4kw system outputting 2.5kw. Why?

[Solarisium]

Issue: VMP voltage is low.
System:
Grid tied single string using Sunny Boy 3.8kw us-41 inverter. Designed and installed diy. Finishing final steps before requesting inspection.

Location: Southern California
9 Solarever panels, each at 455w. Connected in series with APSMART single rapid shutdown devices one pet panel.
Tilted 23degrees facing south
From array to inverted: less than 15 ft, 10G wires
Inverter: sma Sunny boy 3.8 US41
From inverter to service panel : less than 10ft. 10G wire.

Module VMP: 41.51VDC
Expected PV Array VMP voltage 9x41.51:~370VDC. Actual: received at the inverter: 270-290VDC
Array max output expected: 3600w, actual at full sun: 2400w
Inverters AC output ~240V. ~9.5 AC Amp

Why my output is low? (Edit in March 2024, two Apamart devices were faulty)
Is 55 to 65% efficiency normal? (Edit in March 2024, No. )
Are APSMART shutdown devices the culprit? (Edit in March 2024, indeed they were the issue)
how to troubleshoot? ( Edit in March 2024, See this post)

Edit on 26March2024: Two APSmart devices were not functioning properly, bypassing two modules, this low voltage. See this post for details:
Post in thread '4kw system outputting 2.5kw. Why?'
https://diysolarforum.com/threads/4kw-system-outputting-2-5kw-why.62392/post-1048446)

Thank you in advance.

I've read @Hedges is the SME at SMA products.
Everyone, Please let me know what you think.


Details including pictures ere in a poat in thread 'Drafted my design- aiming a 4.5KW grid tie system'
here in a poat in thread 'Drafted my design- aiming a 4.5KW grid tie system' hre in a poat in thread 'Drafted my design- aiming a 4.5KW grid tie system'
https://diysolarforum.com/threads/drafted-my-design-aiming-a-4-5kw-grid-tie-system.53251/post-777357



.

 

[42OhmsPA]

How hard would it be to take the front left panel out of the string? I don't think the shadow /shading from the overhead wires is helping anything, removing the panel would easily rule it out.
No expert on grid tied or sma, wait for others to chime in.

 

[Solarisium]

Certainly doable.
IT is hard to believe that much shade is causing nearly 50% loss though.

Going to see it if I can try it now. Edit in March 2024: that shading from the overhead wires caused ~10 to 30w loss. Negligible.

 

[LydMekk]

You ARE aware that a solar installation only delivers what your LOADS are right?
If it is specced to a certain wattage it does NOT deliver that at all times.
Also, clouds/unclear weather...

 

[Solarisium]

You ARE aware that a solar installation only delivers what your LOADS are right?
If it is specced to a certain wattage it does NOT deliver that at all times.
Also, clouds/unclear weather...

I am not aware! So right now my loads are almost less than 200watts. With the amount of sunlight at 5:27pm in Southern California at this time, it is delivering roughly 1kw. Is that what you mean?

Also the sky is very clear, and the panels are not covered by any shades.

Certainly I don’t expect the system to deliver the exact power that it is specced to. But at 2 PM when the sun is at full force with no shade or cloud, I did not expect 55% efficiency.

So maybe a 55 to 65% efficiency is normal for a string inverter.

 

[400bird]

You ARE aware that a solar installation only delivers what your LOADS are right?
If it is specced to a certain wattage it does NOT deliver that at all times.
Also, clouds/unclear weather...

With a grid tied PV inverter it just puts out every watt it can when the grid is present. Not the same as an all in one, off grid, or battery charge controller system.

To the OP, solar production varies over the year and depends on ambient temp and sun to panel angle.
Put your house and PV info in here and see what it shows you.

PVWatts

Estimates the energy production and cost of energy of grid-connected photovoltaic (PV) energy systems throughout the world. It allows homeowners, small building owners, installers and manufacturers to easily develop estimates of the performance of potential PV installations

pvwatts.nrel.gov

I have 7960 watts of panels and just 2 weeks ago we had a hot spell, only hit about 5.5kw at max. It was a perfectly sunny day. Heat (sun with no wind to drive away the heat) brings down efficiency.

