Do home batteries work in the cold? Four makers' documented limits against eight cities' coldest months
Each of the four makers compared here has a current document allowing air as cold as -4 degrees F (-20 degrees C). SolarEdge's 2022 and March 2024 data sheets started at 14 degrees F (-10 degrees C), and its May 2025 warranty, October 2025 installation guide and January 2026 data sheet say -4. Only Tesla's and FranklinWH's documents describe a heater. Whether an unheated garage stays inside a window depends on the city and on which document governs.
Cold is a charging problem before it is a discharging problem. Tesla's Heat Mode paper lists no charging at all for a Powerwall 3 cell below 0 degrees C, with discharging limited, and FranklinWH's winter paper says the charging rate is affected more severely than the discharge rate. NOAA's 1991-2020 normals show how often the air is cold enough to matter: in January the Minneapolis airport averages 30.5 days with a low at or below freezing and 10.0 at or below 0 degrees F (-17.8 degrees C), Fairbanks 31.0 and 25.7, and Boston 24.8 and 0.4. Added over a year, six of the eight cities below average more days with a freezing low than the 109.5 (30 percent of a year) that SolarEdge's warranty and January 2026 data sheet allow a battery location to spend below 32 degrees F.
Heating the cells is cheap in one maker's figure and not in another's. Tesla puts a typical Heat Mode draw at 200 Wh at -20 degrees C, while FranklinWH's logged winter days on its aPower X run from under 0.5 kWh to 2.8 kWh. SolarEdge's warranty and data sheet footnote alone count cold days.
Four makers and their stated limits
The documents behind the table were published between 2022 and October 2026, and the word temperature means different things in them. Tesla, Enphase and FranklinWH state a range for the air around the battery. Tesla and FranklinWH add a second quantity, the cell temperature, because their heaters act on the cells instead of on the air. SolarEdge changed its ambient floor between editions: its 2022 and March 2024 data sheets state 14 degrees F, and its documents from May 2025 on state -4 degrees F. Comparing the ambient numbers is possible, and it is only the first step.
All four share a floor in their newest documents. The Powerwall 3 owner's manual, the IQ Battery 5P data sheet (for charging and for discharging), the FranklinWH manual and SolarEdge's warranty, installation guide and January 2026 data sheet all give -4 degrees F (-20 degrees C). Their ceilings differ: 50 degrees C for Tesla, FranklinWH and SolarEdge, 50 for Enphase charging and 55 for Enphase discharging. Tesla's wording adds two qualifiers to its number. The unit is capable of charging and discharging within the range, and at the extremes of the range it may limit charge or discharge power to improve battery lifespan. A footnote says the unit is designed to operate in all climates and in direct sunlight, and that performance may be de-rated above 40 degrees C (104 degrees F).
Enphase's data sheet carries the most placement guidance of the four. Footnote 8 tells the installer to choose a non-habitable indoor location, with a 2-car garage as its example, or an outdoor location where the ambient temperature and humidity stay within -4 to 131 degrees F (-20 to 55 degrees C) and 5 to 95 percent relative humidity, non-condensing, and to avoid direct sunlight so charging and discharging currents are not de-rated. Full performance, in the same footnote, occurs from 59 to 113 degrees F (15 to 45 degrees C) while charging and from 32 to 122 degrees F (0 to 50 degrees C) while discharging. The practical reading for a cold garage is that Enphase documents a wide permitted range and a narrower range for full performance, and the lower edge of the full-performance range for discharging is freezing.
| Maker and product | Documents and dates | Stated range for the surrounding air | Stated best-performance range | Heater described | Stated storage range |
|---|---|---|---|---|---|
| Tesla Powerwall 3 | Owner's Manual & Safety Instructions, US print edition, published June 2025; Datasheet, 2024; Heat Mode White Paper, 2024 | -20 to 50 C (-4 to 122 F). Charge or discharge power may be limited at the extremes; performance may be de-rated above 40 C (104 F) | The manual points to an "optimum temperature range specified below" and prints none | Yes. Heat Mode, with resistive heating elements at each cell | -20 to 30 C (-4 to 86 F), starting at 25 percent state of energy (datasheet) |
| Enphase IQ Battery 5P | Data sheet DSH-00010-11.0, May 2025; Limited Warranty for storage, effective on or after September 1, 2025 | Charging -20 to 50 C (-4 to 122 F). Discharging -20 to 55 C (-4 to 131 F) | Optimum 0 to 30 C (32 to 86 F). Full performance 15 to 45 C while charging and 0 to 50 C while discharging (footnote 8) | None described in the data sheet | Not stated in the data sheet |
| FranklinWH aPower 2 | System User Manual V1.3, October 19, 2026; Powering Through the Winter white paper, April 2, 2025 (data for the earlier aPower X); System one-pager, February 22, 2025 | -4 to 122 F (-20 to 50 C) in the manual. The system one-pager lists -4 to 131 F (-20 to 55 C) as a system feature | Recommended ambient 32 to 86 F (0 to 30 C) in the white paper, for the aPower X | Yes. An integrated heating blanket; two 220 W blankets on the aPower 2 | Manual: minimum storage temperature -22 F (-30 C), below which non-warrantied cell damage may occur |
| SolarEdge Home Battery 400V | Data sheet DS-000255-NAM, January 6, 2026 (reseller-hosted copy; the May 22, 2025 edition reads the same); Quick Installation Guide version 1.3, October 2025 (reseller-hosted copy); Limited Warranty revised May 2025 (Wayback copy, saved July 8, 2025). Older data sheets, dated 2022 and March 5, 2024, state 14 F | -4 to 122 F (-20 to 50 C) in the January 2026 data sheet (footnote 8) and in the warranty, with time below 32 F limited to 30 percent of the days in any rolling 12 months. The installation guide says not to operate the battery above 122 F or below -4 F. The 2022 data sheet gives 14 to 122 F (-10 to 50 C); the March 2024 data sheet gives 14 to 122 F, with 32 to 104 F required for at least 95 percent of the warranty period | Not stated | None described in the data sheets or the installation guide | January 2026 data sheet: up to 12 months since shipment, 14 to 86 F (-10 to 30 C); limited period, -22 to 140 F (-30 to 60 C) |
The SolarEdge row holds documents from different years, and a later section sets them side by side. The warranty condition is the one the warranty throughput guide also tables: a location where the ambient temperature falls between -4 and 122 degrees F, with the time below 32 degrees F limited to 30 percent of the days in any rolling 12-month period. The January 2026 data sheet carries the same wording in its footnote 8, and the version 1.3 installation guide lists an updated minimum operating temperature in its revision table and warns against operating the battery below -4 degrees F. The 2022 and March 2024 data sheets start higher, at 14 degrees F, and are older editions.
