PV O&M Diagnostics

Solar String Current Deviation Calculator

Q: What does a low solar string current mean?

If a string current is lower than the current of nearby strings under the same conditions, it usually means less irradiance or current loss in that circuit. Common causes include shading or tracker misalignment, dirty modules or bird droppings, a loose or damaged connector, a blown fuse or tripped isolator, or a module or sub-string fault such as an activated bypass diode. The calculator flags the string for inspection and the field team checks the string to find the actual cause.

Q: How do I tell whether a string fuse has blown?

If a string shows 0.00 A while nearby strings are producing normally, the string may have a blown fuse or an open circuit. However, the current reading alone does not confirm a blown fuse. Isolate the string, check fuse continuity with a meter, measure string Voc at the DC disconnect, and inspect the MC4 connectors and trunk cable. Before replacing the fuse, check for an earth fault or short circuit that may have caused it to blow.

Q: What is a normal current difference between PV strings?

A current difference of about 2–3% between PV strings in the same combiner box is normal when the modules are clean and receiving similar sunlight. This calculator uses a ±5% limit by default: differences below 3% are considered normal, 3% to less than 5% should be checked, and 5% or more is flagged for inspection. A difference of exactly 5% is flagged, while exactly 3% falls into the check range. These are screening values, not manufacturer limits, so always compare them with the module datasheet and your plant's O&M criteria.

Q: Why are zero-current strings left out of the baseline?

Because a 0 A reading is a fault condition, not a healthy reference. Including a zero-current string would lower the reference current and could make other weak strings look normal. Zero-current strings are still shown in the results with their full deviation and a critical status; they are excluded only from the reference current calculation.

Q: Can I compare string currents on a cloudy day?

You can run the calculator on any reading set, but results taken in unstable irradiance are not reliable. Passing clouds, low sun angles and tracker movement change each string's current at different rates, so percentage deviations become less meaningful. The calculator caps its confidence rating at low when the group average current is very low. For decisions that lead to a work order, compare strings during stable, clear-sky conditions.

Q: What is the difference between the fixed ±5% band and the MAD method?

The fixed band compares every string with the same group reference current and applies identical percentage limits to all of them. The MAD method instead measures how far each string sits from the group's median spread, so it adapts to how tight that particular array actually is. On small groups MAD is more sensitive, and the same string can land in a watch band under the fixed method and be flagged by MAD. Run both and treat disagreement as a reason to look at the string.

Q: Can this calculator prove that a module or string is faulty?

No. It identifies strings that need a physical check. Confirming a fault requires irradiance and temperature context, a calibrated DC clamp measurement, a string I-V curve trace with fill factor, thermal imaging, or insulation-resistance testing. Equal current between two strings also does not by itself prove equal energy production, because voltage and operating temperature still differ.

Q: What data does the calculator need, and where does it go?

Three columns: a combiner box or MPPT group ID, a string ID that is unique within that group, and the measured DC current in amperes. Optional voltage and timestamp columns are accepted. Everything runs locally in your browser and no reading or file is uploaded. A group needs at least two strings to compare, because a single string has nothing to benchmark against.

Compare DC current readings from strings connected to the same combiner box or inverter MPPT, and identify readings that need field inspection.

Reviewed by Aman Yadav, Solar Plant Performance Engineer · Last updated 27 Sep 2026

Input Currents

Group (MPPT)	String ID	Current (A)

Baseline & Flagging Thresholds

Healthy Group Benchmark:

Zero-current readings (≤ 0.05 A) are excluded before calculating the benchmark.

Fault Detection Method:

Deviation Threshold (±%):

Strings outside ±5% are flagged for inspection; deviations from 3% up to under 5% (below or above the reference) enter the Watch band.

Modified Z Threshold (|M|):

Standard Iglewicz-Hoaglin outlier cutoff (|M| ≥ 3.5). 0.6× threshold (|M| ≥ 2.1) enters Watch Band.

Readings taken near solar noon under clear sky & stable irradiance (improves diagnosis confidence).

