Forum Replies Created
-
Posts
-
A point list should define every point well enough for a contractor to price, wire and program it: tag, description, point type, signal and range, alarm and trend requirements, and whether it is hardwired or read over a network. Point counts alone are not enough.
Recommended columns
- System and equipment tag (e.g. AHU-01), point tag and plain-language description.
- Point type: AI, AO, DI, DO, or network (integrated) point.
- Signal: 4-20 mA, 0-10 V, PT1000/NTC, volt-free contact, pulse, with range and engineering units.
- Alarm limits and priority, trend interval, display on graphics, schedule, and safe or fail position (e.g. heating valve opens on failure for frost protection).
- Controller or panel, and field device reference (to the valve or sensor schedule).
- Hardwired or integration (BACnet object, Modbus register map reference).
Worked example: AHU with supply and return fans on VFDs
- DI (8): supply fan status, return fan status, supply VFD fault, return VFD fault, pre-filter differential pressure, bag filter differential pressure, frost thermostat, fire alarm shutdown.
- DO (2): supply fan start, return fan start.
- AI (6): supply air temperature, return air temperature, outdoor air temperature, supply duct static pressure, return air humidity, return air CO2.
- AO (5): supply fan speed, return fan speed, heating valve, cooling valve, mixing damper.
Total 21 hardwired points. With 20 % spare, rounded up per type: DI 8 x 1.2 = 9.6, so 10; DO 2.4, so 3; AI 7.2, so 8; AO 6.0, so 6. The controller must therefore provide at least 27 I/O. Many VFD values (power, current, running hours, fault codes) are better read over the network than wired, and should be listed as network points.
Pumps
Pumps are specified as separate duty and standby units, not twin-head pumps. Each pump gets its own start command, run status and fault input, plus speed command and feedback if on a VFD. The list should state duty rotation (e.g. weekly or by running hours) and automatic changeover to the standby pump on fault or loss of differential pressure.
Spare capacity and general items
- Specify spare I/O per controller (commonly 10-20 %) and spare space in panels.
- Include virtual points that need trends or alarms, such as setpoints, valve positions and calculated energy.
- Add meters (heat, electricity, water) with their interface type.
Common mistakes
- Giving counts without signal types, so the contractor cannot select controllers.
- Missing safety points: frost thermostat, fire alarm shutdown, smoke damper feedback.
- Wiring points that could be integrated, or integrating safety functions that must stay hardwired.
- No fail positions for valves and dampers.
Related: HVAC commissioning checklist, AHU sizing quick rules. For deeper study: HVAC systems design course.
Check the edition of the code that applies to your project.
Use BACnet (BACnet/IP preferably) for equipment that needs alarms, schedules and many meaningful points, such as chillers and boilers; Modbus is acceptable and common for simple devices like VFDs and meters, provided the register map is supplied and the polling load is checked.
The practical difference
- BACnet (ASHRAE 135, ISO 16484-5) describes data as objects with names, units, status flags, alarms, schedules and trend logs. Devices can report changes of value instead of being polled. Interoperability is defined by device profiles and the protocol implementation conformance statement (PICS).
- Modbus is a simple master-slave register protocol. A register is just a number; meaning, scaling, units and data type (16-bit, 32-bit, float, byte order) come only from the manufacturer’s register map. There are no native alarms or schedules, and data is read by polling.
Physical layer
- BACnet MS/TP and Modbus RTU both use RS-485: daisy-chained twisted pair, up to about 1200 m (4000 ft) per segment at lower baud rates, 32 unit loads per segment without repeaters, end-of-line termination and correct shield earthing.
- BACnet/IP and Modbus TCP run on Ethernet and are much faster, but need IT coordination: addressing, VLANs and cybersecurity.
Worked example: Modbus RTU polling time
Twelve devices (VFDs and meters) on one RS-485 segment at 19,200 baud, 11 bits per character. Each device needs 3 read requests of 10 registers.
- Request 8 bytes + response (5 + 2 x 10) = 25 bytes, total 33 bytes = 363 bits = 363 / 19,200 = 18.9 ms on the wire.
- Silent interval of 3.5 characters before each frame: 38.5 bits = 2.0 ms, twice = 4.0 ms.
- Assumed device response delay 30 ms: about 53 ms per transaction.
- 36 transactions x 53 ms = about 1.9 s per full scan.
Two seconds is fine for monitoring and slow control, but too slow if the BMS must also write speed commands to many drives for fast loops. Larger systems should be split into several segments or use Modbus TCP or BACnet/IP. Note that real device response delays vary widely; ask the supplier.
What to specify
- Protocol and physical layer for each equipment item in an interface schedule.
- Points to be integrated per item (read/write, units, alarms), not “full integration”.
- For BACnet: PICS and required BIBBs or device profile. For Modbus: register map with data types and scaling, delivered before shop drawings.
- Who supplies the interface card or gateway and who tests the integration point by point.
- Network addresses, baud rate and segment layout on the BMS architecture drawing.
- Safety functions (fire shutdown, frost protection, plant trips) remain hardwired.
Common mistakes
- Gateways that expose hundreds of unused points and slow the network.
- No termination resistors or wrong biasing on RS-485, causing intermittent communication.
- Scaling or byte-order errors in Modbus floating point values found only at commissioning.
