New PV Panels Making only 72% of Nameplate Rating

[MWeiss]

For background context, in April of last year, I installed an array of sixteen Sungold Power 450W panels (bottom photo, right hand two front most arrays).
I quickly found that the new panels were my biggest energy producers in all weather conditions. I measured the power to the batteries on the Growatt SC48120 controller and calculated that they were putting in 93% of nameplate rating.

By comparison, my old Trina solar 240W panels in two groups of 20 (top photo before upgrade) were making about 75% of nameplate. I was seeing 3200-3600W from that 4800W array. In addition, in Dec.-Jan., the bottom third of the Trina panels were in shade from the array in front of it. Those are on a Growatt SC4880 controller.

Naturally, I thought it would be a good idea to upgrade the Trina panels which were decades old (bought used) when Sungold had a 25% discount last November. I ordered a palette of sixteen panels, intending to swap out all of the Trina panels in two arrays.

Based on 93% expected output, I predicted that 8 550W panels (4400W nameplate) would make 4050W in practical terms. So I did the swap out 2 weeks ago. Second photo is the new array.

I was stunned that the new panels are only making 72% of nameplate rating.

I used a Fluke IRR-1 to measure irradiance at the peak solar hour for this array and it was 1027W/sq mtr. So based on that, I should be seeing 4400W output. However, I am seeing only 3200W and less. Furthermore, the overcast performance of the new panels was even worse, falling behind the other Trina panels by 34% on dark cloudy days. (160W vs 280W) My other Sungold array makes over 700W in the rain.

As such, I am hesitant to replace the other 20 Trina panels, as I would be losing some 10-20% of production.

When I built the 7200W array last year, they were the only panels that were connected to the SC48120 charge controller.

However, this array in question had the 20 Trina panels connected in a 5S 4P configuration for about 2 years. Then I shut down the controller, did the swap out and the new array is 4S 2P. At noon, I have 11 amps of current measured by current probe on each leg of the two parallel groups of panels. I should be seeing well over 13 amps. The open circuit voltage is lower than expected too, at 180V with the controller off. Should be close to 200V being each panel is 49.9V open circuit and I have over 1000W/sq mtr at the time of measurement.

I tried washing the new panels and checking all connections.

The next test was to disconnect all panels and test each one with a 2 ohm resistive load and measure current and voltage. The panels produced about 13.2-13.7 amps at around 27V each, individually. I know this is not maximum power point, but just to see if any panels are weak or underperforming.

On one hand, I'm pondering if the Growatt controller 'trained' itself on the old panels and is not adjusting to the new configuration, so not making optimal power?

I've got Sungold support on the case, but they're as baffled as I am.

PHOTO: OLD TRINA PANELS


PHOTO: NEW SUNGOLD 550W PANELS



PHOTO: RIGHT SIDE FRONT ARE FIRST SUNGOLD ARRAY OF TWO ROWS OF PANELS. BACK IS NEW REPLACED SUNGOLD 55O0W PANELS

 

[sunshine_eggo]

Irradiance is only part of the equation. CELL temperatures must be 25°C, and they never are given that the sun is heating them up.

When you have 1000W/m^2, expect 8-9% decrease on temperature alone.

Even so, I would expect more than 72%.

1. Is your battery charger in absorption phase and restricting current due to voltage?
2. Can you get the array to produce more if you increase loads?
3. Is the Growatt tracking to the series Vmp of the array?
4. Have you used a clamp DC ammeter to confirm each string is producing the expected current?

 

[MWeiss]

Oh yes, definitely I observed when the system needs maximum charging current. And for certainty, the wife was running a load of wash through the electric dryer, on top of other loads. When our battery bank hits 98%, the charge controllers start to taper off. These tests were done well before that point. And I have lots of data from the old array at the same time of day, showing that PV output is much higher with the old panels.

As mentioned, I put a clamp on ammeter on each of the two series strings and got 11 amps each, for a total of 22 amps going to the controller. My first thought was to expect one leg to be at zero, but not the case, unfortunately.

My first group of Sungold panels make the vast majority of the power out of all the groups I have. So I naturally expected an upgrade to their largest panels for my back array would give comparable results and outperform the old panels. I'm in disbelief that they are my WORST performing panels in reality. No idea if they're a bum batch of panels or my charge controller still thinks it's reading the old panels.

South array, old Trina panels:


South array after upgrading to new panels:

 

[sunshine_eggo]

1. Is the Growatt tracking to the series Vmp of the array?

How does 11A compare to the Imp value of the panels?

