Low Solar Production and Peak vs Rated

[v4556]

I have a new 8.4 Kw Tesla Solar Panel system (21 panels) with 2 PW+. It's been up and operational for over 2 months. The peak solar production I've ever seen this system produce is 5.6Kw and a total of 35.6Kwh so far. I live in Texas and we've had some pretty good sunny days and this is the peak I've seen this system generate. Is this normal? There are no obvious obstructions and the system is facing South (13 panel string) and West (8 Panel string).

I used PyPowerwall to start digging into see if both strings are generating power properly and this is what I got from the string data:

{'A': {'Current': 9.11, 'Voltage': 271.2, 'Power': 2449.0, 'State': 'PV_Active', 'Connected': True},

 'B': {'Current': 9.01, 'Voltage': 270.3, 'Power': 2400.0, 'State': 'PV_Active_Parallel', 'Connected': True},

 'C': {'Current': 0.0, 'Voltage': -1.0, 'Power': 0.0, 'State': 'PV_Active', 'Connected': False},

 'D': {'Current': 0.01, 'Voltage': -1.299999999999999, 'Power': 0.0, 'State': 'PV_Active', 'Connected': False}

 }

Is it normal for the state of the second string to be 'PV_Active_Parallel'? Some of the other discussions I've seen on this forum have state as 'PV_Active'. Anyone familiar what parallel means for state of the string? And if this string data makes sense to you?

 

[wwhitney]

A little photo interpretation follows that may or may not be useful; you may need to blow up the image to see everything. If you are certain you could do so safely, taking a picture of the inverter wiring from a somewhat different angle would provide more information. An unobstructed view of the terminal blocks for the DC conductors (near the J markings in the image below) would show if there is any miswiring there.

So below I have just the wiring inside the cabinet, with some letter markings in blue. I cropped out the diagram just above the image, but that diagram tells you how to wire the DC side.

The wires labeled R1, R2, G, and B1 are coming from the top conduit entry and are DC conductors from the arrays. I can't locate B2 in the picture, so maybe it is directly behind B1 throughout the space. There's an apparent widening of B1 to the upper right of the ? where I labeled "? B2", so that's probably the case. Apparently the convention is that red insulation is DC +, black is DC -, and the two circuits are differentiated by one have wraps of red tape on them (which I've called R1 and B1), the other not.

For reference, the red, white, and black wires originating in the green terminal block behind my B1 label are the AC side conductors that leave through the lower conduit.

I wrote a J near a red jumper. Behind that is the DC red (DC +) terminal block. It has 6 connections, a column of 4 on the right (I'll call them 1-4, with 1 on top) and two more on the left (5 and 6). Terminals 1,2, and 5 represent one MPPT input with two connection points (1 and 2); likewise 3, 4, and 6. If you have 2 identical strings (same length and orientation) for which you've brought separate DC conductors to the inverter, you can parallel them by connecting them to 1 and 2, and leaving 5 empty. Or if you've already paralleled them closer to the array and just brought 1 pair of conductors to the inverter, you are supposed to use the factory jumper between 1 and 2 and land the red DC conductor at 5.

For the red terminals, my interpretation of the photo is 1 and 2 are jumpered, and R2 lands on 5. Then R1 lands on 3 or 4, and 4 or 3 (respectively) is empty, along with 6. Which is plausible but seems to leave only half of the capacity of the 2nd MPPT in use; not sure why R1 wasn't landed on 6 with 3 and 4 left jumpered. Perhaps for an 8 panel string half the capacity would be sufficient, not sure.

On the black terminals, it's much harder to make out what is going on, along with the apparently missing B2. The J? marking is on a label that might be obscuring a black jumper (such a jumper would be required for parallelism with the DC+ configuration). You can clearly seen 1 unused red jumper at the bottom of the picture, and at least 1 unused black jumper, but perhaps there are 2 black jumpers there?

Best case on the black terminal B1 is going to 3, and 1 and 2 are jumpered, and the obscured B2 is going to 5. That would be wired analogously to the red terminals. Any deviation from that would be a sign of miswiring at the inverter.

