Mostly yes. A solar garden light in full sun runs about even: a good summer day puts back roughly what the light burns that night, with almost no margin. Cut the light to 15% and you are not running a dimmer lamp — you are running a battery down. Here is what every common spot actually delivers, measured.
What each sky is worth, in watts and milliamps
These are one-minute measurements at solar noon from the NOAA SURFRAD station at Penn State, Pennsylvania (40.7°N), across 2025. The milliamp column assumes a typical stake-light panel that puts out about 40 mA in full sun, which is what the box means by “6 to 8 hours of direct sunlight” — 40 mA for 7.5 hours is 300 mAh, one AA NiMH cell. If your light carries a 600 mAh cell, which some do, double the last column.
| Condition | Measured irradiance | % of full sun | Charge current | Hours to fill a 300 mAh cell |
|---|---|---|---|---|
| Clear sky, summer noon | 950–1,100 W/m² | 100% | 40 mA | 7.5 |
| Thin or broken cloud | 400–900 W/m² | 40–90% | 16–36 mA | 8–19 |
| Bright overcast | 200–450 W/m² | 20–45% | 8–18 mA | 17–38 |
| Heavy overcast, summer | 78–110 W/m² | 8–11% | 3–4 mA | 75–100 |
| Heavy overcast, December | 26–60 W/m² | 3–6% | 1–2.4 mA | 125–300 |
| Open shade, clear sky, whole sky visible | 110–230 W/m² | 11–23% | 4–9 mA | 33–75 |
| Darkest reading of the year (31 Jan, storm) | 16 W/m² | 1.6% | 0.6 mA | 500 |
Two things jump out. Bright overcast is far better than its reputation — you lose about two thirds, not everything. And open shade on a blue-sky day is worse than overcast. That surprises people, because open shade looks bright. It is bright to your eye because your pupils opened. The panel has no pupils.
Three more spots, from long-term irradiation data for Garden Grove, California (33.8°N), where we ship from. Figures are annual totals on the surface in question, against a panel tilted 30° to the south:
| Surface | Annual (kWh/m²) | vs. 30° south | December | December vs. 30° south |
|---|---|---|---|---|
| Tilted 30°, facing south | 2,244 | 100% | 143 | 100% |
| Flat (how your light ships) | 1,969 | 88% | 87 | 61% |
| Vertical, south-facing wall | 1,388 | 62% | 148 | 104% |
| Vertical, west-facing wall | 1,159 | 52% | 58 | 41% |
| Vertical, east-facing wall | 939 | 42% | 51 | 36% |
| Vertical, north-facing wall | 294 | 13% | 12 | 8% |
A north-facing wall gets 13% of the light a properly aimed panel gets, and 8% in December. That is not “charges more slowly.” That is a light that quits around Halloween and does not come back until March.
Dappled shade under a tree is the hardest one to pin down, because it depends on the tree. Open-grown urban shade trees measured by the US Forest Service ran a leaf area index of 2.25 to 6.29, mean 3.69. Run that through the standard Beer’s-law canopy attenuation and you get roughly 5% to 30% transmitted, around 16% for the average tree — in the same band as open shade. But dappled shade behaves much worse than that number suggests, for the reason in the next section.
Why a shadow on one corner costs far more than its area
This is the part nobody explains, and it is the most useful thing on this page.
Your garden light’s panel is not one cell. It is four to twelve narrow cells wired in series, because one silicon cell makes about half a volt and you need roughly two volts to push charge into a 1.2 V NiMH cell through a diode. Series wiring means every cell carries the same current. So the whole panel’s current is set by whichever cell is getting the least light.
The PV Education group at UNSW states the consequence flatly: shading one cell to half cuts the whole module’s power to half, and completely shading one cell drops the module’s output to zero no matter how many cells are in the string.
Work it through on a real panel. A 2 V garden panel is four cells, each a quarter of the glass:
- A leaf covering one whole cell — 25% of the panel area — takes output to roughly zero. A 4× penalty.
