308 watts from a 270 panel???? HUH?

[Murby]

So my array consists of an SMA SunnyBoy 6.0 US inverter (3 mppt) with 28 SolarWorld SW270 panels.

My strings are set up as MPPT: A=8 panels , B=8 panels and C=12 panels.

I would expect that channels A and B should both be putting out similar power with the max at 8 panels times 270 watts = 2160 watts.. Of course, it would never expect to see all that power..
I would also expect that channel C (12 panels) should be maxed out at 3240 watts.. again, I wouldn't expect to actually see that much power... and yet.. take a look at the DC output of the C channel in the screen grab below... 3644 watts divided by 12 panels = 303 watts per panel.. And here's the kicker.. It was 3800 watts just before I was able to get a screen shot of it... and as I type this, its holding at 3400 to 3600 watts and bouncing around.

My child's playset has a stainless steel slide.. see the next picture. I polished it to a bright finish so she can pick up some speed going down.. and it does produce a glare.. Is it possible that could be the cause???


3644.jpg


Slide.jpg

[Mike90250]

Warranty. When new, panels tend to overproduce for a year or so. That's so in 10 years they can still meet the warranty spec for the model !
Heat. The cooler the panel, the more power it will put out.
Photons. The more light (reflections from snow, sand, slide) , the more power.

So if it was a chilly day, and the sun just came out from behind clouds at noon, you will get great output for 3 minutes, till the panels heat up.

25 watts spread over 8 panels does not seem like a great difference to me.

[Murby]

Warranty. When new, panels tend to overproduce for a year or so. That's so in 10 years they can still meet the warranty spec for the model !
Heat. The cooler the panel, the more power it will put out.
Photons. The more light (reflections from snow, sand, slide) , the more power.

So if it was a chilly day, and the sun just came out from behind clouds at noon, you will get great output for 3 minutes, till the panels heat up.

25 watts spread over 8 panels does not seem like a great difference to me.

The A and B channels are doing what I would have expected.. Its the C channel and 330 watts from a 270 watt panel that had me a bit baffled.

[Murby]

MPPT Channels A and B consist of 8 panels on each channel.. Even under the brightest days, I have not seen them produce more than 220 watts per panel...

[max2k]

MPPT Channels A and B consist of 8 panels on each channel.. Even under the brightest days, I have not seen them produce more than 220 watts per panel...

Something is off. It's not the absolute numbers that would bother me but mismatch between strings: since they are all part of the same array and supposedly irradiated the same I'd expect ratio between A/B to be around 1.0 +- 5% and C/A or C/B to be 1.5 +- 5%. My own 2 strings feeding individual A inputs of 2 separate SMA 3.8 inverters are within 0.5% from each other which I admit is much closer than I expected. In your case this ratio is 2.8 which is almost twice as what it is supposed to be. Check your wiring and if nothing sticks out give SMA tech support a call/email.

how those strings are physically located in the array relatively to each other? For example, their currents don't make much sense- 12 panel string produces more than twice current of 8 panel strings. If they are made of the same type of panels, part of the same array, no shading this doesn't make sense. Besides current from string C exceeds Isc rating of your panels. I have some trouble to believe you managed to randomly select all 'bad' panels for strings A & B but all 'good' ones ended up in string C. Simple way to check would be to disconnect string C and switch over one of the other strings to its input and compare input C against the remaining one, they should be pretty close as both would be fed from independent 8 panel strings.

[wienerdog]

Is the short circuit current on the modules even 9.99 amps?

[Gdwats]

Shouldn't array C be dummied down to arrays A and B when strings are hooked together?

[ButchDeal]

Shouldn't array C be dummied down to arrays A and B when strings are hooked together?

The strings are not connected together. The inverter has three separate MPPT inputs.

[sensij]

I'm not sure you are looking at this correctly.

