Ducting a Bathroom Extractor Fan So It Does Not Drip

What the fan is being asked to move

Before anything is judged, the run needs a target, and there are two legitimate ways to arrive at one. The first is the code minimum for the room, which is a fixed rate rather than a calculation: ANSI/ASHRAE Standard 62.2, Ventilation and Acceptable Indoor Air Quality in Residential Buildings, sets local exhaust rates for bathrooms with separate figures for intermittent and continuous operation, and Approved Document F, Ventilation, does the same job in England in litres per second. Both distinguish a fan that runs on a switch from one that runs permanently, and the continuous figure is always the smaller of the pair. Which document applies, and which edition of it, is a question about the adopting jurisdiction rather than about the bathroom.

The second route is the room's own volume against an air change rate, and it is the one worth running when the room is unusual: a large wet room with a steam shower, a bathroom with a three-metre ceiling, an internal shower room with no window at all. A prescriptive rate written for a typical bathroom quietly assumes a typical volume, and a room at twice that volume clears at half the pace on the same fan.

Take the larger of the two answers, not the more convenient one. A fan selected on the code minimum in a room that needs the volumetric figure will satisfy an inspector and still leave a fogged mirror, and the householder is judging it on the mirror.

The data sheet is quoting a different fan

Fan literature is written in whichever units the factory ships in, and a domestic extract job routinely has three sets in play at once: a duty in litres per second from a metric specification, a carton figure in cubic feet per minute, and a performance curve with static pressure on the vertical axis in either pascals or inches of water column. Nothing about a bathroom is complicated enough to justify getting this wrong, and it goes wrong constantly, because the two airflow units are close enough in magnitude to look interchangeable on a schedule and are not.

The number to distrust is the big one on the front of the box. That is the free-air figure, measured with nothing attached, and no installed fan ever sees it. The Home Ventilating Institute's Certified Products Directory lists rated airflow at more than one static pressure precisely because of this, and the higher-static figure is the one a real run will approach. Some codes and utility programmes name which of those ratings a selection has to be made against; confirm which, because selecting on the free-air number and installing against four metres of duct is how a compliant fan produces a non-compliant flow.

Sound belongs in the same conversation and is rated in sones on the same certificate. A fan chosen at the top of its curve is a fan chosen at the top of its noise rating, and a bathroom fan that wakes the room next door gets switched off at the isolator within a fortnight. Standard 62.2 caps rated sound, with the tighter cap on continuously running units, for exactly this reason.

Convert once, write the converted figure on the drawing, and select against that. Two people converting the same duty independently at two stages of a job is how a 15 litre per second bathroom ends up with a fan chosen for 15 cubic feet per minute.

The route was fixed before the fan was chosen

On a bathroom refit the duct route is settled at first fix by whoever is holding the hole saw, usually weeks before anyone reads a fan curve, and it is the single decision that most often makes the job impossible afterwards. A fan sitting a metre from an outside wall with a straight run through it is a different appliance from the same fan asked to climb into a roof space, cross four rafter bays over a purlin and come out at a tile vent, even though the invoice describes both as supply and fit one extractor fan.

Bends are the expensive part, and their cost is expressed as the length of straight duct that would resist the air just as much. The general method for that arithmetic, and the fitting tables it reads from, belong to the ductwork layout guide rather than here. What is specific to a bathroom is the scale: a domestic extract fan has a very small pressure capability to begin with, so two ninety-degree bends and a short crossing can represent more resistance than the entire straight length they interrupt, and there is no larger system to absorb it.

Two placement rules pay for themselves on every job. Leave a straight length off the fan spigot before the first change of direction, because a bend hard against the outlet meets air that has not developed yet and costs more than any table says it should. And where the framing forces a change of direction, buy it with a swept bend or two forty-fives rather than a moulded square elbow. Neither costs meaningfully more to fit; both are decided at first fix and cannot be revisited once the ceiling is closed.

One hundred, or one hundred and twenty-five

Most domestic extract kit is supplied with a 100 mm spigot and most installers duct it in 100 mm, which is defensible on a short straight run and marginal on anything else. The bore of a 125 mm duct is a little over half again the bore of a 100 mm one, and on a run with any length or any bends in it that difference is the whole margin between a fan that clears the mirror and one that does not. Where the fan body offers the larger spigot, taking it and ducting the whole way in the larger size costs the price difference on a few metres of pipe.

The reduction that undoes all of this is the adaptor. A 125 mm fan outlet stepped down to 100 mm duct a hand's width from the housing throws away the reason the larger fan was specified, and it is easy to do accidentally because the boxed accessories often include the reducer. The same applies to flat channel duct, which is sold against a round equivalent: the equivalence is one of cross-sectional area, and area is not resistance. A rectangle carries more wetted perimeter than a circle of the same area, so the flat section costs more pressure for the same flow, and each of the two adaptors that convert to and from it costs more again.

