Hello, I ran a simulation with a centralized inverter and DC optimizers. in the loss diagram, the mismatch loss (modules and strings) is close to 0%. that is fine. I ran a similar simulation with string inverters instead and still DC optimizers. in the loss diagram, the mismatch loss (modules and strings) is higher than 3%, as if there is still a module mismatch ... I don't understand why? Could you help me? thanks