Solar Powered Minisplit Heating – Real World Results

We’ve been working on staying off grid at Rally Creek while keeping some modern amenities. Our first project is keeping Blue Skies Cabin 100% off grid while having a comfortable place to stay. A primary criteria is heating and cooling especially during the South Carolina summers. To meet that goal I found the most efficient way to heat and cool on 100% sun-power is to use a high efficiency minisplit heat pump. A small 3/4 ton (9,000 BTU) is more than enough for our 12×16 one room cabin and as a bonus it runs on 120V AC power.

Minisplit Heating With Solar

At this time of year the days are perfect in the high 60s but at night it can still drop into the 30s. This past week it even dropped down below 40 degrees for the entire evening, 11PM until 8AM … it was an unusually cold early spring night. Time to test the minisplit.

The summary of the experience — the minisplit kept the cabin at a comfortable 62 degrees for MOST of the night. I say MOST because the power shut down sometime around 6AM, about 2 hours before the sunshine started hitting the solar panels enough to start charging. The means in the real world I saw about 7 hours of heating from the battery storage.

I had anticipated the heat would EASILY make it through the night with plenty to spare. I even was confident enough to figure the batteries would refill on the next sunny day to make it through another 40-degree evening the next night if I were to stay. Turns out I was wrong.

This is one of the main objectives of our Rally Creek project — learning with real world experiences on what it means to live off grid in a tiny home. Not the marketing bullshit so many companies produce, but the real world experience and finding the science behind it.

Real world on a 38-degree night has the heat pump working likely closer to 15 minutes on , 10 minutes off. Thankfully easy to fix and highlights the need for insulation. However the battery ran out early and waking up to a 52-degree cabin wasn’t horrible but wasn’t nice and comfy either. The outside temperature made a huge difference in a cabin with only insulated walls.

Side bar — the first night it only dropped to the high 40s. The heat pump was on early in the evening from 8PM until 10PM on 62 degrees when it was 55 outside. I turned of off as it was too hot for my comfort while sleeping. It didn’t go back on until 4AM when the cabin was down to 52 degrees. The cabin was quickly up to 62 degrees and stayed there until 7AM when I woke up to a 62-degree cabin and 60% battery storage the first night.

Running Out Of Power

So why did I run out of power and how were the calculations so far off? The math… it all comes down to the math. Turns out my calculations came up an hour short, party due to some incorrect assumptions (what do they say about assume?) as well as missing some details. The details matter. Let’s dig in…

The Math Of Solar Powered Heating

Ultimately it all comes down to the math of the electrical system. Let’s recap the basic numbers for the system for the past week.

Usable Battery Capacity : 540 Ah * 12V = 5.56 kWh

The lead-based gel batteries were swapped out for lithium batteries. That changes the math significantly — lead batteries can be drained to 38% meaning you can only use 62% of the capacity (theoretical claims 50% for maximum health, real world shows 62% before they shut down). The lithium batteries can be drained to 10% meaning 90% of the purchased capacity can be used.

Old lead batteries: 400ah * 62% = 320ah usable power.
New lithium batteries: 600ah * 90% = 540ah usable power.

Since these are 12 volt batteries the 540 Ah battery holds 12V * 540AH = 5560 Wh or 5.56 kWh of usable power.

Consuming The Power

To figure out how long this power stored in the batteries will last, assuming we end a sunny day at 100% battery capacity — which we did this past week — means we need to figure out how much our devices connected to the batteries consume. There are some key elements to this equation.

First thing to consider is ANYTHING that consumes AC power needs to add 25% to the listed power consumption numbers to account for “system losses”. What are these losses? The inverter claims 90% efficiency, so it loses 10% of the power in the batteries as it converts the electrons from steady state DC to “doing the wave” AC output. In another “real world vs. theory” example, the Renogy 2000W Pure Sine Wave manual says “we are super efficient at 90%” but then says “do all your AC math at 75% efficiency; this accounts for heat losses, transmission line losses, and more. Even with our short 20′ power runs we still realize at least 20% loss between the battery and devices in the cabin.

For simplicity and to have a margin of error we will use a 25% multiplier on any AC devices to calculation consumption.

Minisplit : 100A on 12V (1200 W)

Minisplit Heat : rated at 8A (+25%) @ 120V = 8A * 1.25 @ 120V = 10A @ 120V , we can rephrase this as 100A @ 12V = 1.2 kWh

Minisplit running non-stop (not likely) means we can keep that system running for 5.4 hours (540Ah / 100A on a 12V system).

This is where the first assumption comes into play. With the insulated walls (but not floor & ceiling yet), I assumed from a very small sample set of tests that the minisplit would run for 20 minutes of every hour, about 33% of the time.

The actual heating run time was closer to 10 minutes on / 10 minutes off = 50% power consumption. And waking up the morning to a 52-degree cabin explained why — no floor or ceiling insulation (yet) and the cold air was refrigerating the floor and flowing in through the ridge cap down into the cabin.

Starlink Internet : 7.9A on 12V (95 W)

The other thing we setup to allow for remote 24×7 monitoring of power generation — WHEN THE RENOGY APP WORKS (they broke it recently, right before I bought the Renogy One Core with an app update last week , Renogy is notorious for sub-par customer support and breaking things whenever they make changes!). It does this by logging power generation data to the cloud via the Internet. We also have on-site security cameras as we are off property for weeks at a time — that also works best with 24×7 Internet.