280 V / 8 panels = 35 V per panel
365 V / 12 panels = 30.4 V per panel

Total DC = 1303 + 1275 + 3644 = 6222 W

My guess is that the inverter is "clipping" inputs A and B, which drives up their voltage and lowers their current.

The fact that current C is right at the input limit (10 A) suggests there could a small amount of clipping there, but 30.5 V is not far from where the mpp voltage would be for those panels, especially on a cool day.

I don't know enough about that inverter to know the logic it uses when clipping is needed, but that it would treat inputs A and B differently than input C doesn't seem impossible, since those two inputs are allowed to have paralleled strings and the third input is not.

The 300 W per panel over 5-10 min periods (with even more dynamic spikes) is not impossible on partly cloudy days, where edge of cloud effects can drive up irradiance to 1200 W/m2 or more.

[Murby]

As it currently stands, at 4:15pm in Michigan, Channel A = 1750 watts (8 panels = 219 watts per panel), Channel B = almost exactly the same, and channel C = 2650 which is 12 panels or 220 watts per panel.

It must have been the playset slide reflection that caused the jump because it only lasted for about 30 minutes.. and by visual inspection, I'd say the sun was reflecting directly at the C channel array.....

[sensij]

As it currently stands, at 4:15pm in Michigan, Channel A = 1750 watts (8 panels = 219 watts per panel), Channel B = almost exactly the same, and channel C = 2650 which is 12 panels or 220 watts per panel.

It must have been the playset slide reflection that caused the jump because it only lasted for about 30 minutes.. and by visual inspection, I'd say the sun was reflecting directly at the C channel array.....

I doubt it. Please read the post I just made. At 4:15pm, what were the voltages per panel?

[max2k]

As it currently stands, at 4:15pm in Michigan, Channel A = 1750 watts (8 panels = 219 watts per panel), Channel B = almost exactly the same, and channel C = 2650 which is 12 panels or 220 watts per panel.

It must have been the playset slide reflection that caused the jump because it only lasted for about 30 minutes.. and by visual inspection, I'd say the sun was reflecting directly at the C channel array.....

I think sensij explanation is more feasible- you were simply observing inverter clipping behavior as it was above its total rated power, something gotta give. When clipping passed it returned to more expected values. Nothing to worry about.

[Murby]

I'm not sure you are looking at this correctly.

280 V / 8 panels = 35 V per panel
365 V / 12 panels = 30.4 V per panel

Total DC = 1303 + 1275 + 3644 = 6222 W

My guess is that the inverter is "clipping" inputs A and B, which drives up their voltage and lowers their current.

The fact that current C is right at the input limit (10 A) suggests there could a small amount of clipping there, but 30.5 V is not far from where the mpp voltage would be for those panels, especially on a cool day.

I don't know enough about that inverter to know the logic it uses when clipping is needed, but that it would treat inputs A and B differently than input C doesn't seem impossible, since those two inputs are allowed to have paralleled strings and the third input is not.

The 300 W per panel over 5-10 min periods (with even more dynamic spikes) is not impossible on partly cloudy days, where edge of cloud effects can drive up irradiance to 1200 W/m2 or more.

That is what I wanted to know.. I get the clipping part and I see it often.. But watching a 270 watt panel generate over 300 watts was a bit confusing.

[sensij]

That is what I wanted to know.. I get the clipping part and I see it often.. But watching a 270 watt panel generate over 300 watts was a bit confusing.

Does the SMA system log those string voltage and current values? I'm interested in the correlation between strings A and B, and whether there is occasionally more spread between them on cloudier days, during hours that the inverter is not clipping.

[Murby]

Does the SMA system log those string voltage and current values? I'm interested in the correlation between strings A and B, and whether there is occasionally more spread between them on cloudier days, during hours that the inverter is not clipping.

The SMA interface seems to be lacking features.... and it would seem its lacking data too..

Looks like I'm missing two days of data (missing blue bars)..

MissingDays.jpg