Related: HVAC commissioning checklist. For deeper study: HVAC systems design course.
Check the edition of the code that applies to your project.
The designer should define what is controlled and how it must behave: architecture, control schematics, point list, sequences of operation, interface schedule and field device schedules. The contractor then selects products, designs panels and wiring, and programs to that definition.
Deliverables at tender stage
- BMS architecture diagram: head end, network levels, controllers per plant room or floor, integration interfaces, power supply (UPS) and riser routes.
- Control schematics per system: each AHU, chiller plant, boiler plant, pump set and terminal unit type, showing every sensor, valve, damper and safety device in its real location.
- Point list: point type, signal, alarms, trends, fail positions.
- Sequences of operation: written step by step, including start-up, normal, alarm, failure, fire mode and restart after power loss. ASHRAE Guideline 36 is a useful model for AHU and VAV sequences.
- Interface schedule: each third-party system, protocol, points and responsibility.
- Field device schedules: control valves (Kvs, size, actuator, fail position), sensors (type, range, accuracy), dampers.
- Testing and commissioning requirements: point-to-point tests, functional tests and trend review periods.
Worked example: control valve in the valve schedule
AHU heating coil, 60 kW, water 70/50 °C (20 K).
- Flow = 60 / (4.186 x 20) = 0.717 kg/s, approximately 2.58 m³/h.
- Design pressure drop across the valve chosen as 20 kPa (0.2 bar): Kv = Q / √Δp = 2.58 / √0.2 = 5.77.
- Select the next standard Kvs of 6.3. Actual valve drop = (2.58 / 6.3)² = 0.168 bar = 16.8 kPa.
- If coil and branch piping between the valve and the differential pressure control point lose 15 kPa, authority N = 16.8 / (16.8 + 15) = 0.53, above the usual target of 0.5.
Putting this on the schedule stops the common habit of selecting the valve the same size as the pipe, which usually gives an oversized valve with poor control.
What can be left to the contractor
- Controller selection and I/O allocation, within the spare capacity specified.
- Panel layouts, wiring diagrams and cable sizes, submitted for review.
- Graphics design, to the designer’s minimum content list.
Standards worth referencing
- EN ISO 16484 series for BACS design, functions and protocols.
- EN ISO 52120-1 (which replaced EN 15232) for BACS efficiency classes A-D; state the class the design targets.
- ASHRAE 135 (BACnet) and ASHRAE Guideline 36 for sequences.
Common mistakes
- Sensors drawn in the wrong place, e.g. a differential pressure sensor at the pump instead of at the index circuit.
- No fire mode sequence coordinated with the fire alarm cause and effect.
- Undefined responsibility for integration of packaged plant controls.
- Pump sets shown as twin-head pumps; specify separate duty and standby pumps with changeover logic.
Related: Fire alarm cause and effect matrix, HVAC commissioning checklist. For deeper study: HVAC systems design course.
Check the edition of the code that applies to your project.
Measure pipe along its centreline, by system, material and size, continuously through fittings and including all vertical runs, then add fittings, valves, insulation and supports as the measurement rules of the BOQ require. Most shortfalls come from missed risers and drops, not from plan lengths.
Measurement rules in practice
- Agree the method first. UK NRM2 practice measures pipes over fittings and enumerates fittings as extra over the pipe. Many national unit-price systems instead have pipe items that include fittings in the rate. Read the item description and measure to it.
- Separate items by system (cold, hot, recirculation, heating flow/return), material, size, and location if rates differ (in shaft, in ceiling, buried, exposed).
- Include vertical lengths from sections, riser diagrams and fixture connection heights: risers, drops to fixtures, rises to radiator valves, offsets around beams.
- Fittings and valves are enumerated by type and size (elbow, tee, reducer, union, valve).
- Insulation is measured by pipe size and insulation thickness, usually to the same lengths as the pipe it covers.
- Supports are measured either as an item in the pipe rate or enumerated from the spacing in the specification.
- Waste: do not add waste to the measured BOQ quantity if the measurement rules say “net”; the contractor prices it. For a material estimate, 3-5 % on pipe is a common allowance.
Worked example: a DN25 galvanised distribution branch
- Horizontal length measured on plan: 42.6 m.
- Six drops to manifolds of 0.8 m each: 6 x 0.8 = 4.8 m.
- Net measured length: 42.6 + 4.8 = 47.4 m (the BOQ quantity).
- Material estimate with 5 % waste: 47.4 x 1.05 = 49.8 m, ordered as 50 m.
- Fittings counted on the drawing: 14 elbows, 6 tees, 2 isolating valves.
- Supports at an assumed 2.5 m spacing from the specification: 47.4 / 2.5 = 19 spaces, plus extra supports at each drop and valve.
Here the drops are 10 % of the plan length. On a typical floor with many fixture connections, vertical lengths can be a larger share still.
Practical tips
- Measure by layer or system on a copy of the drawing and colour what has been measured.
- Keep a take-off sheet with drawing number, revision and location for every line, so revisions can be updated.
- Check scale on scanned or PDF drawings against a known dimension before measuring.
- From BIM models, export by system type and size, but check that fittings are not being counted as pipe length and that unconnected or duplicate elements are removed.
Common mistakes
- Double counting at drawing match lines or at the boundary between floor plans and riser drawings.
