Is there a problem with my PV setup? Dual MPPT w different orientations & sizes show near identical power.

[PVReeper]

I’d appreciate a few expert eyes looking over our solar PV setup. We’re in the midlands and have a system consisting of a 4kW inverter and 16x250W panels, giving 4kW, installed 12 years ago. The inverter was dying, so we had it replaced & moved as part of a house electrical re-wiring. We were quoted for a 3kW Growatt min3000tl-x, but actually had a 2.5kW growatt min2500tl-x installed. I *think* that should be OK for 4kW 12-year old panels and possibly give us more generation overall, or should we push for it to be replaced with the 3kW to be on the safe side? We don't seem to have any documentation on the make/model of the panels and the firm that installed them has since gone bust.

The main question I have is about the two different arrays. We have 3 panels (750W) South-south-west and 13 panels (3250W) East-south-east, both at 35 degree roof pitch. The growatt min2500tl-x has dual MPPTs, so I would have expected the power generated by each to be independent and peak at different times. I.e. as shown by this diagram:


https://www.solarpowerworldonline.com/2014/02/dual-mppt-defined-understanding-mppt/

The East array should generate higher power earlier in the day and the South array should generate higher power later in the day. However, on many days the power generated is identical, which seems very odd. E.g. a few day’s comparison of the power (W) generated by MPPT1 vs MPPT2.






Today was a day without clouds, so it should give a cleaner picture of what’s going on. These are the predicted Solcast curves for each array (note the different power scales):



Here are the voltage, current & power of today's output:


The voltages are reasonably consistent while the currents fluctuate, so I believe that means the array panels are each wired in parallel.



Today's power, full sunshine. Sunrise 08:13, fully aligned with the SSW array 13:35.

There's a definite earlier increase in the power of the 3250W East array today (lightest blue). Around 11ish the panel power levels converge, which is when the sun is exactly half way between them., i.e. 45 degrees off both of them. The South 750W array peaks around 12:30 and they then both decline together, despite the sun being fully on the S array 13:35.

Is there a problem here or is this expected behaviour for separate MPPTs with separate orientations and very different sizes to have very similar power levels during the day?

 

[PVReeper]

Added to that, here's the Solcast Estimated Actual PV (calculated from weather satellites measuring the actual cloud cover at that point in time) plus the GivenErergy measured PV (which is cruder than the Growatt inverter reading, as it's a clamped measurement).

Odd that the Actual PV measured by the GiveEnergy peaks well after when it should be declining. Looks like Solcast predicts much higher generation earlier in the day. None of the roof is shaded btw and this was in full sunshine.

Could the MPPTs be cross-wired, so they are limiting each other?

 

[JBertok]

It looks likely that the panels have been connected to the MPPT inputs with even distribution. 8 on the first, 8 on the second. Your roof has 3 panels on one facing and 13 on the other - crucially with different orientations. This is breaking design rules:
1. All panels connected in series need to be at identical angle orientations.
2. All panels on a single MPPT input need to be at identical angle orientations.

There are a few specs on the inverter MPPT in play here:
- Maximum input current = 13.5A
- Maximum DC voltage - 500V
- Nominal Voltage - 360V
- Maximum number of strings per MPP tracker - 1

The "nominal voltage" spec is the preferred array voltage, and 360V aligns with 8 panels in series. So, it seems likely this is wired 8S per MPPT. That creates a huge problem with 3 panels on one facing and 5 on the other - that is absolutely not okay.

Fixing this:
The ideal solution would be to move 5 panels over to the side with 3, or to move the 3 odd ones over to the side with 13.

Moving panels is an inevitability - you've got odd numbers on each facing, impossible to properly distribute between the MPPTs. The voltage from 3 panels in series is too low, and the voltage from 13 in series is too high for the MPPT, because remember different facing panels cannot share an MPPT or be wired in series.

 

[PVReeper]

We previously had a 4kW SunnyBoy inverter that could take 3 strings and it was wired with 3 panels on one string and 13 panels on the other, or so we were told. It was a Sunny Boy SB 3800/V 0153 with 3 strings and one MPPT and it's working producing for over 12 years, with an average yearly yield of 2372kWh.
The panels are attached to the roof and there's only four wires coming into the house, so I can't tell how the panels are wired.
You say about them being in series, but aren't' they in parallel, as the the voltage is roughly level through the day & the current varies?

