Testing: polarity, Voc, insulation resistance, earth resistance

Testing is how you prove the system is safe and built as designed, before it is connected to the inverter and the grid. It also gives you a baseline: a record of how each string and earth pit measured on day one, which makes later fault-finding much faster. Inspectors, commercial clients and insurers increasingly ask for test sheets, and PM Surya Ghar uploads need earthing values.

The standard sequence

IEC 62446-1 (testing, documentation and maintenance of grid-connected PV systems) sets out a basic set of tests for every system. In order:

1. Continuity of protective earthing and bonding conductors
2. Polarity of every string
3. String open-circuit voltage (Voc)
4. String current (short-circuit or operating current)
5. Functional tests of switchgear and the inverter
6. Insulation resistance of the DC circuits

Add earth electrode resistance (IS 3043) and the AC-side tests required for the building wiring. Record every result on a test sheet with the date, time, irradiance, module temperature and the instrument serial number.

1. Earth continuity

With a low-resistance ohmmeter (or the continuity range of an insulation tester), measure between the main earth bar and: each structure section, sample module frames from each row, the inverter body, and the DCDB and ACDB enclosures. Expect a fraction of an ohm. A high or open reading usually means a painted or anodised contact surface, a loose lug, or a broken daisy chain.

2. Polarity

At each string's loose end (still disconnected from the DCDB), check that + is + with a DC voltmeter. A reversed string connected in parallel with others can drive large currents and damage connectors or the inverter input. Some inverters survive reversed polarity; many do not.

3. String Voc: predict, then measure

Voc depends mainly on cell temperature, and only slightly on irradiance above about 200 W/m². So you can predict it closely.

Worked example

String of 10 modules, Voc(STC) 49.5 V, β(Voc) −0.27 %/°C. Module back-sheet temperature measured at 45 °C.

Expected Voc per module = 49.5 × [1 − 0.0027 × (45 − 25)] = 49.5 × 0.946 = 46.83 V Expected string Voc = 10 × 46.83 = 468.3 V

What different readings tell you:

Compare strings with each other too. Identical strings on the same roof should be within about 1–2% of each other.

4. String current

Measure the short-circuit current (with a proper PV short-circuit test device or a multifunction PV tester) or the operating current with a DC clamp meter once the inverter is running.

Expected Isc scales with irradiance:

Isc(site) ≈ Isc(STC) × G ÷ 1,000

With Isc(STC) 13.85 A and a plane-of-array irradiance of 820 W/m²: 13.85 × 0.82 = 11.36 A. A reading within about 5–10% is fine; a string well below its neighbours points to shading, soiling, a damaged module or a high-resistance joint.

5. Functional tests

Operate every switch, isolator and MCB. Test RCCBs with their test button and, where you have the instrument, with an RCD tester. Check SPD status windows. Confirm the inverter starts and stops as expected (lesson 4).

6. Insulation resistance

This test finds damaged cable insulation, water in connectors or junction boxes, and pinched cables under clamps: faults that cause shocks, inverter "isolation fault" errors and fires.

Procedure (array side, isolated from the inverter): 1. Open the DC isolator; remove string fuses; make sure the strings are disconnected from the inverter and SPDs (SPDs would conduct during the test). 2. Connect the tester between string negative and earth, apply the test voltage, read after it stabilises. 3. Repeat between string positive and earth. 4. Discharge the circuit through the tester before disconnecting the leads.

Test voltage and minimum values (IEC 62446-1):

The minimum is a pass/fail limit, not a target. A new small array typically reads tens or hundreds of MΩ. Record the actual value; if one string reads much lower than the others, investigate even if it passes. Results fall in humid monsoon mornings, so note the weather.

7. Earth electrode resistance: fall of potential

1. Disconnect the electrode from the system at its test link (otherwise you measure everything in parallel).
2. Drive the current stake C in a straight line, well away: for a 3 m electrode, 30 m or more.
3. Drive the potential stake P at 62% of the E–C distance (about 18.6 m for 30 m).
4. Read the resistance. Move P to 52% and 72% and read again. If the three readings agree within a few percent, the 62% value is valid. If not, move C further out and repeat.
5. Reconnect the test link.

Clamp-on testers are quick but only work where the electrode is part of a loop with other earths in parallel (for example a multiply-earthed network). On a single isolated pit they give a meaningless reading.

Measure in the dry season if you can; note the soil condition on the sheet.

Key points

• Follow the IEC 62446-1 order: earth continuity, polarity, Voc, current, functional tests, insulation resistance; then earth electrode resistance.
• Predict Voc from module temperature; a shortfall of one module's Voc means a missing or shorted module, two modules' worth means a reversed module.
• Isc scales with irradiance; compare strings with each other.
• Insulation resistance: 1,000 V test and at least 1 MΩ for systems above 500 V; record actual values and investigate outliers.
• Measure earth resistance by fall of potential with the electrode disconnected, checking at 52%, 62% and 72%.

Check yourself

1. A 12-module string (Voc 50.0 V, β −0.28 %/°C) is measured at 40 °C module temperature. What Voc do you expect?
2. The same string reads about 479 V. What is the likely fault?
3. Why must the test link be opened when measuring an earth electrode?