- Missing equipment connections in plant rooms, which are dense with fittings and valves.
- One item for all sizes, making the BOQ impossible to price properly.
- Insulation forgotten on valves and fittings where the specification requires it.
Related: MEP quantity take-off tips, How to draw a clear MEP riser diagram.
Check the edition of the code that applies to your project.
Give priority to what is hardest to move: structure and gravity drainage first, then large ducts, cable trays and busbars, sprinkler mains, large pressurised pipes, and small pipes and conduits last. Agree this order and the clearance rules in the BIM execution plan before the first clash run.
A practical priority order
- Structure and architecture fixed elements (beams, columns, ceilings).
- Gravity drainage and rainwater pipes: their falls cannot be changed without changing their level everywhere upstream.
- Large ductwork and smoke extract ducts.
- Cable trays, busbar trunking and ladders: limited bending radius, and spacing from wet services.
- Sprinkler mains and branch lines (sprinkler heads have deflector distance rules from the ceiling).
- Large chilled water, heating and domestic water mains, including insulation.
- Small pipes, conduits and flexible connections.
Worked example: drainage fall budget in a ceiling void
A DN100 drain (110 mm outside diameter) runs 18 m at 1 % fall in a 600 mm void between slab soffit and ceiling.
- Fall over the run: 18 x 0.01 = 0.18 m = 180 mm.
- Start: top of pipe 50 mm below the slab for supports, so pipe bottom at 50 + 110 = 160 mm below soffit.
- End: pipe bottom at 160 + 180 = 340 mm below soffit, leaving 600 – 340 = 260 mm to the ceiling.
- A 400 mm deep duct cannot cross under the drain near its downstream end; it must cross near the upstream end (440 mm available) or be re-routed.
Doing this kind of check before modelling saves many clash cycles; it shows that drains, not ducts, decide the void zones.
Setting up useful clash tests
- Run tests discipline against discipline (e.g. ductwork vs structure, drainage vs ductwork), not “everything against everything”.
- Use a small tolerance to ignore touching elements and modelling noise, and soft-clash (clearance) tests for maintenance and insulation zones.
- Model insulation thickness, or test with an equivalent clearance.
- Group clashes by element or location so one rerouted duct closes dozens of clashes.
- Assign each group an owner and a due date, and track status between runs.
Beyond hard clashes
- Access and maintenance zones: filter withdrawal, valve handles, damper access panels, ceiling access hatches.
- Spacing from cable trays to wet services as the specification requires.
- Supports and hangers: room for threaded rods, unistrut and seismic bracing if required.
- Installation sequence: can the large duct actually be lifted into position after the pipes?
- Penetrations: sleeves and fire stopping coordinated with structure before casting.
Common mistakes
- Models at too low a level of development for the clash run, e.g. ducts without flanges or pipes without insulation.
- Resolving clashes by moving sprinkler branches below ceilings or drains to lower falls without approval.
- Ignoring the ceiling height when “solving” clashes vertically.
Related: MEP coordination clash checklist, BIM LOD for MEP models explained, Drainage falls and cleanouts.
Check the edition of the code that applies to your project.
Use one documented office standard based on an international scheme: ISO 13567 or the US National CAD Standard (AIA layer guidelines) for layers, ISO 19650 naming fields for files, ISO 128 line widths and ISO 3098 text heights; then follow any client or project standard in the BIM execution plan where it overrides yours.
Layers
- ISO 13567 structures layer names into fields: agent responsible, element, presentation and optional fields such as status and sector.
- US NCS / AIA uses a discipline designator followed by major and minor groups, e.g. a mechanical HVAC group with separate minor groups for supply, return and exhaust.
- Whatever scheme is chosen: separate systems (supply, return, cold water, hot water, drainage), equipment, annotation, dimensions and hatches; keep new, existing and demolished on separate layers; do not draw on layer 0 except inside blocks.
Sheets and files
- ISO 19650 file naming uses fields such as project, originator, volume or system, level or location, type, role and number. Agree the codes once per project.
- Sheet numbers by discipline and series, e.g. mechanical plans, sections, details, schematics in separate number ranges, with one sheet list per discipline.
- Sheet sizes from ISO 216: A1 594 x 841 mm and A0 841 x 1189 mm for plans; keep one size per drawing set where possible.
Worked example: text height and line weights
Minimum plotted text height of 2.5 mm (ISO 3098 series 1.8, 2.5, 3.5, 5, 7 mm) for notes, 3.5 mm for room names and tags, 5 mm for titles. In a model drawn in millimetres at 1:1:
- At 1:100, 2.5 mm plotted text = 2.5 x 100 = 250 mm in the model; 3.5 mm = 350 mm.
- At 1:50, 2.5 mm = 125 mm; 3.5 mm = 175 mm.
Annotative text or view-scale text avoids recalculating this by hand. For line weights, ISO 128 uses a series with a ratio of about √2: 0.13, 0.18, 0.25, 0.35, 0.5, 0.7, 1.0 mm. A common MEP practice is to plot the architectural background thin and grey (0.13-0.18 mm), services in 0.25-0.5 mm, and the main system being described heavier so it reads first.
Title block and revisions
- Project name and location, client, consultant, drawing title, drawing number, scale(s) at the stated sheet size, date.
