The Bedroom That Was Only Cold With the Door Shut
Why this matters
A comfort complaint that takes four visits costs the shop four truck rolls and the customer's patience, and the first three visits in this case were done well. The furnace was healthy, the balancing was competent, the register swap was reasonable. Each tech measured the right things with the bedroom door standing open, because that is how you walk into a room with a flow hood. The complaint only existed with the door shut. Nobody wrote down which way the door was, so nobody noticed that all three visits had tested a room the customer never sleeps in.
This case is reconstructed backwards from the four service tickets, because the gap in the record is the finding.
The complaint
A two-story house, about 1,900 square feet, gas furnace in the basement, one central return grille in the upstairs hallway and one on the first floor. The back bedroom upstairs, over the garage end of the house, has one supply register and no return. The owner's teenager sleeps with the door shut. Complaint, first call, early December: "the back bedroom is freezing at night."
Ticket 1: the equipment
The tech checked what a no-heat-to-one-room call usually starts with. Filter loaded, replaced. Temperature rise across the furnace measured 55 F, inside the nameplate range. Burner and flame sensor normal. Thermostat in the upstairs hall reading within a degree of a separate thermometer. Supply plenum warm, the branch duct to the back bedroom warm at the boot. Note on the ticket: "equipment OK. Back bedroom 1-2 deg cooler than hall at time of visit. Recommend airflow balance."
What the ticket proves: the furnace makes heat and the branch delivers it. What it does not say: the time of day, or that the door was open during the visit. A room 1 to 2 F cooler than the hall at 2 in the afternoon, door open, is normal.
Ticket 2: the balancing
A week later a balancing tech measured every register with a flow hood. The back bedroom read 112 cfm against a design figure of 120 on the original duct layout, within 7 percent and better than several other rooms. He trimmed the dampers on two first-floor branches to push a little more air upstairs and re-read the back bedroom at 116 cfm. Ticket note: "balanced to design within 5%. Back bedroom 116 cfm."
The measurement was correct. Every flow hood reading on this ticket was taken door open, because a tech holding a flow hood over a ceiling register is standing in the room with the door behind him, and the ticket has no field for door position.
Ticket 3: the register
The owner called again. A third tech swapped the bedroom register for a higher-throw model, reasoning that the air might be short-circuiting along the ceiling, and offered a quote for a duct booster fan or a zoning damper. The ticket notes the owner's description now: "worst after the kid goes to bed." That was the first written clue, and it read as a time-of-day observation rather than a door observation.
Ticket 4: the reading that settled it
The fourth visit went to a senior tech, who asked one question on the phone before driving out: "Is the door open or shut when it's cold?" Shut. He arrived with a flow hood, a digital manometer and a plan to measure everything twice.
UPSTAIRS PLAN (blower on high, all readings)
+-----------------+-------------------------+
| BACK BEDROOM | HALLWAY |
| supply [S] | |
| door open: | central return [R] |
| 116 cfm | (only return upstairs) |
| door shut: | |
| 79 cfm | |
| | |
| room vs hall, | |
| door shut: | |
| +6.0 Pa | |
| no return | |
+-------| door |--+ |
| 3/4 in undercut |
+-----------------------------------+
bedroom wall shares one interior partition
with the hall; 2x4 at 24 in o.c.
Door open, blower on high: 116 cfm, the same as ticket 2. Door shut: 79 cfm, and the manometer, with its tube passed under the door, read the bedroom 6.0 pascals above the hallway. The undercut, about 3/4 inch across a 30 inch door, is roughly 22 square inches, nowhere near enough to pass 116 cfm without building pressure.
What the 6.0 pascals means. The supply duct pushes air into the room, the room has nowhere to send it but the crack under the door and the leaks in its own exterior walls and ceiling, so it pressurizes until the outflow equals what the duct can still deliver against that back pressure. The duct, pushing against a pressurized room, delivers less: 116 down to 79 is a loss of 37 cfm, 32 percent of the door-open flow. That is a third less heat, every hour of every night. The air that does arrive leaves partly through the exterior wall and ceiling leaks, taking its heat with it rather than circulating back through the house. See related: Return Air Starvation and the Closed Bedroom Door, which owns the mechanism.
