NVQ L4 Refrigeration & AC common mistakes
NVQ L4 Refrigeration & AC · Understand the traps and how to avoid them.
Writes a generic two-line risk assessment ("hazards identified, control measures applied, PPE worn") for a task such as brazing in a plant room or recovering refrigerant from a chiller, instead of naming the actual hazards and controls for that job.
Candidates learn a template during site inductions and reproduce it from memory rather than walking through the specific task, so they never demonstrate task-specific hazard identification or the hierarchy of control.
What to do instead
Before writing, list the task steps in order and for each step force out at least one hazard (e.g. hot work near flammable insulation, refrigerant displacement of oxygen in a basement plant room, stored pressure energy, manual handling of a 60 kg recovery cylinder). Then apply controls in hierarchy order - eliminate, substitute, engineer, administrate, PPE last - and state the residual risk. Practise on five different job types so the pattern becomes automatic.
Describes pressure testing or system isolation without mention of a permit to work, physical lock-off/tagging, exclusion zones, venting/equalising stored pressure before breaking containment, or draining trapped liquid refrigerant from low points.
Candidates conflate 'switching off' with 'isolating' and have usually only seen live working done quickly by experienced engineers, so stored energy from compressed nitrogen, hot water, or liquid refrigerant legs is invisible to them.
What to do instead
Write isolation as an explicit sequence: identify all energy sources including pressure and trapped liquid, isolate, lock and tag, prove dead/zero pressure at the test point, establish the exclusion zone with barriers and signage, and only then break containment. Add a line naming who authorises the permit and who verifies the isolation. Rehearse the sequence verbally until it appears unprompted.
Selects condensers, evaporators or condensing units using catalogue nominal duty only, without applying correction factors for the actual design ambient, evaporating/condensing temperature difference, altitude, part-load profile, or the pressure drop over the actual pipe run and fittings.
Nominal ratings are printed in large type and candidates are used to picking 'the next size up' on site; correction tables and the interaction between TD and capacity are the part of design they practise least.
What to do instead
Build a fixed design calculation sheet: state design conditions, apply the manufacturer's correction factors for ambient and evaporating temperature, deduct allowances for suction/discharge line pressure drop and fouling, check part-load performance against the load profile, then check the corrected figure meets the load plus margin. Show every step in the answer so the marker can see the reasoning even if one figure is wrong.
Produces a programme that is a numbered list of activities with durations but no logic links, no critical path, no float, no allowance for permit lead times, refrigerant supply lead times, scaffold or crane bookings, or client shutdown windows.
Candidates think of scheduling as 'what order do we do the work in' rather than resource and dependency management, and they underestimate the administrative lead time on F-gas and electrical isolation paperwork.
What to do instead
Always deliver a linked programme: list activities, state predecessors, mark the critical path, show float on non-critical activities, and add a separate 'approvals and lead times' column for permits, procurement and third-party bookings. Then test the plan by asking what happens if the crane is delayed a day. Practise on a two-week chiller replacement scenario.
Fills in commissioning and witness sheets with 'satisfactory' or a tick without recording the measured value, the instrument used and its calibration status, the acceptance criterion, or the ambient conditions at the time of test.
On site the paperwork is often completed after the fact and from memory, so the habit carries into the exam; candidates also do not distinguish between a test result and a judgement.
What to do instead
Use a three-column discipline: measured value with units, acceptance criterion/tolerance from the specification or standard, and pass/fail. Add instrument ID and calibration date, plus ambient temperature and date/time. Practise filling a nitrogen pressure test record and a superheat/commissioning record from a written scenario until every line has a number.
Jumps straight to 'replace the compressor' or 'the system is short of gas, top it up' without first recording operating pressures, superheat/subcool, amp draw against FLA, temperature differentials, sight glass/level and control fault codes, or comparing against a known baseline.
Diagnosis is taught as pattern recognition from remembered symptoms rather than as a data-gathering method, and under exam time pressure candidates write the conclusion instead of the evidence trail.
What to do instead
Write the diagnosis as a numbered evidence sequence: safety isolate, gather operating data, compare each reading to expected values, form a hypothesis, state the test that would confirm or eliminate it, then the repair. Never state a conclusion without the measurement that supports it. Practise three fault scenarios where the obvious answer is wrong.
Omits key fields from the refrigerant record - cylinder serial/tag numbers, opening and closing weights, quantities charged and recovered, refrigerant GWP and CO2e tonnes, leak check dates and findings, and the identity of the operative - or records a bulk total with no traceability.
Candidates copy the headings of a job sheet without understanding that the record must be auditable and reconcilable, and they confuse refrigerant weight with CO2e.
What to do instead
Working from a scenario, complete the record line by line and then reconcile in and out: opening stock plus purchases minus recovered minus returns must equal closing stock. Convert every charge to CO2e by multiplying the mass in tonnes by the GWP of that specific refrigerant. Check the sheet reads correctly if the regulator picked it up cold with no knowledge of the job.
Applies one leak-check frequency (usually 'every 12 months') to every system regardless of charge size and refrigerant, instead of deriving the interval from the CO2e threshold and adjusting for a detected leak, a recently repaired leak, or leak detection equipment being fitted.
The frequency table is remembered as a single number rather than as a calculation that depends on GWP, and candidates do not routinely convert kilograms to tonnes of CO2e first.
What to do instead
Make the calculation a fixed first step: mass in kg x GWP / 1000 = tonnes CO2e, then read the interval off the threshold band, then apply the modifiers (halve the interval after a leak is found until repaired, recheck shortly after a repair, extended interval where a fixed detection system is installed). Show each of these steps explicitly in the answer.
Answers leadership questions in abstractions - 'I motivate the team', 'I promote a safety culture', 'good communication' - with no named situation, no specific actions (toolbox talk content, delegation with competence check, performance conversation, handover briefing) and no evidence of follow-up or outcome.
Candidates assume leadership is a personality trait being assessed rather than a set of observable behaviours, and they have not rehearsed the STAR structure that the assessor is looking for.
What to do instead
Answer every leadership question as a short case: situation, the specific behaviour (what was said, to whom, in what forum), the check that it landed (competence sign-off, observation, audit result), and the outcome. Prepare three real examples in advance - a near-miss briefing, an underperforming team member, and a handover - and reuse them flexibly.
Writes corrective action as 'retrain the technician and monitor' with no root cause analysis, no interim containment, no defined owner or date, and no verification step to show the problem has actually gone away.
Candidates treat quality as filling in a non-conformance form rather than as a closed loop, and they hold back from naming process causes because it feels like blaming the system.
What to do instead
Structure every corrective action as: contain the immediate risk, root cause by repeated 'why', the process change that removes the cause, an owner and a date, then a verification activity with a measurable acceptance point at a stated interval, and a review of whether the same issue recurs elsewhere. Show the loop closing, not just the training session.