Very low solar generation - 26k System - 3 Inverter - 3 PW - Less power on sunny days

[aimetti]

Let me preface this by saying my apologies up front , this is going to be a semi long post with as much detail as I can provide. Some of you may have seen my prior post regarding the oddities in my system reporting (which seemed to be the CTs misconfigured by my installer , which was fixed by tesla)

Found here : New 26kw system - some weird usage results after a month of use.

Essentially the neurino / cts were doubling the actual solar output and was causing the system to be doubling the actual net grid usage at the same time. Fast forward to present day and now that I have what looks to be accurate readings I am noticing what I believe is very very low solar generation.

My system is a 26kw , 3 7.3kw inverter , 3 powerwall. With the incorrect readings it was showing solar as high as 33kw in the summer. With the change of seasons and angle of sun and such I expected to see a pretty solid decrease but not to the extent of what im seeing currently. Last Saturday / Sunday I peaked at 5.3 KW generating. Each day was about 21KW total. Now this is where it gets weird. Yesterday was a completely cloudy day , drizzle , and some rain. I actually peaked at 10.1KW and generated 29KW. On a far worse day of sun. I cant wrap my head around why this is the case. I have read about strings in parallel / shade decreasing output etc, but I am not smart enough when it comes to electrical work to understand how I am setup.

I am going to attach some detail on the system here. Showing the sitemap with compass , roof setup , and the string detail. Followed by my inverter snapshots for both the cloudy and sunny days at around 1pm est. And then finally the wiring of each inverter. Perhaps someone with the know how / intelligence with the system can give me any idea why I am not seeing the output I thought I would get. And maybe it is just shade or something from trees but the cloudy/rainy day with more output is throwing me off. If someone could give me any indication to what could be the problem I can then make an educated support request to tesla. I appreciate the help!



House Layout

String / System Detail

Inverter 1 Cloudy Day

Inverter 1 Sunny Day

Inverter 2 Cloudy Day

Inverter 2 Sunny day

Inverter 3 Cloudy Day

Inverter 3 Sunny Day

 

[aimetti]

Inverter 1 Wiring

Inverter 2 Wiring

Inverter 3 Wiring

 

[Redhill_qik]

@aimetti Solar production is highly dependent on your location and the azimuth and tilt of the panels. The azimuth and tilt are listed in your site drawing, but you haven't provided any information on your location. You need to provide a rough area, within 50 miles, of where you are located to be to provide any analysis. Tesla should have provided you with an annual estimate for production in kWh, what is that number?

Edit: For a non-cloudy day after your CT fixes, please upload an app screenshot of a full day's solar production.

 

[aimetti]

for alittle bit of privacy how about Norwich, CT 06360
They also estimated on the low end for 11000 kwh , prior to me knocking down a bunch of trees. So not sure how accurate that would be now.

I messed with pvwatts alittle and tried to just do my back roof with the 47 azimuth / 34 pitch and roughly 8kw and that shows as 7200kwh alone. I could narrow it down and do each roof section and add it all together. I have to figure out how to calculate the kw per roof. Id imagine i could find that out based on the amount of panels on each section?


Before I had tesla come to house my usage was not accurate but i was fully off grid in june / july before pto. So in the summer I definitely was producing a good amount considering my house was using 2500ish kwh a month. Seems like its been pretty bad ever since they started tinkering w/ it or maybe just time of year?

 

[jgleigh]

You're generating less power when it's sunny than when it's cloudy. Something is really off.

To start with, the plans don't match the actual wiring. The plans show a total of nine MPPTs being used, but in reality you're using all twelve. Inverter 2 has MPPT 1/2 in parallel and inverter 3 has MPPT 3/4 in parallel. The pictures of the wiring and the inverter reading confirm those parallel channels. But hard to make sense of much else. The plans normally would tell you how much power to expect from each string but I'm not sure you can trust them until they match the actual wiring on your house.

 

[aimetti]

You're generating less power when it's sunny than when it's cloudy. Something is really off.

To start with, the plans don't match the actual wiring. The plans show a total of nine MPPTs being used, but in reality you're using all twelve. Inverter 2 has MPPT 1/2 in parallel and inverter 3 has MPPT 3/4 in parallel. The pictures of the wiring and the inverter reading confirm those parallel channels. But hard to make sense of much else. The plans normally would tell you how much power to expect from each string but I'm not sure you can trust them until they match the actual wiring on your house.

