Back to Blog
solar designing 22 min read

Solar String Design Guide

Step-by-step solar string sizing: NEC 690.7 voltage limits, ASHRAE temperature math, MPPT range checks, bifacial corrections.

Rainer Neumann

Written by

Rainer Neumann

Content Head · SurgePV

Keyur Rakholiya

Edited by

Keyur Rakholiya

CEO & Co-Founder · SurgePV

Published ·Updated

A Phoenix installer strings 20 panels in series. At STC, open-circuit voltage is 49.5V × 20 = 990V — 10V under the 1000V inverter limit, so the installer ships the job. January arrives, ambient temperature drops to 5°C, and the cold-corrected Voc climbs to 1047V. The inverter shuts down every morning until the array warms up. The installer gets a warranty dispute call and a permit re-inspection.

Three numbers govern every string design: the cold-corrected Voc (which must stay below the inverter’s maximum DC input), the hot-corrected Vmp (which must stay above the MPPT minimum start voltage), and the MPPT current maximum (which caps how many strings you can put in parallel on one input). Skip any one of them and the system either violates NEC 690.7, clips output through every hot afternoon, or draws more current than the inverter’s input is rated for.

This guide walks the full calculation chain with real datasheet numbers. It covers NEC 690.7 voltage tiers, ASHRAE 99.6% heating design temperature methodology, the 3-check sizing framework, a complete worked example using a 400W module with a Fronius Symo 8.2-US in Phoenix, bifacial current corrections under NEC 690.8(A)(3), multi-MPPT string assignment, and the 10 most expensive mistakes found on permit-rejected drawings. Every formula is shown step by step, and every number is reproducible from public datasheets.

TL;DR — Solar String Design in 4 Rules

Cold-corrected Voc must stay below your inverter’s max DC input (600V residential per NEC 690.7; 1000V commercial). Hot-corrected Vmp must stay above the MPPT minimum start voltage. NEC 690.8(A)(3) requires a 1.25× current multiplier for bifacial modules. Use ASHRAE 99.6% heating design temperatures — not all-time record lows — per the 2017 NEC. Fronius data shows mismatched strings on one MPPT cost up to 0.82% annually.

In this guide:

  • What solar string design is and why it fails in the field
  • NEC 690 voltage rules and the ASHRAE temperature methodology
  • The 3-check framework: cold Voc, hot Vmp, MPPT current
  • Full worked example: 400W module + Fronius Symo 8.2-US in Phoenix
  • Why most installers use the wrong temperature input
  • Multi-MPPT assignment and mismatch loss data
  • Bifacial module string design and the hidden current problem
  • 10 common mistakes with field consequences and severity ratings

What Is Solar String Design — and Why It Fails in the Field

A solar string is a series-connected chain of PV modules wired to a single MPPT input on an inverter or charge controller. In series wiring, each module’s voltage adds to the next while current stays constant across all modules. String design is the process of choosing how many modules go in each chain — and which chains connect to which MPPT inputs — so the system operates within every electrical limit the inverter, the code, and the modules impose.

Definition: Solar String

A solar string is a series-connected group of PV modules forming a single DC branch circuit. All modules in a string share the same current. String voltage is the sum of individual module voltages at the operating point. One string connects to one MPPT input (or, in parallel-string arrangements, multiple strings of equal length may share one MPPT input).

In theory, string design is straightforward arithmetic. In practice, most failures come from four sources: ignoring temperature effects on voltage, using the wrong temperature reference, skipping the hot-side check entirely, or paralleling strings on the same MPPT without checking for mismatch. The table below maps failure type to root cause and field consequence.

Failure TypeRoot CauseField Consequence
Inverter trips on cold morningsVoc exceeds max input at low tempProduction loss, warranty void
Zero output during peak heatString Vmp too low for MPPT startComplete string dropout
Nuisance breaker tripsBifacial Isc not corrected × 1.25Conductor damage, AHJ rejection
Clipping loss above 3% annuallyDC/AC ratio too high for bifacialRevenue loss over system life
AHJ plan check rejectionVoltage calc missing temperature correctionPermit delay, redesign cost

Permit rejections for missing temperature corrections are among the most common plan-check failures on residential solar drawings in jurisdictions running 2017 NEC or later. Inverter over-voltage trips are the most common field complaint on systems installed in Sun Belt states without proper cold-day voltage checks — counterintuitive but consistent across service call data. The hot-side failure is rarer but more damaging to production because it recurs every hot afternoon rather than just on the coldest days.

All of these failures are preventable at the design stage. Solar design software that integrates ASHRAE weather data, live inverter specs, and NEC 690.7 voltage rules catches every one of these violations before the permit package leaves the office.


NEC 690 String Voltage Rules — What the Code Actually Says

NEC 690.7 sets the maximum DC system voltage for PV installations by occupancy type. The numbers are absolute — the cold-corrected Voc of any string in the system must stay at or below these limits across all expected temperature conditions.