Why charging gives way before discharging
Lithium-ion cells take in energy less readily as they cool. FranklinWH's winter paper names the mechanism: low temperatures slow intercalation, the movement of lithium ions into the graphite layers of the anode, and that slowing shows up as reduced capacity, slower charge rates and higher internal resistance. Tesla's Heat Mode paper makes the same point in different words, noting that cell operating parameters such as voltage and charge current vary with cell temperature, so each cell needs to be inside an optimal range to charge and discharge effectively.
Both makers publish the consequence as a table keyed to cell temperature, not to the air. The two tables have the same four bands, and they are the most useful numbers in either paper for judging a cold mount. Tesla's table shows charging limited between 0 and 10 degrees C, none between -20 and 0, and discharging limited between -20 and 0. FranklinWH's shows charging derated in both of those bands with discharging normal until the cell reaches -20 degrees C. The FranklinWH table belongs to the aPower X, the earlier product, and the paper says the aPower 2 has not been through a full winter.
| Cell temperature | Tesla Powerwall 3: charging | Tesla Powerwall 3: discharging | FranklinWH aPower X: charging | FranklinWH aPower X: discharging |
|---|---|---|---|---|
| Above 10 C (50 F) | Normal | Normal | Normal | Normal |
| 0 to 10 C (32 to 50 F) | Limited | Normal | Derated | Normal |
| -20 to 0 C (-4 to 32 F) | None | Limited | Derated | Normal |
| Below -20 C (-4 F) | None | None | None | None |
The FranklinWH document is not fully consistent with itself. A passage under the cell derating chart says low charging current is protected between 35.6 and 41 degrees F (2 and 5 degrees C), and that charge power falls to 0 kW below 35.6 degrees F (2 degrees C), a curve that "shows the cells' performance without the heat blanket operation." The table printed after it lists charging as derated, not stopped, for a cell between -4 and 32 degrees F (-20 and 0 degrees C). The paper does not reconcile the two. It does state that the heat blanket extends the charge power window down to -4 degrees F ambient, so a buyer who needs to know how a cold cell charges should ask FranklinWH which of the two applies below 2 degrees C.
A second FranklinWH table is keyed to ambient temperature with the blanket on: charging and discharging normal above -4 degrees F (-20 degrees C), and, when the ambient is temporarily between -4 and -22 degrees F (-30 degrees C), charging temporarily derated with discharging normal. The FAQ adds the edge case. If the cell temperature itself falls below -4 degrees F, the entire system shuts down and returns when the cell warms past that point. Tesla's table has no equivalent row for ambient air; its ambient limits sit in the manual and the datasheet.
How each document handles a heater
Tesla's Heat Mode is described in a 2024 white paper, and it is not mentioned in the owner's manual or the datasheet. Each Powerwall 3 cell has a dedicated resistive heating element, and the unit uses solar and battery history from the previous 7 days to decide when and how much to heat, preparing for sunrise so it can charge as soon as solar begins producing. Below 0 degrees C ambient the paper says Heat Mode holds the internal cell at 0 degrees C for discharge and heats further ahead of expected solar surplus. When the outside air is between 0 and 10 degrees C and peak charge power is needed, it keeps the cell above 10 degrees C. A unit first commissioned in sub-zero cold heats at an average of 6.7 degrees C an hour to bring the cell into the range that allows charging, so a 20 degree C rise takes about 3 hours at that average rate.
The power for heating comes in a stated order. On grid with the battery above its backup reserve, the order is solar, then battery, then grid. The paper gives a separate order for off-grid operation, and says that off grid with the state of energy below 10 percent, Heat Mode deactivates to avoid bricking the battery until solar or grid power returns. When the unit is on grid the heating appears in the app's power flow diagram as a home load. Heat Mode is standard from firmware 23.40, which rolled out to the existing fleet in November 2023. In a February 2024 histogram of 100 Powerwall 3 sites across Colorado, Utah, New York, New Jersey, Connecticut and Massachusetts, the paper reports daily minimum cell temperatures above 0 degrees C, apart from a small group at full state of energy where the unit does not heat because it cannot charge further.
