Installing a Wood Stove: The Chase, the Draught and the Woodpile

Framing a Box Around a Tested Distance

A chase is the least interesting thing a framer builds and one of the easiest to build wrong. It carries no load, it is four studs and some plates, and its inside dimension is set by a document the framer has usually not been handed. Whoever frames it is working to a figure that came off a chimney manufacturer's installation sheet, and when that sheet is still in somebody else's van the opening gets framed to whatever looked about right.

The arithmetic is not the hard part. The rough opening is the outside dimension of the chimney across the middle, plus the required clearance once on each side — and that doubling is what gets missed. A two-inch clearance around an eight-inch casing wants a twelve-inch opening, not a ten-inch one. Round pipe in a square hole is measured the same way, on the diameter, because the clearance has to hold at the closest point and the closest point is where the circle faces the middle of each stud.

The gap is then maintained by parts, not by intention. Firestop spacers at every floor and ceiling penetration, a support box at the appliance end where the system calls for one, a proprietary shield where the run passes through a ventilated loft — each is a listed component whose job is to hold the pipe central and stop somebody filling the space later. On an air-cooled system that space is the insulation. It is not a void waiting for a bag of loose fill, and treating it as one is the defect a sweep reports most often.

It is also a defect somebody built deliberately. A chase framed correctly and then insulated out to the sheathing line by a crew who saw a cavity has had its clearance removed by a trade that never read the chimney sheet and had no reason to. Mark the chase on the drawing, mark it on the framing itself, and tell the insulation contractor in writing that this cavity is not theirs.

The top of the chase is a small roof in its own right: a pan sloped away from the pipe, a storm collar bedded on the flue above it, counterflashing lapped by whatever the chase is clad in, and a cap sized to the system. Termination height comes from the same documents as the clearances — the chimney extends a stated distance above the point where it passes through the roof, and a further distance above any part of the building within a horizontal radius of it, the three-two-ten arrangement in NFPA 211 and the IRC. A chase built tall enough to look right and not tall enough for that rule gets rebuilt in October, when the stove will not draw.

How Much Stove the Room Actually Wants

Wood stoves are oversized far more often than they are undersized, and the oversized one is the worse outcome. Run below its design output it smoulders, the flue runs cool, tar condenses in the top third of the chimney where the gas is coldest, and the glass blacks over on the first still evening. The occupant closes the air down further because the room is too hot, which accelerates all of it. A stove that has to work near its rated output through most of the winter burns clean and stays clean.

The connected volume decides the size, not the room the stove stands in. An open plan where the living space, the stair and a landing all share air behaves as one large space with poor distribution. A cottage of small rooms behind closed doors behaves as one small space with the rest of the house unheated. Neither is described by the floor area on the plan, which is why a rule-of-thumb figure per square foot needs reading with the layout in front of you and the manufacturer's own rated heating area beside it.

Insulation shifts the answer further than climate does. The forty BTU per square foot that lands about right on a moderate-climate house of ordinary construction is generous for a recently retrofitted envelope and thin for a solid-wall cottage with single glazing. Treat the output as a bracket to test a shortlist against, then let the rated heating area and the burn time between reloads choose between two stoves inside that bracket.

The Building Is Competing for the Same Air

A naturally aspirated stove has no fan. It takes combustion air out of the room and relies on the buoyancy of hot flue gas to pull it, which puts it in competition with every other opening in the envelope and loses it to anything with a motor. A range hood at boost, a dryer, two bathroom fans and a central vacuum can between them hold a tight house several pascals negative, while a stove flue develops only a few pascals of draught when it is warm and close to none when it is cold.

The failure is not a smoky room, or not only that. It is spillage: flue gas leaving through the door seal and the connector joints instead of going up the chimney, carrying carbon monoxide into the space with it. It arrives at exactly the wrong moments — a cold flue at start-up, a reload with the door open, a windy evening with the extract running while dinner cooks. A carbon monoxide alarm is required by the adopted residential code wherever a fuel-burning appliance is installed, and it is the last line of defence, not the design.

Screen it with arithmetic long before anyone books a test. A blower-door result, its reference pressure, and a total of everything that can exhaust at once feed the envelope power law, and that law extrapolated down from the test pressure estimates how negative the building goes. A house that needed five hundred litres per second to hold fifty pascals, then asked to give up two hundred and fifty litres per second with nothing coming back, lands near seventeen pascals negative — against a cap that in the CSA F326 lineage is a small single-digit number. That is not a marginal result to argue about.

Only three remedies exist and all of them are structural. Interlock a powered makeup-air unit to the exhaust appliance that causes the problem so it runs whenever that appliance runs. Move to a sealed-combustion stove drawing its air from outdoors through a direct-air kit the appliance and the chimney are both listed for. Or reduce the exhaust. What is not a remedy is a passive hole in a wall with no damper on it, and neither is a line in the handover pack asking the occupants to open a window before lighting the stove — that is a control which depends on a person doing something in an order nobody wrote down.

