This is the report of the only comprehensive study of the behavior of outdoor air supplies for natural draft woodburning fireplaces.  You can get a paper copy from the CMHC Information Centre by calling (613) 748-2367, or you can download a copy in pdf format here  from the Masonry Heater Association site.

Executive Summary

Research at Canada Mortgage and Housing Corporation (CMHC) through the eighties showed that excessive house depressurization can cause the spillage of combustion products from fuel-burning appliances into the indoor air. Operating fireplaces (especially open masonry ones) can be major sources of air exhaust from houses, and can cause this excessive depressurization. Fireplaces also can be a source of indoor air pollution themselves when house depressurization causes them to spill. This project involved the investigation of factory-built fireplace air demands, pressure limits, and air supply strategies, as well as an effort to find ways to isolate house and fireplace air. The work was performed at the laboratories of ORTECH International.

Fireplace Air Requirements

Prepared for
The Research Division
Policy Development and Research Sector
Canada Mortgage and Housing Corporation


ORTECH International, Mississauga
Scanada Consultants Ltd., Oakville
Sheltair Scientific Ltd., Vancouver

Principal Consultants:
Colin A. McGugan, ORTECH
Michael C. Swinton, Scanada
Sebastian Moffatt, Shelair

CMHC Project Manager: Don Fugler

December 6, 1989

In a test room at ORTECH, five factory-built fireplaces were taken through test burns to establish: their resistance to spillage under various room depressurizations, their chimney flow rates, and the flow rates in their specified fresh air intakes. Separate tests were carried out to determine the airtightness of the glass doors and fireboxes, and the flow characteristics of the air intakes and chimneys. Thermal characteristics of the fireplaces and chimneys can be calculated from the data.

The results show that most of the factory-built fireplaces tested would not act as major house exhausts nor would they be likely to spill, under normal operation. Chimney flow rates were relatively low when the fireplaces were operated with closed doors, but open door testing showed significantly higher flows. Fresh air intakes proved to be of variable utility, supplying close to all required air in some fireplaces and less than 25% in others. Those air intakes which were connected to the circulation air plenums were usually ineffective. Those directly connected to the firebox could match air requirements but could be dangerous in reverse flow incidents (when combustion products flow out through the intended intake). Note: the frequency of occurrence of such reversals has yet to be established. All fireplaces would spill, during fire diedown, if a room depressurization of roughly 10 Pascals was maintained. This is a rare level of depressurization in most existing houses, although it is attainable, especially in mechanically exhausted dwellings.

The study also describes the use of the fireplace simulation computer program, WOODSIM, to translate the laboratory results to other types of fireplaces. The report outlines some fireplace design guidelines, based on the study results.


(from page 33)

There are some limitations on the scope of the testing which may affect the widespread application of the results to all fireplaces. The fireplaces that were used for the testing were all factory-built models that had been tested and certified to ULC requirements. They are required to be installed with prefabricated metal chimneys that are tested with the fireplace.

Most of the testing was done with fireplace doors closed, as this was found to be the condition most resistant to spillage.

The information from the testing may not be directly applicable to masonry built fireplaces, or to fireplaces that are operated without doors. The results may also not be applicable where fuel other than dry split hardwood is used.

[Note: A subsequent study carried out for CMHC by Virginia Polytechnic Institute performed similar tests on a masonry fireplace and found the same results at -10 Pa depressurization: ". . . the CO spillage sometimes increased and sometimes decreased as the external air supply opening was adjusted from closed to intermediate to open." — The Effects of Glass Doors on Masonry Fireplace Spillage and Surface Temperatures, CMHC, 1994]

It should be noted that tests were done with charges of at least 6 kg of wood. Small short fires might be more susceptible to spillage during diedown due to less storage of heat in the fireplace and chimney for maintenance of draft during diedown.

The results show that the fireplaces tested were more resistant to spillage than had previously been expected. As was expected, it is difficult to start a fire without spillage when the room is under a negative pressure and there is a flow down the flue. However, if the negative pressure is removed (eg. by opening a window in the fireplace room), draft can usually be readily established. Once a good draft is established, the fireplaces were relatively resistant to spillage as long as the fire is burning well. One condition where draft might be difficult to establish is the case where a chimney has been backdrafting for an extended period of time in cold weather. If the stack is cooled significantly below the house temperature, it may act as an opening below the neutral pressure plane of the house. This condition was not included in the tests because of the difficulty in maintaining the chimney exhaust chamber at cold temperatures for long periods.