System TypeMax DC System VoltageCode Reference
Residential (1- and 2-family)600VNEC 690.7
Commercial / multifamily1000VNEC 690.7
Ground-mount utility-scale1500VNEC 690.31(G)

The 600V residential limit is a hard stop. It applies to the entire conductor system, not just the inverter’s rated input. A 600V-rated inverter fed by strings that reach 620V on a cold morning is a code violation regardless of what the inverter’s input protection circuit does.

How NEC 690.7 Handles Temperature Correction

The 2017 NEC added ASHRAE 99.6% heating design temperatures as the standard method for calculating maximum Voc. Before the 2017 revision, many designers used NEC Table 690.7(A), which provides correction multipliers by ambient temperature. That table is still valid and still used by AHJs that run pre-2017 NEC — but the 2017+ path using module temperature coefficients and ASHRAE temperatures is now the preferred method because it uses site-specific data rather than a generic lookup.

NEC Table 690.7(A) correction factors (for jurisdictions using pre-2017 NEC or as a cross-check):

Ambient Temp (°C)Voc Correction Factor
251.06
201.07
101.10
01.13
−101.16
−201.20
−301.23
−401.25

NEC 690.11 — DC Arc Fault Protection

NEC 690.11 requires DC arc-fault circuit interrupter (AFCI) protection on any PV system conductor operating at 80V DC or more (2023 NEC). This applies to the DC home-run conductors between the array and the inverter. The 80V threshold means virtually every string system in the field requires AFCI. Omitting it is an AHJ rejection on any jurisdiction running 2020 NEC or later.

ASHRAE 99.6% — What the Number Actually Means

The “99.6% heating design temperature” is the temperature threshold that a location’s hourly temperatures fall below only 0.4% of the time annually. In a standard 8,760-hour year, that is approximately 35 hours. The 2017 NEC language reads: “…or the lowest expected temperature as specified in ASHRAE 99.6 percent design temperatures in the ASHRAE Handbook…”

A common question: “Can I use my own temperature estimate based on local weather station data?”

No. The code specifically requires the ASHRAE-listed value. Using a local weather website’s record low, a conservative round number, or a tool’s built-in default is not compliant on 2017+ NEC jurisdictions. Only the ASHRAE Handbook value satisfies the code reference.

Why this matters practically: Phoenix’s ASHRAE 99.6% design temperature is 5°C. Phoenix’s all-time recorded low is −9°C. Those two temperatures give different maximum string lengths on a 1000V commercial system — and the ASHRAE value is the one the code requires you to use. The next section shows exactly what difference that makes.


The 3-Check String Sizing Framework

Every string design requires three sequential checks in this order. Skip the sequence and you risk designing to one limit while violating another.

Check 1 — Cold-Side Voc Maximum (NEC Compliance Check)

The cold-side check determines the maximum number of modules per string. The principle: crystalline silicon modules have a negative voltage temperature coefficient, meaning Voc rises as temperature drops. The coldest expected condition produces the highest open-circuit voltage — and that value must stay below the inverter’s maximum DC input voltage.

V_oc_cold = V_oc_stc × [1 + (α_Voc / 100) × (T_min − 25)]
Max modules per string = floor(V_inverter_max / V_oc_cold)

Note: α_Voc is negative for crystalline silicon (typically −0.25% to −0.35%/°C). When T_min is below 25°C, the term (T_min − 25) is negative, and multiplying two negatives gives a positive correction — so Voc at cold temperatures is higher than at STC.

Pro Tip

Always use the temperature coefficient from the module datasheet, not a generic industry default. A 0.05%/°C difference in α_Voc can shift the maximum string length by 1 module on a 10-module string at 600V. Using −0.30%/°C when the datasheet says −0.29%/°C seems trivial — at 1000V with 19-module strings, it can mean the difference between 19 and 18 modules per string across an entire commercial roof.

Check 2 — Hot-Side Vmp Minimum (MPPT Start Check)

The hot-side check determines the minimum number of modules per string. Hot cell temperatures depress the maximum power point voltage (Vmp). If the string’s Vmp at peak cell temperature falls below the inverter’s MPPT minimum start voltage, the inverter cannot track the string and produces zero output.

T_cell_max = T_ambient_max + 25   (rule of thumb for rack-mounted modules)
V_mp_hot = V_mp_stc × [1 + (α_Vmp / 100) × (T_cell_max − 25)]
Min modules per string = ceil(V_mppt_min / V_mp_hot)

The +25°C adder for rack-mounted modules is a standard approximation from IEC 61215. Flush-mounted roof installations may run 5–8°C hotter — use +30°C if modules are mounted with less than 3 inches of air gap behind them.

MPPT voltage range by inverter class:

Inverter ClassMax Input VoltageMPPT Start Voltage
Residential string600V150–200V
Commercial / industrial1000V200–500V
Utility-scale1500Vvaries by inverter

Check 3 — MPPT Current Maximum (String Count per Input)

The current check determines how many strings can be paralleled on a single MPPT input. Each MPPT input has a rated maximum input current; paralleling more strings than the input supports draws current above spec.

