Solving a Chimney Moisture Problem

Joe Ponessa is a faculty member with Rutgers Cooperative Extension, specializing in housing and energy.

Wall damage that had appeared on the second floor around the chimney gradually worsened over time. It would reappear, even after the new roof was put on.

When ceiling damage appeared on the first floor, reinforcements were called in to help.

The chimney contains two exhaust flues, one serving the fireplace on the first floor and one serving the furnace and water heater in the basement.

After the leaky roof (left) had been repaired (right), ice was visible emerging from beneath the flashing on the chimney exterior. This occurred even during a relatively dry period.

Deteriorated mortar can be seen in this inside view of the chimney at the attic level. The new sheathing and old framing are also visible. A small stream of steam, not visible in this photo, was emerging from the chimney.

Although the ceiling damage had been repaired once before, it had reappeared and had been getting progressively worse, with wetness episodes coinciding with heavy rainfalls. It was obvious that rainwater leakage was a factor. We proceeded to make observations over the next several weeks. During this period, the weather was unseasonably warm most of the time, and rainfall was unusually low; in fact, it was the second-driest fall on record. But the problems continued.

Fischer installed a new roof in October and early November. That project included a tearoff of asphalt shingles and below that, cedar shingles installed on lath. It also included installation of new roof sheathing. During this work, it was noted that the flashing around the chimney was in bad condition. Also, the cricket, a small structure on the roof behind the chimney that deflects rainwater, showed considerable water damage. It was rebuilt.

With a new roof in place, the problem should have been solved, but it was not. The damaged plaster on the inside walls had been repaired, but within the next few weeks, dampness and stains appeared again.

A Roof Leak?

By mid-November, the moisture problem had reappeared in full force. While we all had confidence in Fischer's repair job, we felt that a continuing roof leakage problem could not be ruled out. However, the reappearance of dampness after the roof had been replaced indicated that the problem was not linked so closely with rainfall. We thought that perhaps there were smaller leaks, which would take longer to cause damage on the inside.

Another possibility we considered was that rainwater was soaking into the parging (mortar covering) on the brickwork above the roofline and migrating downward. This seemed unlikely, but just in case, Coyle applied waterproofing to the parging, which was inspected for cracks and treated to a bit of caulk. This didn't help.

Condensation within the furnace flue became another option to consider. The natural gas furnace appeared to be about 15 years old, and was not a high-efficiency unit. The furnace is rated at 100,000 Btu input. It vents through a 6-inch fluepipe that connects to the brick chimney. The chimney has two flues with terra cotta liners; one flue serves the fireplace, and the other serves the furnace and the gas-fired water heater (see Figure 1). A chimney cap had been added to the chimney above the furnace flue when the roofing had been done to complement the existing cap above the fireplace flue.

Condensation can occur when ascending combustion gases meet cooler temperatures in higher parts of the flue, finally dropping below the dew point. It tends to occur in very cold weather, and/or when the chimney is mostly exposed to the outside. We thought condensation was unlikely here, since the weather had not been severely cold and the chimney was located on the inside of an exterior wall, providing some protection from outside temperatures. However, if the chimney was blocked, what otherwise would have been slight condensation could accumulate into heavy condensation because of the obstruction. We were inclined to rule out a blockage in this case, since we were occasionally able to view wisps of vapor emerging from the top of the chimney.

As if we needed further complications, there was a low spot in the guttering right next to the chimney area, where large amounts of water tended to overflow during heavy rainfalls. Although it was probably responsible for some water damage to the soffit there, this overflow did not wet the wall, nor did it saturate the foundation wall (if it had been drenching the wall, the water might have been conveyed to the chimney plaster). We decided to rule this out as a factor in the problem.

All things considered, getting clues from the appearance of moisture was difficult because of the apparent time lag between moisture production and its appearance on the finished wall.

Tests and Observations

In the attic, a kneewall blocked observation of the chimney. In early January, we cut an access to observe the chimney and the underside of the roof deck. The new decking was mostly dry, except for droplets of moisture, apparently condensate (not leakage from the roof), on a few nails.

However, the chimney was a different story. It was immediately evident that the chimney behind the kneewall was quite wet. The bricks were wet to the touch, as if they had been doused with a hose, and the mortar was wet and somewhat crumbly. Also, there was a fair amount of wet, sandy mortar on top of the attic insulation adjacent to the chimney, indicating that the mortar was deteriorating. Since the entire height of the chimney in the kneewall area (about 3 ft) was wet, we began to think of the chimney as a giant wick, conducting moisture from its source--wherever that was--to most other regions of the chimney. The exposed portion of the chimney in the basement was not wet, however.