You'll never see rated panel output, outside of an occasional blip, panels normally perform at 80% at best. Less if the angle to the sun is less than perpendicular.

 

[Solarisium]

With a grid tied PV inverter it just puts out every watt it can when the grid is present. Not the same as an all in one, off grid, or battery charge controller system.

To the OP, solar production varies over the year and depends on ambient temp and sun to panel angle.
Put your house and PV info in here and see what it shows you.

PVWatts

Estimates the energy production and cost of energy of grid-connected photovoltaic (PV) energy systems throughout the world. It allows homeowners, small building owners, installers and manufacturers to easily develop estimates of the performance of potential PV installations

pvwatts.nrel.gov

I have 7960 watts of panels and just 2 weeks ago we had a hot spell, only hit about 5.5kw at max. It was a perfectly sunny day. Heat (sun with no wind to drive away the heat) brings down efficiency.

You'll never see rated panel output, outside of an occasional blip, panels normally perform at 80% at best. Less if the angle to the sun is less than perpendicular.

Your actual 5.5kw out of nominal 7.96kw is almost 70%! Very nice!
My actual 2.5kw out of nominal 4kw is only 62.5%. I think I’ll move forward with this to get the PTO and will troubleshoot later.




Per the NREL calculator, my 4kw system at 15% loss should be able to generate 658kw in May, almost 21kw per day. Surprisingly the SMA portal expects the same amount. I might be calculating inaccurately!

I am close to the beach so the breeze keeps the panels at a good temperature. Tested multiple times during the day, the panels surface doesn’t go over 90F on a hot day. Also tilted at 23degrees facing south with no shade.

The actual power generated was 2.5kw at 2pm. In order to generate 21 kw per day, I would need 8.5 sun hours.

I have not tried a full day. We’ll try it tomorrow. To find out how much it delivers.



I was expecting 80% of the nominal power (i.e. ~3.2kw) on a cool but fully sunny day, but I guess 60% is normal (2.5kw)! I wont take it ! Will troubleshoot later!
I have a feeling that my APSmart rapid shutdown devices are lowering the voltage. I should have test driven the entire system without APsmart devices while the panels we’re still on the floor! It’s much harder now to do any troubleshooting on the rapid shutdowns!

 

[Hedges]

Grid tied single string using Sunny Boy 3.8kw inverter. Designed and installed diy. Finishing final steps before requesting inspection.

Location: Southern California
9 panels, each at 455w. Connected in series with APSMART single rapid shutdown devices one pet panel.

Disconnecting those would be one test, see if they are causing problems.

Tilted 23degrees facing south

"Southern California" - San Diego latitude 33 degrees.
Earth's tilt is 23.5 degrees
I'm too lazy to figure out how close to optimum angle your panels are for the sun today, but I think it is near maximum.

From array to inverted : less than 15 ft, 10G wires

Too little to matter.

Inverter : sma Sunny boy 3.8 US41

I haven't used this model yet, but will be buying them for my latest projects.
At 290V and part power, should be around 97% efficient

From inverter to service panel : less than 10ft. 10G wire.

Too little to matter.

Arrays Nominal voltage 9x49:~440V.
Actual received : 250 to 290 v.

By "Nominal" do you mean Vmp? or Voc?
What model panel? Link to data sheet?

290V/440V = 66%. Other panels Vmp is around 84% of Voc

With less light I expect less current, but not significantly lower voltage.

If some panels are shaded, MPPT has two choices: Take the lower current, for all panels. Or draw more current which pulls voltage lower, and activates bypass diodes.
In other words, the loss from partial shade on a panel isn't the portion of a panel which is shaded. It is either that entire panel (at least the diode bypassed portion.) Or, that portion of all panels.
Pulling full current and bypassing panels would explain reduced voltage.
(so would bad connection getting hot)

Arrays nominal watts: 4090w. Actual shown on the inverter display :~2400W

59% of STC. That is quite low, unless not full sun or excessively hot.