- Status and suitability code (e.g. for information, for approval, for construction).
- Revision table: revision code, date, description, drawn, checked, approved; revision clouds on the drawing for the current revision.
- Drawn, checked and approved names, and a north point and key plan on plans.
Common mistakes
- Text that is readable on screen but below 2.5 mm when plotted at A3 for site use.
- Colours used to mean systems on screen but plotted in black, so systems cannot be told apart; use line types and labels as well.
- Legends that do not match the symbols actually used on the sheet.
- External references with absolute paths that break when files are exchanged.
Related: MEP drawing standards: layers, sheet numbers, title blocks, MEP shop drawing review checklist.
Check the edition of the code that applies to your project.
All of the listed limits govern at the same time, and they are manufacturer- and model-specific: check every one against the data book of the exact outdoor unit, then apply the length and height capacity correction before you confirm the selection.
The limits to check
- Longest pipe, actual and equivalent – from the outdoor unit to the farthest indoor unit. Typical published values are in the range of 150-200 m actual and somewhat more as equivalent length.
- Length after the first branch – often about 40 m as standard, extendable (commonly up to about 90 m) only if the manufacturer’s conditions are met, such as upsizing certain pipes or limiting the difference between the nearest and farthest unit.
- Total piping length – sum of all liquid (or gas) pipe runs, typically several hundred metres up to about 1000 m on large systems.
- Height difference outdoor-indoor – usually different for outdoor unit above (often 50 m, more with options) and below (often 40 m).
- Height difference between indoor units – commonly 15-30 m. This is the limit most often forgotten on multi-storey buildings.
Equivalent length
Equivalent length = actual length + (number of elbows x equivalent length per elbow) + (branch kits x their equivalent length) + any other fittings listed by the manufacturer. Per-fitting values depend on pipe size; take them from the data book, not from water-pipe tables.
Worked example
Roof outdoor unit, 56 kW nominal (two equal 28 kW modules). Indoor units total 54 kW nominal, so the connection ratio is 54 / 56 = 96 %, within the 100 % maximum. Assumed data-book values: 0.5 m per elbow, 0.5 m per branch kit.
Check Design Example limit Result Longest equivalent: 118 m + 22 x 0.5 + 5 x 0.5 131.5 m 190 m OK After first branch 58 m 40 m (90 m with conditions) Check conditions Total piping 640 m 1000 m OK Outdoor above indoor 26 m 50 m OK Indoor to indoor 22 m 15 m Fails Two fixes are needed. The 58 m after the first branch is acceptable only if the extension conditions are satisfied; otherwise move the first branch closer to the far end of the system. The 22 m indoor-to-indoor height exceeds this model’s 15 m, so split the building into two systems by floor zone (or choose a model with a higher allowance).
Capacity: if the manufacturer’s correction chart gives, say, 0.90 for 131.5 m equivalent and 26 m height, the available cooling capacity is 56 x 0.90 = 50.4 kW. Compare that with the diversified block load of the zone, not with the sum of indoor nominal capacities. Do not fix a shortfall by raising the connection ratio above 100 %; select the next outdoor unit size, keeping the modules in a group equal or at most one step apart.
Common mistakes
- Checking only the longest run and ignoring indoor-to-indoor height.
- Using plan lengths without risers and drops.
- Forgetting that additional refrigerant charge grows with liquid pipe length, which then affects the room leak-concentration check.
- Applying one manufacturer’s limits to another manufacturer’s equipment at tender stage.
Related: VRF branch kit selection, 10 common VRF design mistakes. For deeper study: VRF/VRV system design course.
Check the edition of the code that applies to your project.
Divide the total refrigerant charge of the circuit by the volume of the smallest occupied room it serves; for R410A in a normal occupied space the result must not exceed the EN 378-1 practical limit of 0.44 kg/m³. If it does, you must add a measure such as a larger effective volume, leak detection with shut-off, or a smaller circuit.
Steps
- Total charge of the circuit: factory charge of the outdoor unit(s) plus additional charge calculated from the liquid pipe lengths and sizes. Use the final as-designed pipe lengths, not an estimate.
- Smallest room: find the smallest room that contains an indoor unit or refrigerant piping of that circuit. Room volume = floor area x height of the occupied space (the ceiling void counts only if it is not sealed from the room).
- Classify: EN 378-1 sets the allowed charge from the refrigerant safety class, the access category (a: general, b: supervised, c: authorised only) and the location class of the system. For a direct VRF system with indoor units in rooms used by the public or office staff, the simple rule is: maximum charge = practical limit x room volume.
- Compare and, if it fails, apply additional measures.
Worked example
Total charge: 11.8 kg factory + 18.4 kg additional = 30.2 kg R410A. Smallest room: meeting room 3.0 x 4.0 x 2.7 m = 32.4 m³.
- Concentration = 30.2 / 32.4 = 0.93 kg/m³ > 0.44 kg/m³ – fails.
- Maximum charge for this room = 0.44 x 32.4 = 14.3 kg.
- Room volume needed for 30.2 kg = 30.2 / 0.44 = 68.6 m³.
Note that simply splitting into two equal circuits of about 15.1 kg still gives 15.1 / 32.4 = 0.47 kg/m³, marginally over the limit. The circuit containing that room must be reduced to 14.3 kg or less, or another measure applied.