Max strings per MPPT = floor(I_mppt_max / I_sc_module)

For bifacial modules, NEC 690.8(A)(3) adds a separate current multiplier. The conductor sizing calculation requires two distinct steps — they are not combined into one:

  • Step 1: Apply the bifacial factor from NEC 690.8(A)(3): I_bifacial = I_sc × 1.25
  • Step 2: Apply the continuous current factor from NEC 690.8(A)(1): I_design = I_bifacial × 1.25

The result: conductor ampacity must be rated at or above I_sc × 1.5625 for bifacial systems.

The bifaciality factor φ (rear efficiency / front efficiency) typically runs 0.65–0.80 for PERC bifacial modules. At a DC/AC ratio of 1.3, bifacial rear-side gain can increase annual clipping losses by 110% compared to an equivalent monofacial system at the same ratio, per MDPI Energies 17(22), 5658 (2024). This is why bifacial system design requires a lower DC/AC ratio than monofacial.


Full Worked Example — 400W Module + Fronius Symo 8.2-US in Phoenix, Arizona

This section runs the complete 4-step calculation for a real commercial project in Phoenix. Every number comes from a publicly available datasheet or the ASHRAE Handbook. The math is shown in full so you can replicate it for any module and inverter combination.

ParameterValueSource
ModuleGeneric 400W mono PERCDatasheet
V_oc_stc49.5VDatasheet
V_mp_stc41.2VDatasheet
I_sc_stc9.98ADatasheet
α_Voc−0.29%/°CDatasheet
α_Vmp−0.35%/°CDatasheet
InverterFronius Symo 8.2-USSpec sheet
Max DC input voltage1000VSpec sheet
MPPT voltage range200–800VSpec sheet
Max input current per MPPT18ASpec sheet
MPPT inputs2Spec sheet
SitePhoenix, AZProject
ASHRAE 99.6% T_min5°CASHRAE Handbook
T_ambient_max45°CPhoenix climate data

Step 1 — Cold-Side Voc Calculation

V_oc_cold = 49.5 × [1 + (−0.29 / 100) × (5 − 25)]
          = 49.5 × [1 + (−0.0029 × −20)]
          = 49.5 × [1 + 0.058]
          = 49.5 × 1.058
          = 52.37V per module

Max modules per string = floor(1000 / 52.37) = floor(19.09) = 19 modules

The ASHRAE 99.6% temperature of 5°C for Phoenix gives a maximum of 19 modules per string on a 1000V system. Now compare what happens if the designer uses the all-time record low of −9°C instead:

ASHRAE vs. All-Time Record Low — Phoenix

Using Phoenix all-time record low of −9°C instead of the ASHRAE 99.6% value of 5°C:

V_oc_cold = 49.5 × [1 + (−0.0029 × (−9 − 25))] = 49.5 × 1.0986 = 54.38V

Max modules per string = floor(1000 / 54.38) = 18 modules

ASHRAE gives 19 modules per string. The record-low approach gives 18. On a 100-string commercial project, that is 100 additional modules — at no added equipment cost — that the record-low method would leave off the roof.

Step 2 — Hot-Side Vmp Calculation

T_cell_max = 45 + 25 = 70°C

V_mp_hot = 41.2 × [1 + (−0.35 / 100) × (70 − 25)]
          = 41.2 × [1 + (−0.0035 × 45)]
          = 41.2 × [1 − 0.1575]
          = 41.2 × 0.8425
          = 34.71V per module

Min modules per string = ceil(200 / 34.71) = ceil(5.76) = 6 modules

Valid string length range: 6 to 19 modules

The optimal target for this inverter is 17–18 modules per string. At 17 modules, string Vmp at STC is 41.2 × 17 = 700.4V — 87.5% of the 800V MPPT top. At 18 modules, it is 741.6V — 92.7%. Both are within the 80–90% target band.

Step 3 — MPPT Current Check

Max strings per MPPT = floor(18 / 9.98) = floor(1.80) = 1 string per MPPT

The Fronius Symo 8.2-US supports a maximum of 18A per MPPT input. One string draws 9.98A — well within limit. Two strings in parallel would draw 19.96A, which exceeds the 18A maximum. Before paralleling two strings on one input, check the inverter datasheet for short-term overload tolerance. The Fronius Symo 8.2-US does not list an overload tolerance that covers this case, so two strings per MPPT is not permitted here.

With 2 MPPT inputs, the maximum system size is 2 strings × 19 modules × 400W = 15,200W DC. The inverter is rated 8,200W AC.

Step 4 — String Count and DC/AC Check

Two strings at 17 modules each:

DC capacity = 2 × 17 × 400W = 13,600W DC
AC capacity = 8,200W
DC/AC ratio = 13,600 / 8,200 = 1.29

A 1.29 DC/AC ratio falls within the 1.15–1.35 standard range. For Phoenix’s high-irradiance profile, annual clipping losses at this ratio stay below 1.5% (industry-observed range for south-facing arrays at 1.25–1.30 DC/AC in high-irradiance climates; modeled in NREL SAM). The design passes all three checks and leaves no production on the table.

