When Should an Architect Bring HVAC Into the Drawing? Architects and HVAC Coordination in Kerala, 2026

The ceiling problems that survive into a finished interior almost never start as HVAC problems. They start as coordination gaps, usually ones that were deferred past the point where they could be resolved cleanly.

By the time a site team flags an issue, the schematic ceiling is framed, the lighting layout is committed, the joinery is fabricated, and the only available response is negotiation over what can be forced to fit. That negotiation produces the outcomes that end up in defect lists and snagging rounds: a diffuser position that compromises the ceiling geometry, a return grille that sits where no one wanted it, an access panel that reads as an afterthought because it was one.

These are not outcomes that better contractors produce. They are outcomes that earlier coordination prevents.

Where HVAC sits in a typical design process, and why that is the problem

In most residential and commercial projects, HVAC enters the design process late. The concept is developed first. The reflected ceiling plan gets laid out around the lighting logic. Interior elevations get resolved. Joinery details get issued for fabrication. Then, somewhere in the documentation or tender stage, the HVAC consultant or contractor is brought in and asked to work within what has already been decided.

On straightforward projects with conventional ceiling configurations and standard wall-split units, that sequence is workable. On projects where the design intent depends on concealed ducted systems, slot diffusers, integrated return air paths, or any configuration that requires the ceiling to do mechanical work behind a finished surface, it is not. The spatial requirements of a properly functioning concealed HVAC system are fixed by physics, not by programme. They do not compress to fit what the concept has left available.

What typically gets lost in that late-entry sequence is not any single element but the relationship between elements: the plenum depth that duct sizing requires versus the ceiling drop the section can afford; the return path that air balance needs versus the joinery run that the interior elevation assumes; the access point that maintenance requires versus the finished surface that the design intent protects. Each element in isolation might appear manageable. Together, when they arrive at site simultaneously, they produce compromises that no amount of contractor coordination can fully recover.

The load is decided before the HVAC is

There is a useful way to think about what a cooling system is actually doing, put well at a whole-building design workshop run by the Indian Institute of Human Settlements: cooling a building is pumping heat back out against the direction it naturally wants to flow. Every watt the envelope lets in is a watt something has to pump out again, every hour of every cooling season, for the life of the building.

That reframes the order of the work. Heat ingress is an architectural output before it is a mechanical input, and it is set by decisions that belong to the drawing rather than to the plant room:

• Orientation and plan geometry. A long facade facing east or west takes low-angle sun straight into the glass. Turning the long axis to face north and south is the cheapest cooling measure available on any project, and it costs nothing after it is drawn.
• Window-to-wall ratio. Around 30 to 40 per cent is the band where daylight and heat gain trade off sensibly. The simulation exercise from that workshop modelled an office at 70 per cent glazing with no external shading at all, and solar gain accounted for 45 per cent of the sensible peak in the worst-oriented ground floor room.
• External shading. A projection factor of zero means the glass specification is doing all the work on its own.
• U-value and SHGC of wall, roof and glass, and the thermal lag a heavier wall buys before the afternoon peak arrives indoors.
• Where the core sits. A central service core shields occupied space on all sides; a side core can be placed to take the worst solar exposure itself.

None of that is the HVAC consultant's decision, and all of it sets the size of the system the project will pay for twice, once in capital and then every month in running cost.

Two honest qualifications belong with that. The first is that a full high-performance specification is expensive, and most projects will not buy one. What is nearly free is the first tier: orientation, glazing ratio, shading and honest zoning, decided while the drawing is still open. Those are the low-hanging fruit, and they are worth taking even on a project with no sustainability ambition at all, because they reduce the plant the client has to buy.

The second is that several of these measures are visible. Deeper reveals, external shading and a lower glazing ratio all change the elevation, and on some buildings that is a gain and on others it is a loss the architect will not accept. That is a legitimate design judgement rather than a technical one, which is exactly why it belongs in the architect's hands early rather than arriving as a mechanical demand late.

Treat each elevation as a different problem

"Turn the long axis north and south" is sound and most sites will not allow it. The more useful principle is that the orientation you cannot change is one you specify around, elevation by elevation, rather than wrapping the building in one curtain wall and hoping the glass specification carries all four sides.

A Grade A office in Mumbai, on a coastal climate far closer to Kerala's than a Deccan one, shows what that looks like done deliberately. The plan is close to square with the service core in the middle, so the whole perimeter is workspace and every elevation is occupied. The facade is then differentiated: generous glazing to the north, where daylight arrives with little direct gain; vertical fins on the exposed elevations; and punched windows east and west, where low-angle sun is hardest to shade and the cheapest defence is simply less glass. Same building, four different answers.