The extreme-cold passage is the one to read before choosing an outdoor or detached location. Heat Mode holds a 0 degrees C cell minimum in ambient air as low as -20 degrees C. Below that, it still activates, but depending on the depth and duration of the cold it may not heat as fast as heat is lost. If prolonged cold below -20 degrees C drags the cell temperature under that level too, the unit enters a deep sleep for up to 30 days to prevent bricking. The paper adds that prolonged exposure below -20 degrees C poses no safety hazard but may harm performance and longevity, and the same answer closes with this sentence: "Powerwall should not be installed outside in areas where it is common to have daily high temperatures of < -20°C."
FranklinWH takes a different route. Its paper says the company chose a silicone rubber pad heater, the heating blanket, over cell self-heating, and that each aPower battery has one built in. The aPower X has four pads rated at 100 W each, 400 W in all. The aPower 2 has two blankets of 220 W, and the paper says the aPower 2 had not been through a full winter, so its data are expected to be similar or better, not measured. The blanket draws from solar first, then from the grid when solar is absent, and during an outage from the batteries. The app does not show the blanket separately; its energy appears as a load. The October 2026 manual repeats the -4 to 122 degrees F range, says low cell temperatures affect charging more than discharging, and recommends indoor installation for regions with prolonged extreme cold.
The Enphase and SolarEdge documents describe no heater. The IQ Battery 5P data sheet lists charging and discharging ranges, an optimum range and a footnote on location, and it describes its cooling as passive, with no moving parts or fans. The SolarEdge Home Battery 400V data sheets (2022, March 2024 and January 2026) list an operating range and two storage ranges and give no heater, and the installation guide describes none. These are statements about those documents. They do not say whether the products contain a heater that the documents leave out, and the answer to that is a question for the installer, in writing.
Heater energy in kilowatt-hours
Tesla and FranklinWH both put a number on what heating uses, and at the high end the numbers are an order of magnitude apart. Tesla says that at -20 degrees C ambient, Heat Mode "typically will only consume 200 Wh to prepare the cells to accept daily charge from solar." FranklinWH logged real days at aPower X sites and reports a draw under 0.5 kWh a day at its outdoor Vermont systems on the coldest nights, 2.4 kWh for 6 hours of heating during a January 15 to 17, 2024 cold spell in Michigan with the air down to 5 degrees F, and 2.8 kWh a day for 7 hours of heating in South Dakota, where the January ambient reached -22 degrees F.
The two kinds of figure do not measure the same thing. Tesla's is a typical quantity for readying the cells, stated for one ambient temperature. FranklinWH's are logged consumption on named days at named sites, and its 2.8 kWh day is 14 times Tesla's 200 Wh. The 6-hour and 7-hour figures both average 400 W over the heating time, which matches the 400 W rating of the four pads running continuously. At the aPower 2's two 220 W blankets, 440 W, the same hours would be 2.64 kWh and 3.08 kWh. That last pair is arithmetic on a rating, not a measurement, because FranklinWH has not published winter data for the aPower 2.
Set against capacity, the heating energy is small on Tesla's figure and larger on FranklinWH's longest logged days, where it reaches 20.6 percent of the capacity the paper gives. Both papers put solar first in the order of power sources. In an outage with no sun, FranklinWH's blanket draws on the batteries and Tesla's off-grid order does the same, so a battery held as backup spends part of its reserve on keeping itself warm. No dollar figure is attached to any of this, because electricity rates belong to the sister site EnergyBS.
| Document and condition | Heater energy stated | Capacity in the same document | Share of capacity |
|---|---|---|---|
| Tesla Heat Mode white paper: Powerwall 3 at -20 C ambient, "typically" to prepare the cells for the day's solar charge | 200 Wh | 13.5 kWh (datasheet, nominal) | 1.5 percent |
| FranklinWH white paper: aPower X, outdoor systems in Vermont, coldest nights | Under 0.5 kWh a day | 13.6 kWh (white paper) | Under 3.7 percent |
| FranklinWH white paper: aPower X, Michigan, January 15 to 17, 2024 cold spell, ambient down to 5 F (-15 C), heating on 6 hours | 2.4 kWh | 13.6 kWh | 17.6 percent |
| FranklinWH white paper: aPower X, South Dakota, January 13 to 15, 2024, heating on 7 hours (January ambient at the site reached -22 F, -30 C) | 2.8 kWh a day | 13.6 kWh | 20.6 percent |
| FranklinWH white paper: rated heater power | aPower X: four pads of 100 W (400 W). aPower 2: two blankets of 220 W (440 W) | Not applicable | Not applicable |
Eight cities in their coldest month
NOAA's National Centers for Environmental Information publishes the U.S. Climate Normals, each a 30-year average of a measured variable, and the 1991-2020 monthly files carry the monthly mean of the daily low and high temperatures as well as the average number of days a month on which the low reached stated thresholds. The table uses one airport station per city. Fairbanks and Minneapolis stand for the deep cold, Burlington, Chicago and Boston for the Northeast and Midwest, Denver and Salt Lake City for the interior West, and Atlanta as a mild control. Seven of the eight cities have a coldest-month mean low under 32 degrees F; Atlanta's is 35.6 degrees F.
