The Board You Are Replacing Is Evidence
Open the old unit and read it for twenty minutes before quoting anything. Brown heat staining around one breaker's line terminal, three conductors crushed under a neutral screw designed for one, a bus stab that has arced itself grey, four brands of device in a board listed for one — each of those is telling you something about how the installation was extended and by whom. None of it appears on the plan the customer gives you.
Then check the directory against reality. Switch each way off and walk the house. A schedule that says KITCHEN and actually feeds the kitchen, the garage freezer and an outside socket is the version of the truth the next person will inherit, and you are the only one on the job with the outage window to establish it. Expect roughly half the labels to be wrong or partial in anything over twenty years old.
The expensive discoveries live in the existing installation, not in the board. Lighting circuits with no circuit protective conductor at the switch drops. Cable colours from two eras landed in the same terminal. A borrowed neutral running between two circuits that were never meant to share one, which stays invisible until the two circuits land on separate residual current devices and one of them refuses to stay in. Rubber-insulated tails that crumble when you flex them.
So price the discovery separately from the fitting. A changeover quoted as a fixed half-day, with an insulation resistance test done for the first time after the old board is on the floor, is how a straightforward job becomes a house with no power and an argument about who pays for a rewire. Test the outgoing circuits while the old unit is still energised and still someone else's, and put whatever fails in writing before you order the new one.
Counting the Load That Is Really There
Two defensible numbers exist for the same house, and they answer different questions. The calculated load, assembled from a floor-area lighting allowance, the small-appliance and laundry circuits, and the nameplate rating of every fixed appliance, tells you what the code says the service must be able to serve. The measured load, taken with a logger clamped on the incoming conductors, tells you what this household actually draws. The first is what an inspector checks; the second is what the conductors experience.
For an existing dwelling the calculated figure usually lands high, because it assumes an appliance mix that nobody operates simultaneously. That gap is legitimate headroom on paper and dangerous headroom in practice, since the household composition changes and the installation does not. Where a code offers an optional or measured-demand method for existing services, it also sets conditions on how long the recording has to run and what has to be added on top for the new load — read the conditions, not the summary.
The loads worth arguing about are the sustained ones. Electric space heating, an induction range, a heat pump, an unvented cylinder with a direct immersion, a hot tub, and a vehicle charger all draw for hours at a stretch, and several of them peak in the same evening hours of the same cold week. Two of those together on a service sized decades ago for a gas-heated house with a 3 kW immersion is a different installation from the one the meter was fitted for.
Whatever number you settle on, write down which method produced it and which document you took the demand factors from. A year later, when somebody adds a second charge point, that note is the difference between reusing your assessment and starting again.
Run the calculated method first so you know what the paperwork expects, then set your logged demand beside it — the gap between the two is the argument you will be having with whoever wants the service upgraded.
The total livable floor area of the dwelling used for general lighting load.
20A kitchen/dining small-appliance circuits, each counted at 1500 VA.
Dedicated 20A laundry branch circuits, each counted at 1500 VA.
Sum of the nameplate VA for fixed appliances such as the water heater, range, oven, and dryer.
The nominal service voltage supplying the dwelling.
Estimated required service amperage
38 A
This is a SIMPLIFIED illustrative version of the NEC 220 standard method, omitting several real-world factors (HVAC load, EV chargers, largest motor load addition, optional calculation method). Fixed appliances are carried at 100% because this page takes a VA total and not a count; 220.53 permits 75% where four or more fastened-in-place appliances are present, and a range or a dryer has its own table (220.55, 220.54) that this does not apply. It is NOT a substitute for a complete NEC Article 220 load calculation performed by a licensed electrician, which determines your actual required service size and is required for permitting. Always round up to the next standard breaker/service size (100A, 125A, 150A, 200A) and consult your electrician.