2400W/290V = 8.3A
But your meter when picture taken said 9.48A, 274.4V for 2600W
What are PV panel specs?

Inverters output ~240V. ~9.5 AC Amp

Why my output is low?
Is 55 to 65% efficiency normal?

No, not in full sun. Could be with obscured sun or off angle.

Are APSMART shutdown devices are the culprit?
how to troubleshoot?

Possibly. I haven't heard of RSD causing power loss. Have heard of optimizers not working well.
If some RSD don't turn on, those panels output would be lost.

To troubleshoot, take them out of the circuit and just connect panels themselves in series.

Thank you in advance.

I've read @Hedges is the SME at SMA products.

Happy to help out, or at least bang my head against it.


I think your problem is the low voltage. Don't know if it is that slight shade on a panel, something defective, or ??

 

[400bird]

For a 4 kW system in LA at 23° tilt, facing south,PV watts estimated 2477 watts on May 5th at 2 pm.

With perfect sun angle it comes to 3131 watts.

 

[400bird]

I hadn't noticed you PV voltage as Hedges pointed out. That's a good place to start.
But, based on this post I thought that was corrected?

Drafted my design- aiming a 4.5KW grid tie system

Hi Solar folks! I just started my journey! Yep it is a journey for a true DIYer. Here is design (still a draft): I have a clear idea of what I am going to install: a 4.5 KW SMA grid tie system. no power optimizer, no micro inverter, no battery or storage devices. Plain and simple. I will...

diysolarforum.com

 

[Hedges]

4 kW STC? Or PTC or NOCT?
I wouldn't expect PV Watts to deal with that part, which varies between cells.

 

[Hedges]

I hadn't noticed you PV voltage as Hedges pointed out. That's a good place to start.
But, based on this post I thought that was corrected?

Drafted my design- aiming a 4.5KW grid tie system

Hi Solar folks! I just started my journey! Yep it is a journey for a true DIYer. Here is design (still a draft): I have a clear idea of what I am going to install: a 4.5 KW SMA grid tie system. no power optimizer, no micro inverter, no battery or storage devices. Plain and simple. I will...

diysolarforum.com

"For some reasons the inverter capped the DC voltage at 290v when I incorrectly connected 120v AC to the inverter."

I don't get that. If 120Vrms connected with 240V expected, inverter simply wouldn't connect to grid.

 

[Solarisium]

Disconnecting those would be one test, see if they are causing problems.

Will try that during the weekend.

I'm too lazy to figure out how close to optimum angle your panels are for the sun today, but I think it is near maximum.

For Los Angeles,it's near maximum.

By "Nominal" do you mean Vmp? or Voc?
What model panel? Link to data sheet?

Voc. Solarever 455W
Here is link to datasheet: https://149795217.v2.pressablecdn.com/wp-content/uploads/Solarever-455-datasheet.pdf

290V/440V = 66%. Other panels Vmp is around 84% of Voc
With less light I expect less current, but not significantly lower voltage.

I observed low voltage. The voltage shown on the inverter was not a consistent value, rather it was changing from 255v to 291v at 2pm full sun. This was the puzzling part.

If some panels are shaded, MPPT has two choices: Take the lower current, for all panels. Or draw more current which pulls voltage lower, and activates bypass diodes.
In other words, the loss from partial shade on a panel isn't the portion of a panel which is shaded. It is either that entire panel (at least the diode bypassed portion.) Or, that portion of all panels.
Pulling full current and bypassing panels would explain reduced voltage.
(so would bad connection getting hot)

Thanks for explaining this. The only shaded area was Tha shadow of the power cable shown in this picture:

59% of STC. That is quite low, unless not full sun or excessively hot.

It was full sun. Not cloud whatsoever. Panels get cool down by the ocean breeze. Max Panel temp was at 37C (100F) yesterday when I was checking.