The number worth holding

Facade performance becomes arguable the moment it is only described in adjectives. The metric that settles it is peak facade heat gain per square metre of envelope area, and a reasonable target for a project of this kind is under 25 W/m². It is checkable at design stage, it covers glass and opaque wall together, and it gives the mechanical side a figure to design against rather than a finished elevation to react to.

Two supporting checks belong with it. On glazing, around 1.7 W/m²K is a good U-value to be holding, and the number to interrogate is the one for the assembly as built rather than the centre-of-glass figure a brochure may quote. And the Energy Conservation Building Code sets tiered compliance, ECBC, ECBC+ and SuperECBC, each with its own minimum efficiency requirements. Deciding early which tier the project is actually aiming at prevents the familiar late discovery that the envelope cannot support the tier the client has already been promised.

Set the performance targets yourself

The pattern worth copying from that project is not a facade detail, it is a sequencing habit. The sustainability consultant usually arrives after the massing and often after the facade is fixed, at which point the advice is expensive or unusable. Design teams that end up with good buildings tend to set their own internal targets early and regardless of whether the brief asks for them: daylight autonomy, annual sunlight exposure, peak facade heat gain, and thermal comfort in the occupied zone.

Modelling follows the same progression. A shoe box model answers massing and orientation questions while those are still cheap to change; the detailed model comes later, once the geometry is real, and is where the facade target is verified rather than estimated.

The perimeter zone is where this becomes an HVAC problem

There is one consequence worth knowing about before it becomes a complaint. In a plan with a central core, the desks are at the glass, and the perimeter zone routinely feels warmer than the room's air temperature says it is. The air may be at setpoint while the inner surface of the glass and the adjacent wall are considerably warmer, and an occupant sitting a metre away feels that surface as radiant heat regardless of what the thermostat reads.

It is the same category of failure as a room that is cold and still feels clammy: the measured number is correct and the experience is not. The difference is that humidity is the mechanical side's problem and this one starts at the facade. It is also why perimeter and core deserve separate zoning rather than a single thermostat, and why a facade that hits its heat-gain target quietly makes the HVAC design easier for the whole life of the building.

The practical consequence for coordination is simply that the cooling conversation should start while orientation, glazing ratio and shading are still open, not after they are fixed.

Schematic design: the stage where HVAC decisions are cheapest to make

In schematic design, the architect is establishing the spatial logic of the project: structural grid, ceiling heights, primary volumes, interior zones. This is also the stage at which the fundamental HVAC decisions carry the least cost to make and the greatest cost to defer.

The indoor unit type is the first decision that affects everything else downstream. A wall-split unit makes almost no demands on ceiling coordination. A cassette unit requires a ceiling grid that accommodates it and a return path that works within that grid. A concealed ducted unit requires plenum space, a duct route, a supply terminal strategy, a return strategy, access provisions, and a condensate drainage path, all of which have direct implications for the section and the ceiling composition. That is not a detailed design question. It is a schematic design question, because the answer determines what kind of ceiling the project can have.

Slot diffusers are the other decision that belongs at this stage. Architects specify slot diffusers for the right reasons: they read as architectural elements rather than mechanical ones, they align with linear ceiling and lighting geometries, and they reduce visual clutter in finished interiors. The decision to use them is usually sound. What the schematic stage needs to establish is whether the plenum configuration behind them can support the air distribution performance they imply, because a slot diffuser fed by a crude duct arrangement or an inadequate plenum will whistle, throw unevenly, or stain the ceiling line. The terminal looks resolved and the result fails.

The slot count is part of that conversation rather than a later detail. Slot diffusers are made in a range of widths, ours from two to four slots and all flanged for a clean ceiling junction, and the number of slots changes both the air the terminal can move and the throw pattern it produces. Fixing the ceiling line first and asking the terminal to fit it afterwards is how a run ends up either starved or noisy.

Design development: where the section has to be honest

Design development is where ceiling sections get drawn in earnest, interior elevations get resolved, and joinery details begin to take shape. It is also the stage where the gap between a well-coordinated project and a late-stage compromise becomes visible, if anyone is looking.

The section is where HVAC coordination either gets done properly or gets deferred in a way that the plan drawings conceal. A concealed unit drawn in plan inside a bulkhead or above a joinery element looks straightforward at almost any scale. The section tells a different story: whether machine clearance is actually available; whether there is a physical return air path or just an assumed one; whether the drain can fall in the direction the drainage route requires; whether the unit can be accessed for servicing without dismantling what surrounds it.