- Total connected load (before demand factor)
- 13,030 VA
- General lighting and receptacles, after the 220.42 factor
- 5,110.5 VA
- Fixed appliances, at 100%
- 4,000 VA
- Demand load
- 9,110.5 VA
They open the calculator with your figures already in it
Home Electrical Service Panel Amperage Sizing Calculator: 37.96 A — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 38 A — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- Amps are not spaces. This sizes the SERVICE and says nothing about the panel's physical capacity — a dwelling can sit comfortably inside 100 A of calculated demand and still need a new panel because every breaker position is taken, the busbar is rated below what is being added to it, or the enclosure will not accept the breaker type the new circuits require. Panels get replaced for spaces at least as often as for amperage.
- The amperage is where the rest of the sizing starts, not where it ends. Service-entrance conductors, the grounding electrode conductor, the meter base, the mast or lateral and the utility's own drop or transformer all follow from this number and none of them fall out of it — and a dwelling service gets its own conductor allowance under NEC 310.12 rather than the general ampacity table, so the wire for a 200 A house is not the wire that table would give it.
- One dwelling unit. A second unit on the same service, an ADU, or a detached garage or workshop on a feeder are not added by piling their loads into these boxes: a multifamily service runs through 220.84's own demand table, and a feeder to an outbuilding is calculated as its own load under 220.40. Stacking two dwellings into this page understates both of them.
Five Parts Behind One Cover
A distribution board is a listed assembly, and the listing covers the whole thing together: enclosure, busbar, main device, outgoing devices and bars, tested as one unit. That matters more than it sounds. A breaker that fits the stab and carries the correct rating is not automatically a legal device in that board, because what was tested was the combination. Manufacturers publish the devices their assembly was evaluated with, and the label inside the door is the authority for what may land on that bus.
The busbar rating and the main device rating are separate figures and both belong on your check. A board with a 100 ampere bus and a 100 ampere main is fully committed on day one; the same enclosure with a larger bus leaves room for the service to grow without a second changeover. On three-phase boards the bus rating is also what governs how much you may load a single phase before the assembly, not the breaker, becomes the limit.
The bars are where the cheap failures live. One conductor per terminal unless the terminal is marked for more, no doubling of a neutral and an earth into a bar meant for one material, and every unused way blanked so the enclosure keeps the ingress rating it was listed with. A neutral bar with two conductors squeezed under a single screw will pass a visual inspection for years and then produce an intermittent fault that reads exactly like a loose connection somewhere out in the house.
Look at what the assembly is made of as well. Where the adopted rules require a non-combustible enclosure in a dwelling, or set conditions on siting a board under an escape route, that requirement is about the enclosure material and not about the devices inside it, and swapping a plastic unit for a steel one late in the job means a new backplate, new gland positions and possibly a new tail length.
What a distribution board is made of
- Dead front and door — the face an occupant touches, and the plane the required clear working space is measured out from Electrical Equipment Working Clearance Space Calculator
- Protective devices — one module width per pole, so the way count and the spare capacity are the same arithmetic Electrical Panel Circuit Directory Slot Calculator
- Busbar chassis — carries the whole downstream load and is rated separately from the main device sitting on it Busbar Ampacity Calculator (Current Density Method)
- Neutral and earth bars — one conductor per terminal, and the landing point for every main protective bonding conductor Equipment Grounding Conductor Sizing Reference Calculator
- Enclosure back box — sets the ingress rating, the gland positions and, where the rules demand it, the fire performance of the whole unit
Way Count, and the Ways You Will Need in Five Years
Count outgoing ways in module widths, not in circuits. A single-pole device takes one width, a two-pole device two adjacent widths, and a residual current breaker with overload protection may be one or two depending on the range. Add the main switch, add any surge protective device, add the isolators for anything that needs local isolation inside the board, and the enclosure that looked generous on the schedule is suddenly full.
Spare capacity is not a courtesy. A board with no free ways forces every future addition into a tandem device the assembly may not be listed for, a sub-board hung beside it, or a full second changeover. Two or three free module widths in a domestic unit costs almost nothing at the point of ordering and is the cheapest thing you will ever sell a customer who later wants a charge point, a garden supply and a heat pump.