Two cautions go with the station choice. An airport station is open ground, and a home lot can differ from it in either direction. Boston's station sits 4 meters above sea level at the harbor, which a hillside lot does not share. A mean daily low also averages 30 years of nights, so many individual nights in the coldest month are colder than the figure, and the day counts in the last three columns say more about the edges than the mean does. NOAA states that its threshold counts are based on the raw temperature measurements, not the homogenized values behind the means.
The day counts are NOAA's average number of days in the month on which the daily minimum was at or below each threshold: 32, 10 and 0 degrees F. They count days on which the low touched the threshold at any hour, which is not the same as being below it all day. The table uses each station's month with the lowest mean daily low, which is December for Denver and January for the other seven.
| City (station elevation) | Coldest month | Mean daily low | Mean daily high | Days with a low at or below 32 F | Days at or below 10 F | Days at or below 0 F |
|---|---|---|---|---|---|---|
| Fairbanks, AK (132 m) | January | -17.2 F (-27.3 C) | 0.6 F (-17.4 C) | 31.0 of 31 | 29.0 of 31 | 25.7 of 31 |
| Minneapolis, MN (266 m) | January | 8.8 F (-12.9 C) | 23.6 F (-4.7 C) | 30.5 of 31 | 16.7 of 31 | 10.0 of 31 |
| Burlington, VT (101 m) | January | 12.9 F (-10.6 C) | 28.9 F (-1.7 C) | 28.7 of 31 | 14.0 of 31 | 7.3 of 31 |
| Denver, CO (1650 m) | December | 18.4 F (-7.6 C) | 44.0 F (6.7 C) | 29.2 of 31 | 6.3 of 31 | 1.9 of 31 |
| Chicago, IL (201 m) | January | 18.8 F (-7.3 C) | 31.6 F (-0.2 C) | 28.1 of 31 | 8.9 of 31 | 3.5 of 31 |
| Boston, MA (4 m) | January | 23.1 F (-4.9 C) | 36.8 F (2.7 C) | 24.8 of 31 | 3.9 of 31 | 0.4 of 31 |
| Salt Lake City, UT (1288 m) | January | 24.2 F (-4.3 C) | 38.6 F (3.7 C) | 27.4 of 31 | 3.2 of 31 | 0.4 of 31 |
| Atlanta, GA (308 m) | January | 35.6 F (2.0 C) | 54.0 F (12.2 C) | 12.8 of 31 | 0.1 of 31 | 0.0 of 31 |
Cold days against each lower limit
The newest documents of all four makers give a lower limit of -4 degrees F (-20 degrees C); SolarEdge's 2022 and March 2024 data sheets gave 14 degrees F (-10 degrees C). NOAA's day counts use thresholds of 0, 10, 20 and 32 degrees F, so none sits exactly on either edge. The 0 degrees F column is warmer than -4 degrees F, which makes it an upper bound on the days that fall below the floor, not a count of them. For the older 14 degrees F edge, the 10 degrees F column is a lower bound on the days that fall below it. The table below turns the mean low into distance from each edge, and the last column carries the upper-bound day count.
Fairbanks is the only city where the coldest-month mean low, -17.2 degrees F, is itself below the -4 degrees F floor, by 13.2 degrees. It has 25.7 days of 31 with a low at or below 0 degrees F, and a mean January high of 0.6 degrees F (-17.4 degrees C), only 4.6 degrees F above the line in Tesla's sentence about places where daily highs are commonly under -20 degrees C. The normals do not carry a count of days with a high that cold, so the sentence cannot be tested exactly, but the mean sits close enough to the line that a cold snap crosses it. Of the eight cities, Fairbanks is the one where that installation sentence applies closest to its letter.
Minneapolis has a mean low 12.8 degrees F above the floor and still counts 10.0 January days at or below 0 degrees F, and every one of those nights is within 4 degrees F of the floor or past it. The same count is 7.3 days for Burlington and 3.5 for Chicago. In Denver, Boston and Salt Lake City it is 1.9, 0.4 and 0.4 days, and in Atlanta 0.0. For the four makers sharing the floor, the air temperature alone puts an outdoor mount in those last four cities inside the window on nearly every night. It does not settle the cell question. Air at 32 degrees F is 0 degrees C, the line that separates limited charging from none in Tesla's table, and the coldest month in seven of the eight cities has at least 24 days with a low at or below it. On those nights, where a maker describes a heater, the heater or the warmth of a battery that has been cycling is what would hold the cell above the line; the Enphase data sheet describes no heater.
