Advice sought to upgrade old PV system

[cpa]

Thanks in advance to all who respond!

We currently have two PV systems on our roof. The original installation was in 2011. It consists of 14x185 watt Suntech panels with a Sunny Boy 3000 string inverter in our garage.

After we purchased our Model S in 2014, we added 7x270 watt (unsure of the American manufacturer) panels with micro inverters.

I have been mulling whether or not to increase our PV generation because we bought our Model 3 in January, and we will soon be dragged kicking and screaming into TOU metering. Those of you familiar with the Central Valley realize that we have very hot summers. In fact, we are in the midst of I-don't-know-how-many consecutive days >100 degrees. It seems like close to 15, with no end in sight. But I digress.

The 14x185 panels get varying amounts of shade from about 3:00 onwards starting around the autumnal equinox. Generation drops from ~1,200 watts to ~350-400 watts in about thirty minutes due to the shading and series-wired panels going into the inverter. I estimate that we lose a good 2.5kWh per day come the first day of fall because of shading.

A local solar company suggested that they replace the string inverter (he said that after seven years they begin to lose a lot of efficiency) with micro inverters for the 14 panels. The micro inverters would be more efficient in and of themselves, plus they would also generate more electricity because there would no longer be the limitations inherent with installation in series.

I ask: Is this a sensible approach? Are there better options out there?

Thanks much again!

 

[SageBrush]

AC or DC micro inverters ? The former have higher resistance losses due to the high currents when all panels are producing.

A photo of the shade would be most helpful. You can get one now from a solar pathfinder.

 

[nwdiver]

Is this a spare no expense type scenario or are you looking for something that's cost effective?

 

[cpa]

The micros are AC per the company.

Diver, this is not a spare no expense situation. Our true-up this past year was $377. I anticipate that this will increase by about $100-$150 now that we have a Model 3 in addition to the S. (My wife does not drive much these days.) I do not know how much the forced switch to TOU billing will have on us due to the peak period time frames. Overnight lows of 73 and highs of 103+ with no breeze makes for early AC usage!

I felt that a grand or so to increase our output from the PV panels would likely offset the additional usage by the 3; that amount amortized over our time in our home would make it sensible.

I appreciate your questions, Sage and NW.

 

[nwdiver]

Diver, this is not a spare no expense situation. Our true-up this past year was $377. I anticipate that this will increase by about $100-$150 now that we have a Model 3 in addition to the S.

Do you know how your panels are wired on the string inverter? How old is the string inverter? I don't buy the 'less efficient' argument. Some efficiency might be lost but it would largely be negligible. If you have 2 or more strings it might be worthwhile to replace the old string with a new string that has a separate tracker for each string, new inverters are ~97% efficient older models were ~95%. The new inverters run ~$1k for 3kW to ~$1500 for 7.7kW. But... no matter what... micros or a new inverter I doubt production would improve by >10%.

I don't think micros are a cost-effective approach to generate more energy if it's current on a string inverter. But if you go through the trouble of adding micros might as well swap out the panels with higher wattage versions....

Do you have room for more panels? Depending on how many panels you have you could buy a SMA 7.7kW inverter with 3 trackers. Put the existing panels on one tracker than add more to the other 2 trackers. Each tracker is independent so you can mix and match.

 

[cpa]

The string inverter is 7 years old. Don't know how they are wired, but I think that there is but one string for all 14 panels.

Again, it is the shading from mid-September until the leaves drop from the deciduous trees (usually around mid-December) that curtails a lot of our afternoon production. When production on the 14 panels drops from 1,200+ watts to 350 watts in thirty minutes at 2:30-3:00, (earlier by November 1) it is because of the trees. Sunset is still three hours away when production drops. As I said above, I estimated that we lose ~2kWh per day due to shading in the fall. We also lose during spring when we have (gasp!) clouds in the sky that will sometimes cloak one panel entirely, although the cloud cover is transitory.

If we did not have the shading issue, I would just keep the old Sunny Boy 3000 until it crapped out. But I think we are losing out on ~150+kWh per year; maybe more due to the shading of the 14 panels.

 

[nwdiver]

The string inverter is 7 years old. Don't know how they are wired, but I think that there is but one string for all 14 panels.

Again, it is the shading from mid-September until the leaves drop from the deciduous trees (usually around mid-December) that curtails a lot of our afternoon production. When production on the 14 panels drops from 1,200+ watts to 350 watts in thirty minutes at 2:30-3:00, (earlier by November 1) it is because of the trees. Sunset is still three hours away when production drops. As I said above, I estimated that we lose ~2kWh per day due to shading in the fall. We also lose during spring when we have (gasp!) clouds in the sky that will sometimes cloak one panel entirely, although the cloud cover is transitory.

If we did not have the shading issue, I would just keep the old Sunny Boy 3000 until it crapped out. But I think we are losing out on ~150+kWh per year; maybe more due to the shading of the 14 panels.

Do you know what voltage your SB is operating at when there's shading vs no shade? There was a long and unfortunately rather tribal debate over the effect of shade. If you have a single string and your panels are working properly the shade shouldn't have much of an effect vs micros. There are 'bypass diodes' in the panels, these allow current to shunt around shaded sections. It's possible you have failed diodes, if there are diodes that fail to shunt current this would cause a DRAMATIC voltage drop when the panels start getting shaded.

I don't think there's any solution to get more energy that would provide a payback of <20 years...

 

[SageBrush]

But I think we are losing out on ~150+kWh per year

That is the generation of about 1/3 of a panel = 100 watts, and would cost around $50 these days.
Any chance you can just add a panel and call it a day ?

