Also SDG&E customer and watching this thread
As for other thoughts - For utmost in theoretical efficiency, DC panel to battery makes sense. BUT, one was to consider higher voltage risk (and wiring cost), possible homeowners insurance implications, etc and panel placement options, and any chimney, tree, or other shade considerations. I could see a scenario where the DC to DC connection efficiencies are outweighed by cost and other factors. My main feedback is for OP to not assume avoiding micro-invertors/AC coupling) is best choice. I had wanted a DC coupled system for similar reasons, but in the end, an AC coupled system made more sense. it depends
It is my opinion that because of NEM 3.0, you should not sell to SDGE. You should only use your solar and batteries to self consume what you generate and avoid the high TOU rates. Furthermore, in San Diego county, there are several large corporately owned battery facilities being built, online already and planned where they charge up in the day at wholesale rates and sell back to SDGE at night when there is no PV so SDGE doesn't have to burn natural gas for power. Everyone wins except the rate payer in this system.
However, in regards to the above reply, I disagree. NEM 3.0 intentionally encourages (as it should, and I'm no fan of CPUC) time-shifting, as CA already has excess solar mid-day for many months of the year. Those of us old enough recall when time of use rates started and mid-day M-F was the most expensive... We all knew (or should have, if one was paying attention), that CA NEM was to encourage solar installs, which it was successful at, and that a further adjustment to NEM 2.0 was going to be eventually required. NEM 3.0 was late, imo, and I'm definitely
not going to say CPUC got it right, BUT the switch to self-storage and time-shifting kW to avoid grid consumption during peak hours was always the eventual target. Unfortunately, battery tech is taking longer than desired, and building codes (lumping all batteries into same risk category as Li-Ion) always take a years longer to adjust.
So, I'd argue, for OP and risk-averse spouse, and self-stated grid-connected full compliance setup, the target is best financial return, however that works out to be. Unfortunately, the ROI calculation got more complicated with NEM 3.0. But, we are still some time away from volume residential solid-state (or other longer life) battery technology, so over-buying on battery now just to avoid an occasional sell-back to local util co makes no financial sense to me. I suspect, but would require more info to confirm, that buying enough battery now to avoid Peak 4-9p TOU consumption/rates (accounting for recommended max charge and discharge settings (ie something like charge to ~80%, discharge to ~20%... or whatever specific battery chemistry and mfg recommendation is), would probably be most financially advantageous. Then, as battery tech improves and consumption patters adjust, possible electrification of natural gas appliances as well as home EV charging, combined with production and consumption monitoring from newly installed system, add batteries as it makes sense. A full PV residential lifecycle challenge is adding panels (from a non-DIY perspective) gets tricky (potentially problematic) as pre NEM 3.0, many solar installers I talked to couldn't be bothered to add panels to an existing system (much more profit in full system install). As inverters are expensive, I'm inclined to think over-sizing them may be appropriate as well, but I don't know reasonable life expectations for a quality hybrid inverter.. so that would be for someone else to chime in on
So, in a scenario where assuming paying a contractor for a full solar system (PV panels, inverter, etc), it may make financial sense to over-provision panels and inverter up-front, even knowing you won't get much, if any credit, for excess production *IF* one reasonably expects significant increases to electricity consumption in not-distant future years.
The 'trick' is to get your SDG&E kWh consumption data (simple download) for a year (probably in the 15 min increments), noting any significant changes during that time (ex adding an EV, heat pump, etc), and provide that to a knowledgeable solar installers who can analyze that consumption info and provide system sizing recommendations. Previously, one could make a good estimate based on a simple monthly statement and overall monthly kWh consumption data. Now, such simplistic summarized data isn't good enough, but the required data for system sizing is freely available. The other obvious challenge is finding the right contractor/solar installer. I'd think if solar sales rep isn't asking for such detailed consumption data, that would be a good sign company/sales rep isn't sophisticated enough to fine-tune system sizing.
I too am curious if OP opportunity passed? or still in progress? or system config selected and contracted for?