SolarEdge's older 14 degrees F data sheet edge would have changed the order. Minneapolis, Burlington and Fairbanks have mean lows below 14 degrees F, by 5.2, 1.1 and 31.2 degrees, with 16.7, 14.0 and 29.0 days a month at or below 10 degrees F. Chicago and Denver sit 4.8 and 4.4 degrees above the edge on the mean, and 8.9 and 6.3 days at or below 10 degrees F still put some nights under it. Boston and Salt Lake City count 3.9 and 3.2 such days. Only Atlanta, at 0.1, has essentially none. If that edge governed, an exposed battery would meet nights below it in seven of the eight cities. SolarEdge's newer -4 degrees F wording is the more permissive of the two and puts its air-temperature test where the other makers' tests are, so what remains specific to SolarEdge is the freezing-days rule in the next section. Which edition governs a particular unit is a question for SolarEdge or the installer, answered in writing.
| City | Mean daily low | Against the -4 F (-20 C) floor | Against the 14 F (-10 C) edge in SolarEdge's 2022 and 2024 data sheets | Days at or below 0 F |
|---|---|---|---|---|
| Fairbanks | -17.2 F | -13.2 F | -31.2 F | 25.7 of 31 |
| Minneapolis | 8.8 F | +12.8 F | -5.2 F | 10.0 of 31 |
| Burlington | 12.9 F | +16.9 F | -1.1 F | 7.3 of 31 |
| Denver | 18.4 F | +22.4 F | +4.4 F | 1.9 of 31 |
| Chicago | 18.8 F | +22.8 F | +4.8 F | 3.5 of 31 |
| Boston | 23.1 F | +27.1 F | +9.1 F | 0.4 of 31 |
| Salt Lake City | 24.2 F | +28.2 F | +10.2 F | 0.4 of 31 |
| Atlanta | 35.6 F | +39.6 F | +21.6 F | 0.0 of 31 |
SolarEdge's freezing-days rule, city by city
The warranty's temperature condition, repeated in footnote 8 of the January 2026 data sheet, is a counting rule, which suits the NOAA data. A SolarEdge Home Battery 400V is to be installed where the ambient temperature falls between -4 and 122 degrees F, and the allowed time below 32 degrees F is limited to 30 percent of the days in any rolling 12-month period. Thirty percent of 365 days is 109.5 days. Adding the twelve monthly averages of days with a low at or below 32 degrees F gives each city's expected count for a year.
Six of the eight cities exceed 109.5 days: Fairbanks (221.9), Denver (156.2), Minneapolis (147.2), Burlington (144.7), Chicago (121.3), Salt Lake City (111.7). Boston (91.6) and Atlanta (35.8) stay under. Salt Lake City is over the line by only 2.2 days, a margin small enough that one milder or colder winter would move it across. Fairbanks has a freezing low on 60.8 percent of days, about twice the 30 percent the clause allows.
Three details keep the comparison honest. The normals count a day when the low is at or below 32 degrees F, and SolarEdge's wording says below. Neither the warranty nor the data sheet footnote says whether a day counts when the air dips under freezing for an hour or only when it stays there, so the day count is the cautious reading: a time-based reading would show a smaller share. The data are airport air, which stands in for an exposed outdoor mount or a detached, uninsulated garage and overstates the cold of an attached garage that shares a wall with the house. And the clause tests the ambient temperature of the battery's location, which is why the location, and not the city, decides the answer.
| City | Days a year, low at or below 32 F | Share of 365 days | Against the 109.5-day limit | Months with more than 30 percent of days |
|---|---|---|---|---|
| Fairbanks | 221.9 | 60.8 percent | 112.4 days over | 8 of 12 |
| Denver | 156.2 | 42.8 percent | 46.7 days over | 6 of 12 |
| Minneapolis | 147.2 | 40.3 percent | 37.7 days over | 5 of 12 |
| Burlington | 144.7 | 39.6 percent | 35.2 days over | 6 of 12 |
| Chicago | 121.3 | 33.2 percent | 11.8 days over | 5 of 12 |
| Salt Lake City | 111.7 | 30.6 percent | 2.2 days over | 5 of 12 |
| Boston | 91.6 | 25.1 percent | 17.9 days under | 4 of 12 |
| Atlanta | 35.8 | 9.8 percent | 73.7 days under | 1 of 12 |
Show the day counts as a table
| City | Days a year at or below 32 F |
|---|---|
| Fairbanks | 221.9 |
| Denver | 156.2 |
| Minneapolis | 147.2 |
| Burlington | 144.7 |
| Chicago | 121.3 |
| Salt Lake City | 111.7 |
| Boston | 91.6 |
| Atlanta | 35.8 |
What a garage does to the numbers
None of the four sets of documents says how much warmer an unheated garage runs than the air outside, and the NOAA files do not measure garages. A figure for the offset would have to come from a logger in the actual building, and anything else is a guess. What the sources support is a bracket. The outdoor normals above are the cold bound for a mount that is exposed or in a detached, unheated, uninsulated structure. A garage that shares walls with the heated house tends to run warmer than the outdoor air, and by how much is a property of that building.
The battery's own temperature adds a second offset. Tesla's white paper says the cell is insulated from ambient swings by the cell's thermal mass and by self-heating as a by-product of normal daily operation, and it includes a chart of cell and ambient temperature at a site without Heat Mode. FranklinWH says that in winter, if the aPower X has been operating instead of sitting in standby, the average cell temperature is usually higher than the ambient. Both statements describe a battery that is cycling. A unit in standby on a cold weekend has less self-heating to count on than one that cycles every day.
