rushed it to market. I hope they can get it corrected.

The other issue they have now is, it will only sell back and invert 13k of PV. It caps out. Not only it caps out, but it does the things above I just described.

Maybe some of this is a recent software issue that plagues both the 12k and 15k? Who knows at this point.

@Devin
@SolarRich

 

[Rennit]

I've had issues similar to this with my 12k.
Issue 1: PV1 and PV2 would read different voltages, amps and watts after a firmware update. I had sol-ark roll back the firmware version and it fixed the problem.

Issue 2: In the app, the graph under the plants tab would show something similar to what you are seeing with my battery voltage but under the equipment tab everything graphed normal. I wrote it off as a monitoring issue. I think that only happened once.

 

[SolarRich]

I've learned a few things about the spikes I'm seeing on the PV inputs.
1. It appears that the voltage spikes to Voc (or higher) when the inverter stops taking in power from the panels.
2. The AC inverter output also drops at the same time.
3. This issue is not related to the power output of the inverter. It happens on cloudy days as well as full sun days. So no issue with the grid-tie wiring, etc. Normal AC voltage here runs from 120-122 per leg at night. During the day the voltage is 121-123 most of the time.
4. The SMA inverter that is here also feeds into the same inverter output panel. And it has no problems.
5. The inverter that the Sol-Ark replaced never had this problem.
6. Sol-Ark support has been in the loop with this. They asked to turn on ARC detection as it was off by default. No ARC errors.
7. Both PV arrays have been basically ruled out.
8. There have been other small issues during this time that were solved but did not seem to be part of this problem.

At the end of the day I have to point to the Sol-Ark inverter as the reason for the spikes. Let's see if Sol-Ark support has anything more to say. Still looking for problems here - other than the Sol-Ark.

From what I've read about these Sol-Ark inverters is that they are coming from a factory somewhere and being rebadged as Sol-Ark for sale in the US. I'm sure that Sol-Ark has done alot of work to get UL listed (if there were not already) and have the ability to test units and make software changes.
At this time, my take on it, is that these units were just not fully tested during product development. So now that there are a bunch of these out in service the problems are showing up. I don't have much evidence to support this train of thought - just what I read here. We'll see how things unfold. Will more people start reporting problems? If the inverters turn out to have problems, will Sol-Ark admit it?
If I was not looking at the charts I'd think all is well. And that may be the majority of owners.

 

[Clint-L]

I've learned a few things about the spikes I'm seeing on the PV inputs.
1. It appears that the voltage spikes to Voc (or higher) when the inverter stops taking in power from the panels.
2. The AC inverter output also drops at the same time.
3. This issue is not related to the power output of the inverter. It happens on cloudy days as well as full sun days. So no issue with the grid-tie wiring, etc. Normal AC voltage here runs from 120-122 per leg at night. During the day the voltage is 121-123 most of the time.
4. The SMA inverter that is here also feeds into the same inverter output panel. And it has no problems.
5. The inverter that the Sol-Ark replaced never had this problem.
6. Sol-Ark support has been in the loop with this. They asked to turn on ARC detection as it was off by default. No ARC errors.
7. Both PV arrays have been basically ruled out.
8. There have been other small issues during this time that were solved but did not seem to be part of this problem.

At the end of the day I have to point to the Sol-Ark inverter as the reason for the spikes. Let's see if Sol-Ark support has anything more to say. Still looking for problems here - other than the Sol-Ark.

From what I've read about these Sol-Ark inverters is that they are coming from a factory somewhere and being rebadged as Sol-Ark for sale in the US. I'm sure that Sol-Ark has done alot of work to get UL listed (if there were not already) and have the ability to test units and make software changes.
At this time, my take on it, is that these units were just not fully tested during product development. So now that there are a bunch of these out in service the problems are showing up. I don't have much evidence to support this train of thought - just what I read here. We'll see how things unfold. Will more people start reporting problems? If the inverters turn out to have problems, will Sol-Ark admit it?
If I was not looking at the charts I'd think all is well. And that may be the majority of owners.

I agree with you 100% and I’m seeing the exact same behavior. Solark support is in the loop with my issue as well.

I really believe in Solark, but I wish they would be more transparent on something of this magnitude. I also agree that many folks, especially ones that just had the system installed and doesn’t really understand it, may have these issues and not even know it.

 

[Clint-L]

Here’s another thread about more of the 15k issues.

Thread 'Sol-Ark 15k will only invert 13k of DC PV'
https://diysolarforum.com/threads/sol-ark-15k-will-only-invert-13k-of-dc-pv.49211/

 

[SolarRich]

Today the trackers were left flat to minimize power input to the Sol-Ark. The SunnyBoy was turned off - both AC and DC. So it was inactive. I wanted to see if having the SunnyBoy on made any difference to the Sol-Ark. The answer is 'no'.
Here's some charts from today:
DC input current:



DC Input voltage:



Still waiting for Sol-Ark to get back to me on these voltage spikes and corresponding drop in input and therefor output current. At the end of the day can I rely on this? How long will it last?

AC output current:

 

[Clint-L]

My 15k only has this spike/drop issue when it hits the 13,200w of PV and clips out. It will clip all day and voltage run high on DC during that time. When it first clips, the PV drops to zero and resets, ramps back up. It then seems to stabilize through the day and stays connected and produces with a flat clipping of about 13.2kw of PV.

 

[Devin]

Here’s another thread about more of the 15k issues.

