Home Photovoltaics Xantrex Inverters SMA Inverters After 7 Weeks After 18 Months

SMA Inverters


The charts on this page were derived from kilowatt-hour readings indicated by the respective inverters. The Xantrex inverters present readings on a local LCD screen. The Sunny Boy readings were collected via RS-485 data communication interface. These charts are size-reduced for fit. Click on any chart that you'd like to see more clearly.

Xantrex introduced an improved version of the STXR2500 in January, 2003. This model provides greatly improved performance that compares well to that of the Sunny Boy.


After two disappointing months of service with my Xantrex inverters, I acquired a single SMA Sunny Boy inverter to use for comparison. The results were encouraging - in fact downright exciting. The chart below shows the percentage improvement provided by the Sunny Boy. The biggest gains occurred on days with passing clouds. Days like this emphasize the meager MPPT algorithm of the Xantrex inverters.


Below is a one-month chart of the daily energy production from the four Xantrex inverters. By this time I had replaced the three original ST2500's with the newer STXR2500 model. The ST2000 was supposed to be an ST2500 but the wrong model was shipped. The PV clusters 1, 2, and 3 each consisted of 20 panels. Cluster 4 was only 12 panels so I normalized the energy output by a factor of 1.67 for comparative tracking. Its interesting to see how much less was produced by the ST2000 on cluster 1, even though it was fed from an equal number of PV panels.


Next the enemic ST2000 serving cluster 1 was replaced with a Sunny Boy SW2500U inverter. The three STXR2500 inverters remained on clusters 2, 3 and 4. Again the cluster 4 data is scaled by a factor of 1.67 for normalization.

Interestingly cluster 1 with the Sunny Boy shows no sign of substandard performance, which removes any capacity doubts about the PV panels of this cluster.


The output of our PV plant is metered by a single kilowatt-hour meter which is considered the standard reference. The summation of the inverter indications is continuously compared to this meter. The chart below reveals an interesting phenomenon.

This chart's time scale covers four phases of a 4-1/2 month period. First there were three ST2500 and one ST2000 inverters, then the ST2000 and three STXR2500. When inverter comparison began there was a period when one of the Xantrex inverters was replaced with a Sunny Boy, and finally the current configuration in which all of the Xantrex inverters were replaced by three Sunny Boys.

The chart shows the deviation between the system reference kilowatt-hour meter and the summation of values reported by the inverters. During the first phase, the inverter-reported values agreed closely with the reference meter. This agreement scattered noticeably as ST2500's were replaced with STXR2500's, however the mean error was not far from zero.

However as the Xantrex inverters were replaced with Sunny Boy inverters, the error makes a clear positive excursion.

From this it can be concluded that in my installation the Sunny Boy inverters report optimistic values for energy production, approaching an error value of about +3%.