Years ago, shortly after putting solar panels on my house, I ran some tests to see if cleaning my panels of pollen made a difference. At the time, I found minimal difference in the output of my entire system so I concluded that I didn’t really have to worry about keeping the panels clean. Plus, most of the on-line info I found supported this, stating that rain naturally washes away most of the debris that settles on the panels.

However, this past year, I upgraded my system with solar optimizers. Due to tree growth, my array was having more problems from shading, causing the entire system performance to degrade. Optimizers allow each panel to operate more independently, reducing the negative effect of shading on one panel. In addition, the optimizers were connected to the internet, allowing me to monitor the output of each individual panel. Most modern systems provide this capability. It’s an extremely powerful troubleshooting tool.

As an energy geek, I wanted to see how my panels were performing after installing the optimizers, since I paid a good chunk of change for the upgrade. Sadly, when I compared my system output to past years, I didn’t really see much, if any improvement! Why would this be?

This discovery took me down the rabbit hole of energy analysis. Before the optimizers, I could only see the total energy produced by the array, so I didn’t have a basis for comparing one panel’s output to another. However, by viewing the production at noon, when all the panels were fully illuminated, I could eliminate the effect of shading and compare full system output from previous years. What I found was upsetting – every year, the output from the system diminished measurably. This was both total output and peak output (the power at high noon on a clear day). Were my panels “degrading?” Or, was something else at work?

Fast forward. I was recently talking with my brothers about their solar installations. Brother Chris noted that his newly installed panels were getting coated with a black substance that was coming from a nearby oak tree. He’d come up with a cleaning formula that worked to remove the gunk from the panels. But we all wondered how much of an effect this tree-scum had on our solar systems.

When I returned home, I was reminded of the big oak tree that had grown a branch that hung over my panels. I climbed up on my roof to inspect my array and was horrified to discover my array completely covered by this hard, black substance! The newest panel which was installed just a couple years ago as a replacement for one that had gone bad was the worst – it was totally black!

Clearly, I needed to clean the panels! Yuck.

This raised two questions:
– how much of an effect does this have on the panel’s output?
– why the heck was the newest panel so much worse than the older panels?

I don’t have an answer for the second question. The only thing I can guess is that the new panel has a different glass surface or coating than the old ones. It’s unfortunate, but it’s obvious, the newer Sunpower 230W panels have to be maintained much more frequently than the older ones.

How much of an effect does ‘dirt’ reduce the panel’s output?

Take a look at the power from my panels at noon, before cleaning. I picked a day where the output was maximal, indicating full sun and optimal generation conditions for this so as to be fair.

The substantial variation in the individual panel’s power is a huge red flag. Either a panel is going bad, or the panel is not getting as much energy from the sun. If it was just a single panel, then I’d think the panel was defective. But in this case, there’s a wide variation in power from the panels on the left to those on the right.

Guess where my oak tree is?

Sure enough, the panels on the right (east) side of the array are the ones that are nearer the giant oak tree. And as you can see from the photo, they’re really dirty! Compare the output of the panels on the left (west) side of the array with those on the right – that’s serious performance degradation!

Since my roof is relatively flat and low, I scurried up and cleaned some panels. Amazingly, the dirty panel in the photo at the top of this post jumped up over 200W production after cleaning! I cleaned it more a day later because I wasn’t able to get all the scum off the panel. It is now producing 236W. Pretty amazing since the panel is rated at 230W. I guess they’re conservative in their ratings.

Improvements from other parts of the array weren’t as impressive, but after spending many hours cleaning, the average peak panel output of the panels I cleaned increased by 19.4%. That’s huge.

Note that I hadn’t completed cleaning when this snapshot was taken. You can see panels B8, C1-C5, F1, D5-D8 and E1 haven’t yet been cleaned.

Note: I will not go into details regarding cleaning the panels. It’s dangerous to climb up on the roof, even a low slope roof like this one. There are plenty of other sites that you can refer to for more information. My suggestion is to leave it up to the pros!

I am now a firm believer in solar panel cleaning! Had I not had the discussion with my brothers, I many have still been in the dark about why my solar system’s output was so disappointing this year. Now I know – tree scum kills panel performance! With just a couple of days of work, I was able to improve the system’s performance by about 20%. That’s amazing and was well worth it. Even better, I didn’t fall off the roof or break any of the panels in the process 🙂

A few final notes

You know the saying “the devil is in the details”? While cleaning and checking panel output, I noticed something odd – one of the high-output panels’ production had dropped substantially. I thought I knew why….
A quick trip to the roof confirmed my suspicions

Usually, they move vent stacks to prevent this type of problem. Unfortunately, this one couldn’t be moved without tearing apart the roof. Fortunately, we have the optimizers, so one panel with diminished output doesn’t kill the bunch. But it does show how important shading is. Even though just a tiny fraction of the panel is shaded, the entire production is reduced. In this case, by 30%.