C&G 2922-34 Solar PV common mistakes
C&G 2922-34 Solar PV · Understand the traps and how to avoid them.
Quotes the G98/G99 capacity threshold (16 A per phase / 3.68 kW single-phase) but cannot state when and how the DNO must be notified, or assumes any installation under the threshold can simply be connected and forgotten.
Candidates memorise the kW figure from marketing material and never learn the notification process, so they treat G98 as 'no paperwork' rather than 'simplified notification'.
What to do instead
Learn the sequence as a timeline: confirm connection type (single or three phase), check the inverter's G98/G99 compliance certificate, submit the G98 notification to the DNO before energising, and send the commissioning notification within the permitted period (28 days for G98). Be able to say which form and which body for each route.
Commissions the system and walks away with only an Electrical Installation Certificate, omitting the PV-specific schedule of test results, the DC-side inspection schedule and the handover/O&M pack.
Candidates revise BS 7671 Chapter 71 as theory but never rehearse producing the actual document set, so they don't realise a PV installation needs extra schedules plus building regulations and MCS paperwork.
What to do instead
Build a mental checklist of deliverables: EIC with the additional PV schedule of inspections and test results (including DC string values), building control notification where applicable, datasheets, commissioning record, labelling schedule, and user instructions covering shutdown procedure and maintenance intervals.
Sizes the string using the module's STC open-circuit voltage and compares it directly to the inverter's maximum DC input voltage, without applying the Voc temperature coefficient at the lowest expected cell temperature.
The datasheet headline Voc is at 25 °C; cold bright mornings push Voc up by roughly 10 %, and candidates forget the array will be coldest when it is producing most.
What to do instead
Always calculate Voc(cold) = Voc(STC) × [1 + (αVoc × (Tmin − 25))] using a realistic minimum cell temperature for the site, and compare that figure with the inverter maximum DC voltage; similarly check Vmp at high cell temperature against the MPPT start-up and minimum operating voltage.
Designs strings by counting modules to fit the roof and connecting panels of different orientation, tilt or shading condition onto one MPPT, ignoring the resulting mismatch losses.
Candidates think in terms of module count and total kWp rather than electrical series/parallel behaviour, and surveys often don't record per-plane orientation and shading.
What to do instead
Group modules by identical orientation, tilt and shade exposure; only place electrically similar strings on a shared MPPT (or use separate MPPTs/optimisers). Record orientation, tilt, azimuth and shade timings per roof plane at survey and show the string layout on the design.
Leaves module frames and mounting rails unbonded, arguing that the DC circuit is isolated from earth so no protective earthing is needed.
Candidates learn 'PV arrays are isolated' and over-apply it, forgetting that exposed-conductive-parts of the array and mounting system still require earthing and equipotential bonding under BS 7671 Section 712.
What to do instead
State the rule: all module frames and metallic mounting structures are exposed-conductive-parts and must be connected to the protective earth by a suitable conductor and listed earthing components (toothed washers or purpose-made lugs), with continuity verified and recorded, unless the equipment is Class II throughout.
Routes standard AC flex or single-insulated cable in the DC array wiring, or runs DC and AC circuits in the same conduit with no separation, and secures cables so connectors sit in inaccessible or water-trapping positions.
Candidates treat the DC side as low-voltage wiring and reuse familiar cable and methods, forgetting continuous high DC voltage, UV exposure, thermal cycling and the need for segregated systems.
What to do instead
Specify PV-rated double-insulated DC cable (e.g. H1Z2Z2-K) with UV and temperature ratings, keep DC and AC circuits segregated (separate containment, or a partition), support cables to avoid strain on connectors, keep connectors accessible and off the roof surface in a drip loop, and record the cable route and derating factors used.
Selects a Type AC or Type A RCD for the inverter AC circuit without checking the inverter manufacturer's DC residual current declaration, or omits RCD protection altogether because 'the inverter already has one'.
Candidates apply the generic socket-outlet RCD rule and are unaware that smooth DC residual current can blind Type A devices, so the device type depends on documented inverter data.
What to do instead
Check the inverter's declaration of DC residual current (typically <6 mA) - where confirmed, Type A is acceptable; where not confirmed or where DC residual current can exceed 6 mA, a Type B RCD is needed. State the device type, rating and the evidence used on the design and test schedule.
Connects the inverter output into an existing consumer unit without checking spare ways, busbar rating, the rating of the main switch/fuse, or whether the existing installation has capacity for the added generation.
Candidates focus on the new circuit only and skip the existing-installation assessment that the connection actually depends on.
What to do instead
Before selecting the point of connection: verify a spare way exists, confirm busbar and main switch ratings against total load plus generation, check the supply earthing arrangement (PME needs specific attention - be able to state the requirements for bonding and any DNO/EAWR constraints), and size the dedicated AC isolator and protective device for the inverter's maximum output current with the applicable 1.25 factor.
Performs insulation resistance testing on the DC side with the inverter still connected, or with modules left in circuit, and records a meaningless or damaging result.
Candidates copy the domestic dead-test sequence without adapting it to a live DC source that cannot be switched off at the array.
What to do instead
Test the DC installation in the correct order: isolate AC, isolate DC at the isolator, disconnect the inverter, and test each string/array to earth at the specified test voltage before reconnecting. Then record string Voc and polarity, reconnect in the documented sequence, and note that array-side conductors remain live whenever irradiance is present.
Measures open-circuit voltage and short-circuit current at the array and declares the system faulty (or healthy) by comparing directly with datasheet STC values, without correcting for irradiance and temperature.
Candidates treat STC figures as expected field readings and don't realise a 600 W/m², 5 °C morning can look 'wrong' on voltage while the array is fine.
What to do instead
Record irradiance and module temperature at the time of measurement, apply the Voc temperature correction and scale current with irradiance, then compare with the expected value; use an I-V curve trace where available. Write the readings and correction basis on the commissioning record so the fault decision is defensible.