After the chimneys were drafting well, no problems with spillage were noted, even when the room was depressurized to a constant -5 Pa. Room depressurizations of  -10Pa did result in spillage from the fireplaces towards the end of the fire when coals were burning. This is a potentially hazardous situation, since the spillage flow is usually high in CO concentration, which is odourless and does not contain any smoke particles. It was found that an ionization smoke detector would respond to spillage during startup of a fire, however it would not respond to spillage during diedown of a fire.

These fireplaces do not appear to have a high potential for depressurization of a house during their operation. They operate well with a supply flow rate of about 20 L/s, and appear to have a maximum flow rate on the order of 50 L/s for the sizes tested. Larger models with higher burn rates may draw higher amounts of air.

The 100 mm diameter combustion air ducts connected to the circulation air plenum do not supply the total air requirements of the fireplace at a house pressure of -5 Pa during the medium to high burn rates. They do provide some measure of protection against excessive depressurization in a tight house, however their ELA is about 0.005 m2 as compared to an estimated ELA of 0.020 to 0.030 m2 for tight houses. To supply 20 L/s at a 5 Pa differential pressure, an ELA of 0.012 m2 would be required. This indicates that combustion air inlets have to be roughly 2-3 times as large in order to match the fireplace exhaust rate at low bum. The ELA of the fireboxes in the units tested ranged from 0.001 in2 on the tightest unit to 0.027 m2 for a unit with loose fitting glass doors. The ELA of the chimneys used was about 0.044 in2 .

The 83 x 254 mm rectangular air intake on Unit B (Pressurizer) was much more capable of providing the combustion air requirements of the fireplace, especially when assisted with a circulation fan. The only problem encountered with this intake was reverse flow of heated air through the intake duct. A fan forced air supply, with a capacity of 20 to 40 L/s connected to the air circulation plenum, would appear to have potential as a combustion air supply.

The 100 mm combustion air duct connected directly to the fire chamber can supply the total air requirements for a low burn fire, once a draft of 15 to 20 Pa has been established, and if the firebox can be sealed tightly from the room. Once operating, this type of fireplace is relatively insensitive to house pressures, and would work well in houses where intermittent high depressurization occurs. The major problem with this type of intake is the potential for reverse flow of hot gases through the air intake when a large negative pressure is applied to the air intake. This condition could occur in a strong wind if the intake were in a leeward area. The pressure in the firebox (15 to 20 Pa) must be overcome in order to produce reverse flow. In comparison, air intakes connected to the circulation plenum will show reverse flow at pressures as low as 3 Pa. Therefore circulation plenum intakes are far more likely to reverse but without major repercussions. If a direct coupled intake is to be used, it must be treated as a flue gas duct, and be appropriately isolated from combustible materials.

[Note: This discussion of outdoor air connections directly to combustion chambers (fireplaces D and E) failed to clarify that, "All fireplaces would spill, during fire diedown, if a room depressurization of roughly 10 Pascals was maintained."  (from executive summary)  Therefore, the direct combustion air supplies did not control or prevent spillage when the room was depressurized.]

One possible method to reduce potential hazards from reverse flow of hot gases through the air intake would be to install a backflow prevention damper in the intake duct. A test, using a draft control damper of the type normally used on oil furnaces, showed that reverse flow could be kept to a minimum using this strategy. The long-term reliability of this approach would need to be investigated before it could be relied on to provide complete protection against reverse flow.

A combined air intake was briefly studied as a potential solution to some problems. This intake consists of a central duct connected directly to the combustion chamber, surrounded by a larger duct connected to the circulation air plenum. Figure 14 shows how this could be set up. This combined duct would allow the fireplace to draw air directly from outdoors for most of the time during operation, and would also help to maintain the pressure in the room in which the fireplace is located closed to the outdoor pressure. If a negative pressure occurred at the air inlet, and flow reversal took place, the air space surrounding the combustion air intake duct would act as an insulator to keep the outer surface of the air intake duct cool. It may be possible to design an intake system so that the fireplace would draw air from the house at times when the air intake was under negative pressure.