Inverter Detaching from Grid - Wide voltage differential between L1 and L2

[DIYrich]

I have a new Sol-Ark 15k, and it seemed to randomly go off grid for a few minutes. I finally figured out it was because of L1 and L2 voltage changes.

Normally, L1 and L2 track within 0.5v of each other, varying between 120.0-121.0v on average. When the inverter detaches from the grid, the spread between L1 and L2 abruptly widens out to 3v, but the total still seems to be around 240v (L1 could be 119.0 and L2 would be 122.5v).

Do other people see similar situations? I'm tinking it is ok to stay off grid becuase the grid may be having issues when the divergence happens. There may also be something funky about the power that it sees that is not evident in the voltage log. Note: Frequency stays rock solid.

 

[Hedges]

You may have a poor connection in neutral.
I wouldn't expect that slight voltage variation itself to be a problem or cause disconnect. It may instantaneously jump much farther.

If you were to apply a large 120V load (e.g. a few space heaters), then you could probe with DMM, look for voltage drop across contacts. After a long time, it could also show up as heat.

 

[DIYrich]

You may have a poor connection in neutral.
I wouldn't expect that slight voltage variation itself to be a problem or cause disconnect. It may instantaneously jump much farther.

If you were to apply a large 120V load (e.g. a few space heaters), then you could probe with DMM, look for voltage drop across contacts. After a long time, it could also show up as heat.

Where do I look. The grid side L1-L2 differential continues even after Sol-Ark disconnects from the Grid. Whatever is happening is either:
1) Grid side; or
2) Inside the Sol-Ark where is measures Grid Voltage.

I could expose the Grid Lines and take a reading the next time it happens. See if it is real, or Sol-Ark problem.

 

[Hedges]

Do you have loads which don't pass through SolArk? If so, you could try loading heavily, see if it changes a lot, try to track down where.

The problem could come from grid if neutral has a problem there. Ideally have continuous monitoring which logs highest/lowest voltages seen.

Does SolArk log/report what it saw which was out of spec?

 

[Quattrohead]

My first thought is bad neutral from grid.

 

[DIYrich]

Sol-Ark was reporting Grid Voltages of 120 and 124 for L1 and L2 (on its screen). I measured grid voltages on the Sol-Ark Grid Terminals and saw 122 and 122. So there was a problem with the Sol-Ark's Grid Circuits.

I called my installer, and he was stumped (he was able to pull the inverter logs and status). He said to call Sol-Ark. It was after hours on Friday, so I couldn't talk to anyone from support. Since it was needlessly draining the batteries, tried to turn off the sol-ark, but it would just turn off, and stayed disconnected from the grid. I powered back up, and during a short period where it reconnected to the Grid, I flipped the sol-ark battery breakers to turn off the battery. At that point, it stayed connected to the Grid and objected (had no where else to get power from). A few hours later, the problem resolved itself, warning message went away, and I reconnected the batteries to charge them back up.

I talked to Sol-Ark Monday morning. They said I had one of the early 15k's with the original firmware. He said there was a known grid problem in the software, and said an update would take care of it. I asked for the details (what conditions caused the software glitch), but he didn't know. We shall see.

 

[dreancestere]

Old thread, but has this problem come back for you? I've got a year old Sol-Ark 15k with fully up-to-date firmware with the same issue. There doesn't seem to be any rhyme or reason for it (no big loads). Periodically, the inverter will see the voltage difference between L1 and L2 at the grid go from something like 0.5v to 3v and disconnect from the grid for a minute or so. Kind of hard to troubleshoot with no error messages or way to reproduce the issue.

 

[DIYrich]

I have not yet setup logging, so I can't say if the problem has persisted. It does randomly go off grid for a few minutes. But most of the time it is during a weather event, so it might be working as expected. I haven't had the prolonged problem (off grid for more than a few minutes when the grid was up), so that problem has gone away.

 

[dreancestere]

Sol-Ark's data logging finally captured an instance of this 1 minute grid disconnect. Since I finally have some kind of data, I sent in a report to Sol-Ark.

 

[dreancestere]

Today, Sol-Ark emailed me and told me they had changed my Grid parameters to, hopefully, fix this issue. The problem is that I'm selling energy back to my utility and I was under the impression that I HAD to be using UL1741SB as my Grid Mode to do that AND changing parameters only works under General mode. To accommodate that, he changed me to Grid mode.

