Most solar garden lights run between 2 and 12 hours, and the spread is not luck. It is arithmetic you can do in about thirty seconds once you know which number actually limits your light. For almost every stake light sold in America, that number is not the battery. It is the panel.
The short answer, by season and latitude
A typical budget stake light runs about 10 to 11 hours on a June night anywhere in the lower 48, and about 2 to 6 hours on a December night depending on how far north you live. Here is the same fixture, same battery, same LED, in four places:
| City | December runtime | December darkness | June runtime | June darkness |
|---|---|---|---|---|
| Seattle, WA | 1.6 h | 14.9 h | 11.2 h | 6.6 h |
| Minneapolis, MN | 2.5 h | 14.1 h | 11.1 h | 7.1 h |
| Phoenix, AZ | 5.9 h | 13.2 h | 15.5 h | 8.7 h |
| Miami, FL | 6.3 h | 12.6 h | 11.0 h | 9.6 h |
In June every one of those lights covers the whole night with hours to spare. In December not one of them gets close. A Minnesota reader and an Arizona reader own the same product and have completely different experiences of it, and both of them are right.
The formula
Runtime is stored charge divided by draw. For a solar light the stored charge is whatever the panel managed to put in that day, so:
Runtime (hours) = (Panel current in mA × Peak sun hours × 0.7) ÷ LED draw in mA
… capped by the battery’s capacity in mAh, which, as you will see, almost never binds.
Three inputs. Panel current is printed on the panel or in the manual: the commodity garden-light panel is 2V 40mA, better fixtures use 2V 100mA or 4V 100mA. Peak sun hours is your location’s daily solar energy, which is 5 to 8 in summer and 0.8 to 3.4 in December depending on latitude. LED draw is the current the whole fixture pulls from the cell.
The 0.7 covers two real losses: NiMH cells accept roughly 75% of the charge you push into them, and a garden light wires its panel straight to the battery with no maximum-power-point tracking, so the panel never runs at its best operating point. That factor is an engineering estimate, not a specification. Treat it as a good approximation, not gospel.
If you would rather work in watt-hours, convert first: mAh × volts ÷ 1000 = Wh. A 600 mAh AA NiMH cell at 1.2 V holds 0.72 Wh. Most guides skip this step, which is why their formulas don’t work on a product you can actually buy.
A worked example
Take a mid-range path light in Minneapolis, and run it twice.
- Panel: 80 mA. Cell: 800 mAh AA. Draw: 55 mA.
- June. Minneapolis gets 5.95 peak sun hours on a flat surface. Charge banked = 80 × 5.95 × 0.7 = 333 mAh. Runtime = 333 ÷ 55 = 6.1 hours. Darkness that night is 7.1 hours, so it dies about an hour before dawn.
- December. Minneapolis gets 1.36 peak sun hours. Charge banked = 80 × 1.36 × 0.7 = 76 mAh. Runtime = 76 ÷ 55 = 1.4 hours. Darkness is 14.1 hours. It is dark for the last twelve and a half.
Now look at what the battery did. In June it held 333 mAh of an 800 mAh capacity. In December it held 76 mAh. The cell was never more than 42% full. That is the whole story.
Why a bigger battery will not help you
Because the panel is the bottleneck, not the cell. Run the budget light through the formula with four different batteries on a Minneapolis December day and you get 2.5 hours every single time, whether the cell is 600 mAh or 2800 mAh. The panel banks 38 mAh either way, and 38 mAh at 15 mA is 2.5 hours no matter how large the container holding it is.
This is the most expensive misunderstanding in solar lighting. People buy high-capacity replacement cells expecting longer nights and get nothing, because they solved a problem their light did not have. We wrote about this specifically in can I use a higher mAh battery in solar lights. A larger cell helps in exactly one situation: when your light genuinely runs the battery flat in good summer sun, which means it is battery-limited rather than charge-limited. That is rare.
What does help is giving the panel more light. Cleaning it, moving it out of afternoon shade, and tilting it toward the southern sky. Tilting is worth more than most people expect: at Minneapolis latitude, tilting a panel 45 degrees lifts December collection from 1.36 to about 3.0 peak sun hours, which more than doubles runtime. See how to clean the solar panel on garden lights and whether solar lights need direct sunlight.