I thought that was the case as well. But thats where I was getting confused. And yea lol , as the inverter strings show the total output even on those screenshots at 1pm (same time each day) on both cloudy and sunny days. Some of the strings show 0A or very low compared to the cloudy day. Not sure how that makes sense.

 

[aimetti]

Grasping at straws here but im worried about the strings circled here 12 , 13 , 14 are on a garage facing north west pitched down away from the sunlight and string #3 is on the otherside of the pitch of the garage that gets blocked by the house. If any of those are running parallel with other strings could that be impacting the entire inverter or just the other strings w/ them? It looks like maybe 12,13 are together ? 12 is on the garage , 13 is on front facing of house towards sun. 14,15 look together on their own which are on the same roof. String 3 looks like its possibly with 2 , also front facing of the house in parallel

 

[Redhill_qik]

for alittle bit of privacy how about Norwich, CT 06360
They also estimated on the low end for 11000 kwh , prior to me knocking down a bunch of trees. So not sure how accurate that would be now.

I messed with pvwatts alittle and tried to just do my back roof with the 47 azimuth / 34 pitch and roughly 8kw and that shows as 7200kwh alone. I could narrow it down and do each roof section and add it all together. I have to figure out how to calculate the kw per roof. Id imagine i could find that out based on the amount of panels on each section?


Before I had tesla come to house my usage was not accurate but i was fully off grid in june / july before pto. So in the summer I definitely was producing a good amount considering my house was using 2500ish kwh a month. Seems like its been pretty bad ever since they started tinkering w/ it or maybe just time of year?

This is what I have pulled together for you data for zipcode 06360, late October, at 1:00pm:

String	Azimuth	TIlt	Tiles	DC kW	Annual AC kWh	Inverter	String PVwatts AC kW Sunny​	String PVwatts AC kW Cloudy​	Inverter PVwatts AC kW Sunny​	Inverter PVwatts AC kW Cloudy​
MP5	137	34	8	0.574	737	1	0.244	0.150
MP15	229	34	33	2.368	2958	1	1.595	0.838
MP12	227	34	80	5.741	7215	1	3.883	2.039	5.72	3.03
MP16	49	34	41	2.942	2450	2	0.139	0.379
MP6	47	34	109	7.822	6429	2	0.370	1.008	0.51	1.39
MP11	227	30	30	2.153	2707	3	1.409	0.750
MP7	317	34	70	5.023	3985	3	1.225	0.995	2.63	1.75
			371	26.623	26481		8.865	6.159

You said that Tesla estimated 11,000 kWh annually, but PVwatts is estimating 26,481 kWh which is really far off, so this is the first thing that I would be looking to understand. I am calculating 26.623 kW for your Solar Roof based on the internet saying that it is 71.76W per panel.

Looking at it on an inverter basis and comparing to your screen shots I get this table:

Inverter​	Inverter PVwatts AC kW Sunny​	Inverter Actual AC kW Sunny​	Inverter PVwatts AC kW Cloudy​	Inverter Actual AC kW Cloudy​
1​	5.72​	2.97​	3.03​	2.73​
2​	0.51​	0.67​	1.39​	1.53​
3​	2.63​	0.97​	1.75​	1.62​

Inverter 2 seems to match well against PVwatts even though on the cloudy day numbers are higher than for the sunny day. Before anyone pipes up that the weather factor is subjective, I looked at data for October 25-27 and selected the 25th as day as "Sunny" and the 26th as "Cloudy", so it is the same weather data being applied to all panels for each condition. Why are the panels facing 47/49 degrees producing more on a "Cloudy" day than on a "Sunny" day? My best guess is that is back scattering from the clouds to these northeast facing panels from the sun that is in southwest.

Both inverters 1 & 3 look like they are underproducing in the "Sunny" condition, but are closer in the "Cloudy" condition.