- A fence-post shadow covering half of one cell — 12.5% of the area — takes output to about 50%. Also 4×.
- The same shadow lying across all four cells equally, covering 25% of the panel, costs about 25%. A 1× penalty — proportional, fair, survivable.
The worst-case multiplier equals the number of cells in series. On a 5.5 V panel, that is nine or ten. Roof modules survive this because they carry bypass diodes that route current around a shaded group. A garden light is far too small and too cheap to carry them; the one diode inside is a blocking diode in series with the battery, and it does nothing whatsoever about shading.
Two practical rules fall out of this:
- A shadow that sweeps across the whole panel costs you its area. A shadow that parks along one cell costs you many times its area. Look at your panel in sunlight and find the fine lines dividing the cells. Then orient the light so that whatever casts a shadow — a railing, a stem, a downspout — throws that shadow across the lines rather than along one of them.
- Dappled shade is the worst of all shade. Averaged over the day a tree canopy might pass 16%, but the leaf pattern is exactly the case this penalty punishes: part of the panel lit, part dark, all day. A spot under a tree that meters 16% of full sun will behave like 4%. If you want light under a tree, use a fixture with the panel on a cord and put the panel in the open — our solar spot lights and tree uplights are built that way for exactly this reason.
The cliff: deep shade gives you zero, not “slow”
Every article on this topic treats charging as linear. It is not, at the bottom end.
A panel has to beat the battery’s voltage plus the blocking diode’s drop before a single electron moves — call it 1.45 V for a NiMH cell part-charged, plus 0.3–0.7 V for the diode, so somewhere around 1.75–2.15 V. A four-cell panel makes about 2.4 V open-circuit in full sun. Open-circuit voltage falls logarithmically with light: roughly 60 to 120 mV per cell for every tenfold drop in irradiance, which for four cells is 0.24 to 0.47 V per decade.
Drop that panel to 1% of full sun — around 10 W/m², which is a dark storm afternoon or genuinely deep shade — and its open-circuit voltage lands between about 1.5 V and 1.9 V. Under load it is lower still. You are at or below the floor. Not a trickle. Nothing.
That is why “it charges a little in the shade” is bad advice. Above roughly 5% of full sun you get a proportional trickle. Below it you fall off a cliff, and the battery then discharges every night with nothing coming back. A few weeks of that and the cell is flat and sulking, which is how most “dead” garden lights actually die. If yours already has, start with the usual suspects before buying a battery.
Your panel is lying flat, and flat is wrong from September to April
Garden lights ship with the panel horizontal because it is cheap to mold and it looks tidy. Horizontal is the correct angle for a solar panel only at the equator. Here is what tilting the panel 30° toward true south buys, from long-term irradiation records at three US latitudes:
| Location | Flat, annual | 30° south, annual | Annual gain | Flat, Dec | 30°, Dec | December gain |
|---|---|---|---|---|---|---|
| Garden Grove, CA (33.8°N) | 1,969 | 2,244 | +14% | 87 | 143 | +64% |
| Kansas City, MO (39.1°N) | 1,572 | 1,824 | +16% | 57 | 96 | +70% |
| Seattle, WA (47.6°N) | 1,230 | 1,428 | +16% | 25 | 44 | +80% |
All figures kWh/m². The annual gain is a useful 14–16%. The December gain is 64–80%, and December is when you need it — the nights are longest, the sun is lowest and the battery is coldest.
Now the part that explains why manufacturers get away with it: in June, flat actually beats 30° tilt at all three latitudes (220 vs 198 in Garden Grove, 199 vs 187 in Kansas City, 169 vs 166 in Seattle). A flat panel is optimized for the one season when the nights are shortest and you needed the least.
If your light has an adjustable head — most spot lights and wall lights do — set it to about 30° and leave it. If you want to chase the seasons, steepen to 45–50° from November to February; December output at 45° beats 30° everywhere in the lower 48. More on the arithmetic in our guide to optimum panel angle, and on cold-weather behavior in do solar panels work in winter.