Concertina duct is not duct

Aluminium concertina flexible duct is what most domestic extract runs are actually built from, and almost everything that goes wrong with them starts there. Its resistance when pulled fully taut is already a multiple of smooth-walled rigid duct of the same diameter, because the corrugation that lets it bend is also what the air has to climb over. The compression research behind ACCA Manual D, and the installation practice set out in the Air Diffusion Council's Flexible Duct Performance and Installation Standards, put a fully extended flex duct somewhere between one and a half and three times smooth pipe, and an ordinary field installation with sag in it higher again; a run left visibly compressed or kinked leaves that range altogether.

The habitual fault is buying a three-metre length for a two-metre crossing and leaving the surplus coiled above the ceiling. That surplus does three separate kinds of damage. It multiplies resistance along its whole length. It introduces a bend radius tighter than anything the manufacturer permits. And, worst for this particular job, it puts a low point in the run where water can stand.

That low point is the mechanism behind the complaint that only ever presents in the first minute of a run. Condensate collects in the belly of the sag between showers. When the fan starts, part of that reservoir is picked up and carried, and because the fall from the sag back to the fan is shorter than the fall onward to the terminal, some of it arrives at the grille as droplets. The householder describes it as the fan spitting water, which is exactly what is happening.

Pull flex taut, cut it to the measured length rather than coiling the remainder, and support it at the spacing the manufacturer states with a strap wide enough not to bite. Where a run is long enough to need real support, the honest answer is to build it in rigid pipe and use flex only for the short connection at the fan, where its ability to absorb vibration is genuinely worth something.

Where the water is made, and which way it runs

A shower puts air into the duct at close to saturation and at a temperature well above the room. That air then travels through a roof space which, on the mornings when the complaints arrive, is within a degree or two of outdoors. The inside face of an unlagged duct sits at the temperature of the space around it, which puts it far below the dewpoint of the air passing over it, and the airstream gives up water onto the bore continuously for as long as the fan runs. Nothing has failed for this to happen; it is what an unlagged duct in a cold space is for.

Lagging works by moving the bore's surface temperature away from the roof space and towards the airstream. That is why the wrap has to be continuous, and why the two places it is habitually left off, the last half metre at the terminal and the connection at the fan housing, are the two places condensate is most often reported. The psychrometric relationships underneath this are set out in the ASHRAE Handbook - Fundamentals, and BS 5250, Management of Moisture in Buildings, is the document that states the practice for ducts crossing unheated spaces.

The outer jacket is doing a second, separate job, and it is the one that gets skipped. Insulation with an open or torn outer face lets humid air reach the cold duct wall from outside, where it condenses inside the insulation, wets it, destroys its performance and eventually drips onto whatever is beneath. Lap the jacket, seal every lap and both terminations, and treat a wire hanger cutting through the jacket as a defect rather than a detail.

Then give the water somewhere to go. Set the run so it falls continuously away from the fan and towards the outside terminal, so that whatever does condense drains out of the building rather than back into it. Where the geometry forces a rise, such as a fan below a duct that has to climb to a roof terminal, the manufacturer's in-line condensate trap exists precisely for that arrangement and is not optional in it. A run with a fall in one direction for two metres and the other direction for one is a run with a reservoir in it, whatever the intention was.

The roof space is not a termination, and a powered ventilator is not the cure

A duct that discharges into the roof space is the worst outcome available and it is not rare, because it happens two ways: deliberately, by someone who thought the loft counted as outside, and accidentally, when a flexible duct pulls off a tile vent spigot and lies on the insulation for a winter still running. Either way the roof space receives the entire moisture output of every shower in the house, and the first evidence is usually mould on the underside of the sheathing at the coldest end of the roof.

The reflex fix, once a wet loft has been found, is to ventilate the loft harder, and a powered roof ventilator is the product that gets reached for. It is worth seeing what such a fan is asked to move before deciding, because the airflow a rule-of-thumb sizing produces for a roof space is an order of magnitude above what the extract fan puts up there, and moving that much air out of a roof space means drawing the same amount in. If the intake at the eaves cannot supply it, the fan takes it through the ceiling instead, which pulls conditioned and humid air out of the house and can steal flow from the extract duct it was installed to compensate for.

The repair is the duct, every time: reconnect it, clamp it, lag it, fall it, and terminate it through the roof covering or the wall. How the roof space's own intake and exhaust are balanced afterwards, and the net free area arithmetic behind that, is the roof ventilation guide's subject and not this one's.