Active / Total Strings 0 / 0 Near-zero readings excluded

Healthy Benchmark 0.00 A Active median current

String Variation (CV) 0.0% Coefficient of variation

Underperforming 0 Requires field inspection

Reading Quality & Confidence High Confidence

Optimal screening conditions: Sufficient string sample size and stable current values.

Current Distribution & Threshold Profile

Normal (±3%)

Watch Band (3% to under 5%, below or above the reference)

Underperforming (≤ -5%)

Zero Current — Check Fuse

High / Check Sensor (≥ +5%)

String Inspection & Action List

← Scroll horizontally to view full diagnostics →

String ID	Current (A)	Deviation (%)	Stat Score	Field Status	Screening Result & Recommended Action

Estimated Energy & Revenue Loss (Insolation-Normalized Model)

Unlike a simplistic snapshot multiplication (ΔI × V × hours), which tends to overestimate string losses, this calculator applies deficit fractions against standard Peak Sun Hours (PSH) to reflect the diurnal irradiance curve.

Rated String DC Power (kWp): e.g. 28 modules × 540 Wp

Daily Peak Sun Hours (PSH): Site solar resource (kWh/m²/day)

PPA / Tariff Rate ($/kWh): Energy sale or offset value

How Solar String Current Comparison Works

Peer Group Benchmarking

This calculator compares DC current readings from strings connected to the same combiner box or inverter MPPT. Enter a group identifier, a unique string ID, and the measured current in amperes for each string. Readings in the same group are compared with a reference current calculated from the active strings. For background on why string-level checks catch problems that plant-level PR alone can miss, see String-Level vs Plant-Level Monitoring.

Zero-Current & MAD Screening

Readings at or below 0.05 A are treated as zero-current conditions and are excluded from the active baseline so they do not lower the comparison reference. The results show each string’s percentage difference, screening status, and recommended field follow-up. The calculator can also use a robust MAD-based method to identify unusually low or high readings.

Root Cause Indicators

Low current may be associated with shading, soiling, an open circuit, a fuse or connector problem, or module damage. High current may indicate sensor scaling errors, incorrect wiring, or unequal operating conditions.

Verification & Safety: These results are screening clues only. Confirm them with stable irradiance measurements, a DC clamp meter, V_oc testing, thermal inspection, or an I–V curve trace. The calculator does not replace electrical safety procedures or a complete PV performance assessment.

When String Comparisons Are Not Valid

Only compare strings built from the same PV module model and the same number of modules in series. If either differs, put those strings in separate comparison groups.
Keep strings with the same azimuth, tilt, and tracker position in the same group. Strings on different roof faces or tracker positions can produce different current even when both are operating correctly.
Leave out strings that are intentionally isolated, under maintenance, or disconnected by design. Those are not faults.
Review the as-built electrical drawing before grouping. Retrofit wiring, parallel branches, or Y-connectors can make one current channel represent a different number of strings.
If the group configuration is uncertain, confirm the module count, wiring, and operating status against plant records before relying on the comparison.
A combiner box that also measures each string is normally called a String Monitoring Box (SMB); the enclosure itself is the String Combiner Box (SCB). Plants label the same hardware either way, so group strings by the channel they physically land on, not by the document's terminology.

Step-by-Step Guide: How to Use the Application

A practical operational handbook for Solar PV O&M engineers, site technicians, and asset managers to screen DC string currents, flag suspected fuses, screen partial shading, and prioritize field work orders.

Workflow A: Manual Data Entry (On-Site String Current Check)

1

Set Combiner Box or Inverter MPPT Group

Enter your combiner box (SCB/SMB) identifier or Inverter MPPT channel name (e.g., SMB-04 or INV01-MPPT2) in the group column. All strings in the same group are compared against each other.
2

Enter String IDs and Current Readings

Provide the String ID (e.g., SMB-04-Str01) and the measured DC Current in Amperes (A). Ensure each String ID within the same group is unique. The input strictly accepts only valid positive numbers.
3

Add Strings or Load Sample Plant Data

Click "Add String Reading" to add more string channels, or click "Load Sample Data" to test the calculator with an 8-string combiner-box example.
4

