Wiring Non-Identical PV Panels Michael S. Elliston and Tom Stockebrand c1992 Michael S. Elliston and Tom Stockebrand At Carrizo Solar we are frequently asked the best way to wire our used ARCO 4-volt M52L panels in large arrays. This question is important not only with our panels but any non-identical panels connected in large arrays. For small systems you may not have any wiring alternatives but you will for larger arrays. There are two ways to wire panels for large arrays. Series-Parallel The first way is to wire the panels in series banks, then put the series banks in parallel. I will call this the series-parallel wiring. For example with 16 4-volt panels charging a 12 volt system this wiring would look like this: DIAGRAM INSERTED HERE Parallel-Series The second way is to wire your panels in parallel groups, then wire these parallel groups in series. I will call this parallel-series wiring. For example with 16 4-volt modules charging a 12 volt system, this wiring would look like this: DIAGRAM INSERTED HERE Non-identical PVs If all your panels were identical, the two wiring methods would not result in any major differences except for the number of connections. However wiring does become an issue when you are working with non-identical panels such as our M52Ls. Although they are graded into power categories, the amperage of the M52L will vary within each of these categories. Why is this a problem? It is because when you series connect strings of panels, the current in each string is limited to the current produced by the lowest producing panel. It is just like a chain that is no stronger than its weakest link. How does this make a difference between the two wiring methods? DIAGRAM INSERTED HERE For illustration each figure shows 8 panels varying in current carrying capacities from 5 to 6.75 amps in 1/4 amp steps. Figures 3 and 4 show series-parallel wiring while Figures 5 and 6 show parallel-series wiring. High-Low Pairing The array in Figure 3 yields the lowest output current since, unluckily, the two lowest output panels are in separate strings limiting both strings to the lowest possible current, 10.25 amps. There is a 50/50 chance of this happening. The best that can be done would be to have all the low panels in one string and high panels in the other string for an output of 10.75 amps as shown in Figure 4. If the same panels are connected with parallel-series wiring, as shown in Figure 5, then the worst case occurs when the two low panels are in the same pair as shown. The chance of this occurring is only one in six. You have a much higher probability of low output using series-parallel wiring than with parallel-series wiring. Note that if you could test each panel, you would want to pair the highest with the lowest, next highest with next lowest, etc. This technique equalizes the current in each pair and gives you the peak performance as shown in Figure 6. Testing the Bronze To verify this practice, we took 16 of our Carrizo Bronze modules and wired them as shown in Figures 1 and 2. We performed two tests on two separate days using two different sets of Bronze modules. The tests were conducted at our plant in the Carrizo Plain. It is located at 2000 ft. above sea level, half way between Bakersfield, CA. and San Luis Obispo, CA. The panels were mounted on adjustable rack and pointed directly at the sun and allowed to heat up. The cell temperatures were recorded from a probe taped to the back of the panel. We use a PVI Inc. curve tracer to record the power output. In June 1990 this machine was checked against Southwest Institute's curve tracer and found to be about 10% lower. The global normal insolation and horizontal global insolation were taken off pyronometers located in the middle of solar field a quarter mile to the south of the test site. First Test First test was conducted on October 29, 1991. The panels were wired series-parallel first and readings were taken at 12:30 P.M. The panels were then rewired parallel- series and readings were taken at 1:30 P.M. The weather was clear with occasional breezes. TEST RESULTS INSERTED HERE Second Test The second set of tests was performed on November 12. A different set of Bronze panels were used. The set up was the same as the first set of tests. The panels were wired series-parallel and readings were taken at 11:30 A.M. The parallel-series readings occurred at 12:30 P.M. The weather was clear with a slight haze low in the sky and no wind. TEST RESULTS INSERTED HERE As you can see our tests confirmed that wiring in parallel-series will produce higher power. Our estimate is from 10% - 20% more after compensating for temperature and insolation. Conclusion Interpanel variations arise whenever you build an array. New panels added to an existing array, even if new panels are from the same manufacturer as the old ones, will be different, having come from different lots. Also the old ones will probably have degraded with time. Used panels have a much greater variation in current output from panel to panel but they are often combined since they are a low cost solution. Panels of entirely different current ratings can also be combined using this principle as long as the panel output voltages are similar. Although you will have more connections to make wiring your panels parallel-series it appears this is more than offset by the gain you receive in power. For 4 volt ARCO M52L this analysis holds true for any combination of 4 charging a 12 volt system or any combination of 8 (or 7 depending on climate and panel rating) charging a 24 volt system. Access Authors: Michael S. Elliston, President Carrizo Solar Corp., 800-776-6718 and Tom Stockebrand, PE LGK Corporation, POB 10239, Albuquerque, NM 87184-0239 USA