Off-grid cabin, large system, starting from scratch

This is a difficult question to answer... Don't know the size (from 1,000 WH per day or 3,300 WH, or 10,000 WH). But some general suggestions...Theses are all my opinion and can be "wrong" if somebody makes other assumptions, and/or there are changes in what is available 5-10 years from now...

• My suggestion is that you can up or down size a system by 2x or 1/2x.... More than that, (3-10x or more up sizing), really a new system design and hardware.
• Specifically, AC inverters and DC loads (lights, radio/electronics, etc.) are usually fixed input voltages (12/24/48 volts typical for home solar). If you change battery bank voltage, these loads generally need to be changed out too.
• PWM solar charge controllers can be operated at 12/24 or 12/24/48 volts. This will require rewiring the solar array (Vmp~18/36/72 volts) as the battery voltage changes (you can rewire what you have... I.e., 12 in parallel or 2x 6 in series parallel or 4x 3 in series parallel). Or, you may choose to add more panels in series (or series parallel) to increase the array wattage (with matched panels).
• MPPT solar charge controllers... Also typically available in 12/24 or 12/24/48 volt ranges (roughly). You can run (for example) a 80-100 volt Vmp array, and (no rewiring of the array) simply switch out/rewire your battery bank voltage.
• Adding Batteries... Lead Acid batteries can be more of a pain to add new batteries to an old bank. Li Ion batteries with integrated BMS, may be perfectly OK to add more new batteries to an old bank (need to check with vendor).
• And there is the issue of changes in technology... Solar panels (Vmp/Imp/Pmp) have been changing evey 2-4 years. Trying to find a few panels to "match" (Vmp and/or Imp within 10% or better) can be very difficult. There is no "standard" anymore that makes mixing/matching panels "easy" (depends on what you have, your controllers, and what is in the market 5+ years from now).
• There is also changes in technology for charge controller--Communications buses, capacity, voltage ratings, etc. can also change. And trying to match a "small" Controller with a new "large" controller--You may not be able to bus them together (unified computer control, centralized data harvest, etc.). Almost always, you can parallel two or more controllers on a single battery bank--And there is no real issues with doing this (each controller makes its own charging decisions--mode changes will not be synchronized between controllers--No big deal).
• If you increase system current (larger inverter, more batteries, larger charge controllers, etc.), you will either increase wire AWG and/or increase battery bank voltage (P=V*I ... twice the current equals twice the power, OR twice the voltage equals twice the power).

We can give you rules of thumb:

• 12 volt battery bank--Roughly 1,200-1,800 Watts maximum inverter and solar array power
• 24 volt battery bank--Roughly 2,400-3,600 Watts maximum DC bus power
• 48 volt battery bank--Roughly 2,400-3,600 Watts to 10,000 - 20,000 Watts DC bus power

I use a 3,300 WH per day system as a "near normal" electrical existence for a very conservation minded home/family (full size refrigerator/freezer, LED lighting, Laptop Computer, LED TV, clothes washer, "solar friendly" well pump.

A 3x larger system or 10,000 WH per day (10 kWH per day or 300 kWH per month) is not much power in Air Conditioning land...

Next post, I will give you the basic math for making a calculation... Starting with a 3.3 kWH per day system:

-Bill

Hi Tom,

Take your time and ask questions... I try to write the math like an English sentence. And use "your numbers" (or my guesses) rather than using variables A, B, C, or AC_Watts, Panel_Eff, etc.

This way you can see where my fudge factors are (solar array efficiency, inverter efficiency, battery losses, etc.) are all entered.

Lithium are much more efficient (probably close ot 99%, where FLA batteries are closer to 80-90% efficient. Some of the efficiency numbers are dependent on exactly how you operate the system (day/night, 10% of rating, 50% of rating, etc.). Much of this stuff would drive you nuts to take everything to the 4 digit accuracy.

So, instead, I try for +/- 10% accuracy for guesstimates. Going for "more accurate or more precise" just is not worth it. The amount sun can vary from 10% to >100% of average harvest (weather, bird droppings on panels, etc.). And for normal folks, who knows their daily or hourly loads within 10% accuracy. If you are within a factor of 2 or 1/2... The is just how things work out. Some days you will use more power (fresh food in fridge, making ice, doing the wash), and other days you will use less energy... You can look estimate how much sun/battery storage and decide to turn on the "optional loads", start the genset, or wait for the next sunny day.

I suggest that the system should work for what you need without having to manage loads except for gross issues (been cloudy last two days--Conserve power or startup the genset). And if you have spouse, kids, visitors--More than likely they will not do 5% of the power management that you would do (and if you are not there, they can drain and kill your battery bank--leaving loads/tv on, nobody home).

Can you run a mini-split system on a solar power system with a "small/LiFePO4" battery bank--Yes... But what about the rest of your power needs (Lights, refrigerator, laptop, washing machine, well pump, TV/Radio/Cell phone).

Designing a power system optimized to run just a mini-split (and you have modulate the power consumption of the mini-split)--It can be done, and folks are out there that are integrating Solar+Mini-Split systems with Arduino or Raspberry Pi single board computers.

If you are into to these type of projects... They are very powerful and a great education.

If you want the system just to work... You would want to find somebody that has already done the hardware of integration (perhaps some folks here can suggest sources).

Or you can also find (don't know how many are out there) HVAC systems designed to run from solar only (and can use Grid or Battery backup power):

pberrypi.org/documentation/raspbian/updating.md

I don't know anything about the above website--Just a starting point for your research.

You can get DC refrigeration compressors that run from solar power, or even complete refrigerators that run from solar/DC power directly. They are not cheap, and tend to be smaller systems. Danfoss DC refer compressors are well known.

https://sundanzer.com/product-category/household/

They tend to be on the expensive side... Make sense if this is your only electrical load...

If you are planning on a larger Solar power system (lights, refrigerator, well pump, washing machine, laptop/TV/phone charging/etc.), generally it ends up being less expensive to simply add up all your AC loads and design an off grid solar power system to run it (suggest 9+ months out of a year occupation). Just pick very efficient AC appliances and don't waste power.

If you are looking at a Weekend/seasonal place--A large AC solar power system (fridge, etc.) usually just does not make economic sense... Design a small (500-1,000 WH per day system) for lights/laptop/tablet/phone charging/12 volt water pump... And use propane for fridge/hot water/heating/cooking.

There are so many battery types/designs/etc. out there... The above numbers are suggested starting points. As always, need to confirm the battery mfg/manuals need to be reviewed... There are some MPPT Solar Charge controllers that can limit battery charging current and yet deliver "excess power" to other loads on the DC bus (and AC loads on the inverters).

In engineering you cannot really say "something cannot be done"... With the electronics and computer/system controllers that are available today--Much of what we dream can be done. But has somebody already turned that into a product, or does the user have enough knowledge to design and integrate their controllers into existing products...

What end of the spectrum are you (hard core DIY, or you want the stuff to "just work" without daily interactions)... Everyone is different. No right answer.

-Bill