Pro Tip

This 4-step calculation runs automatically in solar design software that integrates ASHRAE weather data, inverter spec sheets, and NEC 690.7 voltage limits. The software flags any violation during module layout — before the permit package is assembled — and exports the full string sizing calculation to the permit documentation set.


ASHRAE Temperature Methodology — The Calculation Most Installers Get Wrong

The most common string sizing error in the field is not the calculation itself — it is the temperature input going into the calculation. Designers routinely pull a “lowest expected temperature” from weather history websites, use the default 0°C built into their string sizing tool, or apply a conservative round number. All three approaches are wrong under 2017+ NEC, and all three produce incorrect string lengths.

What “99.6%” Means

The ASHRAE 99.6% heating design temperature is the value below which a location’s hourly dry-bulb temperatures fall only 0.4% of the time. In a standard 8,760-hour year, that is approximately 35 hours. The selection is intentional: the NEC wants a temperature that represents real cold-day conditions at the site, not the theoretical worst case that might occur once per decade.

The contrast with all-time record lows is substantial in Sun Belt climates:

CityAll-Time Record LowASHRAE 99.6% T_designModules Gained (1000V system)
Phoenix, AZ−9°C5°C+1 per string
Denver, CO−29°C−18°C0
Chicago, IL−27°C−18°C0
Miami, FL−3°C10°C+1 per string
Seattle, WA−16°C−4°C0

For cold-climate cities like Denver and Chicago, the ASHRAE 99.6% temperature is close enough to the record low that the string length does not change. For warm-climate cities like Phoenix and Miami, the ASHRAE value is 14–13°C warmer than the record low — and that difference adds one module per string on a 1000V system.

Where to Find the ASHRAE Value

The authoritative source is Chapter 14 of the ASHRAE HVAC Applications Handbook. The same values appear in NREL’s PVWatts weather files and the TMY3/TMY4 datasets available from the NREL Climate Data portal. Most permit-grade solar design software databases include ASHRAE heating design temperatures indexed by weather station.

Why This Matters at Scale

At 100 strings on a 400W utility project, gaining one module per string at no added equipment cost translates to 400 × 100 = 40,000W of additional DC capacity from a single correction in the temperature input. The permit and design cost is zero. The revenue over a 25-year project life is material.


Multi-MPPT String Assignment — Reducing Mismatch Loss

MPPT stands for maximum power point tracking. An inverter’s MPPT input constantly adjusts the operating voltage of connected strings to find the voltage at which the array delivers maximum power. That tracking algorithm picks a single operating voltage for all strings on the same input.

When strings on the same MPPT input have different optimal voltage points — because they have different module counts, orientations, shading profiles, or irradiance levels — the tracker settles on a compromise. Every string that deviates from the compromise operating point loses output.

Mismatch Loss Data

Fronius has published string mismatch loss data from field installations in their whitepaper “Implications of Different String Lengths for an MPP Tracker”:

Mismatch ScenarioAnnual Energy Loss
1 deviating string out of 14 on same MPPT0.14%
5 vs. 9 unequal strings on same MPPT0.82%
East-west strings properly on separate MPPTunder 0.1%

A 0.82% annual loss sounds small, but on a 100 kW commercial installation producing 150,000 kWh/year, that is 1,230 kWh/year — at $0.12/kWh, that is $148 per year, every year, for the life of the system. Proper MPPT assignment costs nothing.

String Assignment Decision Rules

ConditionAssignment Decision
Same pitch, orientation, lengthSame MPPT
East vs. west facesSeparate MPPT inputs
South + shade-affected north stringsSeparate MPPT inputs
Strings of unequal lengthSeparate MPPT inputs
Bifacial high-albedo vs. standard zoneSeparate if Isc difference above 5%

When Unequal String Lengths Are Unavoidable

Roof geometry sometimes produces one section with 8 modules and another with 10. The correct approach: assign the 8-module strings to one MPPT input and the 10-module strings to the other. Never mix the two lengths on the same input when the inverter offers a separate one.

If all MPPT inputs are already committed and you must mix lengths on one input, calculate the voltage mismatch at both STC and hot-corrected Vmp before accepting the design. A one-module length difference on a typical string produces roughly 40V of mismatch at STC — tolerable in some cases, but worth quantifying.

Pro Tip

Solar software with 3D rooftop modeling assigns strings to MPPT inputs during module layout — flagging orientation mismatches and unequal string lengths in real time before the design is finalized.


Bifacial Module String Design — The Hidden Current Problem

Bifacial modules generate power from both the front and rear surfaces. Rear-side generation depends on ground albedo, mounting height, row spacing, and surface reflectivity. The bifacial gain — additional energy from the rear — typically ranges from 5–30% of front-side nominal output, with a bifaciality factor φ (rear efficiency divided by front efficiency) of 0.65–0.80 for standard bifacial PERC cells.

The key electrical consequence: bifacial rear-side irradiance adds to module current (Isc), not voltage. Voc and Vmp stay essentially unchanged from datasheet values. For string voltage calculations, you can use the standard monofacial formulas. But for conductor sizing and overcurrent protection, the current increase must be accounted for.