These questions are not the HVAC contractor's questions to answer alone. They are section-level architectural questions, because the answers determine the built geometry. If the design intent depends on concealed HVAC, the section has to be developed with that intent tested against real spatial requirements, not assumed to work because the plan looks tidy.

Return air is the element that design development most often leaves unresolved. Supply terminals get drawn carefully because they are visible; return air paths get indicated loosely because return grilles read as minor elements. But the return side of an HVAC system affects airflow stability, static pressure, unit performance, and noise in ways that a decorative grille with insufficient free area, or a return path interrupted by a joinery run, will make apparent during operation. The design development stage is where return sizing and routing need to be confirmed, not assumed.

Documentation: locking what site needs to execute correctly

By the time the project reaches documentation, the coordination decisions should be made. The documentation stage is for recording them accurately so that they can be built correctly, not for resolving questions that design development left open.

In practice, the documentation stage is where deferred HVAC coordination tends to surface as problems in the drawing set. Sections that do not show duct depth. Ceiling details that do not account for access. Return grille positions that have not been sized. Condensate drainage paths that have been noted but not drawn. Each of these is a question that site will answer independently, in the absence of a drawing, during construction, under time pressure.

Access provisions are where documentation most consistently under-delivers. Every concealed HVAC installation requires access: for filter cleaning, drain inspection, valve servicing, and balancing adjustments after commissioning. If the documentation does not show how that access is provided, site will provide it in whatever way is available at the time, which is rarely the way the architect would have chosen. Integrating access into the documentation as a design decision, rather than leaving it as a site instruction, is what separates a resolved interior from one that carries visible evidence of its own maintenance requirements.

This is why HVAC firms like HRS prefer to settle service access with the design team rather than leave it for the technician to discover later.

The Kerala climate and what it changes about acceptable outcomes

The coordination decisions above matter on any project. In Kerala, they carry additional weight because the climate removes the tolerance that drier conditions provide.

HVAC systems in Kerala run longer, run harder, and operate in ambient humidity conditions that expose poor air distribution faster than they would elsewhere. Humidity complaints, condensation staining near terminals, ceiling and wall damage from drainage issues, and noise from airflow that is being forced through undersized return paths are common outcomes of layouts that might have passed unnoticed in a temperate climate. A return grille that is marginally small will cause problems here sooner than it would elsewhere. A drain detail that was not properly drawn will leak here sooner. A concealed unit with inadequate access will require a disruptive intervention here sooner.

That is not an argument for a different design process. It is an argument for taking the standard design process seriously at the stages where it is meant to catch these decisions, rather than deferring them in the expectation that the contractor will manage.

What the design process needs to carry forward at each stage

The following is not a checklist for the HVAC contractor. It is a set of questions the architect needs to have answered before each stage closes, because they affect the geometry, section, and documentation of the project.

Schematic design

• Indoor unit type confirmed - wall split, cassette, concealed ducted, floor-standing, or other
• Ceiling height feasibility checked against plenum and duct depth requirements for the chosen system
• Slot diffuser specification tested against plenum viability, not just visual intent
• Return air strategy established in principle, with a realistic sense of where the path runs and how much free area it needs

Design development

• Ceiling sections drawn honestly against real machine clearances and duct routing, not assumed to fit
• Return air sizing confirmed with free area calculated, not indicated loosely
• Access provisions located and integrated into the ceiling and joinery logic before those elements are detailed
• Condensate drainage path drawn with direction and fall shown, not just noted as a contractor responsibility
• Lighting and terminal coordination resolved for linear ceiling conditions, particularly where slot diffusers and feature lighting share the same zone
• Acoustic collar requirements identified for noise-sensitive spaces before the ceiling composition is locked

Documentation

• Access panel positions shown on the reflected ceiling plan and interior elevations, treated as design elements rather than site instructions
• Return grille sizes and free areas specified, not left to contractor discretion
• Duct depths confirmed against the ceiling section with actual dimensions
• Drainage falls noted with dimensions sufficient to build from
• Any special terminal or collar specifications called out clearly enough that they survive the tender process intact

The project does not need the architect to specify HVAC systems. It needs the architect to hold the design process open long enough for those systems to be properly coordinated before the decisions that constrain them are locked.

What that coordination should produce, and what a drawing set has to prove before it is worth building from, is the subject of our companion piece on AutoCAD HVAC design and drawings.

Where HRS fits

The architects who bring us in early tend to keep doing so, because the coordination above stops being their problem once it is handled at the right stage. HRS works alongside design and interior teams across Kerala to settle duct depths, access panels, drainage falls, and outdoor-unit locations while the drawings are still open, so the HVAC supports the architecture instead of quietly compromising it on site. If you want that input while the decisions can still be made cleanly, talk to our team.