Thread 'Sol-Ark 15k will only invert 13k of DC PV'
https://diysolarforum.com/threads/sol-ark-15k-will-only-invert-13k-of-dc-pv.49211/

From that past thread, this concern with the 15K has been brought to the attention of our QA engineers. If I am following this correctly the primary issue is the drop in PV production occurring towards the latter end of the day. And this is occurring with AC+DC coupling regardless of cloud coverage?
If you already submitted a support ticket I will tie it with the past thread and see what trends are occurring.

 

[SolarRich]

Support ticket 65898. Carlos is working on it.
Arrays and output wiring/breakers ruled out.
Now working on the assumption something is going on with the Discovery battery closed loop comms with the sol-ark. If I understand correctly, the idea is that somehow the Discovery batteries(2) are causing the Sol-Ark to cut power briefly. As of today the sol-ark is now doing open loop using battery % instead of closed loop. The Discovery Lynk communications box is disconnected. This is a test for a few days to see what happens.

There is a sunnyboy inverter that is not coupled directly to the sol-ark. It is simply an identical system (same array specs) that feeds AC power into the same 200 amp panel that the sol-ark does and then out to the grid.
The drop in power is quite quick. On the order of seconds. Certainly less than a minute. And it can occur any time of the day but I have to say that it's more likely in the latter half of the day. And at any power level. From a few hundred watts to 7kw.
It has taken a bit of time to get to this point. Testing and ruling out the various things that can cause a problem.
I need to look at the Lynk box and see if any data can be pulled from it while operating. If any data is available it would be thru the Lynk Windows app that is used to setup the Lynk for the Sol-Ark. I don't think there is any data logging ability in the app.

Can Sol-Ark dig deeper into the inverter and pull out more data than is available on PowerView? Like the details of the communications to and from the Discovery batteries? Or can I get that data using rs-485?

 

[Devin]

Support ticket 65898. Carlos is working on it.
Arrays and output wiring/breakers ruled out.
Now working on the assumption something is going on with the Discovery battery closed loop comms with the sol-ark. If I understand correctly, the idea is that somehow the Discovery batteries(2) are causing the Sol-Ark to cut power briefly. As of today the sol-ark is now doing open loop using battery % instead of closed loop. The Discovery Lynk communications box is disconnected. This is a test for a few days to see what happens.

There is a sunnyboy inverter that is not coupled directly to the sol-ark. It is simply an identical system (same array specs) that feeds AC power into the same 200 amp panel that the sol-ark does and then out to the grid.
The drop in power is quite quick. On the order of seconds. Certainly less than a minute. And it can occur any time of the day but I have to say that it's more likely in the latter half of the day. And at any power level. From a few hundred watts to 7kw.
It has taken a bit of time to get to this point. Testing and ruling out the various things that can cause a problem.
I need to look at the Lynk box and see if any data can be pulled from it while operating. If any data is available it would be thru the Lynk Windows app that is used to setup the Lynk for the Sol-Ark. I don't think there is any data logging ability in the app.

Can Sol-Ark dig deeper into the inverter and pull out more data than is available on PowerView? Like the details of the communications to and from the Discovery batteries? Or can I get that data using rs-485?

It may be worth looking into using Solar Assistant to compile better sample data as PowerView is capped to data output every 5 minutes.

 

[SolarRich]

I can pull data out of the Sol-Ark using RS-485. The question is really though - can anyone, either solarassistant, myself or a sol-ark tech pull the data between the batteries and the sol-ark inverter? And then what is the inverter doing with the data? Why is the inverter shutting down for a few seconds?
How many other Sol-Ark owners have the same discovery batteries? Is this happening on these other inverters? Sol-Ark has the ability to look at other's PowerView charts and data. Are there any conclusion that could be drawn for that data?

 

[Devin]

I can pull data out of the Sol-Ark using RS-485. The question is really though - can anyone, either solarassistant, myself or a sol-ark tech pull the data between the batteries and the sol-ark inverter? And then what is the inverter doing with the data? Why is the inverter shutting down for a few seconds?
How many other Sol-Ark owners have the same discovery batteries? Is this happening on these other inverters? Sol-Ark has the ability to look at other's PowerView charts and data. Are there any conclusion that could be drawn for that data?

Not remotely but It may be possible locally with an RS-485 to USB adapter (maybe a solar assistant can access this data as well). The Discover AES is operating over CANBUS. The inverter is listening for commands exclusively when COMMS are enabled. Battery voltage, current, and limits on discharge are all dictated by the battery. The Discover battery COMMs has existed for a few years and uses battery profile 00 shared with several other batteries.

 

[SolarRich]

I don't have the time to debug the problems on the Sol-Ark inverter to the degree that you mention. To actually determine just what is going on in the inverter would require source code, circuit diagrams and the ability to make changes to the source code and print out diagnostic information from within the code running on the inverter. The RS-485 data(the same data that's on PowerView and Solar Assistant) that can be pulled does not get to that level. It's common practice to have diagnostic output available in a product. Debug print statements get put into the code during product development and can be turned on when needed.

We have gone now 1 1/2 days w/o any problems. In a few days we'll know for sure. I'll switch it back to closed loop mode to make sure the problem can be duplicated.
How many other installations are using Discover batteries and are they having the same problem? I'm hoping Carlos can find a solution.

 

[Clint-L]

From that past thread, this concern with the 15K has been brought to the attention of our QA engineers. If I am following this correctly the primary issue is the drop in PV production occurring towards the latter end of the day. And this is occurring with AC+DC coupling regardless of cloud coverage?
If you already submitted a support ticket I will tie it with the past thread and see what trends are occurring.

 

[SolarRich]

Update...
Carlos at Sol-Ark had me setup the Discovery batteries to run open loop. The problem is gone using open loop with the batteries. Put it back into closed loop and the spikes and drop-out re-appear.