Opinions are like *ssh*les, everone has one... So a bit long winded but IMNSHO,
The cheapest/fastest way to
implement working Solar is on-grid with micro-inverters, depending on what your grid connection costs are (control boxes, electricians, inspectors, permits, etc). In this way you just dump all your excess solar output to the LPC when your demand is low, and the LPC buffers your demand when it exceeds your output. That being said, cost savings depends on how much you can lower your demand from the local power company, and how much they may give you for what you push to them. It's actually pretty good here in Arizona.
The minute you push batteries in the mix, if you still want to grid-tie, you may as well get one of the newer all in one's and skip the micro-inverters, just run your panels in series and get string voltages up. This really only helps you flatten your peak demand slightly unless you get enough inverter output to run your whole house. At this point, you can use PV to charge your batteries when your demand is low, and batteries to to cover when your demand is high. This is where I think it get's interesting, and less useful to pay for the additional expense of grid-tie & sell back.
I'm literally putting panels on the roof today/this weekend. It's 108F (3pm) so I quit until later. I will have 14KW of panels hooked to a pair of EG4 18K's. I have 30K of battery, the "18"K is PV input, they each have 12KW of output. I've been load testing, the setup struggles a bit starting the 4T A/C but with the soft-start kit, it just blinks, instead of browning-out for 500ms or so. Note: Grid tie would eliminate that issue. The power company has enough reserve joules, to keep the voltage from dropping too low.
To start with I will not be tied at all. I have enough 'solar/battery generation' ability to output 100A. So for now I have moved everything (All my loads) to a 100A (125 de-rated) panel and I have a 100A transfer switch with feeds from the grid, and from my solar inverters. The goal is to remove all peak usage, that is between 1300-2000 M-F, when in June and July the cost is a little under $0.25/kwh. Off-Peak is a hair over $0.08/kwh. To the point of the other poster, this is really a hobby. But in this case, I will just take myself off-grid as long as I have enough juice. Based on my testing, I can run 4 full hours from the batteries only (Charged off-peak from the grid
![Smile :) :)](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
). My mean load is around 5kw, so at some point in the evening the PV output will drop below 5KW, and I'll have at least 4 hours before I need help from the grid, and that will definitely be well after 8PM, we have some sun until 7PM or so.
Now these 18K's fully support grid tie, and I MIGHT do something with that, but some things must happen before that is cost-effective. Mostly I'll need to actually produce electricity that I am not using, because I don't need a buffer. Pushing my own electrons at my loads is always going to be more cost effective than selling it to the LPC. 14K of panels with about 6 hours equivalent full sun/day is 84KWH at best. Last summer this time my daily usage was closer to 105kw/day based on my bill. That puts me about 20KWH short at best, reality is things are likely to be significantly lower, how much lower I will be finding out shortly.
For me, grid-tie/sell back woutld be after I double my panels and batteries, just to cover the meter charges and whatnot, so I can keep it as a backup. ie get my electricity cost to 0. This is going to be interesting. My July/Aug bills last year were very close to $400. The Winter peaks are 0500-0900,1700-2100 so those are going to be more interesting. I have an SBC running a bunch of sensors and relays to keep my peak demand under 100A. I will have to tie it into the inverters for a bit finer control come November. First upgrade is obviously more panels. Having Fun!