Where the board serves anything with a defined future — a garage with no supply yet, a loft conversion in planning, a solar array with an inverter yet to be chosen — reserve the ways for it and note the reservation on the schedule. Reserved and labelled is a decision; empty and unlabelled becomes the first thing someone fills with a socket circuit.
The physical arrangement matters too. Two-pole devices need adjacent widths, some ranges require the residual current devices at a particular end of the bus, and a surge protective device usually wants to sit as close to the incoming terminals as its connecting leads allow, because the length of those leads is part of its protective performance.
Convert the circuit schedule into module widths with the two-pole devices counted at full width, and you find out now whether the enclosure you specified has the spare ways you promised.
The total number of breaker spaces the panel provides.
Each single-pole breaker occupies one panel slot.
Each double-pole breaker occupies two panel slots.
Available panel slots remaining
10 slots
The breakers entered need 30 slots against the 40 this panel has. Being inside the rating on this one check settles nothing about the rest of the circuit — the conductors, the overcurrent device, and the work as installed are all outside it.
- Slots used
- 30 slots
They open the calculator with your figures already in it
Electrical Panel Circuit Directory Slot Calculator: 10 slots — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- Spaces left is not capacity left. What may be added to a panel is governed by its bus and main breaker rating measured against a calculated load with demand factors applied, not by empty positions. A 100 A panel with twelve free spaces may have no room for a range or a heat pump at all, and filling the spaces anyway is exactly how a bus gets overloaded while every breaker in it holds.
- There is no place here for a three-pole breaker. In a three-phase panel a 3-pole device fills three spaces, and entering it as a double-pole loses one space every time - on a commercial schedule that is several positions gone by the time the count matters. The panel also has a fixed number of positions per phase, which decides where a three-phase load can physically go, not just whether it fits.
- Breaker spaces are not the only thing a panel runs out of. Every grounded conductor needs its own terminal, since NEC 408.41 permits one neutral per terminal, and equipment grounds, AFCI and GFCI pigtails and multiwire circuits eat bar holes faster than they eat positions. A panel with four spaces free and a full neutral bar needs an added bar kit, and the wire bending space in a crowded gutter is a code dimension in its own right.
Grouping Devices So One Fault Does Not Darken the House
How the outgoing circuits are grouped under residual current protection is the design decision an occupant actually experiences. A board split into two banks, each behind a single 30 milliampere device, means an earth fault on any one circuit takes half the house with it — and if the split was made by convenience instead of by function, that half can include all the lighting, the boiler and the freezer at once.
Individual residual current breakers with overload protection remove the problem at the cost of module width and unit price. Where the budget will not stretch to a full set, spend it where the consequences are worst: the freezer circuit, the heating controls, the alarm and any circuit feeding a medical device get their own protection, and the split, if there has to be one, keeps at least one lighting circuit on each side so a trip never leaves someone in the dark on a staircase.
Standing leakage is the reason a correctly wired board trips for no visible cause. Every switched-mode power supply in the house leaks a small current to earth through its input filter, and those currents add. Ten or fifteen milliamperes of accumulated standing leakage behind a 30 milliampere device leaves almost no margin, and the device then trips on a kettle or a shower that is doing nothing wrong. Measuring standing leakage per bank at commissioning turns that callback into a design note.
Arc-fault detection follows the same grouping logic and its own separate requirement. Which circuits need it, and whether combination devices are mandated, depends on the occupancy and the code edition the authority has actually adopted — a point worth confirming in writing, since arc-fault requirements have expanded in successive editions and adoption lags publication in many jurisdictions.
| Arrangement | What one fault takes out | Where it earns its place |
|---|---|---|
| One device for the whole board | Everything, including lighting and heating | Small outbuildings and single-circuit supplies only |
| Split load, two banks | Half the circuits, whichever half the fault is on | Budget work where the split is made by function, not by position on the bus |
| Individual devices per circuit | The faulted circuit alone | Anywhere an occupant will be diagnosing the trip themselves |
| Mixed: individual on critical, banked on the rest | The faulted circuit, or one bank of non-critical circuits | Retrofits where module width is the binding constraint |
Fault Current Is a Property of the Assembly
Every device in the board carries a rated breaking capacity, and the board carries a rating of its own. Both have to equal or exceed the prospective fault current available at that position. On a short service close to a large supply transformer, the available current can comfortably exceed the breaking capacity of a general-purpose domestic device — and a device asked to interrupt more than it was tested for does not simply fail to trip, it fails destructively inside a sealed plastic case an arm's length from whoever opened the door.