This past week we setup our Starlink system. No need to live like cave men skinning squirrels for dinner and making shadow puppets for entertainment. Nothing wrong with streaming a video on a cold rainy night in the woods. You do your off grid, we will do ours…


Starlink Power Consumption: 20W sleep mode, 50-75W data active mode
20W 120V = 25W (+25%) @ 120V = 0.21A which will draw 10x from a 12V battery to convert to 120V = 2.1A
50W 120V = 62.5W @ 120V = 0.52A = 5.2A
75W 120V = 93.75W @ 120V = 0.79A = 7.9A

Other Devices

The details matter when figuring out how long the entire system will run. That means accounting for every little thing on the power system overnight. What else do we have running?

  • Phone/Watch charging pad (low use, but not zero — assume the full 2.5A draw 5VDC, close enough to 6VDC for “easy math” = 1.25A (50% run time with sound mode running at night)
  • Mini Cooler (fridge) : 3.5A on 12VDC , 50% run time
  • Charge Controller: 100mA on 12V = 0.1A
  • Renogy DC One Monitor: 1.9W on 12V = 0.15A
  • Compost toilet fan: 0.1A , 100% run time
  • Netgear wireless router: 1.5A @ 12V via AC conversion (twice).
    This warrants extra conversation. The router notes 1.5A 12V consumption, BUT…
    First the 12V DC battery is converted to AC (25% penalty), then the router converts that via a wall wart back to 12VDC.
    Wall warts are NOT efficient, they tend to run at 80% efficiency, another 20% penalty.
    1.5A * 1.20 (wall wart inefficiency) = 1.8A * 1.25 (our standard AC/DC inverter penalty) = 2.25 A
  • Arlo Hub : 1.5A DC/AC/DC Conversion = 1.5A * 1.2 * 1.25 = 2.25 A

Putting It All Together

Here are the devices and the power consumption for what was running overnight on a 38-degree night with the cabin set to 62 degrees.

Initial Assumptions (33% run time for heat pump)

StorageAmpsWattsRuntime %Run (A)
12V Lithium Battery Bank, max usable power540 Ah5560 Wh
Minisplit heating/cooling (38F outdoor / 62F indoor, heating mode)100 A1200 W3333
Starlink internet7.9 A95 W1007.9
Netgear Wireless Router2.25 A25 W1002.25
Arlo Hub, security cameras2.25 A25 W1002.25
Mini cooler, refrigeration (DC)3.5 A42 W501.75
Mobile device charging1.25 A15 W1001.25
Renogy RVR40 Charge Controller (DC)0.1 A1.2 W1000.1
Renogy One Core , solar system monitor (DC)0.15 A1.9 W1000.15
Toilet Fan (DC)0.1 A1.2 W1000.1

Base on the numbers above our 540Ah battery should last 11 hours (540Ah / 48.75 A).

However we came up way short with the initials assumptions. Why?

Real World (50%)

StorageAmpsWattsRuntime %Run (A)
12V Lithium Battery Bank, max usable power540 Ah5560 Wh
Minisplit heating/cooling (38F outdoor / 62F indoor, heating mode)100 A1200 W5050
Starlink internet7.9 A95 W1007.9
Netgear Wireless Router=5402.25 A25 W1002.25
Arlo Hub, security cameras2.25 A25 W1002.25
Mini cooler, refrigeration (DC)3.5 A42 W501.75
Mobile device charging1.25 A15 W1001.25
Renogy RVR40 Charge Controller (DC)0.1 A1.2 W1000.1
Renogy One Core , solar system monitor (DC)0.15 A1.9 W1000.15
Toilet Fan (DC)0.1 A1.2 W1000.1

Base on the numbers above our 540Ah battery / 63.5h consumption should last 8.2 hours.

The Solar Powered Minisplit Heating

So overall, we came up short by an hour so so on a cold night. Not horrible after all. But why?

It is all in the math.

First — details matter. Every single component is important to track. When you are on a limited supply of power, like a battery bank, and not an unlimited tap like grid power you need to pay attention to where every drop of power goes.

Figuring out how long devices run also matters. There is a big difference between a heat pump running 10 minutes on, 10 minutes of versus 10 minutes on, 20 minutes off. Tracking real world use over time and calculating averages is a good start. Or assume “always on” and you will have over-sized storage, but will likely be happy about it when you have a cloudy day and extra electrons saved up in the battery bank.

Another key element – system efficiency when converting DC to AC power. While the inverter in place is 90% efficient, meaning a 10% power loss thing like wiring length, current and resistance which depend on the type of wiring, and even environmental factors make a difference. Our system uses the value recommended in the inverter manual of 25% overall DC to AC losses. A system that losing 20% of the battery to conversion and wiring is going to yield different numbers than a 30% loss.

And one last sidebar about DC and AC power — if you can stay DC powered end-to-end you will get to use more of the electrons stored in the battery. Especially if you have 12 volt batteries and are using 12 volt systems. As soon as you plug into AC you lose 25% of the power from the inverter. If you plug in a “wall wart”, which most consumer devices like routers, modems, phone chargers, televisions, etc. use then they use low-grade inverters that incur another 20-25% loss changing that newly generated AC power back to DC. In theory you can craft a custom cable to directly wire any 12V device (like our netgear router and arlo hub) directly to the batteries (with a fuse and DC distribution box of course).

Discover more from Rally Creek USA

Subscribe to get the latest posts to your email.

Leave a Comment

Your email address will not be published. Required fields are marked *

Discover more from Rally Creek USA

Subscribe now to keep reading and get access to the full archive.

Continue reading

Scroll to Top