If the black terminals are wired as in the previous paragraph, then either the problem is upstream of the inverter case, or it may be that R1 and B1 should be moved to 6 and the jumpers reinstalled between 3 and 4. Not sure.

Cheers, Wayne

 

[dareed1]

Do you know what Tesla had to do resolve your issue? Did they have to re-wire the inverter or panels ? Any info you could share will be helpful. Greatly appreciate the info you’ve been sharing.

Firstly, my advice is to not mention any similarity to the symptoms/cure that I (some random guy on the interweb) had. I think it is better to describe that you're showing equal voltages and currents from your two strings, and avoid telling them how to fix it, at least at this stage.

Anyway, in our installation, the permit was pulled for a SolarEdge inverter, but they actually installed Delta string inverter. Somehow in the changeover, the engineering drawings called for the two strings to be connected in parallel at the roof. So only one pair of wires were brought to the inverter. The service tech who came initially to diagnose the problem told me that the drawings called for this configuration, "but I can't believe anybody would actually wire it this way." They also used only one MPPT channel. Since the connection was on the roof, they needed to send a crew to pull additional wires through the conduit. At that point, I believe the crew made a mistake and jumpered the two strings at their connection to the inverter. Still this was better than before because we were no longer clipping due to a single MPPT channel current limit.

I could have pulled the erroneous jumper myself, but I let them do it so that the service records would be accurate. We've had no problems since.

 

[v4556]

A little photo interpretation follows that may or may not be useful; you may need to blow up the image to see everything. If you are certain you could do so safely, taking a picture of the inverter wiring from a somewhat different angle would provide more information. An unobstructed view of the terminal blocks for the DC conductors (near the J markings in the image below) would show if there is any miswiring there.

So below I have just the wiring inside the cabinet, with some letter markings in blue. I cropped out the diagram just above the image, but that diagram tells you how to wire the DC side.

The wires labeled R1, R2, G, and B1 are coming from the top conduit entry and are DC conductors from the arrays. I can't locate B2 in the picture, so maybe it is directly behind B1 throughout the space. There's an apparent widening of B1 to the upper right of the ? where I labeled "? B2", so that's probably the case. Apparently the convention is that red insulation is DC +, black is DC -, and the two circuits are differentiated by one have wraps of red tape on them (which I've called R1 and B1), the other not.

For reference, the red, white, and black wires originating in the green terminal block behind my B1 label are the AC side conductors that leave through the lower conduit.

I wrote a J near a red jumper. Behind that is the DC red (DC +) terminal block. It has 6 connections, a column of 4 on the right (I'll call them 1-4, with 1 on top) and two more on the left (5 and 6). Terminals 1,2, and 5 represent one MPPT input with two connection points (1 and 2); likewise 3, 4, and 6. If you have 2 identical strings (same length and orientation) for which you've brought separate DC conductors to the inverter, you can parallel them by connecting them to 1 and 2, and leaving 5 empty. Or if you've already paralleled them closer to the array and just brought 1 pair of conductors to the inverter, you are supposed to use the factory jumper between 1 and 2 and land the red DC conductor at 5.

For the red terminals, my interpretation of the photo is 1 and 2 are jumpered, and R2 lands on 5. Then R1 lands on 3 or 4, and 4 or 3 (respectively) is empty, along with 6. Which is plausible but seems to leave only half of the capacity of the 2nd MPPT in use; not sure why R1 wasn't landed on 6 with 3 and 4 left jumpered. Perhaps for an 8 panel string half the capacity would be sufficient, not sure.

On the black terminals, it's much harder to make out what is going on, along with the apparently missing B2. The J? marking is on a label that might be obscuring a black jumper (such a jumper would be required for parallelism with the DC+ configuration). You can clearly seen 1 unused red jumper at the bottom of the picture, and at least 1 unused black jumper, but perhaps there are 2 black jumpers there?