 

[aimetti]

This is what I have pulled together for you data for zipcode 06360, late October, at 1:00pm:

String	Azimuth	TIlt	Tiles	DC kW	Annual AC kWh	Inverter	String PVwatts AC kW Sunny​	String PVwatts AC kW Cloudy​	Inverter PVwatts AC kW Sunny​	Inverter PVwatts AC kW Cloudy​
MP5	137	34	8	0.574	737	1	0.244	0.150
MP15	229	34	33	2.368	2958	1	1.595	0.838
MP12	227	34	80	5.741	7215	1	3.883	2.039	5.72	3.03
MP16	49	34	41	2.942	2450	2	0.139	0.379
MP6	47	34	109	7.822	6429	2	0.370	1.008	0.51	1.39
MP11	227	30	30	2.153	2707	3	1.409	0.750
MP7	317	34	70	5.023	3985	3	1.225	0.995	2.63	1.75
			371	26.623	26481		8.865	6.159

You said that Tesla estimated 11,000 kWh annually, but PVwatts is estimating 26,481 kWh which is really far off, so this is the first thing that I would be looking to understand. I am calculating 26.623 kW for your Solar Roof based on the internet saying that it is 71.76W per panel.

Looking at it on an inverter basis and comparing to your screen shots I get this table:

Inverter​	Inverter PVwatts AC kW Sunny​	Inverter Actual AC kW Sunny​	Inverter PVwatts AC kW Cloudy​	Inverter Actual AC kW Cloudy​
1​	5.72​	2.97​	3.03​	2.73​
2​	0.51​	0.67​	1.39​	1.53​
3​	2.63​	0.97​	1.75​	1.62​

Inverter 2 seems to match well against PVwatts even though on the cloudy day numbers are higher than for the sunny day. Before anyone pipes up that the weather factor is subjective, I looked at data for October 25-27 and selected the 25th as day as "Sunny" and the 26th as "Cloudy", so it is the same weather data being applied to all panels for each condition. Why are the panels facing 47/49 degrees producing more on a "Cloudy" day than on a "Sunny" day? My best guess is that is back scattering from the clouds to these northeast facing panels from the sun that is in southwest.

Both inverters 1 & 3 look like they are underproducing in the "Sunny" condition, but are closer in the "Cloudy" condition.

First off , thanks so much for taking the time to do all that. Really appreciate it.

Based on the data maybe my roof isnt as far off as I thought it was. Looking at inverter 2 , it looks like all the strings on that are on the back side roofs which makes sense if theres cloud cover it may be alittle higher due to reflectivity or whatever the reasoning is. But they are at least all on the same degree facing side. At least I can pretty much rule out 2 being the issue.

1/3 definitely seem to be underperforming. With them being higher generation w/ cloud cover could they also be getting the reflective property back to the spots that are shaded / blocked by the house etc. If you look at the picture I posted on the one previous to yours , Inverter 1 has string 2 and 3 together. 3 is in top left corner MP2 getting blocked by the main roof and gets very little sun. Almost wonder if I can disconnect the small 8 panel section to see if it has an effect? If they are running in parallel could this be impacting the string 5 facing the sun? Inverter 1 also has 4/5 on the same roof and string 1 facing the same way

Inverter 3 looks to have string 12 (facing back away from sun on garage) and another string 13 on the front of the house. Are those in parallel or their own? I cant tell , if so 12 is probably getting no sun while 13 is getting a ton. String 14/15 are also on that top left garage on the same MP7 roof.


Thanks again!

 

[Redhill_qik]

First off , thanks so much for taking the time to do all that. Really appreciate it.

Based on the data maybe my roof isnt as far off as I thought it was. Looking at inverter 2 , it looks like all the strings on that are on the back side roofs which makes sense if theres cloud cover it may be alittle higher due to reflectivity or whatever the reasoning is. But they are at least all on the same degree facing side. At least I can pretty much rule out 2 being the issue.

1/3 definitely seem to be underperforming. With them being higher generation w/ cloud cover could they also be getting the reflective property back to the spots that are shaded / blocked by the house etc. If you look at the picture I posted on the one previous to yours , Inverter 1 has string 2 and 3 together. 3 is in top left corner MP2 getting blocked by the main roof and gets very little sun. Almost wonder if I can disconnect the small 8 panel section to see if it has an effect? If they are running in parallel could this be impacting the string 5 facing the sun? Inverter 1 also has 4/5 on the same roof and string 1 facing the same way

Inverter 3 looks to have string 12 (facing back away from sun on garage) and another string 13 on the front of the house. Are those in parallel or their own? I cant tell , if so 12 is probably getting no sun while 13 is getting a ton. String 14/15 are also on that top left garage on the same MP7 roof.