A measurement is cheap. A min-max thermometer or a data logger placed at the intended mounting height for one winter, with the coldest morning's low and the number of nights below freezing written down, can be compared with the table in this guide on the same terms the NOAA data use: lows at or below 32, 10 and 0 degrees F. If a December and January of logging shows the garage lows within a few degrees of the outdoor lows, the outdoor normals apply directly. If the garage logs lows well above the outdoor lows, the cold-weather rows of the tables above matter less for that wall.
A garage battery also has to satisfy fire and building codes on location, distances and kilowatt-hour limits, which the guide to where a home battery can go works through edition by edition. Its tables have no temperature column. A location can pass an inspection and still sit outside the maker's operating range, and the reverse holds too, so the two questions need separate answers on the permit drawing and in the maker's documents. A battery added later to an existing array, which the guide to adding a battery to existing solar covers, raises the same question about whichever wall the installer proposes.
A detached garage near Minneapolis, tested against the four sets of limits
A hypothetical owner near Minneapolis plans to mount a battery on the wall of a detached, unheated garage. Every figure below is the airport station's normal, which stands in for the cold bound and is not a measurement of that garage. The January mean daily low is 8.8 degrees F, 12.8 degrees above the -4 degrees F floor that all four makers share, and up to 10.0 of 31 January days have a low at or below 0 degrees F.
For a Powerwall 3 or an aPower 2, the ambient floor is crossed only on the nights when the low reaches -4 degrees F or colder, a count the normals do not give. What the documents rely on for the cell is the heater, so the question for the installer is how the heater behaves with the unit in standby through a gray week. For an IQ Battery 5P the data sheet describes no heater, its stated protection is the -20 degrees C ambient floor, and its full-performance discharging range begins at 0 degrees C; 30.5 of the 31 January days have a low at or below that, so on most nights the air is under the bottom of the full-performance range.
For SolarEdge, the current documents give the same -4 degrees F floor, and the freezing-days rule is the harder test: Minneapolis averages 147.2 freezing-low days a year, 37.7 more than the 109.5-day limit, with 5 of 12 months over 30 percent. On the day-count reading, an exposed detached garage would sit outside the SolarEdge warranty condition, and the owner would need the maker's written confirmation, a warmer location or a different battery. Under SolarEdge's older 14 degrees F data sheet edge, which is 5.2 degrees above the mean low, with 16.7 January days at or below 10 degrees F, the garage would also be outside the range on air temperature. A winter of garage logging would decide it.
Where the documents disagree with themselves
Five more conflicts or gaps show up inside the documents compared, and each changes what a careful buyer should ask for in writing.
SolarEdge's editions give different floors. The North America data sheets dated April 28, 2022 and March 5, 2024 list an operating temperature of 14 to 122 degrees F (-10 to 50 degrees C). The warranty revised in May 2025, data sheet DS-000255-NAM (dated May 22, 2025 and again January 6, 2026) and the installation guide version 1.3 of October 2025 give -4 to 122 degrees F, and the guide's revision table lists an updated minimum operating temperature. The March 2024 data sheet lists part BAT-10K1PS0B-02 and the 2025 and 2026 data sheets list UBAT-10K1PS0B-03, and none of the documents says why the floor moved. The January 2026 data sheet is the newest SolarEdge data sheet found, so -4 degrees F is the figure used here. A buyer in a cold climate should still ask which edition governs a specific unit, and for the answer on the quote.
FranklinWH's system one-pager for the aGate 1.3 and aPower 2, dated February 22, 2025, lists operation in a wide temperature range of -4 to 131 degrees F (-20 to 55 degrees C) among the system's features. The System User Manual V1.3 of October 19, 2026 and the winter paper both give -4 to 122 degrees F (-20 to 50 degrees C) for the battery. The manual is the later and battery-specific document, so 122 is the ceiling used here. The manual also warns that prolonged cold in extreme regions is often far below the minimum operating temperature and below the minimum storage temperature of -22 degrees F (-30 degrees C), where non-warrantied battery cell damage may occur. The word non-warrantied is the one to underline.
Tesla's manual tells the reader that for best performance the average ambient temperature over the system's lifetime should fall within "the optimum temperature range specified below," and then prints one operating range and no optimum range. Another maker's optimum, such as Enphase's 0 to 30 degrees C or FranklinWH's recommended 0 to 30 degrees C, cannot be assumed to apply to Powerwall 3. The Heat Mode paper does supply cell-temperature bands, 0 to 10 degrees C and above 10, which are about the cell and not the air.
Enphase's first page gives a charging range of -20 to 50 degrees C and a discharging range of -20 to 55, while footnote 8 on page 3 asks for an install site within -20 to 55 degrees C and states a full-performance charging range of 15 to 45. The footnote's 55 degrees C site limit is 5 degrees above the charging ceiling in the table; the cold end of both is -20. A reader should take 50 as the charging ceiling.
Finally, the FranklinWH winter paper's cell-temperature percentages are given as 2.6 percent, 11.6 percent and 86 percent in one list, and as 31.8 percent and 53.5 percent in the same list, which add to 85.3, not the 86 percent they subdivide. One figure caption places a South Dakota chart under a Massachusetts heading. These are minor, and no total is built from those percentages here.
What to get in writing before a cold-climate battery quote is signed
- Which of the maker's documents governs the quote, with its edition date, and does it state a different operating range from the data sheet or warranty? SolarEdge's 2022 data sheet and 2025 warranty, for one, do not agree.