Behind a south-facing window: the glass is the smaller problem
Putting the light on a sunny windowsill is the standard suggestion, and it works far less well than people expect for two reasons, in this order.
Geometry first. A window is a vertical aperture in a wall. A south-facing vertical surface in Garden Grove collects 62% of what a 30° panel collects over a year, and only 30% of it in June, because summer sun comes from overhead and a wall cannot see it. Worse, the panel usually lies flat on the sill facing the ceiling, so it never points at the window at all — it sees a patch of sunlit floor and a rectangle of sky through the glass. The sun patch also crosses the sill and moves off the panel within an hour or two.
Then the glass. Plain single-pane glass passes most solar energy. Modern windows do not, by design: they are rated by Solar Heat Gain Coefficient, which the Department of Energy defines as the fraction of the sun’s heat that gets through, on a 0-to-1 scale, and low-emissivity coatings exist specifically to push that number down. Low-e coatings reject near-infrared while leaving visible light alone, so the window still looks bright and clear — but near-infrared is precisely where a crystalline silicon panel does a good deal of its work. Read the NFRC sticker on your window; the lower the SHGC, the less your panel gets. Our explainer on low-e glass covers how the coatings do it.
If you need to top a light up indoors, stand it against the glass facing out rather than flat on the sill, and read charging solar lights without sun — which has the real numbers for bulbs, and the reason the popular incandescent advice is wrong. And no, moonlight does nothing.
The two-minute test that tells you if a spot will work
No meter needed. Pick a sunny day.
- Hold a sheet of white paper flat, exactly where the panel will sit.
- Hold your hand about a foot above the paper and look at the shadow.
- Do this at 10 a.m., 1 p.m. and 4 p.m.
What the shadow tells you:
- Crisp, dark, sharp-edged — direct beam is landing. Near full output.
- Soft, grey, fuzzy-edged — diffuse light only. Roughly 10–20% of full sun.
- No shadow you can make out — deep shade. Assume zero, not “a little.”
Score it: 3 crisp shadows and the spot works year-round anywhere in the US. 2 of 3 works April through September and gets marginal in December. 1 of 3 is a summer-only spot. 0 of 3 will not work at any time of year, and no amount of cleaning, battery upgrading or waiting will change that.
If you own a multimeter, the direct version takes another minute: probe the panel’s short-circuit current in the candidate spot and compare it against full sun. Anything under about a quarter of the full-sun reading will not sustain a nightly run. Our guide to testing solar panels walks through it.
If the spot fails the test
In rough order of how much they buy you:
- Move the light. Ten feet is often the difference between 0 of 3 and 3 of 3. Nothing else comes close.
- Split the panel from the fixture. Spot lights, security lights and shed lights with a panel on 10–16 feet of cable let you light a shaded spot from a sunny one. This is the real fix for shade.
- Tilt the panel 30° south. Free, and worth 14–16% a year and up to 80% in December.
- Turn the light off for a week. A cell that has been walked down needs to gain more than it loses. The on/off switch exists for this.
- Clean the panel. Real, but modest — a few percent on a dusty panel, more on a hazed one. Method here: cleaning garden light panels.
- A bigger battery. Last on the list, and largely inert in a dim spot, because a larger cell takes proportionally longer to fill and a panel that cannot fill the old one will not fill a bigger one either. It does no harm, which matters if you have already bought one. See higher-mAh batteries in solar lights.
The honest summary: a solar light in full sun is roughly break-even on the day. There is no spare capacity to absorb a bad location. Direct sun for four hours beats weak light all day, every time — and if the spot you want cannot deliver it, move the panel rather than the expectations.