NEC 690.8(A)(3) — Bifacial Current Multiplier

NEC 690.8(A)(3) addresses bifacial current explicitly. Conductor sizing for bifacial systems uses two separate multipliers — these are not combined into a single step:

I_sc = 9.98A (from datasheet)

Step 1 — Bifacial factor per NEC 690.8(A)(3):
I_bifacial = 9.98 × 1.25 = 12.48A

Step 2 — Continuous current factor per NEC 690.8(A)(1):
I_design = 12.48 × 1.25 = 15.59A

Conductor ampacity must be rated at or above 15.59A

A standard 10 AWG PV wire is rated 30A, so this is not typically a wire gauge problem. The issue arises with overcurrent protection device sizing: the OCPD must be selected based on I_design, not I_sc. An installer who sizes a fuse at 15A to the string’s nominal Isc of 9.98A and then installs bifacial modules will see the OCPD trip under normal rear-side irradiance conditions.

Bifacial DC/AC Ratio Warning

At a DC/AC ratio of 1.3, bifacial rear-side gain can increase annual clipping losses by 110% compared to an equivalent monofacial system, per MDPI Energies 17(22), 5658 (2024). A system designed as “1.30 DC/AC” with a 20% rear-side gain is effectively running at a 1.56 clipping equivalent during peak production hours.

Recommended DC/AC range for bifacial: 1.15–1.25. Higher ratios should only be used when the mounting configuration limits rear irradiance — dense ground-cover systems, low-clearance rooftop installations, or north-facing secondary strings.

Bifacial Yield Modeling

Use the generation and financial tool for bifacial yield modeling by row and mounting height. The tool accounts for albedo, row-to-row shading on rear cells, and height-dependent bifacial gain factors — inputs that a simple DC/AC ratio estimate cannot capture.


Common String Design Mistakes — and What They Cost

The mistakes below come from permit-rejected drawings, inverter warranty disputes, and production underperformance investigations. Severity ratings reflect the combination of code risk and revenue impact.

#MistakeNEC ImpactField ConsequenceSeverity
1Using STC Voc without temperature correctionNEC 690.7 violationInverter over-voltage shutdown on cold daysCritical
2Using all-time record low instead of ASHRAE 99.6%Over-conservativeLeaves 1+ module per string unrealizedMedium
3Ignoring hot-side MPPT minimum checkDesign flawString produces zero output at peak ambient tempsCritical
4No bifacial current correction for conductorsNEC 690.8(A)(3) violationConductor overheat, AHJ rejectionHigh
5Mixing string lengths on same MPPT inputNot a code issue0.82% annual mismatch loss (Fronius data)Medium
6East + west strings on same MPPTNot a code issue0.5–1% annual loss, production mismatchMedium
7DC/AC ratio above 1.3 with bifacial modulesNot a code issueClipping loss doubles vs. monofacial equivalentHigh
8Omitting NEC 690.11 AFCI on 80V+ conductorsNEC 690.11 violationAHJ rejectionHigh
9Using manufacturer string calculator with wrong temp inputUnder-checkWrong result if tool defaults to 0°CMedium
10Mixing orientations on same MPPT because “it’s close enough”Not a code issueOngoing annual energy loss on every projectMedium

Mistake 1 — STC Voc Without Temperature Correction (Critical)

This is the Phoenix scenario from the opening of this guide, and it comes up in service calls repeatedly in climates that installers consider “warm.” The installer calculates 49.5V × 20 = 990V at STC, observes that this is 10V under the 1000V limit, and calls the design done. What they missed: at Phoenix’s ASHRAE 99.6% temperature of 5°C, each module’s Voc rises to 52.37V. Twenty modules in series reaches 1,047V — 4.7% above the 1000V hard limit.

The inverter does not silently absorb extra voltage. Modern string inverters have input voltage protection circuits that either shut the inverter down or trigger a fault mode when the DC input exceeds the rated maximum. In the Phoenix case, the inverter trips every cold winter morning and restarts once the array warms past the threshold. The installer gets a service call, the owner loses production, and the inverter manufacturer’s response to the warranty claim is: “System was operated outside rated input voltage — warranty void.”

At 600V residential systems, the same failure mode appears on systems strung for warm-climate cities that occasionally see winter cold fronts. A Phoenix residential installer stringing 12 modules at STC Voc of 49.5V reaches 594V — 6V under the 600V limit. At 5°C, those 12 modules reach 628V. Permit rejected, and if already installed, a mandatory redesign.

Mistake 3 — Ignoring the Hot-Side MPPT Minimum Check (Critical)

The hot-side failure gets skipped because the mental model goes: “Phoenix is a hot climate, so if anything we’ll be fine.” The logic is backwards. Hot climates are where the hot-side check matters most, because cell temperatures reach 70°C+ on summer afternoons and Vmp can drop well below MPPT start.

Consider a 5-module string in Phoenix with the same 400W module from the worked example. Hot-corrected Vmp per module is 34.71V. Five modules deliver 5 × 34.71 = 173.6V. The Fronius Symo 8.2-US MPPT minimum is 200V. The string produces zero output during the hottest part of every summer day — exactly when the owner expects maximum production.