2400W/290V = 8.3A
But your meter when picture taken said 9.48A, 274.4V for 2600W
What are PV panel specs?

DC Voltage kept changing. 290v was a max DC voltage I observed and 274.4v was the voltage at that moment what I took that picture. Is it normal for the DC vintage to keep changing?

No, not in full sun. Could be with obscured sun or off angle.
Possibly. I haven't heard of RSD causing power loss. Have heard of optimizers not working well.
If some RSD don't turn on, those panels output would be lost.
To troubleshoot, take them out of the circuit and just connect panels themselves in series.

I checked APSMART rapid shutdown device specs. The range is very wide from 0 to 80 volts. Will try taking them out and connect panels without them.

Happy to help out, or at least bang my head against it.

I think your problem is the low voltage. Don't know if it is that slight shade on a panel, something defective, or ??

Yes, not only low voltage, but variable voltage. A loose connection, a defective panel, a defective rapid shutdown device or it could what you said, the MPPT draws more current with lower voltage (a diode would be bypassed). Yesterday tried using a laser temp gun to identify a hot spot but wasn't able to shoot at all cables underneath.
Looking at the inverter meter right now (7:15am) it reads:
P: 315, 316 310, 307 watts (keeps changing which is expected)
AC: 1Amp, 243V.
DC: 231 to 290 V. It keeps changing.


Thank you for your insights.

 

[Solarisium]

"For some reasons the inverter capped the DC voltage at 290v when I incorrectly connected 120v AC to the inverter."

I don't get that. If 120Vrms connected with 240V expected, inverter simply wouldn't connect to grid.

Initially I incorrectly connected 120VAC instead of 240VAC to the inverter. It did connect to the grid and worked. Initially I thought the reason for low voltage /low power was this. Later i confirmed that even at 240AAC, the DC voltage is still at its maximum at ~290VDC.

 

[Solarisium]

For a 4 kW system in LA at 23° tilt, facing south,PV watts estimated 2477 watts on May 5th at 2 pm.

With perfect sun angle it comes to 3131 watts.

That's almost exactly what I observed. Thanks for confirming. This second data point helped. Is this a calculated value or an actual observation from your system?

 

[400bird]

Initially I incorrectly connected 120VAC instead of 240VAC to the inverter. It did connect to the grid and worked. Initially I thought the reason for low voltage /low power was this. Later i confirmed that even at 240AAC, the DC voltage is still at its maximum at ~290VDC.

If you still have about 290 volts DC max, that's something to look into. As you stated, it should be higher than that.

That's almost exactly what I observed. Thanks for confirming. This second data point helped. Is this a calculated value or an actual observation from your system?

That was all from PV watts. You've just got to download the data. For the perfect sun angle info, I just checked the dual axis tracker option.

 

[Hedges]

Will try that during the weekend.

For Los Angeles,it's near maximum.

Voc. Solarever 455W
Here is link to datasheet: https://solareverusa.com/wp-content/uploads/2021/06/Module-HC-72M-440-450w.pdf

Data sheet gives 440W, 445W, 450W. Not the 455W you mention.

Assuming 450W, Vmp = 41.39
41.39 Vmp x 9 panels = 372.5 Vmp, making 255V to 291V rather low.

Don't use Voc except for max inverter would see (after adjusting for cold) and if inverter is curtailing for some reason.
https://solareverusa.com/wp-content/uploads/2021/06/Module-HC-72M-440-450w.pdf

View attachment 148274

I observed low voltage. The voltage shown on the inverter was not a consistent value, rather it was changing from 255v to 291v at 2pm full sun. This was the puzzling part.

Sunny Boy will dither voltage slightly to stay at peak power, and periodically do a full sweep (which is how it gets off the local maxima of full voltage but reduced current of a partly shaded panel or half-cut panel, finds higher wattage peak at lower voltage.)
I wouldn't expect such a wide voltage from dithering.

