Forms of Separation Explained: Choosing Between Form 1 and Form 4b

Few specification clauses cause as much silent over-spending, or as much quiet risk, as the form of separation. It is written into nearly every LV switchboard specification in the GCC, frequently copied from the last project, and rarely re-examined against what the installation actually needs.

What separation actually does

IEC 61439-2 defines forms of internal separation as the degree to which an assembly is divided internally by barriers or partitions. The purpose is twofold: protection against contact with live parts of adjacent functional units during work on the assembly, and limitation of the consequences of an internal fault, an arc in one compartment is far less likely to propagate into the next.

The forms answer three separation questions: are the busbars separated from the functional units? Are the functional units separated from each other? And are the external cable terminals separated from the functional units?

What higher forms cost you

Each step up adds barriers, metalwork and, critically, space. In our build practice, moving the same feeder schedule from Form 2b to Form 4b typically reduces usable mounting height per section by ten to fifteen percent and can add a full section to the lineup. It also adds assembly hours and material, which is why a Form 4b board can cost meaningfully more than the identical schedule in Form 3b.

The question is never "what is the best form", it is what work will be done on this board while it is live. If a feeder will ever be isolated, terminated or maintained while the rest of the board stays energised (hospitals, data rooms, process plants that cannot stop), high separation pays for itself the first time. If the whole board can be shut down for any intervention, Form 4b is often specification inertia.

Three clauses to write into your next specification

1. State the form and the sub-form letter (3b, 4a, 4b), "Form 4" alone is ambiguous and will be priced differently by every bidder.
2. State whether separation is achieved by barriers or by insulated coverings; the two are not equivalent for maintenance access.
3. Require the manufacturer to demonstrate the form at FAT with covers removed, separation is visible, so witness it.

Why Your Main Breaker Trips First, Selectivity, Explained Properly

The call always sounds the same. A fault on one small circuit, a failed motor, a damaged cable, and instead of one feeder tripping, the main incomer opens and the whole facility goes dark. The fault was worth one machine. The outage cost the building.

What selectivity means

Selectivity (discrimination) is the property of a protection scheme whereby, for any fault, only the protective device immediately upstream of the fault operates, and every device above it stays closed. It is not a property of any single breaker, it is a property of the pairing of devices, and it must be engineered, not assumed.

Why it fails

Two breakers in series both see the same fault current at the same moment. If their operating characteristics overlap in the current region the fault produces, the race is undefined, and large faults sit in the instantaneous region of both devices unless the upstream unit is deliberately configured otherwise. The classic failure modes we find in audits:

• Factory-default settings. Electronic trip units ship at conservative defaults; if nobody applies a settings schedule, the main breaker's instantaneous pickup may sit below the fault level of a downstream feeder fault.
• Catalogue selectivity assumed across brands. Manufacturer selectivity tables are valid for the tested pairs of that manufacturer. Mixed-brand hierarchies need study-based verification.
• Time-graded schemes without short-time-rated hardware. Adding an intentional delay at the incomer requires a breaker with an Icw rating able to carry the fault for that delay, an ACB property, rarely an MCCB one.

How a coordination study fixes it

A proper study models the network, utility fault level, transformers, cables, every protective device, and computes fault currents at each bus. Device time-current curves are then plotted in series pairs and settings chosen so that, at every credible fault current, a defined margin separates the downstream and upstream characteristics. The deliverable that matters is not the report; it is the settings schedule: device by device, pickup by pickup, applied at commissioning and recorded in the handover file.

When full selectivity is not achievable

Sometimes physics wins, device ratings, fault levels or cost make total selectivity impossible at every level. The honest engineering answer is declared partial selectivity: state the current threshold up to which discrimination is guaranteed, confirm it covers the overwhelming majority of credible faults (most real faults are far below the bolted maximum), and let the client accept the residual risk knowingly rather than discover it in an outage.

The FAT Witness Checklist: What to Verify Before You Accept an LV Panel

A factory acceptance test is the single cheapest moment in a panel's life to find a problem. The same defect found on site costs ten times more; found after energisation, a hundred times. Yet many FATs are attended rather than witnessed, a walk around a closed board, a signature, lunch. Here is what a witness who intends to find problems actually checks.

Before you travel

• Request the FAT protocol in advance and check it against the specification, the test you don't ask for won't be performed.
• Request the internal pre-FAT inspection report. A builder with nothing to hide has one.
• Confirm instrument calibration certificates are current and traceable.

Visual and mechanical, doors open

• Nameplate data against approved drawings: ratings, Icw, form of separation, IP.
• Busbar joints torque-marked; ask for the torque record and spot-check one joint against the OEM figure.
• Clearances and creepage at the closest points, tape measure, not eyeball.
• Ferrule numbers on control wiring sampled against schematics (pick five circuits at random; never the first five).
• Separation barriers actually fitted, the form is visible, so look.
• Mechanical operation: every breaker racked/operated, interlocks proven by attempting the forbidden operation, doors and locks.