Choose Comparison Settings

Choose the reference current (Active Median is recommended because it excludes zero-current readings from the baseline). Then select either Fixed Percentage Band ±5% or Modified Z-Score via MAD.
5

Run the String Current Check

Click "Calculate Deviations". The calculator compares the string currents, identifies zero-current strings, and marks strings that need inspection.
6

Review the Chart and Results Table

Review the interactive visual bar chart (hover over bars for string details) and the results table with clear color-coded badges, percentage deviations, and specific field technician recommendations.
7

Estimate Lost Energy and Export Results

Input your string peak DC rating, daily Peak Sun Hours (PSH), and PPA tariff in the Financial Impact drawer to quantify revenue loss. Export results as a filtered CSV work order or audit JSON.

Workflow B: Upload SCADA String Currents (CSV / TSV)

1

Download the CSV Template

Switch to the "Upload CSV / TSV" tab and click "Download Sample CSV Template". This includes the required plant data columns: group_id, string_id, current_a.
2

Populate Data from SCADA / Datalogger

Paste or save your plant SCADA string-level current export into the template. The parser supports comma-separated (CSV) and tab-delimited (TSV) files across multiple combiner boxes.
3

Upload or Drag & Drop File

Drag and drop your file into the designated upload zone or click "Browse Files". Parsing occurs 100% locally in your browser—no confidential plant telemetry ever leaves your computer.
4

Automatic Validation & Duplicate Check

The calculator checks each row for a valid combiner-box ID, string ID, and current reading. Blank lines are skipped, invalid rows are reported, and duplicate string IDs within the same SMB are flagged in the import summary.
5

Multi-SMB Combiner Box Navigation

If your file contains multiple combiner boxes or MPPT channels, interactive navigation tabs appear automatically above the KPI summary cards, allowing instant switching between SMB blocks.
6

Generate Prioritized Maintenance Log

Review the reading-quality notes and identify strings with low current, zero current, or unusually high current. Export the list of strings requiring field checks for your maintenance team.

Field Troubleshooting Reference

Use this guide to interpret the results, identify likely field causes, and select the appropriate inspection or maintenance action.

Scroll horizontally to view entire matrix →

Status & Badge	Detection Criteria	Probable Plant Root Causes	Recommended Field Action Plan
Zero Current — Check Fuse	`I ≤ 0.05 A` Isolated from active baseline	Suspected blown gPV string fuse in the combiner box Tripped DC isolator switch or open fuse holder Disconnected or unseated MC4 locking connector Broken or severed string trunk cable / rodent chew Shorted or activated bypass diode bypassing a module sub-string (confirm with an I–V curve trace or thermography) If this array has a ground-fault alarm, follow the site's approved DC isolation and testing procedure before replacing the fuse — see ground faults in solar arrays	Check fuse continuity with digital multimeter (DMM) Measure string Open-Circuit Voltage (Voc) at DC disconnect Inspect MC4 latching & crimp integrity along module tables Clamp current with handheld DC clamp meter to confirm reading
Underperforming	`ΔI ≤ -5.0%` or M_i ≤ -3.5 MAD	Tracker row-to-row shading or mechanical drive misalignment Heavy localized soiling, bird droppings, or vegetation shading Activated bypass diode dropping submodule current Microcracks, localized hot spots, or snail trails Severe Potential-Induced Degradation (PID) High-resistance crimp contact or terminal heating	Perform aerial drone or handheld thermal IR survey Check tracker table tilt angle against adjacent healthy tables Inspect modules for physical glass fracture or severe soiling Perform string I-V curve trace to evaluate fill factor & Vmp Inspect combiner box fuse clips for thermal discoloration
Watch Band	`3.0% ≤ \|ΔI\| < 5.0%` or 2.1 ≤ \|M_i\| < 3.5	Mild non-uniform dust or pollen accumulation Normal module manufacturing flash-test tolerance spread Slight structural rack tilt/azimuth variation Longer DC home-run cable length (increased ohmic resistance) Early-stage PID or cell degradation	Log string in SCADA watch list for subsequent monitoring cycle Schedule string for module cleaning during routine site wash Verify string module count matches design layout exactly Confirm cable sizing and run length against single-line diagram (SLD)
Normal Baseline	`\|ΔI\| < 3.0%` or \|M_i\| < 2.1 MAD	String operating nominally within healthy manufacturer tolerance Normal irradiance tracking and clean module surface Balanced string impedance and healthy connections	No immediate maintenance action required Baseline benchmark operating as expected
High / Check Sensor	`ΔI ≥ +5.0%` or M_i ≥ +3.5 MAD	SCADA Hall-effect sensor zero-point offset drift Combiner box current transducer (CT) scaling factor misconfiguration Two strings paralleled into a single channel via Y-connector Local ground reflection (albedo) boost on bifacial modules	Inspect SMB wiring to verify two strings are not landed on one fuse Measure actual DC current with calibrated handheld clamp meter If clamp meter reading is normal, re-zero or calibrate SCADA CT channel Verify SCADA channel mapping in plant telemetry database