Options when a room fails
- Serve the small room from a separate small system (or a split unit) with a low charge.
- Increase the effective volume with permanent openings to an adjacent space, only where the standard’s conditions for combined volumes are met.
- Refrigerant leak detection in the room with alarm and automatic shut-off valves isolating the indoor unit branch, plus mechanical ventilation where required.
- Keep refrigerant joints out of small rooms; route pipes through corridors.
R32 and ASHRAE 15
R32 is classed A2L (mildly flammable). Its EN 378-1 practical limit is about 0.061 kg/m³, and for A2L refrigerants the charge limits of EN 378 and IEC 60335-2-40 depend on floor area and the installation height of the unit, not on volume alone, so use those formulae. In the US, ASHRAE 15 uses the refrigerant concentration limit (RCL) from ASHRAE 34, about 0.42 kg/m³ (26 lb per 1000 ft³) for R-410A, reduced by half for institutional occupancies such as hospitals.
Common mistakes
- Using only the factory charge and forgetting the additional charge.
- Checking the largest room on the floor instead of the smallest.
- Counting a sealed ceiling void or the volume of a closed adjacent room.
- Doing the check at concept stage and not repeating it after pipe routes change.
Related: 10 common VRF design mistakes, VRF heat recovery vs heat pump. For deeper study: VRF/VRV system design course.
Check the edition of the code that applies to your project.
Choose heat recovery when a significant part of the operating hours has rooms needing cooling and others needing heating on the same system at the same time – typically deep plans with internal zones, mixed facades, or server and meeting rooms next to perimeter offices. If all rooms on a system swing between heating and cooling together, heat pump VRF with sensible zoning is enough.
How to decide
- Map the zones: perimeter by orientation, internal core, rooms with high internal gains (meeting, IT, kitchenettes).
- Check simultaneous demand: in the mid-seasons and in winter, does any core or south zone need cooling while north or perimeter zones need heating? Hourly load simulation is best; a manual check of a winter morning, a sunny winter noon and a mid-season afternoon is a reasonable minimum.
- Consider the alternative: could heat pump systems be grouped by orientation so that every system has one mode at a time? If yes and the core is small, heat recovery may not pay back.
- Consider control expectations: heat recovery lets every room choose its own mode, which matters for hotels, offices with tenants, and clinics.
Idealised energy example
At one operating point the core needs 40 kW cooling and the north perimeter needs 25 kW heating. Assume EER 4.0 and COP 4.0 for a simple comparison.
- Heat pump VRF, two separate systems: cooling 40 / 4.0 = 10 kW compressor power; heating 25 / 4.0 = 6.25 kW; total 16.25 kW.
- Heat recovery VRF: the cooling side absorbs 40 kW and rejects 40 + 10 = 50 kW of heat; 25 kW of this goes to the heated rooms and the remaining 25 kW is rejected outdoors. Compressor power is roughly the 10 kW of the cooling duty plus some losses, say about 11 kW.
- Saving at this point: about 5 kW out of 16.25 kW, roughly 30 %.
This is the best case; the benefit falls when demands are unbalanced or when only one mode is needed. Multiply the saving only over the hours when simultaneous demand really exists, then compare with the extra capital cost of the selector boxes, the third pipe or larger pipes, and controls.
Sizing the outdoor unit
Heat recovery does not change the basic rule: the outdoor unit capacity must be at least the sum of the connected indoor unit capacities (connection ratio not above 100 %), with modules in a group equal or at most one size step apart. In addition, each branch selector box has its own maximum connected capacity and number of ports, and indoor units that must change mode independently need their own port. Check refrigerant charge and the room leak-concentration limit; the selector boxes and extra piping increase the charge.
Common mistakes
- Putting units that need different modes behind the same selector port.
- Locating selector boxes above rooms without maintenance access or where their noise is a problem.
- Claiming heat recovery savings for hours when the whole building is in one mode.
- Using heat pump VRF for a mixed-mode building and then fighting complaints by changing the master mode.
Related: VRF heat recovery vs heat pump systems, VRF branch kit selection. For deeper study: VRF/VRV system design course.
Check the edition of the code that applies to your project.
Because a low-pressure gas system has very little pressure to spend: every appliance needs a minimum inlet pressure to burn correctly and safely, so the pipe is sized to keep the total pressure loss from the meter regulator to the farthest appliance within the small allowance set by the gas code or utility. Velocity is at most a secondary check.
Why pressure drop governs
- Domestic natural gas is commonly supplied at about 20-21 mbar in Europe and Turkey (about 7 in. w.c. in the US). The appliance burner and its gas valve are adjusted for a nominal inlet pressure; below the minimum, flame stability, ignition and combustion quality deteriorate, and CO can rise.
- The allowable drop is only a small fraction of the supply pressure. The US fuel gas code tables, for example, are built on a specified drop such as 0.5 in. w.c. (about 1.2 mbar) for low-pressure systems. Your local code or utility sets the figure for your project.
- Gas is compressible and its density depends on pressure, so a velocity limit alone does not guarantee the delivered pressure.
The method
- List appliances with their nominal heat input (gross or net, as the code requires) from the data plates.
- Convert to gas flow: flow (m³/h) = heat input (kW) / calorific value (kWh/m³), using the value the utility specifies.