Typical capacity, draw and runtime by fixture type
| Fixture | Panel | Cell | Draw | Light | Summer | Winter N | Winter S |
|---|---|---|---|---|---|---|---|
| Budget stake light, single 5 mm LED | 40 mA | 400–600 mAh AAA/AA | 11–15 mA | 1–3 lm | 10–14 h | 2.5 h | 5.9 h |
| Mid-range path light, one 2835 SMD | 80 mA | 600–800 mAh AA | 55 mA | 8–10 lm | 5.6 h | 1.4 h | 3.2 h |
| Bright path light / bollard | 100 mA | 800–1200 mAh AA | 183 mA | 23–26 lm | 2.1 h | 0.5 h | 1.2 h |
| Step / deck light | 40 mA | 150–400 mAh | 11 mA | 1–3 lm | 14 h | 3.5 h | 8.1 h |
| Spotlight (Li-ion, buck driver) | 2 W | 3000 mAh LiFePO4 | 300 mA | 100 lm | up to 10 h | 3–4 h | 6–7 h |
| Security flood, motion only | 4–10 W | 6000 mAh Li-ion | 1.6 A | 810 lm | 3.7 h continuous | — | — |
Note the bottom two rows changed chemistry. Once a fixture needs more than about 25 lumens it abandons the single NiMH cell, because 183 mA flattens a 600 mAh AA in three hours. Brighter fixtures escape the arithmetic by growing the panel and switching to lithium, not by being cleverer. That is the honest reason solar spot lights and shed lights cost more than stake lights.
Your light gets dim long before it goes out
This is the part almost nobody tells you, and it changes what “runtime” even means. The driver chip in a cheap solar light has no current regulation at all. The QX5252, one of the two chips in most of these fixtures, sets LED power by this published relationship:
PLED = 2.5 × VBAT2 × 10-6 ÷ L
There is no feedback path anywhere in that expression. Output follows the square of cell voltage, and a NiMH cell sags from about 1.35 V freshly charged to 1.00 V at its rated endpoint. That is a 45% loss of light output across the night, arriving gradually, with no step and no cutoff.
| Cell voltage | Light output |
|---|---|
| 1.35 V (just charged) | 100% |
| 1.20 V (nominal plateau) | 79% |
| 1.10 V | 66% |
| 1.00 V (rated endpoint) | 55% |
| 0.90 V (driver minimum) | 44% |
So when a box says “up to 8 hours,” it is making a claim about an unstated brightness floor. The light does not run for eight hours and stop. It fades continuously until you stop noticing it. Ask any manufacturer what brightness the eighth hour delivers and you will not get an answer, because none of them publish one.
Winter versus summer is not only about charge time
Winter attacks from both ends at once, and this is why it feels so much worse than “the days are shorter” suggests.
On the supply side, Minneapolis collects 1.36 peak sun hours on a flat panel in December against 5.95 in June. That is 23% of the summer charge. On the demand side, true darkness runs 14 hours 7 minutes in December against 7 hours 8 minutes in June, so the light must work 98% longer. Divide one by the other and a northern light faces an 8.6-fold mismatch between what it needs and what it gets. Phoenix faces 4.0-fold. Tilting the panel cuts those to roughly 3.6 and 2.1.
Cold makes it worse in both directions. NiMH capacity falls below about 14°F, and more importantly, charging a NiMH cell below 32°F causes permanent damage. Panasonic’s handbook is blunt about the mechanism: below freezing the cell cannot absorb the oxygen it generates, internal pressure rises, the safety vent opens and electrolyte escapes. That is irreversible. A solar light trickle-charges into a full cell on a cold sunny February morning with no charger intelligence to stop it, which is the worst case the chemistry has.
If your lights matter to you in winter, bring them in or accept a few hours. More options in charging solar lights without sun.
Why a two-year-old light runs shorter and dimmer
Because NiMH end-of-life is defined as the point where capacity reaches 60% of nominal, and a solar light reaches it fast. Panasonic’s design guidance in its own handbook says to expect 2 years or 500 cycles under proper conditions. A solar light cycles once a day, 365 times a year, and does not provide proper conditions: it trickle-charges with no termination and no timer, which Panasonic explicitly warns against.
Expect a two-year-old light to give roughly a third to a half less runtime. Aging also shows up as dimness rather than just brevity, because a worn cell has higher internal resistance, which drops voltage, which through that V-squared law costs you light immediately. One caveat worth being straight about: in deep northern winter the panel is already the limit, so a fresh battery changes less than you would hope. Replacing cells pays off most in spring and summer. See how long solar lights last and whether solar lights need batteries.
Is “up to 12 hours” achievable?
Yes, under conditions the box does not mention. Twelve hours needs a low-draw fixture (15 mA or less), a clean unshaded panel, a healthy cell, and a location with at least 5 peak sun hours. That describes a dim accent light in a sunny yard from April to September. It does not describe a 100-lumen spotlight in Ohio in November. The claim is usually honest about the hardware and silent about the conditions.