Best case on the black terminal B1 is going to 3, and 1 and 2 are jumpered, and the obscured B2 is going to 5. That would be wired analogously to the red terminals. Any deviation from that would be a sign of miswiring at the inverter.

If the black terminals are wired as in the previous paragraph, then either the problem is upstream of the inverter case, or it may be that R1 and B1 should be moved to 6 and the jumpers reinstalled between 3 and 4. Not sure.

Cheers, Wayne
View attachment 799470

I just took a few more photos of the inverter wiring in different angles.

You mentioned upstream issues. One other thing I didn’t mention was that initially they intended to put the Tesla inverter outside, but when they changed the design to powerwall plus that comes with an inverter, they added a junction box outside. I took a few pics of the inside of that junction box.

 

[v4556]

Firstly, my advice is to not mention any similarity to the symptoms/cure that I (some random guy on the interweb) had. I think it is better to describe that you're showing equal voltages and currents from your two strings, and avoid telling them how to fix it, at least at this stage.

Anyway, in our installation, the permit was pulled for a SolarEdge inverter, but they actually installed Delta string inverter. Somehow in the changeover, the engineering drawings called for the two strings to be connected in parallel at the roof. So only one pair of wires were brought to the inverter. The service tech who came initially to diagnose the problem told me that the drawings called for this configuration, "but I can't believe anybody would actually wire it this way." They also used only one MPPT channel. Since the connection was on the roof, they needed to send a crew to pull additional wires through the conduit. At that point, I believe the crew made a mistake and jumpered the two strings at their connection to the inverter. Still this was better than before because we were no longer clipping due to a single MPPT channel current limit.

I could have pulled the erroneous jumper myself, but I let them do it so that the service records would be accurate. We've had no problems since.

I completely understand your suggested approach and totally agree. My intent is to not DIY this issue or to tell Tesla what to do. I am just trying to understand the range of possibilities for this, so I am better prepared. Thanks for sharing the info.

 

[wwhitney]

I just took a few more photos of the inverter wiring in different angles.

Yes, the first of those is very helpful, it does show that there is a previously obscured "B2" wire coming in the upper conduit. And it shows that the DC+ and DC- connections that I called 1, 2, 3, 4 are in fact so labeled (as Tesla and I both came up with the obvious labeling). Not sure how what I called "5, 6" are actually labeled.

Anyway, what I can see in the photos is consistent with the following configuration, although I still wasn't able to see it all to confirm it is what you have:

R1 (red with tape) to DC+ 3
DC+ 1 and 2 jumpered
R2 (red no tape) to DC+ "5" (upper of left pair)

B1 (black with tape) to DC- 3
DC- 1 and 2 jumpered
B2 (black without tape to DC- "5" (upper of left pair)

I.e. the diagram above that says "One independent string, One combined String". Although I don't think you'd have any combined strings with a string of 13 and a string of 8.

If by visual inspection you confirm the above is accurate, then I think it's very likely that the root cause of the equal voltages/equal currents is not the configuration of wiring within the inverter.

As to the junction box, the shots are too incomplete. We'd need a view that show the full perimeter of the open box, with the ability to trace each wire as it comes in from one of the conduits to one of the connectors within the box.

Cheers, Wayne

 

[v4556]

Yes, the first of those is very helpful, it does show that there is a previously obscured "B2" wire coming in the upper conduit. And it shows that the DC+ and DC- connections that I called 1, 2, 3, 4 are in fact so labeled (as Tesla and I both came up with the obvious labeling). Not sure how what I called "5, 6" are actually labeled.

Anyway, what I can see in the photos is consistent with the following configuration, although I still wasn't able to see it all to confirm it is what you have:

R1 (red with tape) to DC+ 3
DC+ 1 and 2 jumpered
R2 (red no tape) to DC+ "5" (upper of left pair)

B1 (black with tape) to DC- 3
DC- 1 and 2 jumpered
B2 (black without tape to DC- "5" (upper of left pair)

I.e. the diagram above that says "One independent string, One combined String". Although I don't think you'd have any combined strings with a string of 13 and a string of 8.

