Differentiating between bin class variations and underperformance.

By: Adam Baker

Most people have never seen the incredible process of a solar module being manufactured. With a silicon module, conductive material is laid on top of the glass, and then a semiconductor is put down on top. Another conductor goes on and finally, the back panel that includes leads attached to the conductors.

Basically, it’s a highly automated manufacturing process…but module to module variability is surprisingly high.

These variabilities can cause some issues on solar sites, especially when it comes to trying to figure out exactly if your site is underperforming.

 

The PV module manufacturing processes

Even though it uses the exact same tools, the process might spit out a 215 watt module, and then a 240 watt module. The reason for variations are minor differences in the semiconductor layer (semiconductor purity), which have a significant impact on what the module can produce.

When a module comes off the end of the manufacturing process, it’s unknown how many watts it can produce until after a flash test. A giant light flash helps sense the output from the module to determine its exact rating (e.g., a 317.8 watt module).

Then comes module classification. Silicon modules are classified by 5-watt differences. 300 watts, 305 watts, 310 watts, etc. The bin class is the range from the lower number to the next highest number that module falls into, always rounded down.

So, our 317.8 watt module is actually classified as a 315 watt module A 319.9 watt module is also classed as a 315 watt module.

 

Bin class differences at a PV site

In a string of 315 watt modules, the odds are pretty good that none of them are actually 315 watts. They’re 316, 317.3, 317.9 etc.

Here’s where it starts getting interesting. When you string modules together (doesn’t matter how many), they all take the behavior of the lowest performing module. If a 20 module string has one module of 310.1 watts and the rest are 314.9 watts, guess what? They all act as if they were 310.1 watt modules. The random nature of module packaging means you should assume they will each output 310 watts, but you may get lucky and get a little extra energy if they happen to come out as a series of 314.5 watt modules. Just don’t count on it.

When constructing a PV solar site and placing modules in the field, nobody onsite can tell the exact wattage of any modules. Some will be on the low end and some on the high. Around 25% of sites are designed with two different bin classes (e.g., 8,000 – 320 watt modules, and 8,000 – 330 watt modules).

Great caution should be exercised by installers and construction managers to ensure bin classes are separated from inverter to inverter (at the very least combiner box to combiner box). Otherwise, the site could experience a significant decrease in expected output.

Similarly, if one module fails and must be replaced, it should be replaced with the same wattage bin class from the same manufacturer. If that bin class isn’t available, replace it with a slightly higher wattage from the same manufacturer.

Because the IV curve varies greatly from manufacturer to manufacturer, you shouldn’t mix and match modules in the same string.

 

Telling the difference between bin class variations vs. site problems

315 watts to 319.9 watts is a 1.5% variance between modules. That percentage is significant enough to show up on monitoring systems like SCADA.

The problem is, DC health issues like soiling, loose connections, and cracked cells show up in much the same manner as bin class variations. Just like bin class, small amounts of DC health issues like soiling on a module, or within an area of the plant, will affect the rest of modules connected in the series.

To identify these issues early to nip them in the bud, you should normalize current into inverter data as average current per string.

  • A 1.5% difference in DC current from one combiner box to the next can be attributed to module variation within a bin class. It’s just noise in your system, nothing to worry about.
  • However, if you have more than a 2% difference in average current per string, the problem goes beyond bin class. You likely have a DC health issue that could be improved upon to bump up overall site performance.

 

Need help identifying DC health issues throughout your site? We can help.