Thanks again!

Looking at the line drawing, the string configurations don't seem to match your screen shots. All of the inverter screen shots show 4 strings, but the line drawing shows inverter 1 with four strings, inverter 2 with three strings and inverter 3 with two strings.

The photos that you posted appear to show inverter 1 with four strings, inverter 2 with three strings and inverter 3 with three strings, so I'm confused.

 

[aimetti]

Looking at the line drawing, the string configurations don't seem to match your screen shots. All of the inverter screen shots show 4 strings, but the line drawing shows inverter 1 with four strings, inverter 2 with three strings and inverter 3 with two strings.

The photos that you posted appear to show inverter 1 with four strings, inverter 2 with three strings and inverter 3 with three strings, so I'm confused.

hmmmm i guess its kind of impossible to see whats what then. tesla going to have to come out again

 

[BGbreeder]

hmmmm i guess its kind of impossible to see whats what then. tesla going to have to come out again

Sorry that you are having troubles. Thanks for posting your location; for anyone asking for advice, it does make a difference.

Yes, Tesla should come back out and fix things. What was installed seems to be pretty clearly not what was planned, AND, it isn't functioning correctly.

All the best,

BG

 

[aimetti]

Sorry that you are having troubles. Thanks for posting your location; for anyone asking for advice, it does make a difference.

Yes, Tesla should come back out and fix things. What was installed seems to be pretty clearly not what was planned, AND, it isn't functioning correctly.

All the best,

BG

Of course , this community is pretty solid I must say. Lots of help around here.

 

[tivoboy]

I would see if Tesla can focus on inverter configuration, it might be that 1 or all inverters are artificially throttling max power conversion. I’m not sure where that PV/DC power goes if there is a lot of sun and the inverter is capped, but it would make some sense (in my NON EE way) that one could actually get more through AC power in the end on a less sunny day, than a. Very sunny day.

Is there any battery backup here?

Frankly, I would put the issue more with Invertor 2 and 3, than 1 and 3, but again ME, NO EE

PM’d you about privacy too.

 

[aimetti]

I would see if Tesla can focus on inverter configuration, it might be that 1 or all inverters are artificially throttling max power conversion. I’m not sure where that PV/DC power goes if there is a lot of sun and the inverter is capped, but it would make some sense (in my NON EE way) that one could actually get more through AC power in the end on a less sunny day, than a. Very sunny day.

Is there any battery backup here?

PM’d you about privacy too.

Thanks responded.

and yea got 3 power walls. Right now full sun im pulling in a whopping 4.1 Kw

 

[h2ofun]

Thanks responded.

and yea got 3 power walls. Right now full sun im pulling in a whopping 4.1 Kw

Doing better than me, I am at 500 watts so far

 

[charlesj]

@aimetti You indicated 33kW in your very first post. Is that total production for the day or peak at high solar noon.
I ask since you have 3- 7...kW inverter that would max out at about 22kW peak plus perhaps a bit more before clipping. Just curious.

 

[aimetti]

@aimetti You indicated 33kW in your very first post. Is that total production for the day or peak at high solar noon.
I ask since you have 3- 7...kW inverter that would max out at about 22kW peak plus perhaps a bit more before clipping. Just curious.

Ya 33kw was inflated. Was calculating one inverter as double with the neurino. The ct issues seem to be fixed now and more in line with what it should read. It was also showing my house as using roughly 40% more electricity at the time.

 

[aimetti]

Just alittle update,

I am now only producing 12.5 kwh a day and peaking around 2.8kw. Seems way to low for my system. I have a ticket out with tesla and they will be getting back to me soon. My neurino keeps disconnecting / beeping now as well. Hopefully they find something.

 

[BGbreeder]

Just alittle update,

I am now only producing 12.5 kwh a day and peaking around 2.8kw. Seems way to low for my system. I have a ticket out with tesla and they will be getting back to me soon. My neurino keeps disconnecting / beeping now as well. Hopefully they find something.

If your neurio is complaining, have you done two meter reads 24hrs apart to see if you are really under producing, rather than just under reporting to the app?