If by visual inspection you confirm the above is accurate, then I think it's very likely that the root cause of the equal voltages/equal currents is not the configuration of wiring within the inverter.

As to the junction box, the shots are too incomplete. We'd need a view that show the full perimeter of the open box, with the ability to trace each wire as it comes in from one of the conduits to one of the connectors within the box.

Cheers, Wayne

I got another photo of the exterior junction box.

 

[wwhitney]

I got another photo of the exterior junction box.

That's pretty clear a straight 1-1 splicing, so I don't think that that could be the source of the problem. The only error could be swapping the strings or +/- which I don't think would give the behavior you're seeing.

So most likely when you convince Tesla to send a tech out to diagnose the problem, they will start by looking at those two locations (inverter and junction box), and if I'm right, conclude that the problem is on the roof (assuming the inverter is wired as per the lower right diagram on the label above it). Let us know how it turns out.

Cheers, Wayne

 

[v4556]

That's pretty clear a straight 1-1 splicing, so I don't think that that could be the source of the problem. The only error could be swapping the strings or +/- which I don't think would give the behavior you're seeing.

So most likely when you convince Tesla to send a tech out to diagnose the problem, they will start by looking at those two locations (inverter and junction box), and if I'm right, conclude that the problem is on the roof (assuming the inverter is wired as per the lower right diagram on the label above it). Let us know how it turns out.

Cheers, Wayne

Thank you very much for all the info. Will keep you posted.

 

[v4556]

That's pretty clear a straight 1-1 splicing, so I don't think that that could be the source of the problem. The only error could be swapping the strings or +/- which I don't think would give the behavior you're seeing.

So most likely when you convince Tesla to send a tech out to diagnose the problem, they will start by looking at those two locations (inverter and junction box), and if I'm right, conclude that the problem is on the roof (assuming the inverter is wired as per the lower right diagram on the label above it). Let us know how it turns out.

Cheers, Wayne

I asked Tesla to send me the final drawings, as they made a few last minute changes to my project. Per their drawings, the two strings should be configured as follows:

But I was able to get string readings that looked like this:

{'A': {'Current': 9.11, 'Voltage': 271.2, 'Power': 2449.0, 'State': 'PV_Active', 'Connected': True},
 'B': {'Current': 9.01, 'Voltage': 270.3, 'Power': 2400.0, 'State': 'PV_Active_Parallel', 'Connected': True},
 'C': {'Current': 0.0, 'Voltage': -1.0, 'Power': 0.0, 'State': 'PV_Active', 'Connected': False},
 'D': {'Current': 0.01, 'Voltage': -1.299999999999999, 'Power': 0.0, 'State': 'PV_Active', 'Connected': False}
 }

In other words, per the drawings, I should not get a power reading of more than 2000 from one of the strings. But my string readings don't line up with the drawing. Am I misinterpreting something here? Does that mean there's a mix up somewhere when they hooked up the panels? Or inverter?

 

[wwhitney]

I'm thinking that something is messed up on the roof where they go from 3 individual strings (6 conductors) to 1 individual string and 1 set of 2 parallel strings (4 conductors). Not sure exactly what the mistake is, I haven't dealt with this before.

I mean, if the 3 individual strings are A (5 panels), B (8), and C (8), and X and Y are the down conductors to the inverters, then the on roof junction box should be A to X, and B and C to Y (in parallel). Could it be as simple as A is unconnected, and B, C, X, and Y are all connected? That would give you reduced production (16/21 of expected) and give you the identical voltage and currents, as X and Y will be in parallel.

But that would be pretty stupid.

Cheers, Wayne

 

[v4556]

I'm thinking that something is messed up on the roof where they go from 3 individual strings (6 conductors) to 1 individual string and 1 set of 2 parallel strings (4 conductors). Not sure exactly what the mistake is, I haven't dealt with this before.

I mean, if the 3 individual strings are A (5 panels), B (8), and C (8), and X and Y are the down conductors to the inverters, then the on roof junction box should be A to X, and B and C to Y (in parallel). Could it be as simple as A is unconnected, and B, C, X, and Y are all connected? That would give you reduced production (16/21 of expected) and give you the identical voltage and currents, as X and Y will be in parallel.