Am I allowed to sell back under General mode instead of UL1741SB? If so, what are the repercussions of doing that? I've written them back to ask, but it might be a while before they get back to me.

EDIT: It's full of stuff I don't understand, but this web page:

Smart PV inverter overview: IEEE 1547-2018 and UL 1741 explained

What is IEEE 1547-2018? IEEE 1547 refers to “Standard 1547” as developed by the Institute of Electrical and Electronics Engineers (IEEE) to safely and functionally integrate distributed energy resources (DERs) into the electric distribution grid. The Energy Policy Act of 2005 established IEEE...

solarbuildermag.com

says the differences between (what I believe is) "General" mode and later modes (SA and SB) are:

"UL 1741-SA was published in conjunction with California Rule 21 Phase 1 requirements. These safety tests certify the “smart inverter” grid support functionality needed to modernize the grid through widespread DER integration. The testing requirements for UL 1741-SA are as follows:

• Anti-islanding
• Low/high-voltage ride-through (L/HVRT)
• Low/high-frequency ride-through (L/HFRT)
• Specified power factor (SPF)
• Volt/VAR mode
• Volt/Watt mode
• Frequency/Watt mode
• Ramp rate

UL 1741-SB introduced an interoperability conformance test in accordance with IEEE 1547.1-2020. Conformance can be achieved through either DNP3, IEEE 2030.5, or SunSpec Modbus communications protocols, which are used to store or send information and to control adjustable inverter functions. The testing requirements for UL 1741-SB are more stringent and include the following in addition to the requirements defined above for UL 1741-SA certification:

• Watt/VAR mode
• Voltage magnitude and time trip
• Frequency magnitude and time trip
• EMI
• Surge
• Rate of change of frequency (ROCOF)
• Dynamic voltage support
• Enter service
• Synchronization
• Open phase
• Harmonics
• DC injection
• Ground fault overvoltage (GFOV)
• Load rejection overvoltage (LROV)
• Prioritization of DER responses
• Fault current
• Persistence of DER parameter setting"

But, I don't know what that means in Joe Bag o'Donuts English.

 

[dreancestere]

I'll try to fit Sol-ark's changes to my Grid Settings in the following posts (once again, I don't know what they mean, if they're correct or if they'll solve the 1 minute disconnect issue). I had to break this up to fit into what's allowed here.

Besides the Grid Mode, there's no change on the Grid Selection tab.

Setting	UL1741SB	General
Grid Selection tab
Grid Mode	UL1741SB	General Standard
Grid Frequency	60Hz	60Hz
Single Phase Check Box	Unchecked	Unchecked
120/240V Split Phase Check Box	Checked	Checked
120/208V 3 Phase Check Box	Unchecked	Unchecked
Grid Reconnect Time	60s	60s
Power Factor	1	1
Fixed Q	0.00%	0.00%
Q_Response	10s	10s
Output V	120/240V	120/240V
Output V+	+0V	+0V

 

[dreancestere]

Here's the Connect tab. To me, it looks like they've narrowed the normal connection voltage range and increased the reconnection range. The frequency ranges are both increased:

Setting	UL1741SB	General
Connect tab
Reconnect Grid Vol High	263.8V	288.0V
Reconnect Grid Vol Low	211.4V	120.0V
Reconnect Grid Hz High	61.5Hz	65.0Hz
Reconnect Grid Hz Low	58.5Hz	55.0Hz
Reconnect Ramp Rate	60s	60s
Normal Connect Grid Vol High	288.0V	265.0V
Normal Connect Grid Vol Low	120.0V	185.0V
Normal Connect Grid Hz High	62.0Hz	65.0Hz
Normal Connect Grid Hz Low	57.0Hz	55.0Hz
Normal Ramp Rate	60s	60s

 

[dreancestere]

Here's the IP tab. This one looks weird. Lowered the Over Voltage amount, mostly lowered the HV (high voltage?) numbers, mostly increased the LV (low voltage?) numbers, and increased the acceptable frequency ranges. But, the really odd thing is the change to the number of seconds in those numbers. My old UL1741SB settings show 5 of those times in many multiple seconds. The new General ones are all sub-seconds. HF1 and LF2, for example, went from 299 seconds (basically, 5 minutes) to 0.29s (1000 times less):