Electrical tests to witness live

Before you sign

Every result recorded with the instrument serial number, every punch item written with a closure date, and the dossier, not a promise of one, in your hand or your inbox. At HIKS STAR the signed dossier is a release condition: the panel physically does not ship without it, and its contents live permanently behind the Panel Passport QR on the door.

Busbar Sizing and Temperature Rise at 50 °C: Why Catalogue Ratings Fail in the Gulf

A 2,500 A busbar system that performs faultlessly in a European plant room can run dangerously hot in a Gulf summer switchroom, while carrying exactly the current on its nameplate. The nameplate is not wrong. It was simply rated for a different planet.

The ambient assumption hiding in every rating

IEC 61439-1 verifies temperature rise against a reference ambient, a daily average of 35 °C with peaks to 40 °C. The permitted temperature rises (for example at terminals and accessible surfaces) are limits above that ambient. Put the same assembly in a 50 °C switchroom and every component starts its day 10-15 degrees closer to its limit: insulation ages faster, breaker thermal elements drift toward nuisance operation, and joint resistance, the silent killer, accelerates its slow rise.

What actually determines busbar capacity

• Cross-section and material, copper's conductivity buys smaller bars; aluminium buys cost savings and larger sections.
• Geometry and spacing, two thin bars per phase out-cool one thick bar of the same area; bar-to-bar spacing governs convection.
• Enclosure ventilation and IP, a sealed IP55 outdoor enclosure traps every watt; ventilated IP42 breathes. The same bars can differ by a full rating step between the two.
• Joint quality, a joint at the wrong torque is a resistor. Its heat output grows with the square of current, which is why torque records are not bureaucracy.
• Harmonics, non-linear loads add RMS heating and, with third harmonics, neutral current that "balanced" designs never planned for. Neutral bars at 100% (not 50%) section are the modern default for harmonic-rich buildings.

Doing it properly for a 50 °C site

1. State the design ambient in the specification, "50 °C switchroom, 55 °C peak" changes the design; silence changes nothing.
2. Derate or verify: apply the OEM system's published derating factors for elevated ambient, or verify the specific configuration by calculation/test per the IEC 61439-1 design-verification routes. Rules of thumb are not verification.
3. Engineer the room, not just the board: switchroom ventilation or cooling sized for the assembly's dissipated watts (every kW of losses is a kW of heat load) is cheaper than oversizing every bar.
4. Instrument the risk away: for critical boards, specify joint temperature monitoring or a thermographic survey regime, annual imaging finds the failing joint years before it fails.

AMF vs ATS: Specifying Generator Changeover Without the Confusion

The two terms are used interchangeably in half the specifications we receive, and the gap between what was written and what was meant surfaces at the worst possible moment: the first real power failure.

Two functions, often one panel

ATS, Automatic Transfer Switch, is the switching function: the mechanism and interlocking that transfers the load between two sources safely, ensuring both can never feed the bus simultaneously. AMF, Automatic Mains Failure, is the brain: the controller that detects mains failure against set thresholds, commands the generator to start, supervises voltage and frequency until the set is stable, initiates the transfer, and manages the return, retransfer delay, generator cooldown run, and shutdown. An AMF panel therefore contains a transfer mechanism; a bare ATS assumes something else is managing the generator.

The specification gaps that cause failed outage tests

• Undefined transfer type. Open transition (break-before-make) drops the load for the transfer interval, fine for most buildings, fatal for some processes. Closed transition or delayed-transition options exist but carry utility-approval and synchronising implications that must be stated, not assumed.
• No timing schedule. Mains-failure detection delay (to ride through dips), generator start attempts, warm-up, retransfer delay (typically several minutes, so a flickering mains doesn't cycle the set), cooldown, every one of these is a settable number. If the specification is silent, you inherit factory defaults.
• Interlock philosophy unstated. Mechanical interlock alone, electrical alone, or both? For life-safety and most authority submissions in the region, the answer is both, say so.
• The neutral, forgotten. Whether the neutral is switched (4-pole transfer) or solid depends on the earthing arrangement of the two sources; getting it wrong creates circulating currents or defeats earth-fault protection. This single line is missing from most specifications we review.
• No test provision. An off-load and on-load test facility, and a monthly automatic exercise schedule, should be specified, a generator that is never exercised is a decoration.

What a complete changeover specification states

1. Sources, ratings and fault levels of both supplies
2. Transfer type (open/closed/delayed) and permitted break time
3. Full timing schedule with settable ranges
4. Interlocking: mechanical + electrical, with proof at FAT
5. Neutral switching arrangement, justified against the earthing system
6. Controller functions: event log, comms to BMS, remote status
7. Witnessed outage simulation at FAT, multiple full cycles, not one

Straight answers to the questions we hear most.

And for anything specific to your project, drawings, sites, deadlines, the fastest answer is always a direct one: info@hiks.org or +971 52 249 5822.