Laboratory Offset Three Times Design After a Constant-Volume Exhaust Mis-set
A 1,000 ft² chemistry lab in a research building uses a 100% outdoor-air VAV system with airflow tracking: supply follows exhaust to hold a design offset of −200 cfm, about −0.05 in. w.c. relative to the corridor. Exhaust is a 6-ft fume hood (700 cfm design, 250 cfm low flow) plus a general exhaust box that was specified as a variable-volume tracker at 500 cfm design. During a night setback test the hood drops to low flow and the lab pressure falls to −0.28 in. w.c.; the corridor reads +0.06 in. w.c. relative to the stairwell. Staff report doors that are hard to open and that whistle, plus occasional solvent odour in the corridor. Spot measurements: supply 1,020 cfm, hood exhaust 250 cfm, general exhaust 1,390 cfm. The door is 36 in. × 80 in. with a 3/4 in. undercut and no transfer grille. The supply box is a pressure-independent venturi valve capped at 1,000 cfm.
1. From the spot measurements, quantify the actual offset and separate the airflow error from the pressure-path error.
Read the worked answer
Actual offset = supply − total exhaust = 1,020 − (250 + 1,390) = −620 cfm, more than three times the design 200 cfm.
Two independent faults combine. First, the general exhaust is behaving as a constant-volume box at 1,390 cfm instead of its 500 cfm design: it was evidently left wide open at commissioning and never re-balanced as a tracker. With the hood at low flow, total exhaust barely falls, while the supply valve is already at its 1,000 cfm cap, so the offset is set by the exhaust, not by the control sequence. Second, the transfer path cannot carry even the design offset. A 36 in. × 3/4 in. undercut is 27 in² ≈ 0.19 ft² gross; with a discharge coefficient near 0.6 the velocity at 200 cfm is 200/(0.19 × 0.6) ≈ 1,750 fpm, which requires roughly (1,750/4,005)² ≈ 0.19 in. w.c. Design was 0.05 in. w.c.
Note that the room reaches only −0.28 in. w.c. rather than the several inches the 620 cfm offset would imply through that slot. That tells you the envelope, ceiling penetrations and door frame are passing most of the air - and those paths are also how odour reaches the corridor. Fix the cfm and the path; one without the other will not hold.
2. Size a transfer path that satisfies the design offset at −0.05 in. w.c., and state why a deeper door undercut is not the right answer.
Read the worked answer
For 200 cfm at 0.05 in. w.c.: V = 4,005 × √0.05 = 896 fpm; required effective area = 200/896 = 0.223 ft²; gross area at Cd 0.6 = 0.37 ft² = 54 in². Spread across a 36 in. door that is a 1.5 in. undercut.
A 1.5 in. undercut in a laboratory corridor is unacceptable - it defeats the door's smoke and fire function, is a trip and privacy problem, and is impossible to pressure-test. Provide a dedicated path instead: a transfer grille or short transfer duct sized for the maximum offset plus margin, say 250–300 cfm at ≤0.05 in. w.c. That is roughly 0.45 ft² of free area, for example two 12 × 12 grilles at about 60% free area connected by a short duct, or a door louver with equivalent free area. Size for the worst-case offset (low-flow hood, unoccupied supply minimum), not the average. Verify by pressure test after installation: measure lab-to-corridor differential at both hood conditions and confirm it sits at the design value with the door closed.
3. Rewrite the control sequence and define the commissioning tests that would have caught this.
Read the worked answer
Make the general exhaust a true tracker: hold its 500 cfm design value at the occupied condition and reduce it with the hood so the total exhaust never exceeds what the supply valve can deliver. Set the supply valve range for the real maximum (hood design + general + offset) rather than the value inherited from a different hood, and match the response times - a fast hood valve against a slow supply valve produces exactly the transient spikes staff hear as door whistle.
Sequence: supply airflow setpoint = (hood exhaust + general exhaust) − 200 cfm, with the offset enforced by the supply valve; general exhaust = 500 cfm at occupied and reduced proportionally at low flow; an alarm on room differential pressure at ±0.10 in. w.c. from a sensor with 0.01 in. w.c. resolution.
Commissioning tests that would have caught it: (1) an offset verification at three hood conditions - design, mid and low flow - trending cfm and room pressure simultaneously; (2) a door-swing and smoke-pencil test at the door with the lab at design offset; (3) a cascade check of the whole wing (lab, corridor, stairwell) rather than the lab alone; (4) confirmation that the general exhaust box is actually in tracking mode, not locked at a fixed command. Trend the offset itself, not just the exhaust volumes, because the pressure that matters depends on the transfer path as much as on the cfm.