sahil

Ya. that might be the figures we use to convince the banks to lend us money you are right about the ideal conditions. every Wh is sold to eskom, the plant is built with optimised geometry on an optimised site, and the panels are cleaned quite regularly (not sure how often, because we also have to respect water consumption)

read the wiki on Capacity Factor. I think the utility-scale operators use around 25% for our latitude?longitude? over a year, spring, summer, autumn, winter, rain, shine, hail etc you should see 4panels x 450W/panel x 24hrs x 25% = (4 x 450 x 6) Wh derate that for panel ageing, shading, air pollution, panel cleanliness, and other effects in your area

Not an inverter engineer, but can share some anecdotal evidence from a few years ago. I was commissioning some IT-type servers in a solar farm under construction before the grid tie-in was done. The site manager issued me a small diesel genny to run the servers for some software checks. We did notice that the frequency and voltage were poorly regulated until the load was increased. After we plugged in a 2kW heater the genny was able to properly regulate. Thought / brain-fart: maybe when your house load is small the inverter can't regulate nicely? Try increase your load a lot and see if the problem is alleviated. Maybe also the new firmware will improve the controllers' performances A bit of a 1000m view of voltage regulation (based on XP in 50-100MW solar and 500+MW coal generation (I'm left Eskom 10+ years ago - don't shoot me)). So Voltage is a function of reactive power, which in turn is a function of the phase difference between the current and the voltage sine waves. which is then a function of the firing angle of the thyristor stack. It seems to me that the voltage regulator is not able to find the right firing angle at low loads and keeps over-compensating. Because part of this game is played in polar co-ordinates (not the normal x-y axes) it means that some of the functions have some trigonometry involved (hope you paid attention in high school :D). If the non-linear behaviour is not properly taken in account the controllers can be badly behaved at operating points other than those for which they were optimised (high load vs low load). Thence my brain-fart: increase load or fix the controller Note that grid tied inverters don't suffer these because if they did they would need to move the municipal voltage. Imagine you 6kW machine trying to control the voltage for your whole street... suburb... town... city... The other interesting thing is that this seems to happen only on PV, but not on battery. Another idea may be that the same microcontroller is running the MPPT and AC regulation functions. When both are needed, then there is simply not enough processing power to do both jobs. Finally I want to re-iterate that I am not an expert on inverters and how they work so the above is just spit-balling. I am happy to be corrected