Running Starlink Off-Grid: Sizing Solar and Battery for 24/7 Uptime

Most Starlink off-grid sizing guides make the same mistake. They take the dish’s published wattage, multiply by 24, divide by a solar panel’s rated output, and produce a number. That number is wrong by a factor of two to four, and it is wrong in the direction that leaves you offline in January.

The method below is longer and gives you a bigger system. It is also the one that works in December, which is the only month that matters when you are sizing for continuous uptime.

One conclusion worth stating at the front, because it reframes everything that follows: at northern latitudes, your array size in winter is set by how much sunlight you get, not by how efficient your dish is. Starlink’s V5 halved the average draw of the residential terminal — a genuine improvement — but in a Montana December it halves a number that was never your binding constraint. The constraint is that you get two usable sun hours a day. Plan accordingly.

Step 1: Establish the real load, not the published one

Start with Starlink’s published average for your dish. These are the figures from the specification sheets:

• Starlink Mini — 25 to 40 W average
• Starlink V5 — 35 to 50 W average
• Standard (V4) — 75 to 100 W average

Use the midpoint of the range as your planning figure, not the bottom. The bottom of the range describes a clear day, a clean line of sight, light usage and mild temperatures. You will not have all four at once.

Then add what the specification sheets leave out:

The router. Starlink lists power consumption under the dish’s specifications and does not publish a separate figure for the Router Mini or the Gen 3 router. Budget 3 to 6 watts. The Mini’s router is integrated, so it is already inside the published number.

Obstructions. A partly obstructed sky view forces the dish to re-acquire satellites more often, and that costs power. Field reports commonly put this at 5 to 10 watts. If your site has trees on one horizon, assume you are paying it.

Cold. Below freezing, baseline electronics draw rises even with snow melt disabled. Add 5 to 10%.

Snow melt. This is the one that wrecks budgets, and it deserves its own step. See step 6.

Do not treat any of this as a substitute for measurement. An inline DC watt meter on the battery side of your system costs very little and answers every question in this article for your specific install. Published averages are a design starting point. Once your system is running, the meter is the authority.

Step 2: Choose the power path — this is your biggest lever

Before you calculate anything else, decide how power gets from the battery to the dish. This single decision moves your daily consumption by 10 to 25%, which is more than most people gain from any other optimisation.

Native DC (Mini only). The Mini publishes an input rating of 12 to 48 V at 60 W and takes power through a barrel jack. Wire it to your battery bus through appropriate fusing and you skip conversion entirely. Watch voltage drop on long runs — the Mini reboots when input voltage sags too low, and a thin or long cable on a 12 V system will do exactly that during the boot surge. On a 12 V bus with any real cable length, a regulated step-up to around 24 V is the standard fix and is well within the published input range.

First-party DC-DC (Standard V4 only). Starlink sells a DC-DC power supply for the Standard: 12 to 48 V in, 56 V at 3.57 A out, IP66-rated, and specified from −40°C to 60°C. It is documented and supported. If you are running a V4 off-grid and not using this, you are leaving efficiency on the table for no reason.

Inverter (everything else, and the V5 specifically). Starlink publishes no DC input rating for the V5 and sells no first-party DC supply for it. As of August 2026 there is no confirmed DC path for that dish, so an inverter is the only supported option. Use a pure sine wave inverter — the power supply contains electronics that do not tolerate modified sine well.

The inverter penalty is worse than most people assume, and worse on efficient dishes than on thirsty ones. Inverter efficiency collapses at low output because a roughly fixed overhead is spread across a smaller load. At 50 W of output, plan for 75 to 85% rather than the 90%+ on the datasheet. We work through why this matters more on a V5 than a V4 in our breakdown of the V5’s power figures.

If your inverter would otherwise sleep overnight and has to stay awake purely to feed the dish, count its standing consumption as part of the dish’s load. It frequently adds 8 to 15 W around the clock, which on a V5 is a third of the dish itself.

Step 3: Convert to watt-hours per day

Multiply your planning watts by hours of operation. For continuous uptime, that is 24.

Midpoints of published average ranges plus router allowance, continuous operation, clear conditions, no snow melt. Assumes 80% inverter efficiency at V5-level loads, 85% at V4-level loads, 92% for DC-DC conversion.

These are your clear-weather baseline numbers. They are not your design numbers. Steps 4 through 6 turn them into design numbers.

Worth noting before moving on: a Starlink terminal is rarely the only load on an off-grid system, and sizing it in isolation produces a system that runs the dish beautifully and nothing else. If you are building from scratch rather than adding a dish to an existing bank, work through a full load calculation first and treat Starlink as one line in it.

Step 4: Size the battery bank

Battery capacity buys you autonomy — the number of days the system runs with no solar input at all. That is what carries you through a storm front.

The calculation:

Bank capacity (Wh) = (daily consumption × days of autonomy) ÷ usable depth of discharge
Bank capacity (Ah) = bank capacity (Wh) ÷ system voltage

Usable depth of discharge is 80% for LiFePO4 and 50% for lead-acid or AGM. Those are not interchangeable numbers — a 200 Ah AGM bank and a 200 Ah lithium bank are not the same battery, and treating them as equivalent is the single most common sizing error in this space.

Two days of autonomy is the working minimum for a system you depend on. Three is right for northern winters or anywhere a multi-day storm system is normal. One day means any cloudy day is an outage.