But that would be pretty stupid.

Cheers, Wayne

That’s my suspicion. So far, that theory seems to be lining up with all the data I’ve been seeing. Just have to get Tesla to send someone up that roof and take a look at it.

 

[yblaser]

It looks like your 2 parallel strings are working correctly. They are running into ch1 of the inverter and ch1 and ch2 are jumpered together. The 8 panels are consistent with the 270V you are reading on Ch1 and Ch2. The other string is running into ch3 and your inverter is showing nothing connected there.

 

[v4556]

It looks like your 2 parallel strings are working correctly. They are running into ch1 of the inverter and ch1 and ch2 are jumpered together. The 8 panels are consistent with the 270V you are reading on Ch1 and Ch2. The other string is running into ch3 and your inverter is showing nothing connected there.

I am not that familiar with the inverter setup. I have 4 wires coming into the inverter. Are you saying that my inverter is using 3 MPPT channels?

 

[yblaser]

I am not that familiar with the inverter setup. I have 4 wires coming into the inverter. Are you saying that my inverter is using 3 MPPT channels?

Yes that is what your design drawing says and as best as I can see in your pictures that's how it is wired although I can't make out everything. The two 8 panel strings are combined in parallel on the roof and come down as a single pair. The 5 panel string come down as the second pair.

The wires from the two parallel strings are hooked into the MPPT #1 input and then there is a jumper that connects MPPT #1 and #2. The current output of parallel strings would be too high for a single MPPT input. The 5 panel string looks like it is hooked into #3.

 

[dho112]

I'm EXTREMELY new to all of this (and this great group!) having just gotten a Tesla Solar system in March of this year (9.6kW 1 Powerwall+) but it would seem even with your chimney and the placement of your panels that you SHOULD be getting a significantly higher output peak than 5.6kW at least for a tiny portion of the day? Especially in Texas sun? Again, I'm very new here. My 9.6 kW system clips at 7.7kW for almost 5 hours a day and I'm pretty sure it's making more than 8 for that time. (Northern California, 24 panels, south facing 20 degree incline)

Could you post a screenshot of your daily solar production for a good day with no clouds? If the chimney shading some panels is significant, it will be very obvious in the graph and consistent day to day, as will any weird obstruction/shading that you might not be aware of. I always see this strange dip in my early morning solar production (it's a very sharp dip and lasts just a little while) it turned out it is Japanese maple tree of my neighbors, pretty far away, that shades a good portion of my panels during early sunrise with a VERY long shadow that I never could have predicted unless it was graphed

 

[Sylvia Else]

I have a new 8.4 Kw Tesla Solar Panel system (21 panels) with 2 PW+. It's been up and operational for over 2 months. The peak solar production I've ever seen this system produce is 5.6Kw and a total of 35.6Kwh so far. I live in Texas and we've had some pretty good sunny days and this is the peak I've seen this system generate. Is this normal? There are no obvious obstructions and the system is facing South (13 panel string) and West (8 Panel string).

I used PyPowerwall to start digging into see if both strings are generating power properly and this is what I got from the string data:

{'A': {'Current': 9.11, 'Voltage': 271.2, 'Power': 2449.0, 'State': 'PV_Active', 'Connected': True},

 'B': {'Current': 9.01, 'Voltage': 270.3, 'Power': 2400.0, 'State': 'PV_Active_Parallel', 'Connected': True},

 'C': {'Current': 0.0, 'Voltage': -1.0, 'Power': 0.0, 'State': 'PV_Active', 'Connected': False},

 'D': {'Current': 0.01, 'Voltage': -1.299999999999999, 'Power': 0.0, 'State': 'PV_Active', 'Connected': False}

 }

Is it normal for the state of the second string to be 'PV_Active_Parallel'? Some of the other discussions I've seen on this forum have state as 'PV_Active'. Anyone familiar what parallel means for state of the string? And if this string data makes sense to you?

