NABCEP PV common mistakes
NABCEP PV · Understand the traps and how to avoid them.
Plotting a shading-scan horizon or pathfinder reading against a due-south / true-north baseline without first converting compass bearings for magnetic declination, and without recording the site latitude, date, and time of the scan.
Handheld compasses and phone apps read magnetic bearing. Candidates compare that raw reading to a solar path chart drawn in true bearings, so the shade window is rotated by 5-20 degrees and shading losses on the design are wrong (usually understated).
What to do instead
Look up declination for the site, apply it to every bearing before plotting, and write latitude/longitude plus scan date on the survey sheet. Re-check the finished shade profile against a known obstruction (e.g., a chimney) to confirm the orientation is right.
Counting total roof-plane square footage or a simple module-count estimate as 'usable area' and never subtracting ridge, hip, valley, and eave access setbacks, vent/pipe keep-out zones, or the walkway/access pathways a code official will require.
The floor plan and satellite imagery make the plane look larger than the buildable rectangle, and candidates are used to dividing gross area by module area. Setback rules are usually learned as text, not applied as a sketch.
What to do instead
Sketch every plane to scale and draw the setback lines and obstruction keep-outs on it before placing modules. Count only modules fully inside the drawn boundaries, and note the specific access/ventilation requirement you applied.
Sizing string voltage using module Voc at STC (or at an average low temperature) instead of correcting Voc to the record-low ambient temperature for the site using the module's Voc temperature coefficient.
The datasheet headline number is Voc at 25 C, and candidates compare that directly to the inverter's maximum DC input. Cold mornings push Voc far above STC, so the string passes on paper but exceeds the inverter's 600 V or 1000 V rating in the field.
What to do instead
Compute Voc at the design cold temperature: Voc_cold = Voc_STC x [1 + (T_cold - 25) x coefficient], summed over the string, including a bifacial or irradiance allowance if the manufacturer requires it. Compare the result to the inverter maximum DC voltage and keep a margin.
Selecting conductors and overcurrent devices from module Isc or inverter output current directly, skipping the 125% continuous-duty factor and the 125% factor for PV source/output circuits, and skipping the ampacity correction for conduit fill and rooftop temperature.
The arithmetic is simple and candidates stop after the first step. Exam items and real designs both hinge on stacking the factors in the right order and remembering that continuous-load and PV-specific multipliers are separate.
What to do instead
Write the chain explicitly: Isc x 1.25 (PV circuit factor) -> conductor ampacity after conditions-of-use corrections must be at least that, and OCPD at least Isc x 1.25 with 690.9 considerations. Check the table for the conductor's own temperature/derating correction before comparing numbers.
Mating connectors from different manufacturers, or re-using the connectors that came with one module brand with another brand's leads, and tightening lugs or microinverter fittings by feel or with an impact driver instead of a calibrated torque tool.
MC4-style connectors look identical but have different contact geometry and locking dimensions, so the pair is not listed or waterproof as an assembly. Field crews trust hand-tight 'snug' and an impact driver is faster than a torque wrench.
What to do instead
Keep one connector brand/system per array, replace rather than adapt when mismatch is found, and use a calibrated torque wrench set to the manufacturer's value for every termination; record torque values on the commissioning sheet and mark each connection.
Performing the insulation-resistance (megohmmeter) test with the inverter, surge protective devices, or monitoring hardware still landed on the circuit, or skipping the test entirely and recording only open-circuit voltage and irradiance.
Disconnecting every device takes time, and the test is treated as a formality. Energizing a megohmmeter into electronics can destroy them, so crews who have been burned once start avoiding the test altogether.
What to do instead
Isolate the circuit, disconnect or bypass inverters/SPDs, test each conductor to ground and conductor-to-conductor at the manufacturer's test voltage, record the resistance reading and test conditions, then reconnect and re-verify polarity and Voc before energizing.
Pulling and replacing modules on a low-output array before isolating the fault to a string by measuring per-string open-circuit voltage and short-circuit/operating current and comparing against sister strings.
A visibly damaged or soiled module is the most obvious suspect and swapping it is satisfying; string-level measurement requires climbing to the combiner or using a clamp meter in the right place.
What to do instead
Measure Voc and operating current for every string under the same irradiance, flag the outlier, then split that string at midpoints and re-measure to locate the bad module, connector, or blown fuse. Confirm with an IV-curve trace or bypass-diode check before condemning hardware.
Treating the DC disconnect as a de-energizing device and working on array wiring, or opening a combiner, on the assumption that the DC side is dead once the inverter is off.
AC habits carry over: kill the breaker and everything downstream is dead. In PV, the array is a current-limited source that is live whenever there is light, so opening the disconnect leaves the source circuits energized.
What to do instead
Treat all PV source and output circuits as energized at all times. Establish shock-hazard boundaries, use a CAT III-rated meter to verify absence of voltage where it can be verified, and use appropriate PPE and insulated tools; cover or otherwise control irradiance only as a documented last step, never as the primary control.
Describing rapid shutdown compliance by the wrong boundary or the wrong voltage/energy limit - for example citing a 30 V limit inside 5 ft of the array when the applicable code cycle sets a different distance, or forgetting that the controlled conductors must be outside the array boundary and the initiator must be at a readily accessible location.
The requirement changed across code cycles and the numbers are easy to blur; candidates recall the concept ('rapid shutdown') but not the specific limits and locations the edition in force requires.
What to do instead
Anchor on the code edition being tested: state the boundary (inside vs outside the array), the voltage and power limits, the distance from the array, and where the initiating device must be located and labeled. Verify with the adopted edition before answering.
Accepting a verbal scope change from the customer - adding a subpanel, moving an array, upsizing the inverter or adding a battery - and proceeding without a signed written change order that states price, schedule impact, and any effect on the incentive or permit paperwork.
The customer is friendly, the crew is on site, and stopping to write paperwork feels like a delay. Verbal changes later become disputes over who agreed to what, and the as-built no longer matches the permitted design.
What to do instead
Have a standard change-order form on every job. No work proceeds until it is priced, signed, and the permit/utility paperwork impact is noted; keep the signed copy in the job file and update the as-built drawings.