Sources
- NOAA Global Monitoring Laboratory - SURFRAD Network (Penn State station, 2025) gml.noaa.gov Measured one-minute global, direct-normal and diffuse irradiance at solar noon for clear, thin-cloud, overcast and diffuse-only conditions - the entire first table
- PVGIS v5.2, European Commission Joint Research Centre re.jrc.ec.europa.eu In-plane annual and monthly irradiation for Garden Grove, Kansas City and Seattle at 0, 15, 30 and 45 degree tilt and for vertical south, east, west and north surfaces - both the orientation table and the tilt table
- PV Education (UNSW School of Photovoltaic and Renewable Energy Engineering) - Shading pveducation.org That shading one cell to half cuts the whole module's power to half, and completely shading one cell takes module output to zero - the series-string penalty
- PV Education (UNSW) - Mismatch effects in arrays pveducation.org The roles of bypass versus blocking diodes, and why the blocking diode in a garden light does nothing about shading
- PV Education (UNSW) - Open-circuit voltage pveducation.org The diode equation behind the logarithmic fall of panel voltage with irradiance, which sets the voltage floor in the cliff section
- Peper & McPherson (1998), Journal of Arboriculture 24(2), USDA Forest Service research.fs.usda.gov Measured leaf area index of 2.25-6.29 (mean 3.69) for open-grown urban shade trees, the basis for the dappled-shade transmittance estimate
- US Department of Energy - Guide to Energy-Efficient Windows energy.gov The definition and 0-to-1 scale of Solar Heat Gain Coefficient, and low-e coatings as deliberate solar-heat rejection
Questions people ask
Will solar lights work if they are not in direct sunlight?
They will light up, but usually not for a full night and not through the winter. Indirect light on a clear day runs roughly 11-23% of full sun, which works out to about 4-9 mA into a panel that makes 40 mA in direct sun. That is 33 to 75 hours of daylight to replace one night's run. In practice the battery loses ground every day until the light stops coming on.
Can solar lights still charge in the shade?
Open shade with a clear view of the sky, yes, weakly - measurements put it at 110-230 W/m2 against about 1,000 in full sun. Deep shade, no: once the panel drops to roughly 1% of full sun it can no longer exceed the battery voltage plus the blocking diode drop, and the current is zero rather than small. Dappled shade under a tree performs far worse than its average light level suggests.
How do you get solar lights to work in the shade?
Separate the panel from the fixture. Spot lights, security lights and shed lights sold with the panel on 10-16 feet of cable let you mount the light where you want it and the panel where the sun is, which is the only fix that reliably works. Everything else - cleaning, bigger batteries, reflectors, mirrors - moves the number by a few percent when you need a factor of five.
Do solar lights charge on cloudy days?
Better than most people expect. Bright overcast measures 200-450 W/m2, about 20-45% of full sun, so a light gets a partial charge and usually runs a few hours. Heavy overcast is a different story: 78-110 W/m2 in summer and as little as 26-60 W/m2 in December, which is a 1-2 mA trickle. Two or three of those days back to back and the light will not come on.
Can I charge solar lights without the sun?
Only two methods move useful energy: daylight, even poor daylight, and a bright bulb held within about two inches of the panel. A 9 W LED at two inches takes roughly 40 hours to fill a 300 mAh cell; the same bulb a foot away never fills it, because the panel supplies less current than the battery loses to self-discharge. The popular incandescent-bulb advice is wrong - most of that bulb's output is infrared the panel cannot use.
Are you supposed to leave solar lights on all the time?
Normally yes, but not if the light is failing. If a light has been in a poor spot and the battery has walked down, switch it off for five to seven days somewhere sunny so the panel gains without the LED spending it every night. The switch exists for exactly this, and for storing lights over winter without ruining the cell.
Does a north-facing wall work for solar lights?
Barely, and not in winter. A north-facing vertical surface in Southern California collects about 13% of what a 30-degree tilted south-facing panel collects over a year, and only 8% of it in December. It sees diffuse sky and ground bounce, never the direct beam. Use a fixture with a remote panel, or accept a light that works from roughly May to September and quits.