Transformer impedance dominates the arithmetic on short services. A low-impedance transformer with a short run of large service conductors delivers a high fault level; add distance, smaller conductors or a higher-impedance unit and the figure drops quickly. Where the utility will state the fault level at the origin, use their figure. Where they will not, calculate the transformer-limited value and treat it as the pessimistic bound, since neglecting the service conductor impedance can only overstate the current.
Series ratings are a real and legal way to use a lower-rated downstream device, and they are valid only in the exact combination that was tested — that upstream device, that downstream device, that assembly. The combination is listed on the label inside the door or in the manufacturer's published tables. Substituting either device, even for one with a higher individual rating, voids the series rating and there is no field test that recovers it.
Record the available fault current at the board and the date it was established. Several adopted codes require the value to be marked on the equipment, and every future addition to that board depends on it being there and being right.
Where the utility will not quote a fault level, the transformer-limited value gives you a defensible upper bound to check every device rating against, and it takes one nameplate to produce.
The transformer's nameplate kVA rating.
The three-phase line-to-line secondary voltage.
The transformer's nameplate percent impedance.
Estimated maximum available fault current
11,000 A
This is a transformer-only screening approximation (the 'infinite source' method), and it is NOT a worst case in either direction. It ignores upstream utility source impedance and downstream conductor impedance, which reduce the current available at a point further from the transformer. It also ignores two things that push the real figure ABOVE it: running motors feed a fault for the first few cycles at roughly four to six times their own full-load current, and nameplate impedance carries an ANSI/IEEE tolerance of plus or minus 7.5%, so a transformer at the low end of that band delivers proportionally more. Do not select an interrupting rating from this number. This is NOT a substitute for a complete short-circuit study (per IEEE 141/242 or equivalent software) required for proper overcurrent protective device rating, selective coordination, and arc-flash hazard analysis, which must be performed by a qualified electrical engineer.
- Transformer full-load amps
- 601.41 A
They open the calculator with your figures already in it
Transformer-Limited Short-Circuit Current Calculator: 10,935 A — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
Add the equipment this sizes
This result is a specification — 11,000 A — not a quantity. Put the thing it sizes into your project: how many, what you call it, and your supplier’s price.
What this calculation does not cover
- The secondary voltage is used once, as the nominal line-to-line figure in (kVA × 1000) ÷ (volts × √3), and no input asks what the bus was actually sitting at in the moment before the fault, so a system running a few percent above nominal drives proportionally more current than this arithmetic returns — and typing that elevated voltage into the box moves the answer the wrong way, because the term sits in the denominator.
- Entries outside 15 to 2,500 kVA, 120 to 600 V or 2 to 8% impedance are pulled back to the nearest bound and answered there, so a 3,000 kVA unit or a 1.8% nameplate produces a figure lower than the transformer in front of you would actually deliver.
- One division by the impedance decimal serves for every kind of fault, and no input describes the winding connection or the grounding arrangement, so a line-to-ground fault — whose return path those two things govern — is not the event this number describes.
- The transformer full-load amps carried in the breakdown is the dividend of that same division, derived from nameplate kVA, so it is the working shown rather than an independent check, and it is rated capacity rather than the load actually connected to the secondary.
- Nothing in the arithmetic has a time dimension: the output is a magnitude at the instant of the fault, with no duration behind it, so it says nothing about how long the transformer or the bus it feeds can hold that current.