If you have a clamp DC ammeter you could get more real-time reading. AC on output might give same indication (although possibly algorithm just uses capacitor charging ramp to do a rising sweep in which case would only be seen on DC, after pulling voltage down at a different rate using AC. I haven't observed it.)

You can also monitor DC voltage with meter, but of course be very careful not to touch.

Wonder if erratic behavior is due to malfunctioning RSD box or bad electrical contact.

Thanks for explaining this. The only shaded area was Tha shadow of the power cable shown in this picture:

It was full sun. Not cloud whatsoever. Panels get cool down by the ocean breeze. Max Panel temp was at 37C (100F) yesterday when I was checking.

That shadow wiped out most output from several cells. I don't know how they are distributed across the 3 (?) diode-bypassed sections. If wind was moving shadow across cells that would be more erratic.

DC Voltage kept changing. 290v was a max DC voltage I observed and 274.4v was the voltage at that moment what I took that picture. Is it normal for the DC vintage to keep changing?

Only a small amount.

I checked APSMART rapid shutdown device specs. The range is very wide from 0 to 80 volts. Will try taking them out and connect panels without them.

Initially I incorrectly connected 120VAC instead of 240VAC to the inverter. It did connect to the grid and worked. Initially I thought the reason for low voltage /low power was this. Later i confirmed that even at 240AAC, the DC voltage is still at its maximum at ~290VDC.

The only way Sunny Boy can produce low PV voltage is by drawing current.

I'm surprised it worked at reduced voltage. They accept 183V to 264V depending on whether configured for 208V or 240V, which may be automatic depending on whether L1-N and L2-N are 120 or 180 degrees apart. Some models recognize 277V if L1-N is zero volts. Maybe this one supports 120V that way but not documented? (If so, would be a great thing for systems with single Sunny Island.)

How long was it operated that way? After 10 hours of feed-in, certain parameters are locked and require password to change. Off chance this happened and limited watts, but would not necessarily expect erratic behavior (I could continue theorizing why.) How exactly did you connect all the wires? (L1, L2, N, G and Sunny Boy's terminals)

 

[Solarisium]

Data sheet gives 440W, 445W, 450W. Not the 455W you mention.

Assuming 450W, Vmp = 41.39
41.39 Vmp x 9 panels = 372.5 Vmp, making 255V to 291V rather low.

Apparently I sent an incorrect link. My panels are 455W : https://149795217.v2.pressablecdn.com/wp-content/uploads/Solarever-455-datasheet.pdf
Updated the original post as.

That calculation for the panel at 455W would yield: Vmp x 9= 41.51. X 9= 373.6 v.
Connections:
Inverter terminals:

Main service panel:

I commissioned and configured the inverter myself. Chances are that I may have incorrectly configured something., I will look into the configuration later and will send some screenshots. But I can confirm that the system run time at 120 V AC was maybe a couple of hours but certainly not 10 hours.
The systems entire runtime might have just been around 10 hours with majority of it running at 240 VAc.

My ground starts from the array racking and is connected to the inverter’s housing, and then from there to the Earth ground rod.
As shown in the picture, for the AC terminals L1 is connected to one pole of a 20 amp breaker and L2 is connected to the second pole.
N is connected to the neutral bus bar in the main service panel. That neutral busbar is also connected to the ground earth rod.

We’ll send Screenshots from the inverter configuration a bit later.
Thank you so much for taking time

 

[Solarisium]

If you still have about 290 volts DC max, that's something to look into. As you stated, it should be higher than that.


That was all from PV watts. You've just got to download the data. For the perfect sun angle info, I just checked the dual axis tracker option.

View attachment 148285

Yes, I think I have some homework to do to find out why the voltage is lower than expected.
I want to blame it on that small shadow of the cable running over a corner of one of the panels, but then I remember, when that shade is not on the surface of the panel the voltage is still the same.
So possibly, this is a behavior from the inverter I am not a subject matter expert at this, so I don’t know if my conclusion is accurate

 

[Solarisium]

Inverter parameters: