Solar string sizing calculator
Pick a module and an inverter, set the site temperature extremes, and get the string lengths that are actually safe, checked against cold open circuit voltage, hot MPP voltage and current per tracker.
Solar string sizing calculator
1 PV module
Every box is editable. Pick the closest module from the library, then type your own datasheet values over the top. β Voc is negative.
More module data
These rarely change. Max system voltage is 1500 V on almost every module built since about 2018, and α Isc sits near +0.05 on all crystalline silicon.
2 Inverter
3 Site and mounting
Design extremes, not monthly averages. The cold figure decides whether the inverter survives a clear winter dawn, so it is the one worth looking up properly for your own site rather than taking the preset.
Advanced
Datasheets rarely publish βVmp, so it has to be derived. Scaling βVoc by Voc divided by Vmp lands closest to published measured values. Leaving βVoc unchanged understates the hot voltage drop and is the least safe of the three.
1.25 per IEC 62548 and NEC 690.8, covering irradiance above 1000 W/m². Used for the string protection and cable figure, not for the inverter input check.
Where your string sits in the inverter's voltage window
The green band is the MPPT tracking range and the red line is the absolute maximum input voltage. Your string needs its cold Voc left of the red line, and its hot Vmp inside the green band.
Every string length, checked
Voc cold is the number that damages inverters. Vmp hot is the number that costs you yield. Click a row to select it.
| Modules | Voc cold | Vmp cold | Vmp STC | Vmp hot | String kWp | Verdict |
|---|
Sizing follows IEC 62548 and standard inverter matching practice. It is a design aid, not a substitute for the inverter manufacturer's own string configurator, the module datasheet revision for your delivered batch, or a full electrical design covering cable sizing, protection coordination, earthing and arc fault detection. Library values are transcribed from published datasheets and may be superseded.
What string sizing actually decides
String sizing, solar PV array sizing and inverter sizing are different names people use for the same design question: how many modules go in series, how many strings sit in parallel, and whether that arrangement fits the inverter you have chosen. If you have searched for any of those terms, this is the calculation you are looking for.
Solar modules are wired in series into strings, and every module in a string adds its voltage to the total while the current stays the same. How many you put in one string is not a matter of preference. It is bounded on both sides by the inverter, and getting it wrong fails in two very different ways: too many modules destroys hardware, too few quietly wastes yield for the life of the plant. Everything else in an array design, the layout, the cable routes, the mounting, can be adjusted later. String length is set on day one and is expensive to change.
The three numbers that set the limits
A string is valid when all three of these hold at the same time, at the temperature extremes the site actually reaches rather than at the 25 degrees Celsius of the datasheet:
Voc(cold) = Voc × (1 + βVoc ÷ 100 × (Tcold − 25))
Vmp(hot) = Vmp × (1 + βVmp ÷ 100 × (Thot − 25))
Isc(hot) = Isc × (1 + αIsc ÷ 100 × (Thot − 25))
The first sets the maximum string length: multiplied out across the string it must stay below the inverter's maximum input voltage and below the module's own maximum system voltage. The second sets the minimum: the string has to stay above the MPPT window's lower edge when the modules are at their hottest. The third decides how many strings you can put in parallel on one tracker before the input current rating is exceeded.
The full set of rules, written out
Those three corrected values then feed five limits. The whole calculator is these five lines plus the arithmetic above, and the string length it recommends is the largest number that satisfies all of them.
| Limit | Rule | What it protects |
|---|---|---|
| Max series | Nmax = floor( Vlimit ÷ Voc(cold) ) | The inverter, and the module insulation. Vlimit is the lower of the inverter maximum input voltage and the module maximum system voltage. |
| Min series | Nmin = ceil( Vmppt,min ÷ Vmp(hot) ) | Tracking on the hottest afternoons. |
| Tracking ceiling | N × Vmp(cold) ≤ Vmppt,max | Winter yield. Above this the inverter is safe but cannot follow the true maximum power point. |
| Start up | N × Vmp(hot) ≥ Vstart | Waking up in poor light. |
| Strings in parallel | P = min( Pinverter, floor(Imppt ÷ Imp), floor(Isc,mppt ÷ Isc(hot)) ) | The MPPT input stage. |
Round the maximum down and the minimum up, never the other way. Rounding a maximum up is how an inverter gets destroyed. Protective devices and DC cable are then sized on the design current, which IEC 62548 and NEC 690.8 both put at 1.25 × Isc to allow for irradiance above 1000 W/m², a factor you can change in the Advanced panel.
Cell temperature is not air temperature
For the hot check, what matters is the temperature of the cells, not the weather station reading. A module in full sun runs well above the surrounding air, and how far above depends almost entirely on how much air can move behind it. A module lying flush on a roof sheet with no ventilation gap can run 40 degrees above ambient, which in a hot climate means a cell temperature near 85 degrees and an MPP voltage roughly a fifth below the datasheet figure. That single choice can change your minimum string length, which is why the mounting selector sits alongside the temperatures rather than buried in an advanced panel.
The cold case works the other way: it uses the site's minimum design ambient temperature with no addition at all, because at dawn in still air with no sun on the glass a module really does sit at air temperature. The hot case uses the maximum design ambient plus a rise that follows standard practice:
| Mounting | Cell rise above ambient |
|---|---|
| Single axis tracker | +22 °C |
| Open rack, ground mount | +25 °C |
| Roof, 150 mm or more standoff | +30 °C |
| Roof, close mounted | +35 °C |
| Roof, flush or building integrated | +40 °C |
Which cold number to use, and why it is worth getting right
There are two different cold temperatures you could put in, and they are often far apart. Theall time record low is the coldest reading the site has ever seen, sometimes set decades ago. Theminimum design ambient, which in North America means the ASHRAE extreme annual mean minimum, is a statistical value representing how cold a typical year actually gets. NEC 690.7 asks for the second one, and the gap is not small: Albuquerque has a record low around −27 °C against an extreme annual mean minimum near −12 °C. Fifteen degrees is several volts per module, which on a long string is the difference between fitting one more module and not.
The location presets here are built on the design basis rather than the all time record, so they are not needlessly pessimistic, but they are regional starting points and no substitute for looking up your own site. Design values for most of the world are free atashrae-meteo.info. If you cannot find a figure for your site, using the record low instead is the conservative error to make: it shortens the string, which costs a little yield but cannot damage anything.
The coefficient almost every calculator gets wrong
Voc(cold) and Isc(hot) use coefficients printed on every datasheet. Vmp(hot) does not: βVmp is almost never published, so it has to be derived, and the choice moves the minimum string length. Most tools quietly reuse the Voc coefficient in its place, which understates the drop. This calculator offers three derivations and defaults to the first.
| Method | Formula | On a Trina Vertex 575 |
|---|---|---|
| Scaled, the default | βVoc × Voc ÷ Vmp | −0.297 %/°C |
| From Pmax, conservative | βPmax − αIsc | −0.380 %/°C |
| Unchanged, spreadsheet method | βVoc | −0.250 %/°C |
The reasoning for the default is physical. Temperature shifts a cell's voltage by roughly the same number of millivolts per degree wherever on the curve you measure it, so expressed as a percentage that same shift must be larger against the smaller Vmp number than against Voc. Any method that returns a βVmp smaller in magnitude than βVoc is understating the hot voltage drop, and understating it produces a minimum string length that is too short. The second method is a useful cross check because it comes from a different published number entirely: Pmax = Vmp × Imp, so the percentage coefficients add, and βVmp is what is left after taking the current's contribution out. Published dedicated βVmp figures typically land between the two.
A worked example you can follow end to end
Take a Trina Vertex 575 W bifacial module into a small 800 V pump controller, on a Terai rooftop with a 150 mm standoff. From the datasheet: Pmax 575 W, Voc 46.00 V, Vmp 38.70 V, Isc 15.97 A, Imp 14.87 A, βVoc −0.25 %/°C, αIsc +0.04 %/°C. From the inverter: 800 V maximum input, MPPT 300 to 750 V, start up 300 V, 30.6 A per MPPT, one tracker. Site record low 3 °C, record high 45 °C.
Step 1, cell temperatures. Cold case 3 °C, no addition. Hot case 45 + 30 = 75 °C.
Step 2, derive βVmp. −0.25 × 46.00 ÷ 38.70 = −0.297 %/°C.
Step 3, correct the module.
Voc(3°C) = 46.00 × (1 + (−0.25 ÷ 100) × (3 − 25)) = 46.00 × 1.055 = 48.53 V
Vmp(75°C) = 38.70 × (1 + (−0.297 ÷ 100) × (75 − 25)) = 38.70 × 0.8514 = 32.95 V
Isc(75°C) = 15.97 × (1 + (0.04 ÷ 100) × 50) = 15.97 × 1.02 = 16.29 A
Step 4, the series limits. The voltage cap is the lower of the inverter's 800 V and the module's 1500 V, so 800 V.
Nmax = floor(800 ÷ 48.53) = floor(16.5) = 16 modules
Nmin = ceil(300 ÷ 32.95) = ceil(9.1) = 10 modules
So anything from 10 to 16 modules works. Take the top of the range, 16, because a longer string means less current, thinner cable and lower losses. Check it against the other two voltage rules: 16 × 48.53 = 776 V, just under the 800 V limit; and when cold the MPP voltage is 16 × 38.70 × 1.0653 = 660 V, inside the 750 V tracking ceiling. Note that the cold MPP figure uses the derived βVmp, not βVoc.
Step 5, strings in parallel. 30.6 ÷ 14.87 = 2.06, so two strings would fit the operating current, but this controller has one tracker taking one string, so P = 1. The arrangement is16S × 1P: 16 modules, 9.2 kWp.
Step 6, protection. Design current = 1.25 × 16.29 = 20.4 A, which is what the DC cable and any string protection are sized on. With a single string there is nothing to fuse against.
Worth noting what step 4 caught. A spreadsheet that only checks the hot case would apply the 75 °C factor of 0.875 and see 16 × 38.70 × 0.875 = 542 V and 16 × 46.00 × 0.875 = 644 V, both comfortably under 800 V, and would accept 17 modules just as happily. At 17 the cold check gives 17 × 48.53 = 825 V, over the limit and into the territory where inverters fail. The cold case is the one that decides the maximum, and because voltage only ever falls in the hot case it is the check most easily left out of a design sheet.
Why the cold case is the dangerous one
This is the part that surprises people new to PV design. Silicon modules produce more voltage when they are cold, and open circuit voltage climbs by roughly 0.2 to 0.3 percent per degree below 25 degrees Celsius. The worst moment in the year is therefore not a hot afternoon but a clear winter dawn: the modules are at air temperature, the sun has just cleared the horizon, nothing is drawing current yet, and the string sits at full open circuit voltage while stone cold. A string that behaves perfectly all summer can push past the inverter's rating in those few minutes, and inverters are rarely replaced under warranty for it because the fault traces back to the design rather than the hardware.
This is also why the record low matters rather than the average. An average winter minimum understates the extreme by several degrees, and several degrees is worth several volts per module across a long string. Leave headroom on top of that: datasheet Voc itself carries a tolerance of around 3 percent, so a string calculated to land exactly on the limit is already over it for some of the modules in the delivered batch.
The gap this catches in a published design handbook
ICIMOD's Handbook for on-grid rooftop solar PV design optimisation works through a 3 kWp residential system: eight 370 Wp modules in series, Voc 48.3 V each, feeding an inverter rated 450 V maximum with a 50 to 450 V MPPT range. Its compatibility table checks 386.4 V against the 450 V limit and marks it OK, which is 86 percent of the rating and looks comfortable.
That 386.4 V is the string at 25 degrees Celsius. Correct it for temperature with the module's −0.286 percent per degree coefficient and the picture tightens: at a Kathmandu record low of −3 degrees the same string reaches 417 V, which is 93 percent of the limit. Move it to a mid hills site at −15 degrees and it reaches 431 V, or 96 percent. In a cold climate touching −25 degrees it reaches 442 V, 98 percent of the rating, with the module tolerance still to come on top. The string never changed. Only the temperature the design was checked at did. The handbook is explicit that detailed component sizing sits outside its scope, and that gap is exactly what this tool fills.
How to use this calculator, step by step
- 1Pick your module
Choose from the library, or pick the closest one and overwrite the specification boxes from your own datasheet. The temperature coefficients matter more here than the wattage does.
- 2Pick your inverter
Same way. Maximum input voltage, the MPPT window and the current per tracker are the four numbers that decide everything.
- 3Set the site temperatures
Use record extremes, not averages, and use the values for the actual site rather than the nearest large city if it sits at a different altitude.
- 4Choose the mounting
This sets how far above air temperature the cells run in full sun, which drives the minimum string length.
- 5Read the verdict and the checks
Green means every limit is cleared, amber means it will work but something is compromised, red means a hard limit is exceeded and the design should not be built.
- 6Scan the table
Every string length is listed with the cold and hot voltages worked out, so you can see how much headroom each option leaves. Click any row to select that length and watch the checks and the chart update.
Frequently asked questions
What is solar PV array sizing?
PV array sizing covers two separate decisions that often get given the same name. The first is how big the array should be in kWp, which comes from roof area, budget, the load you want to offset and any export limit. The second is how those modules get wired: how many in series per string, and how many strings in parallel, so that the array stays inside the inverter's voltage and current window at both temperature extremes. This calculator does the second one, which is also called string sizing or inverter matching. The first is a commercial decision rather than an electrical limit.
Is string sizing the same as inverter sizing?
They overlap but are not the same thing. Inverter sizing usually means matching the inverter's AC power rating to the array's DC capacity, which is the DC to AC ratio this calculator reports and which normally lands between 1.1 and 1.3. String sizing means matching the DC voltage and current of the array to that same inverter's input window. Both have to be right. Get the DC to AC ratio wrong and you either clip output on sunny days or pay for inverter capacity you never use. Get string sizing wrong and the inverter is damaged on a cold morning or stops tracking on a hot afternoon.
How many solar panels can I put in one string?
As many as fit between two limits. The upper limit is set by open circuit voltage on the coldest morning of the year: multiply the module's Voc by the string length and it must stay under the inverter's maximum input voltage, and under the module's own maximum system voltage. The lower limit is set by MPP voltage on the hottest afternoon: the string has to stay above the inverter's MPPT minimum or the inverter stops tracking. For a typical modern module on a 1100 V inverter that usually lands somewhere between 15 and 26 modules, but it depends entirely on the module, the inverter and the site temperature, which is what this calculator works out for you.
Why does cold weather decide the maximum string length, not hot weather?
Because silicon PV modules produce more voltage when they are cold. Open circuit voltage rises by roughly 0.2 to 0.3 percent for every degree below 25 degrees Celsius, so a string that measures 900 V on a warm afternoon can reach well over 1000 V at dawn on a clear winter morning, before the sun has warmed the modules at all. That is the moment inverters get destroyed. It is counterintuitive, and it is the single most common way an otherwise sensible array design goes wrong.
What temperature should I use for the cold check?
The record low ambient temperature for the site, not the average winter minimum and not the monthly mean. IEC 62548 asks for the lowest expected module temperature, and at dawn in still air with no sun on the glass, module temperature equals ambient temperature. Using an average instead of a record low is how designs end up sitting a few volts under the limit on paper and over it in reality. If your site sits at altitude, check the record rather than assuming, since a Himalayan valley floor can be far colder than the nearest city.
Why is cell temperature higher than the air temperature in the hot check?