A 6-module string at 70°C cell temperature delivers 6 × 34.71 = 208.2V, just 8.2V above the 200V MPPT start. That is a thin margin; a roof with poor ventilation or modules mounted flush without air gap may run 5–8°C hotter than the standard estimate, pushing T_cell to 75°C+ and bringing Vmp back below the threshold. For tight designs near the hot-side minimum, use the +30°C cell temperature adder instead of +25°C.

Mistake 7 — DC/AC Ratio Above 1.3 with Bifacial Modules (High)

A 1.30 DC/AC ratio with bifacial modules effectively operates at a higher clipping equivalent than the nameplate ratio suggests. With a typical bifacial gain of 15–20% during peak irradiance hours, the array’s actual peak power output exceeds the STC DC capacity that went into the DC/AC ratio calculation. At 20% rear-side gain on a 1.30 DC/AC design, the effective peak DC/AC ratio during summer midday hours approaches 1.56.

Annual clipping loss at a true 1.56 DC/AC ratio in Phoenix can reach 2.5–3.5% compared to under 1% at a 1.30 monofacial equivalent. Over a 20-year project life at a modest $0.10/kWh avoided cost, that 1.5% additional clipping on a 200 kW system represents approximately $9,000–$15,000 in lost production (200 kW × ~300,000 kWh/year × 1.5% × 20 years × $0.10/kWh). The fix is to target 1.15–1.25 DC/AC on bifacial systems or to use a per-row bifacial yield simulation that accounts for actual rear irradiance before finalizing the ratio.

Run Your String Calculations in SurgePV

SurgePV’s string sizing engine checks cold-side Voc, hot-side Vmp, MPPT current limits, and NEC 690.7 voltage tiers in one step — with ASHRAE weather data built in.

Book a Demo

No commitment required · 20 minutes · Live project walkthrough


DC-to-AC Ratio — What “Good” Actually Means

The DC/AC ratio (also called the inverter loading ratio or ILR) is the nameplate DC capacity of the array divided by the AC output rating of the inverter. It tells you how much the inverter is loaded relative to the peak output the array can theoretically deliver.

The standard range of 1.15–1.35 reflects a practical tradeoff: below 1.15, the inverter is undersized and adding modules increases annual production at low marginal cost; above 1.35, clipping losses — where the inverter limits AC output because DC input exceeds its capacity — begin to eat meaningfully into annual production.

Three examples with the Fronius Symo 8.2-US (8,200W AC):

  • 18 × 400W = 7,200W DC / 8,200W AC = 0.88 — inverter is heavily oversized; leaving annual production on the table
  • 22 × 400W = 8,800W DC / 8,200W AC = 1.07 — still below the efficient threshold; add 2–4 modules
  • 26 × 400W = 10,400W DC / 8,200W AC = 1.27 — optimal range for a south-facing array in Phoenix

When to Go Higher

West-facing arrays in mild climates see peak production later in the afternoon, when grid electricity prices are often at their highest (time-of-use rates). These systems can tolerate higher DC/AC ratios because the peak production period is offset from true solar noon — clipping losses at DC/AC 1.4–1.5 may still result in better economic returns than a lower ratio on a south-facing array. Low-irradiance climates (Seattle, Amsterdam, Vancouver) can also justify higher ratios because the irradiance profile rarely pushes the array to STC peak output.

When to Go Lower

Bifacial modules in high-albedo environments — white membrane roofs, ground mounts over light gravel, snow-prone sites — should be designed at 1.15–1.25. The rear-side gain effectively increases the peak DC output beyond the nameplate rating used in the ratio calculation.

Annual Clipping Loss by Ratio (South-Facing, Phoenix)

DC/AC RatioTypical Annual ClippingNotes
1.10under 0.2%Inverter underutilized
1.200.4–0.6%Optimal for bifacial
1.301.0–1.5%Standard residential
1.402.0–3.0%Acceptable in low-irradiance
1.503.5–5.0%West-facing only

Use the generation and financial tool for site-specific payback modeling that incorporates actual irradiance profiles, time-of-use rates, and bifacial rear-gain estimates by row.


Series vs. Parallel Wiring — How String Architecture Works

In series connection, voltage adds across each module in the chain while current remains constant throughout the string. Connect 18 modules each rated at V_mp = 41.2V in series and the string voltage is 41.2 × 18 = 741.6V at Vmp, with current held at the module’s rated 9.98A. Series strings produce high voltage, low current.

High voltage and low current is desirable for DC conductors. Resistive loss in a wire is proportional to I² × R. Halving the current reduces conductor loss by 75% for the same wire gauge. A single high-voltage string running from the array to the inverter loses far less energy in the home-run conductor than 18 individual module circuits at low voltage.

This is why string inverters dominate residential and C&I solar: their MPPT inputs are designed for the 200–1000V range that series strings naturally produce. Parallel wiring at module voltage (40–50V) would require either 18 separate conductor runs to the inverter, or a large DC combiner box that adds cost and complexity.