- Apply a diversity (simultaneity) factor only where the code allows it, for example for several dwellings on one riser.
- Measure lengths for each section, including vertical runs, and add fitting allowances as the code requires.
- Size section by section with the code’s tables or formula so that the cumulative drop to the farthest appliance stays within the allowance.
- Apply height correction if the code requires it.
Worked example (flow and sensitivity only)
A flat has a 24 kW combi boiler and a 7 kW cooker. With a calorific value of 9.5 kWh/m³ (use your utility’s value):
- Boiler: 24 / 9.5 = 2.53 m³/h; total at full simultaneous use: 31 / 9.5 = 3.26 m³/h.
- In turbulent flow, pressure drop rises roughly with flow to the power 1.8 and falls with internal diameter to the power of about 4.8. Doubling the flow therefore gives about 21.82 = 3.5 times the drop.
- Going from DN20 to DN25 steel (internal diameter about 21.6 mm to 27.2 mm) cuts the drop to roughly (21.6 / 27.2)4.82 = 0.33 of the previous value.
- Natural gas is lighter than air, so it gains roughly 0.04 mbar per metre of rise; LPG is heavier and loses pressure going up. Many codes specify whether to include this.
The actual pipe sizes must come from the utility’s approved tables or formula, not from these ratios.
Velocity
Some codes and utilities also cap velocity, mainly for noise, dust carry-over and meter or regulator protection, especially in medium-pressure service lines. Treat it as an additional check, not the basis.
Common mistakes
- Sizing each section for the drop allowance instead of the whole route.
- Using appliance output instead of heat input.
- Ignoring future appliances the client has mentioned.
- Not checking pressure at the appliance during commissioning.
Related: Natural gas pipe sizing basics. Always follow the gas utility’s approval procedure and the local gas code for sizing, materials and testing.
Check the edition of the code that applies to your project.
The requirement depends on where the appliance takes its combustion air: open-flued (type B) and flueless (type A) appliances need permanent, non-closable air supply to the room, sized by the local gas code from the heat input; room-sealed (type C) appliances take air from outside through their own flue system and normally need no combustion air opening, although the room may still need ventilation for other reasons.
Appliance types (EN classification)
- Type A – flueless: products of combustion go into the room (e.g. cookers). The room needs ventilation to dilute them and an openable window or extract as the code requires.
- Type B – open-flued: combustion air from the room, products through a flue. Needs permanent air openings; very sensitive to extract fans and to rooms being too airtight.
- Type C – room-sealed: air and flue gas pass through a sealed concentric or twin-pipe system to outside. Most modern combi and condensing boilers are type C.
Principles behind the openings
- Combustion air: natural gas needs roughly 9.5-10 m³ of air per m³ of gas for complete combustion, plus excess air.
- Dilution and cooling: open-flued appliances with a draught diverter draw extra room air into the flue, and plant rooms need air to remove heat from boilers, pipes and pumps. This is why plant rooms usually have both low-level and high-level openings or mechanical ventilation.
- Pressure balance: a kitchen extract hood or bathroom fan in the same space can reverse the flue of an open-flued appliance and cause spillage of combustion products. Codes either prohibit the combination or require a spillage test.
- Location bans: most codes prohibit open-flued appliances in bedrooms, bathrooms and small rooms below a minimum volume.
Small example: the order of magnitude
A 24 kW open-flued water heater at 9.5 kWh/m³ uses 24 / 9.5 = 2.53 m³/h of gas, so it needs about 2.53 x 9.6 = 24 m³/h of air just for stoichiometric combustion. With excess air and draught-diverter dilution the actual air drawn from the room is several times higher. This is why a closed door and sealed windows can starve an open-flued appliance within minutes, and why the openings must be permanent.
Central boiler rooms
For plant rooms, the local code (for example EN 1775-based national rules, BS 6644 in the UK, NFPA 54 in the US, or the utility’s technical specification) typically requires: dedicated combustion air openings or fans interlocked with the burners, high-level openings for the lighter-than-air natural gas, gas detection with an automatic shut-off valve at the room entry, emergency stop outside the door, and in some codes a pressure relief area. Openings must not face escape routes and must not be closable.
What to show on drawings
- Appliance type (A/B/C) and heat input for each appliance.
- Position, free area and level of every air opening, with the code clause used.
- Extract fans in the same room and how the conflict is resolved.
- Gas detection, shut-off valve and CO alarm locations where required.
Common mistakes
- Counting a grille’s overall size instead of its free area.
- Vents later closed by tenants because they cause drafts.
- Specifying open-flued appliances in new airtight dwellings.
Related: Natural gas pipe sizing basics. Final opening areas, locations and permits come from the local gas code and the gas utility.
Check the edition of the code that applies to your project.
Terminal clearances exist so that flue products cannot re-enter the building or the appliance, cannot harm people, and cannot damage the facade. The binding distances come from the local gas code (or the utility’s technical specification) and the boiler manufacturer’s installation instructions; where they differ, the stricter one applies.
The principles behind each clearance
- Re-entry into the building: distance from openable windows, doors, air bricks and ventilation intakes, measured horizontally, above and below the terminal. Openings above the terminal are critical because flue gas rises.