The output from a panel reduces with the angle between the sun and the perpendicular to the panel (it pretty much follows a trigonometric cosine), so it's affected by latitude, panel orientation, time of day, and time of year. Even if there's a time of day/year when the sun is perpendicular to the panel, you'll only so the maximum possible output if you're at the equator, because otherwise the sunlight has to travel through more atmosphere.

Since your panels are not all facing in the same direction, you'll never see the rated output.

 

[v4556]

Tesla finally was able to come and check this. They were able to locate one of the 'disconnected' strings and were able to quickly fix it. Energy production is now in the expected range. I now see power from all the three strings.

Thanks everyone.

 

[Sam88]

I have a new 8.4 Kw Tesla Solar Panel system (21 panels) with 2 PW+. It's been up and operational for over 2 months. The peak solar production I've ever seen this system produce is 5.6Kw and a total of 35.6Kwh so far. I live in Texas and we've had some pretty good sunny days and this is the peak I've seen this system generate. Is this normal? There are no obvious obstructions and the system is facing South (13 panel string) and West (8 Panel string).

I used PyPowerwall to start digging into see if both strings are generating power properly and this is what I got from the string data:

{'A': {'Current': 9.11, 'Voltage': 271.2, 'Power': 2449.0, 'State': 'PV_Active', 'Connected': True},

 'B': {'Current': 9.01, 'Voltage': 270.3, 'Power': 2400.0, 'State': 'PV_Active_Parallel', 'Connected': True},

 'C': {'Current': 0.0, 'Voltage': -1.0, 'Power': 0.0, 'State': 'PV_Active', 'Connected': False},

 'D': {'Current': 0.01, 'Voltage': -1.299999999999999, 'Power': 0.0, 'State': 'PV_Active', 'Connected': False}

 }

Is it normal for the state of the second string to be 'PV_Active_Parallel'? Some of the other discussions I've seen on this forum have state as 'PV_Active'. Anyone familiar what parallel means for state of the string? And if this string data makes sense to you?

Hi,
Then compared to your mine is really off. I have 14 KW system, 34 panels, no powerwall. My highest peak is 6 Kw, that for very short time, less than an hour. Total never higher than 39 Kwh. I live in sunny southern california. I have included the pictures from the inverters portal. Also as you see I have two inverters one at 7.6 kw and the other at 3.8 Kw. I wonder if that is the reason and Tesla have short cut me in inverter size. Any ideas. Thanks

 

[dareed1]

The voltages reported on strings 1 and 2 of the 7.6kW inverter are implausible. But I don't think that the inverter sizes are a problem, because the generation curve shape looks normal. If you can post a single line diagram similar to this example and also the drawing that shows the placement of the panels on your roof, we may be able to offer some possible explanations.

 

[dho112]

Hi,
Then compared to your mine is really off. I have 14 KW system, 34 panels, no powerwall. My highest peak is 6 Kw, that for very short time, less than an hour. Total never higher than 39 Kwh. I live in sunny southern california. I have included the pictures from the inverters portal. Also as you see I have two inverters one at 7.6 kw and the other at 3.8 Kw. I wonder if that is the reason and Tesla have short cut me in inverter size. Any ideas. Thanks

!!! Can you describe your panel config on your roof? With that many panels I’m assuming it must be several arrays on different sections, but your production seems extremely low regardless unless you have some serious shade issues or none of your panels face in even a moderately optimal direction.

I have a 9.6Kw system and in optimal months of May-Jul etc for sun angle I hit the 7.7 Kw max inverter cap for often 4 hours a day and often get close to 70Kwh a day. It was enlightening for me though that super sunny days at the height of summer produced less solar than in spring/early summer, because heat has a much greater effect on production than I thought, sunny cold days with decent sun angle were always the best by far! (spring) I’m in SF bay area and my panels are all south facing.

If you have 7.6 + 3.8 inverters there is no way (barring bad installation) that you’re capped at 6Kw production. I’d love to see a pic of your panel/roof situation.