Balancing Across the Bars
On a three-phase board the single-phase circuits have to be distributed so no one phase carries a great deal more than the others. Imbalance shows up as neutral current, as a voltage difference between phases at the far ends of the installation, and as one phase of the supply transformer working considerably harder than the other two. A board that was balanced on the schedule and wired by whichever way was nearest is a board that has never been balanced at all.
Balance by measured current, not by connected load. Two circuits with identical connected loads can have entirely different duty cycles, and the schedule cannot see that. Clamp each phase at the incomer under normal occupied conditions, note where the imbalance sits, and move devices only after you have the readings.
Two cautions before moving anything. A multiwire branch circuit shares one neutral between two ungrounded conductors that must be on different phases; move one of them onto the same phase as its partner and the shared neutral now carries the sum of both currents instead of the difference, quietly, with no device protecting it. And devices with common trip or a handle tie exist for a reason — separating the poles of a linked device to improve a balance figure is not a trade.
Non-linear loads deserve their own thought. Electronic drivers, switched-mode supplies and variable-speed drives push triplen harmonic currents that add arithmetically in the shared neutral instead of cancelling, so a neutral in a balanced board can still run hotter than any phase. Where the installation is dominated by that kind of load, the neutral conductor and its terminal are sized on that basis and not on the balance calculation.
Feed it the three clamp readings you took under real occupancy and it tells you how far out the board sits, which is the only honest starting point for deciding which devices move.
The measured or calculated current draw on Phase A.
The measured or calculated current draw on Phase B.
The measured or calculated current draw on Phase C.
Phase load imbalance
9.756 %
- Average phase load
- 41 A
They open the calculator with your figures already in it
Three-Phase Electrical Panel Load Balancing Calculator: 9.76 % — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- Current imbalance is not voltage imbalance, and motors care about the second. A small unbalance in supply VOLTAGE produces a far larger unbalance in a motor's phase currents, commonly six to ten times as large, which is why NEMA MG-1 caps voltage unbalance at 1% before a motor must be derated and treats 2% as already costing winding life in extra heat. A panel whose feeders are evenly loaded can still be delivering out-of-balance voltage; that check is made with a meter across the phases, not from this arithmetic.
- One reading is one moment. Circuits are balanced against loads that come and go - a kitchen at midday, chargers overnight, a chiller in August - so a panel reading 5% during the survey can be at 20% eight hours later, and shuffling breakers to flatten one snapshot can make the daily peak worse. What matters is the balance at the time of maximum demand, which comes from a logged reading over days rather than a single clamp.
- Balanced phases do not empty the neutral. On a four-wire wye panel feeding electronic single-phase loads - LED drivers, switch-mode supplies, small drives - the third harmonic and its multiples arrive in phase on all three legs and ADD in the neutral instead of cancelling, so a perfectly balanced panel can run its neutral hotter than any phase conductor. Nothing protects the neutral against that, and nothing on this page will show it; it takes a true-RMS measurement of the neutral itself.
The Metalwork That Must Be Connected Before the Cover Goes On
Every main protective bonding conductor in the building lands on the earth bar of this board, and a changeover is the one moment when all of them are visible at once. Check each conductor is present, correctly sized for the supply arrangement you have, and continuous back to the clamp it claims to serve — the cross-sectional area demanded of a bonding conductor depends on the earthing arrangement of the supply and changes if that arrangement has been altered since the installation was built.
Continuity is the part that gets assumed. A gas bond clamped to a run of pipe that was later interrupted by a plastic section, or a water bond onto a stub of copper downstream of a plastic incoming main, reads as an existing bond and conducts nothing. Test it, do not look at it. The clamp being shiny proves only that someone fitted a clamp.
Where the supply arrangement itself needs establishing or the electrode resistance needs proving, that work belongs to its own sequence of measurements and its own guide; what belongs here is refusing to close the cover with a bonding conductor unaccounted for. A board energised over an unverified bond is an installation whose protective arrangement has never been demonstrated.