Where Parallel Combiners Appear

Utility-scale central inverters aggregate multiple strings before the DC home-run. A typical 1500V central inverter might have 20–24 string inputs per combiner box, with 8–10 combiner boxes feeding a single 1–2 MW inverter. The combiner boxes handle overcurrent protection and monitoring for each string in parallel; the central inverter MPPT then tracks the combined DC bus. At utility scale, the capital cost of a central inverter offsets the complexity of the combiner infrastructure.

Series-Parallel Hybrids

Two strings in parallel per MPPT input is legitimate for string inverters where the MPPT maximum current supports it. From the Phoenix worked example, the Fronius 8.2-US allows 18A per MPPT and each string draws 9.98A — the inverter does not support 2 strings in parallel (19.96A exceeds the 18A limit). But inverters with 30A+ MPPT inputs routinely support 2–3 strings in parallel, provided string lengths are matched to within 1 module.

Longer strings carry a practical production advantage: 18 modules per string vs. 12 modules per string means 33% fewer strings for the same array size, 33% fewer home-run conductors, and lower I²R losses in each run. The solar designing layout tool accounts for conductor lengths and DC resistive losses when comparing string length options.


String Design for Real Projects — Residential vs. C&I

The 600V residential voltage ceiling changes the string design math fundamentally compared to commercial. Using the same 400W module from the worked example in Phoenix:

Residential (600V, 1- and 2-family dwellings):

Max modules per string = floor(600 / 52.37) = floor(11.45) = 11 modules
Min modules per string = ceil(150 / 34.71) = ceil(4.32) = 5 modules
Valid range: 5–11 modules

A maximum of 11 modules per string at 400W gives 4,400W DC per string. A typical residential 10 kW system requires at least 3 strings. Installers often work with 9–10 modules per string on residential jobs to keep voltage in the 80–85% MPPT range.

Commercial (1000V, multifamily and commercial):

The same calculation from the worked example applies: valid range 6–19 modules, with 17–18 optimal.

Flat-roof C&I shade assignment:

Flat commercial roofs typically have HVAC equipment, penthouses, and parapet walls that cast shade across different sections of the array at different times of day. Shade assignment drives MPPT grouping more than geometry does. Modules under the same shade source at the same time of day belong on the same string and the same MPPT input — that way, when the shaded string’s output drops, only that MPPT input adjusts, rather than pulling down an unshaded string on the same tracker. Use solar shadow analysis software to assign modules under the same shade path to the same string before finalizing the flat-roof layout.

Ground-mount (1500V under NEC 690.31(G)):

At 1500V with the same 400W module in Phoenix:

Max modules per string = floor(1500 / 52.37) = floor(28.64) = 28 modules

Twenty-eight modules per string at 400W is 11,200W DC per string — roughly the size of an average residential installation in a single series chain. Long strings at 1500V reduce conductor count dramatically on utility ground mounts and are one of the main cost advantages of 1500V system architecture.


How Solar Design Software Handles String Sizing

Manual string sizing takes 10–15 minutes per inverter for an experienced designer. A commercial rooftop with 12 inverters takes 2–3 hours. One incorrect temperature input invalidates every calculation in the set. If the error is discovered at the AHJ plan check, the redesign starts from scratch.

At batch scale — 50 residential jobs per month, or a 2 MW commercial project with 40 inverters — manual string sizing is not a realistic quality control approach. The error rate on manual calculations is not zero, and permit rejections for temperature correction errors are among the most preventable delays in the permitting process.

What SurgePV Does

Solar design software built for production use handles the full calculation chain automatically:

  • Pulls the ASHRAE 99.6% heating design temperature for the project’s weather station from the built-in database
  • Reads the inverter spec directly from the inverter database (maximum input voltage, MPPT voltage range, maximum input current per MPPT, number of MPPT inputs)
  • Reads the module datasheet from the module database (V_oc, V_mp, I_sc, α_Voc, α_Vmp, bifaciality factor if bifacial)
  • Runs all three checks and returns the valid string length range and the optimal target
  • Flags NEC 690.7 voltage violations in real time during the module layout phase — before the permit package is assembled
  • Assigns strings to MPPT inputs automatically during 3D rooftop modeling in solar designing, grouping modules by orientation and shade path
  • Applies the NEC 690.8(A)(3) bifacial current multiplier automatically when a bifacial module is selected from the database
  • Exports the complete string sizing calculation table — with temperature inputs, intermediate results, and code references — to the permit documentation package via solar proposals

What the installer sees in the layout interface: a string diagram with each string color-coded by MPPT input, violation flags in the sidebar when any string exceeds voltage or current limits, and a summary table showing valid range, actual configuration, and DC/AC ratio.

Pro Tip

Ask your current design tool whether it uses ASHRAE 99.6% temperatures or defaults to 0°C. The difference can mean 1 extra module per string — or a permit rejection. On a 50-job monthly volume with 2 strings per job, that is 100 additional modules per month that the 0°C default approach leaves off the roof.


Conclusion

Solar string design is not a single calculation — it is a 3-check framework run in sequence on every new project configuration. Each check addresses a different failure mode; skipping any one of them creates a real production or compliance problem.