- Re-circulation into the appliance: corners, recesses, balconies with solid parapets and areas under eaves trap the plume, and a room-sealed boiler can draw its own flue gas back in, leading to poor combustion.
- People and property: minimum height above ground, balconies and walkways where people stand; terminal guards where the terminal is within reach; distance from boundaries facing a neighbour’s window.
- Plume nuisance and damage: condensing boilers produce a visible plume and acidic condensate that stains facades, freezes on paving and damages some finishes. Distances from facing walls, car ports and soffits address this.
- Other openings and services: distance from other terminals, gas meters, electrical boxes and drain pipes.
Why condensing boilers need extra attention
Burning methane produces about 1.6 kg of water vapour per m³ of gas (2 m³ of vapour per m³ of methane). A 24 kW boiler using 24 / 9.5 = 2.53 m³/h of gas therefore releases up to about 2.53 x 1.6 = 4 kg/h of water vapour. Part of it condenses inside the boiler and drains away; the rest leaves as a plume. This is why terminal position, plume kits and condensate drainage are part of the design, not site decisions.
Design and coordination steps
- Fix boiler locations with the architect at concept stage, with flue routes shown on plans and elevations.
- Mark the terminal on facade elevations with the clearance zones from the applicable code drawn around it, and check windows, balconies and adjacent terminals on all floors.
- Check maximum flue length and number of bends against the manufacturer’s limits; horizontal flues of condensing boilers are normally laid falling back towards the boiler (commonly about 3 degrees, roughly 50 mm per metre) so that condensate drains to the appliance.
- Where clearances cannot be met, consider a vertical flue, a plume management kit, or a shared flue system approved for the appliance and permitted by the local code.
- Show the condensate drain route to a trapped drain.
Common mistakes
- Terminal directly below an openable window of the floor above.
- Balcony later glazed in by the occupant, turning an outdoor terminal into an indoor one; many codes prohibit flue terminals inside enclosed balconies.
- Terminals in light wells or narrow passages where the plume cannot disperse.
- Using distances from another country’s guide or another manufacturer.
Related: Natural gas pipe sizing basics. The gas utility’s approval and the local gas code govern the final terminal positions.
Check the edition of the code that applies to your project.
A larger ΔT reduces flow in direct proportion, which cuts pipe sizes and pumping energy substantially. The price is larger coils or a lower supply temperature, which costs some chiller efficiency. For most new buildings 6 to 8 K is the better balance, provided every coil is actually selected for that ΔT.
1. Flow rate
q = Q / (c · ΔT), with c = 4.19 kJ/kg·K:
- ΔT 5 K: 500 / (4.19 × 5) = 23.9 l/s (86 m³/h).
- ΔT 7 K: 500 / (4.19 × 7) = 17.0 l/s (61 m³/h), 29 % less.
2. Pumping energy (worked example)
Assume the 5 K design needs 25 m head and pump efficiency 0.70:
P = ρ·g·q·H / η = 9.81 × 23.9 × 25 / 0.70 ≈ 8.4 kW.
Keeping the same pipes at 7 K, friction falls roughly with the square of flow: H ≈ 25 × (17.0 / 23.9)² ≈ 12.8 m.
P ≈ 9.81 × 17.0 × 12.8 / 0.70 ≈ 3.0 kW, about 64 % less. In practice designers take part of the gain as smaller pipes and valves (lower capital cost) and part as lower pump energy. Coils selected for higher ΔT can have a higher water-side pressure drop, so take the actual coil and valve losses from selections.
3. Chiller and coils
- To get 7 K you either lower the supply (6/13 °C) or raise the return (7/14 °C). Lowering supply temperature costs roughly 2 to 3 % chiller efficiency per kelvin; raising the return costs nothing at the chiller but needs deeper coils.
- Coils and fan coils must be selected at the new water temperatures: more rows or larger units for the same sensible and latent duty. Check dehumidification at the higher mean water temperature.
- Many chillers have a minimum evaporator flow; check it against the reduced flow at part load.
4. When 5 K still makes sense
- Existing systems where coils cannot be changed.
- Process loads with tight supply temperature limits.
- Small systems where pump energy is minor and standard fan coil selections are at 7/12 °C.
5. Keep the ΔT in operation
Designing for 7 K is worthless if the plant runs at 3 K (“low ΔT syndrome”). Use two-way pressure-independent control valves, no three-way bypasses at terminals, clean coils, correct supply setpoint and variable speed pumping with separate duty and standby pumps controlled on differential pressure at the index circuit.
Common mistakes
- Changing the plant ΔT on paper while fan coils are still selected at 7/12 °C.
- Three-way valves at terminals, which return supply water and collapse the ΔT.
- Ignoring chiller minimum flow at part load.
- Oversized control valves with poor authority.
Related: Chilled Water Pump Head: A Quick Estimate, Fan Coil Unit Selection Step by Step.
Check the edition of the code that applies to your project.
The trap is a water column that must balance the static pressure in the coil section. For a draw-through (negative pressure) section, the drain pan outlet must sit at least the negative static pressure plus about 25 mm above the trap outlet, and the water seal must be at least half the static pressure. For a blow-through (positive) section, the seal must exceed the positive pressure plus a margin.