Screening thresholds are site criteria, not standardised limits. Measurement, documentation and test requirements for grid-connected PV systems are covered by IEC 62446-1.

Worked Example: Screening an 8-String Combiner Box

This example shows the exact output for one combiner box so you can reproduce every number by hand before trusting the calculator on live plant data. Group: SMB-TEST-01. Reference: Active Median. Both screening methods are shown side by side.

1. Measured string currents

String ID	Measured DC current (A)	Reading type
`SMB-TEST-01-Str01`	12.85	Healthy
`SMB-TEST-01-Str02`	12.90	Healthy
`SMB-TEST-01-Str03`	0.00	Zero current
`SMB-TEST-01-Str04`	12.80	Healthy
`SMB-TEST-01-Str05`	10.45	Low
`SMB-TEST-01-Str06`	12.92	Healthy
`SMB-TEST-01-Str07`	12.30	Slightly low
`SMB-TEST-01-Str08`	13.40	High

2. Group statistics

Active Strings 7 of 8 Str03 (0.00 A) isolated from baseline

Active Median (I_ref) 12.85 A Healthy group benchmark

Active Mean 12.517 A Arithmetic average of active set

MAD (Median Abs Dev) 0.07 A Non-zero (MeanAD fallback not needed)

Governing Formulas: Percentage deviation is ΔI = ((I − I_ref) / I_ref) × 100. The robust score is M = 0.6745 × (I − active median) / MAD.

3. Screening results

Scroll horizontally to view all columns →

String	Current (A)	ΔI (%)	M	Fixed ±5%	MAD (3.5)
Str01	12.85	0.000	0.00	Normal	Normal
Str02	12.90	+0.389	+0.48	Normal	Normal
Str03	0.00	−100.000	—	Zero Current — Check Fuse	Zero Current — Check Fuse
Str04	12.80	−0.389	−0.48	Normal	Normal
Str05	10.45	−18.677	−23.13	Underperforming	Underperforming
Str06	12.92	+0.545	+0.67	Normal	Normal
Str07	12.30	−4.280	−5.30	Watch (Low Marginal)	Underperforming
Str08	13.40	+4.280	+5.30	Watch (High Marginal)	High / Check Sensor

4. Reading the result

Field Diagnostic Interpretation

Str03 reads 0.00 A while seven neighbours produce normally, so it is listed as a critical zero-current string for isolation and testing. Str05 is 18.7% below the reference and is flagged by both methods — shading, soiling, or an open-circuit condition on that row are the usual causes. Str07 and Str08 sit 4.28% either side of the reference: the fixed band places them in the watch bands, while the MAD method flags them, because a 7-string group with a MAD of only 0.07 A makes a 0.55 A offset statistically unusual. Treat that disagreement as a prompt to inspect the string, not as a contradiction between the methods.

5. Energy and revenue screening estimate

The estimate scales each string's fractional deficit to daily energy: δ_i = max(0, (I_ref − I_i) / I_ref) and E_daily = Σδ_i × P_string,STC × PSH × PR. Valid zero-current strings are included with δ = 1.