One unknown factor was the status of the tile furnace flue liner. This looked OK where it exited the top of the chimney, but I learned that sometimes such liners are installed only at the top of chimneys. It is also well known that the combustion products of a gas flame are pretty corrosive, especially if they condense in the flue. While bricks and mortar are most susceptible to the corrosive effects of condensate, it can eventually damage terra cotta flue tiles, as well. We were not able to check the condition of the flue tiles within the chimney.

Closing In

More clues came to the surface after a 3-inch snowfall, followed by a night of freezing rain, caused no noticeable change in the wetness of the wall. This demonstrated to us how unlikely it was that a leak in the roof was sending water along the outer surface of the chimney.

A few days after this precipitation, a cold front moved in, with nighttime temperatures in the teens and daytime temperatures in the low to mid-20s. Observing the chimney exterior at the roofline on the second morning of this cold spell, I noticed a couple of small streams of ice creeping out from beneath the flashing. This ice continued to form and enlarge throughout that day and into the next, while still there was no rain. The continuing accumulation of ice, in the absence of rain, was pretty strong evidence that condensation, and not rainwater, was the source of the moisture.

We returned to the attic--and then we saw it! A droplight had been placed on top of the insulation next to the chimney, the better to observe the chimney and sheathing. In this raking sidelight, shining at a right angle to our line of vision, we happened to notice a small but unmistakable jet of steam blowing out from a small void between the bricks! The vapor was formed as the moist combustion gas met the cold attic air. This was a pretty definite sign of an obstruction in the flue backing up combustion gas and increasing the pressure inside.

It was tempting to think that the obstruction had to be substantial in order to drive a small blast of steam out through a crack in the chimney wall. And yet, when we repeated the match/smoke test at the furnace, we still found no sign of backdrafting. Also, we did observe a faint plume of steam emerging from the top of the chimney, indicating some degree of flow. Perhaps the obstruction created a pressurized pocket that caused the small steam jet. The amount of steam leaking into the kneewall space was trivial, and the space itself was not wet--only the chimney.

It turned out that the chimney was absorbing all the condensate from the inside--that is, because of the blockage, the flue gas moisture was building up inside the brick furnace flue, which basically absorbed it like a sponge (the brick was porous, and the flue liner and mortar were both somewhat degraded). Thus it leaked out, through the bricks, into the walls.

At this point, the owner decided to have a metal liner installed in the furnace flue. During this work, an obstruction was found in the flue. The chimney workers described it as an assemblage of sticks and twigs, along with some leaves. In all likelihood, it was a squirrel's nest.

Ultimately, we realized that there had been two problems: rainwater leakage around the chimney, due to a flashing problem, and a chimney obstruction that led to considerable condensation. We don't know if the condensation problem had existed long enough to have caused damage to the cricket. In any case, both problems are now solved. The wall and ceiling plaster has been repaired again, and it remains dry.

Next Time

The most direct evidence could have been obtained from looking into the flue to check for possible obstructions. This can be done from the top, or from below by removing the flue pipe leading from the furnace and using a mirror to look for a clear path to the sky (this may not be effective if there is an offset in the chimney). Dropping a weighted line through the chimney from the top could also have been useful to identify a blockage. However, given the risk involved in propping an extension ladder against the top of a chimney whose structural integrity is uncertain, a top inspection might be done with a smaller, stepladder on the roof, placed against the chimney.

Using a match or other smoke source to detect a backdraft at the furnace flue has worked for me in the past, at least for a major obstruction. However, in the absence of a strong pull in either direction (after waiting for the flue to warm up) the result should be taken as equivocal. Thus, it would have been better to use some more sophisticated tools and diagnostic equipment (see "Thorough Diagnostic Testing of Unvented Appliances" ).

It would also have been interesting to place a CO detector next to the furnace flue. (CO₂ would be a more sensible gas to measure, but it would require a separate meter.) However, with a clean, properly adjusted burner, CO levels might not be excessive, so this would not always be a foolproof test of backdrafting. Backdrafting would have been, in this case, a definite signal to investigate the chimney further. Other factors (for example, operation of ventilation equipment elsewhere in the house without makeup air) can cause backdrafting, but those factors weren't relevant here.
 
 

Home Energy can be reached at: contact@homeenergy.org
Home Energy magazine -- Please read our Copyright Notice