 

[JBertok]

When wiring the panels in series you can take Voc (Open Circuit Voltage) of the panel rating and multiply it by the number of panels to get a maximum string voltage, and then add a safety margin of 5-10%. (That margin varies depending on climate and installation orientation. It comes from the nature of how panels are rated. The colder a panel gets the higher the Voc. It's also possible to see a panel put out more than rated amperage on scattered-cloudy day when direct sunlight passed between clouds and the surrounding clouds are diffusing the sunlight passing through them.)
The factored open-string voltage cannot exceed the rated maximum DC input of the MPPT. That rated maximum is usually higher than the MPPT maximum. For yours it's actually 500V for both the maximum DC voltage and MPP tracking voltage. For your 250W panels, they are probably 24V nominal. A spec on a multicrystaline 250 panel might have a Vmpp of 30.2V and a Voc of 37.4V. 37.4V * 13 panels is 486.2V, and although technically within the limit of the MPPT, there is no margin of safety and a possibility of exceeding rating exists.
On the other side, with the 3 panels, take that 30.2V * 3 panels and you get 90.6V, and that is above the 80V minimum of the MPPT, but there is a MPPT "starting voltage" rating too, and it's 100V here. 3 panels * 37.4 Voc = 112.2V. That's pretty close to the lower limits of both specs and could be problematic on overcast days and more so as the panels age.

So, based on those panel specs, which could be a ways off from your actual panel specs, it would be possible to have 3 on one input and 13 on the other. The logged data doesn't suggest they're connected that way, however. You've getting an average of 250V on both MPPT inputs. They're pretty close with an explainable imbalance due to the orientation differences of the 3 + 5 panels on the lower-yielding MPPT input. In series, when you have even one single cell shaded or otherwise unilluminated as they would be pointed a different direction, those cell(s) become a resistance to the powerflow. So, if you have 5 panels in direct sunlight and 3 getting sun at a steep angle, output of the entire string will be limited always by the panel of lowest productivity in the string - even if it's close to zero. That's why you're seeing the split array's production coming on so late compared to the other.

Your system works as it is, but could be making more efficient use of the installed equipment with some changes to the wiring. Making any changes to it depends on your needs and expectations to determine if it's worth the effort. If it's falling short there's definitely some opportunity to increase your daily output. You knew something was up because this thread is here. But, please don't take my arbitrary estimations for your PV panel specifications and run with that - we would need the real PV specs to do the math properly to be able to tell you if 13 on one string and 3 on the other would be viable and safe.

 

[PVReeper]

We plan to re-roof the house this summer, when we could change the wiring. For now, all we can sensibly do is with the wires to the inverter and the inverter settings (if anything needs/can be changed).

I'm still a bit puzzled about the wiring of the panels, as I had understood that they were in parallel, as the array currents rises and falls, while the current stays static https://battlebornbatteries.com/solar-panels-in-series-or-parallel/ JBertok, have I misunderstood?

I also don't understand why the power of both arrays seems so linked, with each array closely matching the other for most of the day. I had thought we had a 3 S panel array and a 13 E panel array. JBertok if you are correct that we have actually have a 3S + 5E array and a 8E array, then how can they be producing the same power level when the sun rises, when there is no direct sun on the S panels?

 

[JBertok]

In parallel the voltage remains the same as a single panel and the amperage rises by number of parallel members.
- Panels or strings in parallel require a blocking diode separating each parallel member to prevent reverse current flow - some panels have them built-in. These also prevent a failed panel from shorting the entire output of the string.
- Panels of strings in parallel lose only the output of the affected member when shaded.
- The sum of Isc (Short Circuit Current) amperage of all parallel members is used to size conductor gauge.
- The rated voltage of panels in parallel must match, but rated amperage does not need to.
- All panels in parallel on a single MPPT input do not necessarily need to have matching orientation, however MPP tracking will not be optimized through the day's sun.

In series the amperage remains the same as a single panel and the voltage rises by number of series panel.
- Panels in series do not require blocking diodes between each series panel.
- Panels in series experience an overall output drop of the entire string when even one cell of one panel is shaded. (A gray area here is bifacial or dual zone PV panels, when shading affects only half the overall output of the panel and thus string.) A cell not producing power impedes power flow, it doesn't just resist or fail to contribute to it.
- The sum of Voc (Open Circuit Voltage) of each panel in series, plus a 5-10% margin, is compared against component and MPPT max DC voltage rating.
- The rated amperage of panels in series needs to match, but the rated voltage does not need to.
- All panels in a series string need to have identical orientation.

When combing schemes of both parallel and series wired PV panels into one MPPT input, we first build series strings with matching rated voltage (amperage unimportant) and then parallel together the two or more series strings through blocking diodes. Some MPPT controllers do not support doing this, so specs always have to be referenced. When you see people in the forum referring to the panel wiring as #S, #P, or #S#P refer to Series, Parallel or Paralleled Series strings. For example, each of my four MPPTs has 18 panels in 3S6P because they have a MPPT voltage range. Low voltage needs larger conductors and more parallelism.

To your posted output data, you're seeing (somewhat) similar voltage between the two inputs throughout the day, and similar current outputs toward the end. The split array appears most compromised in the AM with surprisingly similar current in the PM. Wattages (expressed as Power) deviate between the two. The split-orientation array doesn't have the potential to match the homogenously oriented panel array in output because there will always be a portion of panels in that series string holding back the others at the better sun angle.

It's hard to tell from just the output data, but it seems possible the split array is experiencing some degree of shading in the AM hours because that output is so flatlined for so long.