Standing Room in Front of the Door
The clear space in front of the board is a requirement about people, not about equipment. Somebody has to be able to stand square to the dead front, with both hands free, and open or close a device without reaching across an obstruction or standing on something. Codes express that as a depth measured from the live parts, a width tied to the equipment, and a headroom figure, with the depth increasing at higher voltages and where earthed metalwork or exposed live parts sit behind the operator.
The space belongs to the board permanently. A unit fitted correctly in an empty understairs cupboard and then buried behind a vacuum cleaner, a coat rail and four boxes has lost its clearance just as thoroughly as if it had been installed in a doorway. Say this to the customer at handover and put it on the certificate, because you will be the one crouching in there at the callback.
Height and access matter as much as depth. The main switch has to be reachable by the person who lives there, which for some occupants rules out the top of a cupboard entirely. Boards do not belong inside bathrooms or in a position where steam from a shower room or a tumble dryer vents across the enclosure, and siting one directly above a sink or a cooker converts a routine reset into an awkward reach over a hazard.
Leave the enclosure's dedicated space above and below clear of pipework and ducting so future circuits can be brought in through the intended entries. A soil pipe run tight over a board is a permanent obstruction to the only route the next set of cables can take.
Put the equipment width and the depth your voltage and condition demand into it, and you get the floor rectangle to mark out on site — which is the version a joiner or a kitchen fitter will actually respect.
The width of the electrical equipment (panel, switchboard, etc.) requiring working clearance in front of it.
The minimum clear depth required in front of the equipment, from NEC Table 110.26(A)(1) based on nominal voltage-to-ground and installation condition.
Required minimum working clearance area
7.38 ft²
Required depth depends on the equipment's nominal voltage-to-ground and the installation condition (1/2/3, based on what's opposite the equipment) per NEC Table 110.26(A)(1) — look up the correct depth for your specific installation rather than assuming a default. The width floor is stated by NEC 110.26(A)(2) as 750mm (30in) — two code figures, not a conversion of one another, since 30in is 762mm; the area here is worked from the 750mm figure, so if you are building to the imperial column allow the full 30in of width. Height clearance (2.0m/6.5ft minimum) is a separate NEC 110.26(E) requirement not covered by this floor-area calculation.
- Clear width used (NEC 110.26(A)(2) floor: 750 mm / 30 in)
- 30 in
They open the calculator with your figures already in it
Electrical Equipment Working Clearance Space Calculator: 7.38 ft² — shown in imperial, US market. The link sets both, so the result they see is the one on your screen.
What this calculation does not cover
- Gives a floor area, not the volume that has to stay clear. Above a switchboard or panelboard, NEC 110.26(E) keeps that same footprint dedicated to electrical use up to 1.8 m (6 ft) above the equipment or to the structural ceiling — no piping, no ducts, nothing foreign to the installation. A duct routed over the panel fails the room after the clearance in front of it was got right.
- Ignores what swings into the space. NEC 110.26(A)(2) requires the width to allow equipment doors and hinged panels to open at least 90 degrees, and a room door that swings back across the working space eats it. This is a clear floor area, not a strip that people, doors and storage may share.
The Changeover, Hour by Hour
The supply side of the main switch is not yours to isolate. Depending on the jurisdiction and the metering arrangement, removing a service fuse or breaking a meter seal is the distributor's work, the supplier's work, or work you are authorised to do under a specific agreement — and the sequence, the notice period and the paperwork differ between all three. Establish which applies weeks ahead, not on the morning.
Everything testable gets tested while the old board is still standing. Insulation resistance circuit by circuit, continuity of protective conductors, and a note of anything that fails, signed off with the customer before you disconnect a single conductor. Once the old unit is on the floor, every pre-existing defect in the house becomes an argument about whether you caused it.
Work dead, prove dead, and prove the proving unit either side of the test. Lock off, label the lock, and keep the key on you. A second person switching something back on because the house went quiet is the accident this sequence exists to prevent.
- Confirm who is permitted to isolate at the cutout and arrange it in advance, with the notice period the distributor requires.
- Test every outgoing circuit at the old board and record the results with the customer present.