The overnight test made it concrete. The tech left a small temperature logger on the nightstand for two nights, door shut as usual: the room bottomed at 62 F with the hall thermostat holding 68 F. With the door propped open on the third night, 66 F.
What every earlier ticket got right, and what it missed
None of the first three techs made a measurement error. The furnace was fine, the balance was within design, and the register swap did no harm. Each one solved the problem that existed in the room as tested. The miss was a field, not a skill: no ticket asked which way the door was, so no one tested the configuration in which the complaint happened. The owner's "worst after the kid goes to bed" was the door, written down as a clock.
The booster fan in ticket 3's quote would have made it worse. A fan in the branch pushes harder into a room that already has nowhere to send the air, raising the room pressure and pushing more of the extra heat out through the exterior leaks, while costing the owner a fan and its noise.
Where the call would have forked
The 6.0 pascal reading is what made a transfer grille the answer, and it is worth saying what a different reading would have meant, because this room had a second suspect. It sits over the garage end of the house, and a bedroom floor over an unheated garage loses heat downward through its floor and the rim joist above the garage wall, especially where the insulation under it has sagged or was never air sealed at the garage ceiling.
Had the door-shut reading come back at 1 pascal or less, and the door-open logger night still bottomed several degrees below the hall, the door would have been cleared and the floor would have become the suspect: an infrared scan of the bedroom floor and lower walls from inside on a cold night, a look at the garage ceiling for gaps and sagging batts, and a referral to an insulation contractor rather than more ductwork. The two causes can also stack, which is why the tech logged a door-open night before the fix and not only after it: 66 F with the door propped open showed the floor costing a degree or two at most, and the door costing the rest.
The fix: a transfer grille through the shared wall
The options were an offset transfer grille through the partition, a jump duct through the attic, a ducted return, or a bigger undercut. The tech chose the grille: the shared wall is an ordinary interior partition between two living spaces, not a fire separation or a garage wall, the bedroom has no combustion appliance, and the attic route meant a duct through insulation in an attic the owner had just had sealed. See related: Transfer Grilles, Jump Ducts and Door Undercuts, for that choice, and How to Add a Transfer Grille to a Closed Bedroom, for the work.
Sizing, on this house's numbers. The common starting point of about 1 square inch of net free area per cfm of supply, using the design flow of 120 cfm, asks for about 120 square inches. This partition is framed 2x4 at 24 inches on center, so one bay is about 22-1/2 by 3-1/2 inches, about 79 square inches of cross section. One bay carries 79, short of 120; two bays carry about 158. The grilles chosen publish 62 square inches of net free area each, so two pairs give 124 square inches, above the 120 target and below the 158 the bays allow.
After. Blower on high, door shut: bedroom 1.5 pascals above the hall, supply 108 cfm. Door open, 117 cfm. Shutting the door now costs 9 cfm, about 8 percent of the door-open flow, against 37 before. Two nights of logging, door shut: low of 66 F against the hall's 68 F, the same as the room had done with the door propped open before the fix.
What the shop added to its comfort call checklist
Three fields went onto the comfort-call ticket template, each one a direct answer to a gap in this record:
- Door position for every reading, written next to the number: "116 cfm, door open." A reading without a door position is half a reading.
- Room-to-hall pressure with the door shut and the blower on high, for any room with a supply and no return. Over about 3 pascals sends the call toward a return path before anyone touches the equipment.
- When and how the customer uses the room, in the customer's words, with one follow-up question: "door open or shut when it happens?" The answer reroutes the call before the first truck rolls.
The second field costs the tech about two minutes with a manometer that is already on the truck. Over a winter, the shop's comfort callbacks for single bedrooms became a short list instead of a recurring one, because the pressure reading sorts the room-pressure rooms from the equipment rooms on the first visit.
References
- See related: Return Air Starvation and the Closed Bedroom Door; Transfer Grilles, Jump Ducts and Door Undercuts
- See related: How to Add a Transfer Grille to a Closed Bedroom; How to Check a Closed Room for Pressurization
- See related: The House as One Pressure System; What a Pascal Is and Why House Pressures Are Tiny
- Grille manufacturers' published net free area data