Or swap out a couple of the old panels for new ones at the end of a row ?

 

[Uncle Paul]

An alternative to more solar panels is sometimes a whole house fan.

The Quiet Cool type fans are designed to run at night to cool the house by bringing in fresh cool air at in the evenings.

Cost about $1,500 installed.

If you can get your home cooled down at night, then the air conditioning will come on much later in the day.

Out of the box thinking.

 

[cpa]

An alternative to more solar panels is sometimes a whole house fan.

The Quiet Cool type fans are designed to run at night to cool the house by bringing in fresh cool air at in the evenings.

Cost about $1,500 installed.

If you can get your home cooled down at night, then the air conditioning will come on much later in the day.

Out of the box thinking.

Uncle, good suggestion. Alas, at dawn the ambient air temperatures are between 72-76 degrees when we get into these dastardly hot periods like now. Around midnight the temps are usually in the mid 80s We have enough fenestration that I can open up the house around 4 AM to get whatever cool air is available. May-June and late August-September the nights are in the mid 60s or lower, and the house gets cooled off enough so that many days we do not need to turn on the AC.

 

[cpa]

Do you know what voltage your SB is operating at when there's shading vs no shade? There was a long and unfortunately rather tribal debate over the effect of shade. If you have a single string and your panels are working properly the shade shouldn't have much of an effect vs micros. There are 'bypass diodes' in the panels, these allow current to shunt around shaded sections. It's possible you have failed diodes, if there are diodes that fail to shunt current this would cause a DRAMATIC voltage drop when the panels start getting shaded.

I don't think there's any solution to get more energy that would provide a payback of <20 years...

I would have to wait until the shady time of year. I will have to check on that.

Right now here is the reading at 1240PM PDT with the temperature right at 94 degrees in the shade, so the air temperature is much higher.

L1 = 123 V L2 = 124V

VPV = 202V

Don't know what this means, but those are the amounts from the display.

 

[nwdiver]

VPV = 202V

14 panels in series and you're getting ~202v? You probably have 2 strings of 7 or something is wrong.... if you have 2 strings of 7 there would be some benefit to upgrading to a 2 MPPT inverter...

Is this what you have?

You're likely operating outside the MPPT band of the inverter. The old SB3000 has a band of 200-480. The new inverter is 155-480. This can lead to ~10%+ losses too. I think upgrading the inverter could yield ~15% more production annually...

I underestimated the effect of low DC voltage on a ~13kW system I designed. Thought it would settle at 350v if the MPP was 320v because of the heat (band is 350-480). Nope It hunts. Voltage swings from 300 - 400... Production loss is ~10%...

This should be a bell curve.... Your system is probably doing the same thing....

 

[cpa]

The panels per the certificate are Suntech STP185S-24/Adb+, so you are pretty much right on.

Thank you for the graph. You may very well be right. I have no way of graphing the output with this particular model.

What is difficult for me to understand is the loss. Our high temperatures make generation less efficient. The angle of the sun has an effect. It has always fascinated me that during cool weather in the spring with a nice breeze about, those panels have generated over 2,600 watts around noon. On the other hand, during our 100+ degree days with zero breeze output can drop to 1,800 watts or so at the same time.

I will figure something out eventually. Perhaps I shall wait until fall to show the smart people just how much the shade reduces output.

 

[nwdiver]

The panels per the certificate are Suntech STP185S-24/Adb+, so you are pretty much right on.

Thank you for the graph. You may very well be right. I have no way of graphing the output with this particular model.

What is difficult for me to understand is the loss. Our high temperatures make generation less efficient. The angle of the sun has an effect. It has always fascinated me that during cool weather in the spring with a nice breeze about, those panels have generated over 2,600 watts around noon. On the other hand, during our 100+ degree days with zero breeze output can drop to 1,800 watts or so at the same time.

I will figure something out eventually. Perhaps I shall wait until fall to show the smart people just how much the shade reduces output.

I would enter you info into PV Watts. The orientation and angle of your array. See how your actual production compares to modeled production. That's your best bet at determining your true loss.

The most cost effective option is likely to wait until the SB3000 craps out then get the latest and greatest. Even if you increase generation by 500kWh/yr (unlikely) that's still a >10yr payback...

 

[cpa]

Pretty nifty website, NW! Thank you!

 

[Anzir]

Does PV Watts include expected production decrease from weather? According to its calculations I should be producing 2.1MWh/month on a 17kW system. I am 5 days short of my first month of production and I am at 1.67MWh as of 11 am today. But it has been rainy as heck the past 5 days. Decreased average daily production from 75-80kWh down to 61 or 62kWh. Just curious. I was just visiting California last week and I was envious of the cloud-free days.

I would enter you info into PV Watts. The orientation and angle of your array. See how your actual production compares to modeled production. That's your best bet at determining your true loss.

The most cost effective option is likely to wait until the SB3000 craps out then get the latest and greatest. Even if you increase generation by 500kWh/yr (unlikely) that's still a >10yr payback...

 

[nwdiver]

Does PV Watts include expected production decrease from weather? According to its calculations I should be producing 2.1MWh/month on a 17kW system. I am 5 days short of my first month of production and I am at 1.67MWh as of 11 am today. But it has been rainy as heck the past 5 days. Decreased average daily production from 75-80kWh down to 61 or 62kWh. Just curious. I was just visiting California last week and I was envious of the cloud-free days.

Yes. For a typical meteorological year. If it's abnormally cloudy then your production would be abnormally low. It's a fairly complex model. It also takes temperature into account.