- Does the product contain a heater, and if so what is its rated power and what does the document say it draws on a cold night?
- What power source runs the heater when the sun is down and the grid is out, and at what state of charge does it stop heating?
- What does the unit do when the cell temperature, not the air, falls below the floor: derate, shut down or enter a sleep mode, and how does it restart?
- What is the lowest storage temperature, and does any warranty text treat damage below it as outside coverage?
- Has the installer logged or been given a lowest recorded air temperature at the proposed wall, and where does it stand against the table in this guide?
Cold, damage and the warranty text
Tesla's manual cautions that operating or storing Powerwall 3 in temperatures outside its specified range might cause damage to Powerwall; Tesla's warranty text is not part of this comparison. The other three makers tie temperature to coverage in their warranties, and only SolarEdge's wording, in its warranty and data sheet, contains a count of cold days. The Enphase storage warranty effective on or after September 1, 2025 excludes defects in or damage to products "that have been improperly installed, operated, handled or used, including use under conditions for which the product was not designed, use in an unsuitable environment, or use in a manner contrary to the Enphase User Manual or applicable laws or regulations;" as the second item of its exclusion list. FranklinWH's warranty effective October 28, 2026 lists, under external and environmental factors, "Installation in conditions outside of FranklinWH guidance, such as extreme temperatures, humidity, altitude, or corrosive environments."
The Enphase and FranklinWH warranties lean on the maker's own operating guidance, and SolarEdge's states a limit with a counting rule. A battery mounted where the air regularly falls under a stated floor risks becoming a claim the maker can decline on environmental grounds. The battery warranty throughput guide tables these clauses beside the megawatt-hour caps and the years, and cold is one of several conditions there that can shorten coverage. Whether the battery earns its place at all is the separate test in the battery decision guide, and a cold-climate site can add lines to the quote that the battery cost guide itemizes.
Home battery cold weather questions
Do home batteries work in cold weather?
The Powerwall 3 manual, the IQ Battery 5P data sheet and the FranklinWH manual give -4 degrees F (-20 degrees C) as the lowest air temperature, and so do SolarEdge's warranty, installation guide and January 2026 data sheet, so the answer is yes inside those ranges. Charging is restricted before discharging as cells cool: Tesla's Heat Mode paper lists no charging below a cell temperature of 0 degrees C. Tesla and FranklinWH describe heaters to keep the cells above that line; the Enphase and SolarEdge data sheets describe none.
What is the lowest temperature a Tesla Powerwall 3 can operate in?
The owner's manual and datasheet state -20 degrees C (-4 degrees F) for operation, with power possibly limited at the extremes. The Heat Mode white paper says the unit holds a cell minimum of 0 degrees C in ambient air as low as -20 degrees C, and below that it may not keep up. It says Powerwall should not be installed outside in areas where daily highs are commonly under -20 degrees C.
Can I put a home battery in an unheated garage in a cold state?
The documents allow it where the air stays inside the stated range, and Enphase's data sheet names a 2-car garage as an example of a suitable non-habitable indoor location. Whether a given garage stays inside depends on its real winter lows. Log them for a winter, compare them with the tables above, and ask the installer which maker document governs. Code rules on location are separate and are covered in the guide to where a home battery can go.
How much power does a battery heater use?
Tesla says Heat Mode typically uses 200 Wh to prepare the cells at -20 degrees C ambient. FranklinWH's aPower X logs show under 0.5 kWh a day at coldest-night Vermont sites, and 2.4 to 2.8 kWh on cold days in Michigan and South Dakota; the aPower 2 has two 220 W blankets and no published winter data. The documents say heating draws on solar first.
Does cold weather void a home battery warranty?
Cold alone does not void it, but a mount outside the stated limits can lose coverage. The Enphase and FranklinWH warranties exclude damage from conditions outside the product's design or the maker's guidance, Tesla's manual cautions that operating outside the specified range might cause damage, and SolarEdge's warranty and January 2026 data sheet state a numeric condition: a location with ambient between -4 and 122 degrees F and no more than 30 percent of days in a rolling 12 months below 32 degrees F. Inside the stated limits these clauses are not engaged; outside them, the maker may decline a claim.
Is it better to install a battery indoors in a cold climate?
FranklinWH says so in its own documents: for regions with prolonged extreme cold, indoor installation is a best practice, and its winter paper recommends it where cold is prolonged. Tesla's paper says not to install a Powerwall outside where daily highs are commonly below -20 degrees C. Neither document sets the rule for a milder place. The code guide covers the indoor locations that building rules permit.
Do heated batteries still charge from solar on a freezing morning?
Tesla's paper says Heat Mode uses the previous 7 days of solar and battery history to warm the cells ahead of expected solar, and a unit commissioned in sub-zero cold heats at an average 6.7 degrees C an hour to bring the cell into the range that allows charging. FranklinWH's cell table lists charging as derated, not stopped, between -20 and 0 degrees C, although the same paper's cold-cell curve shows zero charge below 2 degrees C without the blanket. Neither is a promise of charge rates on a particular morning.
The mount decides the answer, and the paperwork only frames the question. Each maker's document reduces to a lower edge for the air, a statement about a heater, and a warranty clause that borrows both. The coldest nights on one particular wall, logged for one winter, are the number that sits against them.
