All DC solar system design, unworkable?

[niwatori1]

Howdy all!
I'm in the initial design/ planning stage for an off grid solar system and I had a design idea for an all DC system
I've included a basic schematic outlining my main overview.
Essentially it's a system designed to be current limited by the solar panels themselves.
The battery, and conductors are all intended to be able to handle the solar panels at full output.
system voltage is intended to vary from the open circuit voltage of the panels(full sun,no load, fully charged battery) ~46VDC
down to the lowest safe discharged voltage from the battery(night, full load, discharged battery) ~34VDC?
once battery cutoff is reached a 120VAC to DC PSU will take over feeding DC loads(Not in schematic, listed as control outputs on the BMS)
Why the low system voltage? anything below 48V in the US is considered "Low Voltage" by the NEC and the standards are MUCH simpler for custom systems.
Why Off grid with a grid transfer switch backup? I live in a rural community that uses tiered billing for power. grid tie systems do not make sense in those areas.
Also grid tie involves the city power dept. and getting those approved is nearly impossible. I can think of one so far.
Lets start with potential problems I can already see.
1.I want to use LIPoFe4 chemistry, even with a BMS I'm pretty sure a lack of a charge controller isn't going to do well for the battery.
2.The battery NEEDS to be custom built and sized for the solar panel output voltage(Fully charged at or above panel OC Voltage). Doable, but not easy.
3.Changing or upgrading panels can be a PITA, do all the voltages match? Degradation? Etc...
4.the low system voltage means that the battery charge current and the ampacity of all main conductors HAS to be HUGE. not unfeasible, but difficult.
So....What's the good news?
1.lower system cost(potentially)no need of micro inverters or battery controllers, just a bms, some diodes and fuses.
2.a simpler and more reliable system, with that said solar is already mostly reliable.
3.Individual panel shading isn't a problem. with a diode on each panel, if its shaded and that panels voltage falls below system voltage, it wont back feed.
I'm interested in what you guys think!
thanks for your time!
   

[Kim Christensen]

What the heck is going on with the LM555 timer? 

In all seriousness, you definitely want a proper charge controller. Those batteries are going to be expensive and cheaping out there would be foolish in the extreme.
Your other problem is going to be finding appliances that run on those voltages. An inverter is definitely in the cards for some of the loads you'll want to run.

[niwatori1]

LOL, I didn't specify that LM555, its just a BMS system stand in, not an actual BMS! sorry bout' that!
also you're right, a 48V charge controller is cheap and freely avalable.
for the loads i was thinking most mainstream DC loads would be fed directly off the dc rail and bucked to the right voltage.(computers, networking, lighting, etc.)
A 36-48vdc to 120VAC inverter would handle AC only loads(refrigerator, freezer, etc.)
i was just wondering if a system of such low voltages was viable in a medium to large array without being stupid.
you see them sometimes on RV's and such, but hardly ever above 10 panels or so..

[Faringdon]

Switching DC 48V ...ie the DC current with relays can mean big relays that wear out quick.
This was why the auto community never want to 48V system voltage, but stayed with 12VDC.
Fusing for DC also needs to be more substantial.
As you imply , wire runs are a PITA with LVDC.

[Seekonk]

I have a summer home that works only on a 12V car battery, so it is possible.  Only have an inverter for the fridge, dishwasher and other kitchen needs. Inverters only turn on when needed. Fridge only runs daylight hours. Dishwasher heater element runs off PV array DC.  Hot water is made from any excess PV power.  Clothes washer uses modified MSW inverter feeding array voltage into H bridge, no battery needed. Laundry has own water heater and all wash cycles use hot water. Clothes cleaner than at home. Car battery is nearly always at full charge. Basic system will run even with days of rain. Battery has to have full charge at night to keep me alive.  Electronics cost next to nothing to do this.  It just takes a little thought.

[M0HZH]

The cost in battery lifetime and especially wasted energy is much higher than the cost of a charge controller.

[50ShadesOfDirt]

I would definitely consider doing something like this *as a project*, and as the learning opportunity that it is. This is where you would explore DC-only componentry, and a small-scale project helps you determine what works best, and what works for you (sometimes two different things).

If I understand correctly, you are on-grid ... in this scenario, a "parallel (aka off-grid) system" could be a solution for you. An off-grid system, sized to critical loads you care about, is implemented parallel to grid wiring/components. For example, run the fridge 100% of the time on this parallel system (inverter/battery-bank). Recharge from the grid, using charger appropriate to battery-bank and chemistry (this is a cut-out to the grid); use best off-peak cost/time to recharge. Done, plus you have an emergency backup system in place for those (few, many in your area) times that the grid goes down. No messy "inter-tie" AHJ/codes/components/electricians/etc. to tie solar gear into the grid, as everyone and their dog wants to get a piece of you.

Later on and if needed, add a small inverter-gen to the mix, for longer runtimes when grid is down, or for larger loads. Or, add solar panels, for recharge during the day, and not from expensive grid power at that same time.

Hope this helps ...