Setting	UL1741SB	General
IP tab
Over Voltage U>(10 min. running mean)	276.0V	260.0V
HV3	288.0V	265.0V
HV2	288.0V – 0.16s	265.0V – 0.10s
HV1	264.0V – 13.00s	265.0V – 0.13s
LV1	211.2V – 21.00s	185.0V – 0.21s
LV2	168.0V – 2.00s	185.0V – 0.20s
LV3	120V	185.0V
HF3	62.00Hz	65.00Hz
HF2	62.00Hz – 0.16s	62.00Hz – 0.10s
HF1	61.50Hz – 299.00s	62.00Hz – 0.29s
LF1	58.50Hz – 299.00s	57.00Hz – 0.29s
LF2	57.00Hz – 0.16s	57.00Hz – 0.10s
LF3	57.00Hz	55.00Hz

 

[dreancestere]

Here's the F(W) tab. First, the F(W) Check Box is now UNchecked instead of checked. Does that mean these settings aren't being used? But, the numbers show an increase in the allowed frequency ranges:

Setting	UL1741SB	General
F(W) tab
F(W) Check Box	Checked	Unchecked
Over Frequency Droop F>	40% PE/Hz	40% PE/Hz
Over Frequency Start Freq F	60.50Hz	65.0Hz
Over Frequency Start Delay	0.00s	0.00s
Over Frequency Stop Freq F	60.50Hz	65.0Hz
Over Frequency Stop Delay	0.00s	0.00s
Under Frequency Droop F>	40% PE/Hz	40% PE/Hz
Under Frequency Start Freq F>	59.50Hz	55.00Hz
Under Frequency Start Delay F>	0.00s	0.00s
Under Frequency Stop Freq F>	59.50Hz	55.0Hz
Under Frequency Stop Delay F>	0.00s	0.00s

 

[dreancestere]

Here's the V(W) / V(Q) tab. In both cases, the V(W) and V(Q) Check Boxes are UNchecked. Again, does this mean the settings don't apply? Unfortuantely, I don't understand what these numbers are saying at all:

Setting	UL1741SB	General
V(W) / V(Q) tab
V(W) Check Box	Unchecked	Unchecked
Response_T	5s	10s
V1 / P1	110.0% / 100%	109.0% / 100%
V2 / P2	111.0% / 50%	110.0% / 20%
V3 / P3	112.0% / 0%	111.0% / 20%
V4 / P4	112.0% / 0%	111.0% / 20%
V(Q) Check Box	Unchecked	Unchecked
L.In / L.Out	20.0% / 5.0%	20.0% / 5.0%
V1 / Q1	90.0% / 43%	90.0% / 44%
V2 / Q2	94.0% / 0%	95.7% / 0%
V3 / Q3	106.0% / 0%	104.3% / 0%
V4 / Q4	110.0% / -43%	112.2% / -60%

 

[dreancestere]

And, here's the P(Q) / P(F) tab. No change. But, the P(Q) and P(F) Check Boxes are all UNchecked. Not applied?:

Setting	UL1741SB	General
P(Q) / P(F) tab
P(Q) Check Box	Unchecked	Unchecked
P1 / Q1	20% / 20%	20% / 20%
P2 / Q2	100% / 20%	100% / 20%
P3 / Q3	100% / 20%	100% / 20%
P4 / Q4	100% / 20%	100% / 20%
P(F) Check Box	Unchecked	Unchecked
L.In / L.Out	50.0% / 100.0%	50.0% / 100.0%
P1 / F1	50% / 1.000	50% / 1.000
P2 / F2	100% / 0.800	100% / 0.800
P3 / F3	100% / 0.800	100% / 0.800
P4 / F4	100% / 0.800	100% / 0.800

 

[DIYrich]

Today, Sol-Ark emailed me and told me they had changed my Grid parameters to, hopefully, fix this issue. The problem is that I'm selling energy back to my utility and I was under the impression that I HAD to be using UL1741SB as my Grid Mode to do that AND changing parameters only works under General mode. To accommodate that, he changed me to Grid mode.

Am I allowed to sell back under General mode instead of UL1741SB? If so, what are the repercussions of doing that? I've written them back to ask, but it might be a while before they get back to me.