Sources and verification
- Tesla, Powerwall 3 Owner's Manual & Safety Instructions, US print edition (published June 2025): Care and Maintenance, Environmental Requirements: operating -20 to 50 C, de-rating above 40 C; safety caution on temperatures outside the specified range. Checked November 3, 2026.
- Tesla, Powerwall 3 Datasheet (2024): Operating -20 to 50 C; storage -20 to 30 C at 25 percent state of energy; nominal 13.5 kWh. Checked November 3, 2026.
- Tesla, Powerwall 3 Heat Mode White Paper (2024): Cell-temperature behavior table, 0 C hold, 6.7 C per hour, power-source order, 200 Wh, deep sleep, installation sentence. Checked November 3, 2026.
- Enphase, IQ Battery 5P data sheet DSH-00010-11.0-EN-2025-05-06: Charging -20 to 50 C, discharging -20 to 55 C, optimum 0 to 30 C, footnote 8 on location. Checked November 3, 2026.
- Enphase, Limited Warranty, Storage Systems (effective on or after September 1, 2025): Exclusion for use in an unsuitable environment or contrary to the User Manual. Checked November 3, 2026.
- FranklinWH, Powering Through the Winter white paper (April 2, 2025): aPower X environmental specifications, cell tables, heating blanket power and logged energy, FAQ. Checked November 3, 2026.
- FranklinWH, System User Manual V1.3 (October 19, 2026): aPower 2 and aPower S operating range, heating blanket section, minimum storage temperature. Checked November 3, 2026.
- FranklinWH, System one-pager for aGate 1.3 and aPower 2 (February 22, 2025): Operating range stated as -4 to 131 F. Checked November 3, 2026.
- FranklinWH, System Limited Warranty for End Users (effective October 28, 2026): External and environmental factors exclusion. Checked November 3, 2026.
- SolarEdge, Limited SolarEdge Home Battery 400V Warranty for North America, revised May 2025 (Wayback copy, saved July 8, 2025): Ambient -4 to 122 F; below 32 F on no more than 30 percent of days in a rolling 12 months. Checked November 3, 2026.
- SolarEdge, Home Battery 400V datasheet DS-000255-NAM, USA Domestic Content Eligible, dated January 6, 2026 (reseller-hosted copy): Ambient operating -4 to 122 F (-20 to 50 C); footnote 8: below 32 F limited to 30 percent of the days in any rolling 12 months; storage ranges; part UBAT-10K1PS0B-03; no heater described. Checked November 3, 2026.
- SolarEdge, Home Battery 400V datasheet DS-000255-NAM, edition dated May 22, 2025 (reseller-hosted copy): Same -4 to 122 F range and 30 percent footnote as the January 2026 edition. Checked November 3, 2026.
- SolarEdge, Home Battery 400V Quick Installation Guide, North America, version 1.3 (October 2025) (reseller-hosted copy): Revision table: "Updated minimum operating temperature"; warning not to operate the battery above 122 F or below -4 F. Checked November 3, 2026.
- SolarEdge, Home Battery 400V datasheet DS-000167-NAM, dated March 5, 2024 (reseller-hosted copy): Operating 14 to 122 F; footnote 8: 32 to 104 F for at least 95 percent of the warranty period; part BAT-10K1PS0B-02. Checked November 3, 2026.
- SolarEdge, Home Battery 400V datasheet, North America Spanish edition dated April 28, 2022 (Wayback copy, saved August 12, 2024): Operating 14 to 122 F; storage ranges. Older edition, superseded by the 2025 and 2026 documents. Checked November 3, 2026.
- NOAA National Centers for Environmental Information, U.S. Climate Normals 1991-2020, monthly normals station files: Files USW00026411 (Fairbanks), USW00014922 (Minneapolis), USW00014742 (Burlington), USW00094846 (Chicago), USW00014739 (Boston), USW00003017 (Denver), USW00024127 (Salt Lake City), USW00013874 (Atlanta); variables MLY-TMIN-NORMAL, MLY-TMAX-NORMAL, MLY-TMIN-AVGNDS-LSTH000, 010, 032. Checked November 3, 2026.
- NOAA National Centers for Environmental Information, U.S. Climate Normals product page: Definition of a normal as a 30-year average; 1991-2020 period. Checked November 3, 2026.
- NOAA NCEI, Normals Calculation Methodology 2020: Monthly mean minimum temperature normals; threshold exceedance counts based on raw measurements. Checked November 3, 2026.
- NOAA NCEI, Monthly U.S. Climate Normals 1991-2020 documentation: Definitions of the days-at-or-below-threshold variables. Checked November 3, 2026.
The table records what four makers' documents state, as of the editions dated in the Sources list, and it ranks no battery and makes no claim about how any unit performs. Other makers' products are not in the table. The SolarEdge data sheets and installation guide are cited from reseller-hosted copies, which may differ from the files on SolarEdge's own sites, and no SolarEdge document newer than the January 2026 data sheet was found. NOAA figures are 1991-2020 airport normals, not forecasts and not measurements of a garage. Heater energy figures are the makers' own, from different products and conditions. A document from the maker, the installer's written confirmation and a winter of readings from the actual wall outrank this page.