- Isolate, prove dead at the incoming terminals, lock off and label.
- Transfer circuits one at a time, identifying and labelling each conductor as it leaves the old board.
- Land the main protective bonding conductors on the new earth bar and verify continuity to each service before the cover goes near the unit.
- Torque every terminal to the marked value with a calibrated driver, working line, neutral and earth in a fixed order so none is skipped.
- Complete the dead tests on the new board, then energise, then complete the live tests and the schedule.
Torque, Terminations and the Retighten Nobody Books
Terminal torque is a marked value on modern devices and a calibrated-driver job, not a wrist-feel job. Under-torque leaves a joint that heats, oxidises and heats further; over-torque crushes a conductor, damages the terminal thread and produces a joint that fails later for the opposite reason. Both failures look identical from outside the enclosure: a warm cover and an intermittent fault.
Preparation is half the joint. Strip to the length the terminal expects, so no bare conductor sits outside the clamp and no insulation is drawn inside it. Where a stranded conductor lands in a screw terminal, keep the strands together and land them square — a single strand folded outside the clamp both reduces the contact area and gives a live whisker somewhere inside a board that is about to be closed. Ferrules where the manufacturer calls for them, sized to the conductor, crimped with the matching tool.
Aluminium tails, where the service uses them, carry their own rules: a lug listed for the material, the antioxidant compound the manufacturer specifies, and torque to a marked value with a re-check policy. The joint that anneals slowly over a decade of thermal cycling is the reason those instructions exist.
Book the retighten. On a board carrying substantial sustained loads, a thermographic survey or a torque check at the first maintenance visit catches the joint that settled, and neither one takes an hour. Domestic work rarely gets that visit, which is an argument for getting the terminations right the first time rather than an argument that they do not matter.
Nothing Is Energised Until the Paperwork Says So
The test sequence has an order and the order is protective. Continuity of protective conductors and ring conductors, insulation resistance, and polarity are done dead, because each of them can reveal a condition that makes energising dangerous. Only then does the board go live for earth fault loop impedance, prospective fault current, residual current device operating times and functional checks of every switch and interlock.
Residual current devices get tested at their rated current and at multiples of it, with the trip times recorded against the limits in the adopted standard, and the integral test button pressed as a separate functional check. The button proves the mechanism; it proves nothing about the tripping current or the time, which is why both appear on the schedule.
The certificate is the deliverable. Depending on the jurisdiction that is an electrical installation certificate with a schedule of test results, or an inspection under a permit, and in some places both a notification to the building authority and one to the distributor. Which applies is a matter of local law, and none of it is optional because the work went well.
Finish with the schedule inside the door: every way identified by what it actually feeds, written from the switching you did during the changeover and not from the old directory. That schedule is the only part of this job the next person will read, and a schedule copied from a board you already proved was mislabelled has passed the error on with your signature under it.
Staging a board changeover
Half of this list is established before the van is loaded and the other half is established with the old unit still energised. Nothing on it can be settled once the tails are cut.
- Load assessment, with the method named — Calculated figure and logged demand recorded side by side, with the document the demand factors came from written next to them.
- Way count in module widths, plus reserved spares — Two-pole and linked devices counted at full width; reserved ways labelled as reserved, or they become socket circuits.
- Available fault current at the board — Utility figure where they will state one, transformer-limited value where they will not; it gets marked on the equipment.
- Device grouping decided by function — Freezer, heating controls and alarm on their own protection; at least one lighting circuit on each bank of any split.
- Standing leakage measured per bank at commissioning — Accumulated filter leakage is what makes a correct board trip on a kettle; a number now saves the callback.
- Working clearance marked out on the floor — Depth, width and headroom from the dead front, agreed with whoever is fitting the cupboard around it.
- Pre-changeover test results, signed — Insulation resistance and protective conductor continuity per circuit, recorded with the customer while the old board still owns the defects.
Opens the calculators above on one screen with the dimensions from this article already filled in. Quantities only — this site publishes no price list, because local prices vary too much to publish honestly.