I would phrase it a little differently. If you sell back, then they can make you comply with UL1741SB. More a quid-pro-quo than selling under UL1741SB. What UL1741SB does is requires your inverter to stay connected to the grid during fairly wide grid problems. That is to avoid destabilizing the grid when the Power Companies are trying to balance generation and consumption. For example, if you are exporting to the grid, and the grid voltage drops to 185v, you have to stay connected for whatever time frame is required. They don't want your generation to go offline when Power Generators can produce (or can't wheel) enough power. Similarly, if the grid is over voltage, you have to stay connected to soak up the excess production.

I haven't read UL1741SB, but it seems odd that whether you have to stay connected is not tied to whether you are net consuming or net producing ok to disconnect when undervoltage if you are net consuming, or ok to disconnect when over voltage and you are net producing.

 

[dreancestere]

I guess it's become a moot point. At 10:30 am, I had the same 1 minute disconnect that I had with my UL1741SB settings. So, it doesn't look like these changes have helped.

In their email to me, Sol-ark said it might take 24 hours for the changes to take effect (it's been 16 hours). But, I can't see why that would be. Unfortunately, that mini-outage didn't get captured by the inverter's periodic data update. So, I can't send it to Sol-ark. But, I'm pretty sure I'm just going to flip the Grid Mode back from General to UL1741SB. I assume any parameter changes will revert since they can only be changed under General mode.

EDIT: Yep. When I switched back from General to UL1741SB, it reset all the Grid parameters back to default. So, I'm good. I guess I'll just live with the periodic flicker of lights and computer UPS ramping up when these 1 minute disconnects happen. Of course, if we didn't have batteries connected to the solar system, that would be a different matter.

 

[dreancestere]

It's odd that I always get these disconnects when the inverter shows a 3v difference between the grid legs. I'm wondering if the inverter is seeing a much bigger difference (enough to actually exceed the grid settings), disconnecting because of that, but not showing that actual difference on the LCD or on MySolark. But, then, it also never shows an alert that something out of spec happened and it recovered.

 

[PulseSolar]

Have you looked at the grid frequency? I have seen similar things in these boxes when the grid freq will suddenly show a drop to zero for just a split second. You have to catch it in the logs.

I have also seen what you are explaining when there is a UPS in use in the house. Sol-Ark has also mentioned these can cause some kind of feedback as they put it that can cause a voltage spike like you describe.

The adjustments that Sol-Ark (tried) was to simply broaden the allowable low and high grid frequency before a disconnect. I personally feel this is something they know they have a problem with but have yet to find a real solution since they are so quick to make this change anytime someone calls in.

 

[Oldphile]

I guess it's become a moot point. At 10:30 am, I had the same 1 minute disconnect that I had with my UL1741SB settings. So, it doesn't look like these changes have helped.

In their email to me, Sol-ark said it might take 24 hours for the changes to take effect (it's been 16 hours). But, I can't see why that would be. Unfortunately, that mini-outage didn't get captured by the inverter's periodic data update. So, I can't send it to Sol-ark. But, I'm pretty sure I'm just going to flip the Grid Mode back from General to UL1741SB. I assume any parameter changes will revert since they can only be changed under General mode.

EDIT: Yep. When I switched back from General to UL1741SB, it reset all the Grid parameters back to default. So, I'm good. I guess I'll just live with the periodic flicker of lights and computer UPS ramping up when these 1 minute disconnects happen. Of course, if we didn't have batteries connected to the solar system, that would be a different matter.

My default grid mode is SRD-UL1741 and I'm able to make changes to grid parameters.

 

[dreancestere]

Usually, the grid frequency remains at its normal, solid 60hz. Occasionally, I do see rare instances where the frequency drops to 0. In those cases, I assume the grid actually has hiccupped and gone down momentarily. Of course, part of the problem in figuring out what's going on is that the outages are only 1 minute long. So, they tend not to get captured in the logs where data capture happens only every 5 or 10 minutes. Besides those rare frequency drops to 0, I've never seen any swings in the other electrical characteristics wide enough to cause a disconnect. I sure wish the inverter would record an alert when these happen.

About feedback from the computer's UPSes, I'd actually considered that. But, I just couldn't see a way anything those tiny UPSes would do could be fed back into the house circuits at a level high enough to mess with the inverter's perception of the grid's voltage. It should be monitoring the incoming house circuit levels. But, there shouldn't be anything going back out to it.

@Oldphile I think that's why Sol-ark changed my grid settings from UL1741SB to General mode.

 

[DIYrich]

It could be a glitch at the transformer. Voltage and frequency would drop to zero for a moment.