Because modules sit in full sun. In bright conditions a module runs well above the surrounding air: roughly 25 degrees above ambient on an open ground mounted rack with air moving freely behind it, and up to 40 degrees above ambient for a module lying flush against a roof with no ventilation gap. That rise is why the mounting method changes your minimum string length. A flush mounted roof array in a hot climate can reach 85 degrees cell temperature, and the MPP voltage at that point can be a fifth lower than the datasheet figure.
What actually happens if the string voltage exceeds the inverter's maximum?
In the best case the inverter refuses to start and you lose a morning's production. In the normal case it fails, and the failure is usually not covered by warranty because the fault is a design error rather than a manufacturing defect. Some inverters clamp or disconnect above the limit, some do not. It is not a margin worth testing: the cold Voc figure should sit comfortably under the rating, not a couple of volts under it, because the datasheet Voc itself carries a tolerance of around 3 percent.
What happens if the hot MPP voltage falls below the MPPT minimum?
The inverter stops tracking the maximum power point, and in the worst case shuts down and restarts repeatedly. The frustrating part is the timing: this happens on the hottest, brightest afternoons, which is exactly when the array should be producing most. A string that is too short does not damage anything, it just quietly costs you yield in the season when yield matters. Short strings are also more likely to sit below the start up voltage in poor light, delaying the morning start.
Datasheets do not publish a temperature coefficient for Vmp. What does this calculator use?
By default it scales the published Voc coefficient by the ratio Voc divided by Vmp. Temperature shifts the voltage of a cell by roughly the same number of millivolts per degree wherever you measure it, so the same shift is a larger percentage of the smaller Vmp figure than it is of Voc. Using the Voc percentage unchanged on Vmp therefore understates how far the voltage drops when hot, which flatters the minimum string length. The Advanced panel lets you use the Voc coefficient directly, use the Pmax coefficient, or enter a measured figure if your datasheet publishes one.
What is a good DC to AC ratio for a grid tied array?
Between about 1.1 and 1.3 for most grid connected plants. Oversizing the array relative to the inverter means the inverter runs closer to its efficient operating band for more of the day and captures more energy in poor light, at the cost of clipping the peaks on the brightest days. Below about 0.9 you are paying for inverter capacity that never gets used. Above roughly 1.4 in a sunny climate the clipping losses become significant, though a heavily oversized array can still be the right answer where the roof is free and the inverter is the expensive part.
When do parallel strings need fuses?
Two strings in parallel on one input normally do not need string fuses, because a single healthy string cannot push enough fault current back into a faulted one to matter. Three or more in parallel usually do, since the remaining healthy strings can together drive far more than one string's cable and connectors are rated for. Size string protection and DC cable on 1.25 times the temperature corrected short circuit current, and check the module datasheet's own maximum series fuse rating, which sets a ceiling regardless of what the arithmetic says.
Does the module's maximum system voltage matter, or only the inverter's rating?
Both, and whichever is lower is the one that binds. A 1500 V module on a 1100 V inverter is limited by the inverter. A 1000 V module on a 1100 V inverter is limited by the module, and exceeding it breaks down the module's own insulation and the cable and connector ratings around it, not just the inverter. This calculator checks the cold voltage against both and tells you which limit it hit.
Can I mix different module models in the same string?
Avoid it. A series string carries one current, so the whole string is pulled toward the behaviour of the weakest module in it, and mismatched modules also drift apart differently as they age and heat. If you genuinely have two module types on one site, keep each type in its own strings and, where the inverter allows, on its own MPPT input. The same reasoning applies to identical modules facing different directions: separate orientations belong on separate MPPT inputs, since a shaded or lower producing group drags down anything sharing its tracker.
Does this work for bifacial modules?
The voltage checks work unchanged, because Voc and Vmp on a bifacial datasheet are quoted for the front face and the rear gain does not raise open circuit voltage meaningfully. Current is the part that needs care: rear side irradiance raises operating and short circuit current, typically by 5 to 15 percent depending on ground reflectance and mounting height, and some datasheets quote a separate bifacial gain figure. Add that margin to the current before checking it against the inverter's per MPPT rating and before sizing fuses and cable.