Tomorrow, try each string individually, note power, Vmp and Imp.

 

[yodamota]

Which version of charge controller do you have? the 150VOC one or the 250VOC

 

[MWeiss]

The max power point amperage is supposed to be around 13.5A.
The last sunny day we had, I disconnected every panel and put a 2 ohm load across one at at time and measured about 13.7A at 27V, +/-5% across all eight panels.
I have the 250 VOC Growatt charge controllers.

Interestingly, I got an e-mail response from Sungold offering me 20% refund today.

 

[yodamota]

could you put a link to the 550 panels you are talking about. how are they configured? 5s Xp?

 

[MWeiss]

These are the panels I bought from Sungold:

550 Watt Monocrystalline PERC Solar Panel

Highlight: ★ The long-lasting 550W monocrystalline solar PERC cells, the waterproof design ensures that the panel can be used in all weather conditions. ★ Solar panels are designed for industrial and professional applications using scratch resistant anodized aluminum with a twin wall frame. The...

sungoldpower.com

They are configured 4S 2P into a single channel Growatt SC4880 charge controller.

This is the e-mail I got back from Sungold today. My spidy sense is telling me that perhaps they aren't confident of the performance of these panels. What do you make of this:

"After reviewing your inquiry, we have investigated the issue further. While we have had other customers purchase the same models of solar panels, we have not received any similar reports of performance issues.

In light of the current situation, we would like to offer compensation equivalent to 20% of the total amount of your order, is that OK?

Additionally, we are taking proactive steps to ensure the quality and performance of our solar panels. We will send our solar panels to accredited institutions to undergo tests to identify any potential issues and ensure consistent performance.

We apologize for any inconvenience or disappointment this matter may have caused you, and appreciate your understanding and patience as we work to address it."

 

[sunshine_eggo]

13.7A @ 27V looks to fall right on the 1000W/m^2 curve of the panel's datasheet.

Again,

3. Is the Growatt tracking to the series Vmp of the array?

Tomorrow (now today), try each string individually, note power, Vmp and Imp.

What are your cell temps?

 

[plympton]

The 20% discount looks like a good-faith effort to make you happy, without having to process and take the expense of a return. The other stuff is perhaps them proactively wanting to see if your experience is a one-off or not - perhaps to help them decide whether to continue selling the product or not.

 

[MWeiss]

13.7A @ 27V looks to fall right on the 1000W/m^2 curve of the panel's datasheet.

Again,





What are your cell temps?

Cell temps reported by the IRR1 were 44.6°C.

Can you explain what you mean by 'tracking to the series VMP?" I suspect it might be a controller issue, as in it's been trained on the old array, but didn't retrain on the new one. I found nothing in the manual about this.

 

[sunshine_eggo]

Cell temps reported by the IRR1 were 44.6°C.

You're voltages (Voc/Vmp) should be about 8% lower than spec due to elevated temperature (above 25°C).

Can you explain what you mean by 'tracking to the series VMP?"

What I mean is that the MPPT should be running at or about the Vmp of the string when at max output - adjusted for temperature.

Your Voc is already disturbingly low, so I expect the Vmp is lower than it should be.

I suspect it might be a controller issue, as in it's been trained on the old array, but didn't retrain on the new one. I found nothing in the manual about this.

MPPT shouldn't be "trained" on anything but the array they're connected to UNLESS you can alter the algorithm. Example is the Outback FM80. You can set a "target" % of Voc for the MPPT, and you can also limit the search range though this method of specifying % vs. absolutes has a low risk of problems when changing from one array to another as PV panels all operate very similarly.

With each change in PV conditions, the MPPT algorithm should find the MPP for the attached array.

Tomorrow, try each string individually, note power, Vmp and Imp.

Did you do this? By that I meant run each string individually and see how they performed. If ONE of your panels has an issue or a failed diode, it might drive the array voltage down and further reduce performance. Isolating the strings might give you a clue... or eliminate it as a possibility.

 

[MWeiss]

Oh "duh!", voltage at max power point. I understand now. I'll measure that again when the sun comes out again. We're in a wet rainy period for the next few days.

I misunderstood. I did measure the amperage of each 4S leg with a clamp on ammeter. I did not try disconnecting one leg and seeing what power the controller makes nor the Voc of Vmp readings of each individual string. I will try that next.