Does more lumens always mean shorter runtime?
At a fixed daily charge, yes, and close to proportionally. But not exactly, because LED efficacy improves when you drive it softer. A Bridgelux 2835 delivers 171 lumens per watt at 20 mA and 132 at 80 mA, so driving it gently buys about 30% more light per unit of energy. This is why well-designed dim lights last all night and cheap bright ones do not, and why some fixtures step down after a few hours. If you want brightness without the tradeoff, you need a bigger panel, not a bigger battery. See making solar garden lights brighter.
If your lights are falling well short of these numbers, the problem is probably not runtime at all. Start with solar lights not working, or brand new solar lights not working if they are out of the box.
Sources
- Energizer - Nickel Metal Hydride Handbook and Application Manual data.energizer.com Discharge curve shape and the no-significant-capacity-effect-below-1C finding; temperature derating; trickle charge limits
- Panasonic - Ni-MH Batteries Handbook energy.panasonic.com The 0-45C charge range, the sub-freezing oxygen-recombination and vent mechanism, and the 2 years / 500 cycles design guidance
- QX5252 solar LED driver datasheet mikroshop.ch The LED power law that follows the square of cell voltage with no feedback path, and panel-as-sensor light control
- YX8018 solar LED driver datasheet asset.conrad.com 0.9V minimum input voltage and the inductor-selected current table confirming output is set at manufacture, not regulated
- Bridgelux SMD 2835 0.2W 3V datasheet bridgelux.com Efficacy rising from 132 to 171 lm/W as drive current falls from 80 mA to 20 mA
- US Naval Observatory - Astronomical Applications aa.usno.navy.mil Sunrise, sunset and civil twilight times at both solstices for every city in the runtime table
- NOAA NCEI - National Solar Radiation Database ncei.noaa.gov Horizontal insolation (peak sun hours) by location and month - the supply side of every figure here
- FDK - eneloop HR-3UTG cell specification fdk.com Minimum versus typical rated capacity and the 0.2C-to-1.0V rating convention
Questions people ask
Why do my solar lights only stay on for an hour or two?
Almost always an undercharged panel, not a bad battery. A budget stake light in Minneapolis in December banks only about 38 mAh from a full day of winter sun, which at a 15 mA draw is 2.5 hours. Shade, a dusty or clouded panel, or a fixture that only sees morning sun produces the same result in any season. Clean the panel and check it gets direct midday light before you replace anything.
Are there solar lights that stay on for 12 hours?
Yes, but the conditions matter more than the product. Twelve hours needs a low-draw fixture pulling 15 mA or less, a clean unshaded panel, a healthy cell, and at least 5 peak sun hours that day. That is achievable for a dim accent light in a sunny yard from spring through early fall. No 100-lumen spotlight runs 12 hours on a winter charge in the northern US, whatever the box says.
Do solar lights turn off automatically?
Yes, and in the cheapest lights the solar panel itself is the sensor. The driver chip watches the panel's own voltage and switches the LED on when it falls below a threshold, so there is no separate photocell at all. Better fixtures use a cadmium sulfide photoresistor. Either way you do not need to switch them off, and switching off at the wall does nothing.
Is it okay to leave solar lights on all night?
Yes. They are designed for it, and there is no safety reason to intervene. The real cost is battery wear: a solar light cycles once per day, and NiMH cells are rated for roughly 500 cycles under good conditions, which a trickle-charging garden light does not provide. If a light is decorative rather than functional, using its off switch through the winter will meaningfully extend battery life.
Why do solar lights fail so quickly?
Usually the battery, and usually within two to three years. Manufacturers define end-of-life as the point where a cell holds 60% of nominal capacity, and a solar light gets there fast because it trickle-charges with no termination, which both Panasonic and Energizer warn against. Panel clouding and water ingress account for most of the rest. The LED itself almost never fails first.
Does cold weather reduce solar light runtime?
Yes, on both sides. NiMH capacity drops below about 14 degrees Fahrenheit, so the cell delivers less. More seriously, charging NiMH below freezing causes permanent damage - the cell cannot absorb the oxygen it generates, pressure rises and the vent releases electrolyte. Combined with short days and a low sun angle, this is why northern winter runtime collapses rather than just declining.
How long do solar lights take to charge the first time?
Give them two to three full sunny days before judging performance, and leave them switched off if they have a switch. New cells often ship partly discharged and NiMH needs a few cycles to reach full capacity. If performance has not improved after three good days of sun, the problem is placement or hardware, not conditioning.