1. Why these dimensions
Under negative pressure P at the pan, condensate can only leave when the water in the inlet leg stands P higher than the outlet weir. If the pan outlet is not higher than that, water backs up into the pan and overflows. At fan start, before condensate forms, the existing water redistributes: the inlet leg rises by P/2 and the outlet leg falls by P/2. If the seal is shallower than P/2, it breaks and air is drawn in through the drain.
For blow-through sections, positive pressure pushes the inlet leg down; the seal must be deeper than P so air cannot blow through.
2. Worked example (draw-through)
- Maximum negative static pressure in the coil section, with filters at their final (dirty) pressure drop: 600 Pa.
- In water column: 600 / 9.81 ≈ 61 mm.
- Pan outlet above trap outlet: H ≥ 61 + 25 = 86 mm → use 90 mm.
- Water seal: J ≥ 61 / 2 ≈ 31 mm → use 35 mm.
Add the pipe diameter and the trap body, and the drain pan outlet needs roughly 150 to 200 mm free height below it. This is why AHUs need a plinth or base frame of adequate height; check it on the plant room drawings, not on site.
Units of measure: 600 Pa ≈ 2.4 in wg, so in US terms the inlet leg is about 3.4 in and the seal about 1.2 in. Many manufacturers state the same rule in their installation manuals; follow theirs where it is stricter.
3. Other drain rules
- One trap per drain connection; do not join drains from sections at different pressures before their traps.
- Fall at least 1 % (1/8 in per ft, as in IMC 307.2) to an open tundish or air break; never connect directly to the soil system.
- Fit a cleaning tee or plug and prime traps at commissioning.
- Where traps can dry out in winter, use self-sealing or ball-type traps approved for the pressure.
- Outdoor and roof units: insulate and trace-heat drains against freezing.
- Size the drain for the condensate flow; for example 2 m³/s with a 5 g/kg moisture removal produces about 1.2 × 2 × 0.005 × 3600 ≈ 43 kg/h.
Common mistakes
- Using a standard sink trap with a fixed seal, regardless of fan pressure.
- Calculating with clean-filter static pressure.
- No plinth, so there is no room for the trap.
- Trap on the main drain only, several units sharing it.
Related: Condensate Drain Trap Design for AHUs and FCUs, AHU Sizing: Quick Rules.
Check the edition of the code that applies to your project.
For a quick, defensible figure use the airflow per metre of hood for the duty of the appliances (IMC method) and check it against face velocity; for a detailed design use the appliance heat-load method of EN 16282 / VDI 2052. Supply make-up air for about 80 to 90 % of the exhaust at low velocity so the kitchen stays slightly negative.
1. Methods
- IMC (US): minimum exhaust per linear foot of hood depends on hood type and appliance duty (light, medium, heavy, extra-heavy). For a wall-mounted canopy over medium-duty appliances it is 300 cfm/ft (about 465 l/s per metre).
- Face velocity check: about 0.25 to 0.5 m/s across the hood face area for canopy hoods, higher for heavy duty.
- EN 16282 / VDI 2052: exhaust is built from the convective heat and moisture of each appliance, its simultaneity and the hood arrangement. It often gives lower values for light cooking and higher values for heavy charbroiling.
- ASHRAE 154 follows the hood-type and duty logic used in the IMC.
2. Worked example
- Hood: 3.0 m = 9.84 ft, wall canopy, medium duty (fryers, griddle, ranges).
- Exhaust: 300 cfm/ft × 9.84 ft ≈ 2950 cfm ≈ 1.39 m³/s (5020 m³/h).
- Face velocity check: 1.39 / (3.0 × 1.2) ≈ 0.39 m/s. Acceptable.
- Make-up air at 85 %: ≈ 1.18 m³/s (4270 m³/h).
- Exhaust duct at about 7.5 to 8 m/s: 1.39 / 8 ≈ 0.17 m² → for example 450 × 400 mm (0.18 m², 7.7 m/s).
3. Code checks
- NFPA 96 requires a minimum grease duct velocity of 500 fpm (2.5 m/s); design velocities of 7 to 10 m/s keep grease in suspension, but check noise and pressure drop.
- Canopy overhang: at least 6 in (150 mm) beyond the cooking surface on open sides, and the hood lip no more than 4 ft (1.2 m) above it (IMC).
- Grease ducts: continuous welded or listed, fire-rated enclosure where they pass through other compartments, access doors for cleaning, falling back to the hood.
- Fire suppression over cooking appliances, interlocked to shut off fuel and stop make-up air as the system requires.
4. Make-up air
- Introduce it through low-velocity diffusers or perforated plenums away from the hood edge; keep air velocity near the hood below about 0.25 m/s, or the plume is disturbed and spills.
- Temper it in winter. In large kitchens consider demand control (temperature and optical sensors) to reduce exhaust at low load.
- Interlock exhaust and make-up fans; the exhaust starts first and stops last.
Common mistakes
- Sizing by air changes of the kitchen volume instead of by hood and appliance duty.
- Make-up air through ceiling diffusers blowing straight at the hood.
- Grease duct velocity too low on long horizontal runs.
- Ignoring that the hood length must cover the whole cooking line plus overhang.
Related: Commercial Kitchen Hood Exhaust Airflow: Quick Sizing, Kitchen Hood Wet Chemical Systems: Design Basics.
Check the edition of the code that applies to your project.