Bank sizing at 80% usable DoD (LiFePO4):

For lead-acid or AGM at 50% DoD, multiply the Ah figures by 1.6.

Look at the 12 V column for a V4 at three days: 816 Ah. That is a large, heavy, expensive bank for one appliance, and it is the clearest argument in this article for either a higher system voltage or a more efficient dish. At 48 V the same autonomy is 204 Ah.

Two cautions on batteries in cold climates. Capacity falls as temperature drops — derate usable capacity by 10 to 20% below freezing. And LiFePO4 cells must not be charged below 0°C without integrated heating, which is a hard limit rather than a guideline. If your bank lives in an unheated space, this constrains your design before anything else does. The interaction between temperature, autonomy and array size is exactly the problem we work through in battery bank sizing for shoulder-season use.

Step 5: Size the array

The array replaces what you consume. The calculation:

Array watts = daily consumption (Wh) ÷ (peak sun hours × derate factor)

Peak sun hours is the number of hours per day equivalent to full-rated sunlight at your location. Use your worst month, not an annual average. For continuous uptime that means December.

These are planning approximations for December at roughly latitude tilt. Get site-specific figures from NREL’s PVWatts tool before you buy anything — regional variation within a state can be substantial, and terrain shading is not captured in any regional table.

The derate factor accounts for everything between the panel’s rating and the energy that actually reaches your battery: temperature, soiling, wiring losses, charge controller efficiency, imperfect tilt and orientation, and partial shading. Use 0.75. Manufacturers quote panel output under test conditions you will not experience.

Array sizing at a 0.75 derate:

Read across a single row and the point of this article becomes obvious. The same dish, the same load, needs three times the array in the Pacific Northwest that it needs in Arizona. Dish choice moves you between rows. Latitude moves you between columns, and the columns are wider than the rows.

Step 6: Run the winter stress test

Everything above assumed clear conditions and no snow melt. Now break it.

Starlink does not publish a power figure for the snow melt heater on any terminal. Independent measurements vary widely — some cold-climate users report an additional 20 to 30 watts sustained, others report excursions well beyond 100 watts during active storms. That spread is real and anyone quoting a single figure is guessing.

What is documented is capacity. The V5 and the Standard V4 are both rated to clear up to 40 mm per hour; the Mini is rated to 25 mm per hour. The V5’s average draw halved relative to the V4, but its snow melt capacity did not change. There is no basis for assuming its heater load scaled down with its baseline.

For planning, add 25 to 50% to your winter daily consumption at any site that sees regular sub-freezing precipitation, and size the array to that figure rather than the clear-weather one.

Worked example: a fixed cabin in western Montana, V5 dish

• Clear-weather baseline, V5 on an inverter: 1,425 Wh/day
• Winter uplift, +30% for snow melt and cold: 1,853 Wh/day
• December peak sun hours, ~2.0, derate 0.75: 1,853 ÷ 1.5 = 1,235 W of array
• Battery, 3 days autonomy at 48 V, LiFePO4: (1,853 × 3) ÷ 0.8 = 6,949 Wh = 145 Ah at 48 V

Now the same site in July, with roughly 5.5 peak sun hours and no heater: 1,425 ÷ 4.125 = 345 W of array.

The winter system is three and a half times the summer system, for the same dish running the same hours. That ratio is the whole argument. It is not driven by the dish at all — it is driven by the sun. A Mini on the same site in December still needs around 735 W of array once you apply the same winter uplift. Choosing the most efficient dish available saved you 500 W of panel. Choosing to live at that latitude cost you 900.

Step 7: What to do when the numbers don’t work

Sometimes step 6 produces a system you are not going to build. That is a legitimate outcome, and there are four honest responses.

Cut the hours. This is by far the largest lever available and the most underused. Starlink’s app has a sleep schedule that powers the system down and brings it back automatically. Running 16 hours a day instead of 24 cuts consumption by a third, and for most people the overnight hours are worth nothing. Applied to the Montana example, it takes the December array requirement from 1,235 W down to about 825 W.

Use a standby or paused plan seasonally. If a site is lightly used in winter, low-cost standby options keep the account alive without paying for full service or full power. Confirm the current terms directly with Starlink — plan structures and pricing changed several times through 2026.

Add a generator instead of panels. Sizing an array for the worst two weeks of December means most of that array is idle for the other eleven months. A small inverter generator run for a few hours during a storm week is frequently cheaper than the last 600 W of panel, and it does other useful work.

Do not solve a solar problem with batteries. When December numbers look bad, the instinct is to add capacity. Capacity does not generate anything — it buys you time between charges, and if the array cannot refill it, a bigger bank simply takes longer to die. When the array is the constraint, add array. That is the general case, and we make the argument in full in panels or batteries first.

A note on installation and insurance

Two practical points that sit outside the arithmetic but inside the project.

Fuse everything, close to the battery, on every conductor leaving it. A 48 V DC system is genuinely hazardous and does not forgive the casual approach that 12 V mostly tolerates. If you are not confident specifying overcurrent protection and conductor sizing for the currents involved, have that part done by someone who is.

And if this system is going into a converted vehicle rather than a building, understand that a roof-mounted terminal, a substantial battery bank and a permanent solar array are exactly the modifications that standard vehicle policies handle badly or exclude entirely. A claim after a fire or a theft is not the moment to discover your build was insured at base-vehicle value. This is a well-documented gap and it is worth reading about why standard policies fail on custom and converted vehicles before the install rather than after.