Three decisions determine whether a string design is right:

  • Run the 3-check framework on every new project before opening the inverter spec sheet. Cold Voc sets the maximum string length; hot Vmp sets the minimum; MPPT current sets the maximum parallel strings per input. Run them in that order.
  • Pull the ASHRAE 99.6% design temperature for the site. Do not use all-time record lows, calculator defaults, or round numbers. The ASHRAE value is what the 2017+ NEC requires, and in warm climates it typically adds one module per string compared to the record-low approach.
  • Assign strings to MPPT inputs by orientation before finalizing the roof layout, not after. East and west strings on separate MPPT inputs costs nothing. Leaving them on the same tracker costs 0.5–1% annually for the life of the system.

The calculations in this guide are reproducible for any module and inverter combination using publicly available datasheets and the ASHRAE Handbook. Automated solar software that runs these checks in real time during layout eliminates the manual error rate and produces the permit documentation as a byproduct of the design — not as a separate step.

See String Sizing Automated in SurgePV

Every calculation in this guide runs automatically in SurgePV — ASHRAE temperatures, NEC 690.7 voltage checks, bifacial corrections, and MPPT assignment.

Book a Free Demo

No commitment required · 20 minutes · Live project walkthrough


Frequently Asked Questions

What is a good DC-to-AC ratio for solar?

A DC/AC ratio of 1.15–1.35 is standard for most residential and C&I string inverter systems. Higher ratios up to 1.5 can be justified on west-facing arrays or in low-irradiance climates where clipping losses stay below 1–2%. For bifacial modules, keep the ratio closer to 1.15–1.25 to avoid excess clipping from rear-side gain.

What is the ideal string voltage for a solar inverter?

Target 80–90% of the inverter’s maximum MPPT voltage at standard test conditions. For a 1000V commercial inverter with a 200–800V MPPT range, design strings to hit 700–800V at STC. The cold-corrected Voc must stay below the inverter’s absolute input maximum, and the hot-corrected Vmp must stay above the MPPT minimum start voltage.

How do you calculate the right number of panels per string?

First calculate the cold-corrected Voc: V_oc_cold = V_oc_stc × [1 + α_Voc × (T_min − 25)]. Divide the inverter’s maximum input voltage by V_oc_cold and take the floor — this is your maximum panels per string. Then calculate hot-corrected Vmp: V_mp_hot = V_mp_stc × [1 + α_Vmp × (T_cell_max − 25)]. Divide the MPPT minimum voltage by V_mp_hot and take the ceiling — this is your minimum panels per string.

What is the NEC maximum DC voltage for residential solar?

NEC 690.7 limits residential (1- and 2-family dwellings) to 600V DC maximum. Commercial and multifamily buildings may go up to 1000V DC. Ground-mounted utility-scale systems can operate at 1500V DC under NEC 690.31(G). The 600V limit applies to the entire conductor system, not just the inverter input.

How do bifacial solar panels affect string sizing?

Bifacial modules add rear-side current that can exceed the front-side Isc used in wiring calculations. NEC 690.8(A)(3) requires multiplying the module’s rated Isc by 1.25 when sizing conductors and overcurrent protection for bifacial systems. At high DC/AC ratios around 1.3, bifacial clipping losses can be 110% greater than equivalent monofacial systems, per MDPI Energies 2024.

Can strings of different lengths be on the same MPPT input?

Technically yes, but it creates a mismatch penalty. Fronius data shows that 5 vs. 9 unequal strings on the same MPPT tracker produce 0.82% annual mismatch loss. Match string lengths on the same MPPT whenever possible. Where unequal strings are unavoidable, assign them to separate MPPT inputs if the inverter supports it.

How does temperature affect solar string sizing?

Temperature affects both voltage bounds. Cold temperatures raise Voc — if cold-corrected Voc exceeds the inverter’s max input voltage, the inverter trips or is damaged. Hot temperatures depress Vmp — if hot-corrected Vmp drops below the MPPT start voltage, the string produces nothing. Check both limits using the temperature coefficients from the module datasheet.

About the Contributors

Author
Rainer Neumann
Rainer Neumann

Content Head · SurgePV

Rainer Neumann is Content Head at SurgePV and a solar PV engineer with 10+ years of experience designing commercial and utility-scale systems across Europe and MENA. He has delivered 500+ installations, tested 15+ solar design software platforms firsthand, and specialises in shading analysis, string sizing, and international electrical code compliance.

Editor
Keyur Rakholiya
Keyur Rakholiya

CEO & Co-Founder · SurgePV

Keyur Rakholiya is CEO & Co-Founder of SurgePV and Founder of Heaven Green Energy Limited, where he has delivered over 1 GW of solar projects across commercial, utility, and rooftop sectors in India. With 10+ years in the solar industry, he has managed 800+ project deliveries, evaluated 20+ solar design platforms firsthand, and led engineering teams of 50+ people.

Get Solar Design Tips in Your Inbox

Join 2,000+ solar professionals. One email per week - no spam.

No spam · Unsubscribe anytime