What could account for the low Voc? With no controller load, the Ir loss in the 8AWG wire is nil. I should see the full voltage at the controller.

My expected 4050W of real output takes into account that 8% drop from 4400W rating for the array. Still, I'm seeing 3200W. Something's very wrong with this.

Also, my Trina panels in the other array (yet to be replaced) are making 266W in the dark cloudy rain and the new Sungold panels are making only 160W at the same time. Flat sky lighting conditions. Used to be both arrays made about 250-270W in these conditions. So not only are the new panels not making rated power at full sun, they are that much worse at low light than the old panels. I'm dumbfounded that the same brand of panels in a 550W size are so much less performing than the 450W panels of the same brand I bought the year before. Those 450W panels are my best performing array, making over 700W in the rain and 6680W (controller clipping) in full sun.

 

[sunshine_eggo]

Oh "duh!", voltage at max power point. I understand now. I'll measure that again when the sun comes out again. We're in a wet rainy period for the next few days.

I misunderstood. I did measure the amperage of each 4S leg with a clamp on ammeter. I did not try disconnecting one leg and seeing what power the controller makes nor the Voc of Vmp readings of each individual string. I will try that next.

Just to eliminate the possibility of one string interfering with the other.

What could account for the low Voc? With no controller load, the Ir loss in the 8AWG wire is nil. I should see the full voltage at the controller.

44° cell temp would lower it by 8% from spec, but 180 is about 14% below.

My expected 4050W of real output takes into account that 8% drop from 4400W rating for the array. Still, I'm seeing 3200W. Something's very wrong with this.

Agreed.

Also, my Trina panels in the other array (yet to be replaced) are making 266W in the dark cloudy rain and the new Sungold panels are making only 160W at the same time. Flat sky lighting conditions. Used to be both arrays made about 250-270W in these conditions. So not only are the new panels not making rated power at full sun, they are that much worse at low light than the old panels. I'm dumbfounded that the same brand of panels in a 550W size are so much less performing than the 450W panels of the same brand I bought the year before. Those 450W panels are my best performing array, making over 700W in the rain and 6680W (controller clipping) in full sun.

Something definitely seems off... they're bifacial too, so you should be getting at least a little boost from the backside!

 

[yodamota]

it looks like the VMP of the panels is 41.96 v.

What about shorting out the panels to each other, instead of thru the resistor? than measure with a clamp meter.

what voltage are you seeing at the controller?

heres the specs on the panels put into the midnite calculator

Do you have another charge controller to test the output of the panels on?

 

[MWeiss]

The panels are NOT bifacial, just to be clear.

I haven't tried measuring short circuit current. I can do that next sunny day though.

I'm seeing 180V at the controller with the controller off and not loading the array. If I turn it on, the voltage drops quite a bit and is varying constantly.

Those specs sound about right based on what I expected.

I bought spares of everything. I believe the spare charge controller is an SC48120-MPV. Two 60A channels.

 

[sunshine_eggo]

The panels are NOT bifacial, just to be clear.

Sorry. Not sure where I got that.

I haven't tried measuring short circuit current. I can do that next sunny day though.

Verifying everything you can within reason helps.

I'm seeing 180V at the controller with the controller off and not loading the array. If I turn it on, the voltage drops quite a bit and is varying constantly.

That's what my Vmp does with power ramping up in the morning.

 

[MWeiss]

That chart seems backwards. The current nosedives and the voltage rises as the sun comes up? That looks like my end of day readings. But actually, the voltage doesn't really jump up on my system as the sun goes down.

 

[sunshine_eggo]

That chart seems backwards.

That's morning. Note that I hit absorption at 9am.

The current nosedives and the voltage rises as the sun comes up?

Those are MPPT sweeps where the MPPT tries to pull more power causing voltage drops and increased current. They can happen as often as every 10 minutes, and VRM logging at 1m intervals doesn't necessarily catch them.

That looks like my end of day readings. But actually, the voltage doesn't really jump up on my system as the sun goes down.

If you're still pulling max available power at end of day, you're still looking at that point being very near temp adjusted Vmp.

I didn't make it clear. I was just trying to show you what my Vmp does. It's a little wiggly, and As temps rise, Vmp drops.

 

[MWeiss]

Mine does this in a more gradual manner. But mostly, more sunlight, more voltage and current displayed on the MPPT controller. At end of day, the battery charge is dropping off due to not enough PV input to match the static load of the house + inverter waste energy (about 1500W).