Almost every battery bank that disappoints its owner was sized for July.
That is the whole failure. In midsummer you get six peak sun hours, no heating load, and long daylight that keeps the lights off until nine. In October you get three and a half peak sun hours, a furnace blower cycling all night, and lights on from six. The load goes up and the harvest goes down simultaneously, and the bank that comfortably carried you through August is at 40 percent by the second morning.
So the correct question is not “how many amp-hours do I need?” It is “how many amp-hours do I need on the worst three days I intend to be out there?” Those are very different numbers, and the second one is the only one worth sizing to.
What changes between summer and shoulder season
Four things move at once, and three of them move against you.
Peak sun hours fall, hard. Depending on latitude and local weather, a location delivering 6.0 peak sun hours in June will typically deliver 3.5–4.5 in April and October, and 2.0–3.0 in December. The sun is lower, the day is shorter, and shoulder seasons carry more cloud in most of the US than midsummer does.
Panel angle matters more than it did. In summer, a flat rooftop array is close enough to optimal that nobody bothers tilting. In October at 40° north, a flat panel can lose 25–35 percent against one tilted toward the low sun. This is the cheapest single improvement available to a shoulder-season system and it costs nothing but effort.
Heating loads appear, and the electrical one is the furnace fan. A propane furnace is not an electrically free appliance. The blower and control board draw 5–9A at 12V while running. On a cold night with a 35–45 percent duty cycle over ten hours, that is 250–400Wh of pure electrical consumption — often more than the fridge. A diesel heater is gentler at 1–2A running, but draws 8–10A during its glow-plug start cycle, which matters if it short-cycles.
Lighting and indoor time increase. More hours inside, more hours with lights on, more laptop and entertainment use. Call it 100–200Wh a day.
The one thing moving in your favour: no air conditioning, and the fridge works slightly less hard. That recovers maybe 100–150Wh. It does not come close to offsetting the rest.
Net effect: a rig consuming 1,400Wh a day in summer typically consumes 1,700–2,100Wh a day in shoulder season, while harvesting roughly 60 percent of what it harvested in summer.
The sizing method
Five steps. Do them in order and do not skip the audit.
Step 1 — Establish your real daily consumption
Not the manufacturer’s figures, not a forum average. Your appliances, your usage hours. The load calculation guide walks through the method; the short version is wattage × hours per day for every device, summed.
If you already have a shunt-based monitor, use last season’s actual amp-hour consumption instead. Measured beats calculated every time.
Step 2 — Apply the shoulder-season adjustment
| Adjustment | Typical change |
|---|---|
| Furnace blower or heater | +250 to +400 Wh |
| Additional lighting and indoor hours | +100 to +200 Wh |
| Reduced fridge duty cycle | −80 to −150 Wh |
| Water pump (unchanged) | 0 |
| Net | +270 to +450 Wh per day |
Add that to your summer figure. Call the result your shoulder-season daily load.
Step 3 — Decide days of autonomy honestly
Days of autonomy is the number of consecutive days the bank must carry the full load with negligible solar input. This is a weather question, not a preference question.
- 1 day — you accept that a genuinely overcast day means running the generator or driving. Reasonable for weekend use near services.
- 1.5 days — the common sensible answer. Covers one bad day plus a poor morning.
- 2 days — appropriate for extended remote stays, or the Pacific Northwest in October, where two consecutive overcast days is normal rather than exceptional.
- 3+ days — you are now spending large sums to solve a problem a generator solves for a few hundred dollars. Think carefully.
Step 4 — Divide by usable depth of discharge
Nameplate capacity is not usable capacity.
| Chemistry | Plan on using | Reason |
|---|---|---|
| LiFePO4 | 80% | 90%+ is technically available; 80% preserves cycle life and leaves reserve |
| AGM | 50% | Deeper cycling collapses cycle count |
| Flooded lead-acid | 50% | Same, plus sulphation risk if left partially discharged |
The cycle-life arithmetic behind those percentages, and why the cost per usable kWh diverges so sharply between chemistries, is in the LiFePO4 versus AGM versus lead-acid comparison.
Step 5 — Convert to amp-hours
Required Ah = (shoulder daily Wh × days of autonomy) ÷ (usable DoD × system voltage)Then round up to a purchasable configuration.
Three worked examples
Weekend user, travel trailer, LiFePO4
Summer load 900Wh/day. Shoulder adjustment +300Wh → 1,200Wh/day. One day of autonomy, 80% usable, 12V.
(1,200 × 1) ÷ (0.8 × 12) = 125 AhRound to 200Ah — two 100Ah drop-ins — which gives a comfortable 1.6 days and headroom for a colder trip than planned. A 100Ah bank would technically satisfy the calculation and would be miserable in practice, because a single-battery bank has no margin for the night the furnace runs harder than expected.
Two-week traveller, Class B van, LiFePO4
Summer load 1,600Wh/day. Shoulder adjustment +400Wh → 2,000Wh/day. 1.5 days autonomy, 80% usable, 12V.
(2,000 × 1.5) ÷ (0.8 × 12) = 313 AhRound to 400Ah. This is where the cost of the calculation starts to bite — that is roughly $1,400–$2,800 of cells before anything else.
Full-timer, fifth wheel, AGM
Summer load 2,200Wh/day. Shoulder adjustment +450Wh → 2,650Wh/day. 2 days autonomy, 50% usable, 12V.
(2,650 × 2) ÷ (0.5 × 12) = 883 AhNearly 900Ah of AGM is around 550 lbs of battery and eight to nine group-31 cases. At this point the chemistry decision has been made for you by physics rather than by preference.
The recharge constraint that invalidates most of these numbers
Here is the step that gets left out of every sizing guide, and it is the reason large banks so often disappoint.