1. Total every exhaust device that can run simultaneously — range hood at boost, dryer, each bathroom and utility fan, central vacuum, and the exhaust side of any recovery unit — at real flows rather than nameplate.
2. Take envelope leakage and its reference pressure from the blower-door report for the same envelope, in the configuration the worst-case test will use.
3. Run the screening estimate before the appliance is ordered, while a sealed-combustion model is still a specification decision and not a change order.
4. Where the estimate lands anywhere near the limit, book the worst-case test: envelope closed, every exhaust running, interior doors in their least helpful positions, manometer referenced outdoors.
5. Confirm both the depressurisation limit and the exhaust capacity that triggers mandatory makeup air with the local authority, because both are jurisdictional and neither travels between provinces or states.
6. Retest after any later work that tightens the envelope or adds an extract, since the figure that passed at handover described the house as it stood that day.

A Liner That Grows Inside Masonry That Does Not

Where the stove is being connected into an existing masonry chimney rather than a factory-built system, the liner becomes the component the installation lives or dies on. Clay liners to ASTM C315, firebrick, and cast-in-place refractory all do the same three jobs: carry the gas, keep the products of combustion out of the masonry, and resist the acid condensate that forms whenever the flue runs cool. All three also get far hotter than the brickwork around them, quickly, every time the appliance is lit.

That difference is a movement problem before it is anything else. Fireclay refractory expands at roughly five and a half parts per million per degree Celsius, and a wood flue commonly runs between two hundred and fifty and five hundred degrees in normal operation. Take a six-metre liner from a twenty-degree install temperature through a four-hundred-degree rise and it grows about thirteen millimetres over its length — half an inch of movement inside a stack that has no intention of moving with it.

Thirteen millimetres has to go somewhere. Accommodated, it goes into the joints the liner system was designed with and into whatever gap was left at the termination. Unaccommodated, it goes into cracked sections, a lifted crown, or a liner that has quietly pushed the top course of brick out of line over a decade of heating cycles. Refractory mortar of the kind tested to ASTM C199 is formulated for the heat and for some movement, but it is a mortar and not an expansion joint; manufacturers still specify accommodation gaps on longer runs and at the top. ASTM C1283 sets out the installing practice for clay lining and is the document to work from rather than habit.

Two arithmetic traps sit inside the same short calculation. The first is the temperature field, which wants the rise from installation to peak and not the flue reading, so a flue at five hundred degrees off a twenty-degree install is a four-hundred-and-eighty-degree rise. The second catches anyone working in Fahrenheit: a temperature difference converts by nine fifths with no thirty-two degree offset, so that same rise is eight hundred and sixty-four Fahrenheit degrees, not the nine hundred and thirty-two the flue itself reads. Typing the reading where the rise belongs overstates the movement by about eight per cent, and it is the easier of the two to do without noticing.

Chimney fires sit outside all of it. A flue fire runs far above normal operating temperature and asks the liner to take that rise in minutes rather than over an evening, and an allowance sized for normal service is not what keeps the stack intact through one. That is a reason the sweeping interval matters more than any number on this page.

A Season of Wood, Ordered a Year Early

Fuel is the part of this job the installer hands over and the occupant gets wrong, and the wrong version damages the appliance that was just fitted. Wet wood cannot burn hot enough to keep the flue above the temperature at which tar condenses, so the tar lines the chimney instead of leaving it. A stove fed on unseasoned wood produces a chimney needing twice the sweeping and glass that blacks over in an evening, and the customer blames the stove.

Under about twenty per cent moisture, measured on a freshly split face rather than the weathered outside of a round, is the target that appliance manuals and the EPA's Burn Wise guidance converge on. Reaching it takes a summer for softwood split small and most of two for dense hardwood split large — stacked clear of the ground, open at the ends, covered on top only. Which is why the woodpile is a first-year decision and not an autumn one: whatever is bought in September is being burned the winter after this one.

Buying it is a measurement problem with a legal answer. A full cord is a stacked pile four feet by four by eight, one hundred and twenty-eight cubic feet, and in the United States the method of sale in NIST Handbook 130 recognises the cord and decimal fractions of it for fireplace and stove wood. Face cord, rick and rack are not units at all — they describe a stack four feet high and eight feet long with the depth left to whatever the seller's log length happens to be, so two sellers' ricks can differ by half. Measure the delivery stacked, on the ground, before the truck leaves the drive.

Then turn the stack into seasons honestly. A neat row two and a half metres long, a little over a metre deep and the same again high is not far off a full cord. How many of those a winter takes depends on the appliance output, the hours it runs and above all the species: the spread between dense hardwood and light softwood by stacked volume approaches a factor of two on the published species tables, which is a larger effect than most people expect and the reason a cord that lasted the neighbour until February runs out here in January.