String	Current (A)	δ_i (Fractional Deficit)
Str03	0.00	1.0000
Str05	10.45	0.1868
Str07	12.30	0.0428
Str04	12.80	0.0039
Total Deficit (Σδ_i)	—	1.2335

Daily Yield Loss

66.61 kWh/day

1.2335 × 15 × 4.5 × 0.80

Annualized Energy Deficit

24,312 kWh/yr

66.61 × 365 days

Annual Revenue Impact

$1,458.69 /yr

24,311.6 × $0.06/kWh

Baseline Inputs Used: 15 kW_p string rating, 4.5 Peak Sun Hours (PSH), 0.80 Performance Ratio (PR), and $0.06/kWh PPA tariff rate. If you need to establish that PR from measured plant data first, run the capacity-based Performance Ratio calculator.

This is an operational screening estimate driven by the kW_p, PSH, PR and tariff you enter. It does not model instantaneous irradiance, module temperature, tracker position, clipping, curtailment, or time-series data, and it is not a PVsyst-equivalent yield simulation.

Frequently Asked Questions About Solar String Current Testing

QWhat does a low solar string current mean?

If a string current is lower than the current of nearby strings under the same conditions, it usually means less irradiance or current loss in that circuit. Common causes include shading or tracker misalignment, dirty modules or bird droppings, a loose or damaged connector, a blown fuse or tripped isolator, or a module or sub-string fault such as an activated bypass diode. The calculator flags the string for inspection and the field team checks the string to find the actual cause.

QHow do I tell whether a string fuse has blown?

If a string shows 0.00 A while nearby strings are producing normally, the string may have a blown fuse or an open circuit. However, the current reading alone does not confirm a blown fuse. Isolate the string, check fuse continuity with a meter, measure string Voc at the DC disconnect, and inspect the MC4 connectors and trunk cable. Before replacing the fuse, check for an earth fault or short circuit that may have caused it to blow.

QWhat is a normal current difference between PV strings?

A current difference of about 2–3% between PV strings in the same combiner box is normal when the modules are clean and receiving similar sunlight. This calculator uses a ±5% limit by default: differences below 3% are considered normal, 3% to less than 5% should be checked, and 5% or more is flagged for inspection. A difference of exactly 5% is flagged, while exactly 3% falls into the check range. These are screening values, not manufacturer limits, so always compare them with the module datasheet and your plant's O&M criteria.

QWhy are zero-current strings left out of the baseline?

Because a 0 A reading is a fault condition, not a healthy reference. Including a zero-current string would lower the reference current and could make other weak strings look normal. Zero-current strings are still shown in the results with their full deviation and a critical status; they are excluded only from the reference current calculation.

QCan I compare string currents on a cloudy day?

You can run the calculator on any reading set, but results taken in unstable irradiance are not reliable. Passing clouds, low sun angles and tracker movement change each string's current at different rates, so percentage deviations become less meaningful. The calculator caps its confidence rating at low when the group average current is very low. For decisions that lead to a work order, compare strings during stable, clear-sky conditions.

QWhat is the difference between the fixed ±5% band and the MAD method?

The fixed band compares every string with the same group reference current and applies identical percentage limits to all of them. The MAD method instead measures how far each string sits from the group's median spread, so it adapts to how tight that particular array actually is. On small groups MAD is more sensitive, and the same string can land in a watch band under the fixed method and be flagged by MAD. Run both and treat disagreement as a reason to look at the string.

QCan this calculator prove that a module or string is faulty?

No. It identifies strings that need a physical check. Confirming a fault requires irradiance and temperature context, a calibrated DC clamp measurement, a string I-V curve trace with fill factor, thermal imaging, or insulation-resistance testing. Equal current between two strings also does not by itself prove equal energy production, because voltage and operating temperature still differ.

QWhat data does the calculator need, and where does it go?

Three columns: a combiner box or MPPT group ID, a string ID that is unique within that group, and the measured DC current in amperes. Optional voltage and timestamp columns are accepted. Everything runs locally in your browser and no reading or file is uploaded. A group needs at least two strings to compare, because a single string has nothing to benchmark against.