A battery bank you cannot refill is a battery bank you own once. If your 400Ah lithium bank goes to 50 percent overnight, you need to put roughly 2,400Wh back in the following day. At 3.5 shoulder-season peak sun hours, and allowing 20 percent for controller, wiring and temperature losses, that requires around 850W of array.
If you have 400W on the roof, you will recover about 1,100Wh a day. You are not refilling the bank; you are slowing its decline. The bank buys you three or four days instead of two, and then you are in exactly the same position, only with more money spent.
The rule: for every 100Ah of 12V lithium capacity you intend to cycle deeply, plan on roughly 200–250W of array in shoulder conditions, or an equivalent contribution from alternator charging via a DC-DC charger.
This is why the site takes the position it does in solar array sizing: where you have a chronic shoulder-season deficit rather than an occasional bad-weather one, array watts are almost always the cheaper fix. Four hundred watts of additional panel costs $250–$400 and generates energy every day the sun is up. Two hundred additional amp-hours of lithium costs $700–$1,400 and generates nothing — it only stores what you already harvested.
Batteries solve intermittency. Panels solve deficit. Diagnose which one you actually have before spending, because the two problems present identically at 6am with a flat bank.
Configuration: when 12V stops being the right answer
Amp-hours are a voltage-dependent unit, which makes them a poor way to compare banks. Talk in watt-hours and the comparison becomes honest: 400Ah at 12V and 200Ah at 24V are the same 4,800Wh of storage.
Below about 300Ah at 12V, stay at 12V. The ecosystem is deeper, the appliances are native, and the currents are manageable. Above roughly 400Ah at 12V — or where the inverter exceeds about 2,000W — the case for 24V becomes strong: the same power moves at half the current, which halves the voltage drop percentage and cuts the copper requirement dramatically. The trade is that native 12V appliances then need a step-down converter, and 24V drop-in batteries carry a modest price premium.
Whichever voltage you pick, wire the bank properly. Parallel batteries need equal-length cables to each battery, or diagonal takeoff — positive from one end of the bank, negative from the other. Unequal cable lengths mean the nearest battery does disproportionate work and ages first, which quietly costs you capacity you paid for. This is a five-minute decision at install time and an expensive one to correct later.
When the bank is already undersized
Most people arrive at this article with a bank already installed, and replacing it is not the first option. Four mitigations, in order of cost-effectiveness:
Tilt the panels. Free, and worth 20–35 percent in October at mid-latitudes. If your array is flat-mounted and non-adjustable, even one portable panel that can be angled at the sun changes the shoulder-season picture more than most people expect.
Shed loads in the right order. The furnace is the largest electrical draw and the hardest to give up. A diesel or propane catalytic heater with minimal electrical demand removes 250–400Wh a day at a stroke. After that: reduce inverter idle time, switch to DC charging for laptops and phones rather than running the inverter for a 60W load, and accept a colder night setpoint.
Add alternator charging. A 30A DC-DC charger delivers roughly 360W while driving. Two hours of travel puts around 700Wh into the bank regardless of weather, which is often the single most reliable shoulder-season input available. On a rig that moves every few days this is more valuable than additional panel.
Add array before adding cells. Repeated because it is the recurring answer.
Only after those does a bank expansion make sense, and expansion means replacing the bank rather than adding to it — see the mistakes section below.
Cold weather, and the constraint that is not about capacity
Two temperature effects matter, and they are frequently confused.
Capacity loss on discharge is modest. LiFePO4 delivers roughly 90–95 percent of rated capacity at 0°C and around 80 percent at −10°C. Lead-acid is considerably worse, losing 20–30 percent at freezing. Annoying, manageable, and worth a few percent of headroom in your sizing.
Charge acceptance below freezing is not modest — it is a hard stop. Charging standard LiFePO4 below 0°C causes lithium plating and permanent, cumulative capacity loss. Quality drop-ins have a BMS that simply refuses charge below the threshold, which means your solar array can be in full sun and your bank will not accept a single amp until the cells warm up.
In shoulder-season use this is a real operational problem rather than a theoretical one: a clear October morning at 28°F is precisely when you have a depleted bank and good sun, and precisely when the BMS will not let you use it. The options — self-heating cells, an insulated interior location, or a warm-up load — are covered in lithium batteries below freezing.
Size on the assumption you will lose the first hour or two of charging on cold mornings. It changes the array requirement more than it changes the bank requirement.
The mistakes worth naming
Sizing to the average day. Averages are for accountants. Systems are sized to the worst case you intend to tolerate.
Buying capacity instead of charging. Covered above, and it is the single most expensive error in this category.
Ignoring the furnace. It is the largest shoulder-season load in most rigs and it is invisible because it is “a propane appliance.”
Adding new batteries to an old bank. Mismatched cells in parallel equalise to the weakest one. A new 100Ah battery paralleled with a three-year-old 100Ah battery gives you rather less than 200Ah and shortens the life of the new cell. Replace banks whole.
Treating the bank as a one-off cost. A lithium bank is a consumable with a ten-year-ish life and an AGM bank is a consumable with a four-to-six-year life. Both belong in the ownership cost model rather than the purchase price — the five-year cost of RV ownership breakdown puts numbers on how much of a rig’s running cost is actually deferred replacement of systems like this one.
The short version
Take your measured summer consumption, add 300–450Wh for shoulder-season heating and lighting, multiply by 1.5 days, divide by 0.8 for lithium or 0.5 for lead-acid, then divide by 12. Round up to a whole configuration. Then check that your array can replace half of that in a single shoulder-season day — and if it cannot, buy panels before you buy cells.
The solar and battery sizing tool runs the full calculation including the shoulder-season stress case and the recharge check, and shows the working so you can argue with it.