How to Survey Carbon Monoxide at Openings During a Generator Run

Why this matters

A standby unit that meets every clearance on paper can still push exhaust into a house. Wind wraps around corners, a fresh-air intake draws from the side yard, a crawlspace vent sits lower than anyone looked. Clearance rules are written for the average site and are owned by Exhaust Clearance and CO Safety Standard; this card is the field measurement that tells you whether this site behaves the way the rule assumed. It takes about an hour, it is the only evidence you will have if a CO alarm ever sounds in that house, and the one time it catches something it stops a poisoning rather than investigating one.

Before you start: the people, the alarms and your own meter

You are deliberately creating a CO source next to an occupied building. Before the unit starts:

• Tell the occupants what you are doing, keep children and pets indoors away from the side of the house you are working, and ask everyone to tell you at once if a CO alarm sounds or anyone reports headache, dizziness or nausea.
• Check the house CO alarms. Press test on each and note their locations and ages. A house with no working CO alarm on the floors nearest the unit gets that fixed before the survey runs, not after. A meter in your hand at one point does not protect a person asleep in another room.
• Wear a personal CO monitor, separate from the survey meter. You will stand near the exhaust plume at the outdoor points. If your personal monitor alarms, step out of the plume upwind and stay out until it clears. OSHA's permissible exposure limit for CO is 50 ppm as an 8-hour time-weighted average (29 CFR 1910.1000, Table Z-1); a personal monitor set to alarm well below that keeps you nowhere near it.

And if an indoor reading rises or an alarm sounds: shut the generator down from its controller, open doors and windows on the side away from the unit, get the occupants outside to fresh air, and if anyone has symptoms call 911. That response is Step 7's stop rule; it overrides everything else on this card.

The survey meter is a handheld CO meter reading in ppm with 1 ppm resolution, bump-tested with test gas on the schedule its maker sets. A meter that has not been bump-tested may read zero because it is working or because its sensor is dead, and the survey cannot tell you which.

The points, laid out

Walk the house wall facing the generator and the two walls around the corners from it. Every opening within reach of the exhaust gets a numbered point: operable windows, doors, crawlspace and soffit vents, dryer and bath exhaust terminations (dampers leak), and any mechanical fresh-air intake. Indoors, take the rooms nearest those openings.

 PLAN VIEW, not to scale        WIND noted at each round

          P0 upwind, clean air (meter zeroed here)

 +----------------- HOUSE -------------------+
 |  BEDROOM         LAUNDRY        HALL      |
 |   P5 indoor       P6 indoor               |
 +--[W]------[E]-------[V]----------[D]------+
    P1       P2        P3           P4
    window   fresh-air dryer        side
             intake    vent         door
                 ^
                 |  unit at the maker's setback
          +-------------+
          |  STANDBY    |  exhaust outlet -->
          |  GENERATOR  |  (end away from house)
          +-------------+

Numbering the points the same way every time, and drawing this sketch on the record, is what lets a second survey next year be compared with this one.

The survey, step by step

Step 1: Zero the meter in clean air

Before the generator runs, take the meter well upwind of the unit, away from vehicles, the street and any fuel-burning appliance vent, and zero it or run its fresh-air calibration (P0). Why: a meter zeroed in contaminated air subtracts a real reading from every point that follows. What breaks if skipped: a 3 ppm offset makes every indoor rise look 3 ppm smaller.

Step 2: Take indoor baselines

With the generator still off, read P5, P6 and any other indoor point. Record each. Many houses sit at 0 to 2 ppm indoors from cooking, a gas range, an attached garage. Why: the stop rule in Step 7 is a rise above each point's own baseline, not an absolute number. A gas range that was used an hour ago can leave a kitchen at 4 ppm with the generator off, and without the baseline you would blame the generator. Note what fuel-burning appliances are running in the house; a furnace or water heater cycling during the survey is a variable you want written down.

Step 3: Note the wind and decide whether today is a fair test

Record wind direction and an estimate of speed at the start, and again at each round. The worst case is wind blowing from the unit toward the house, or eddying around the corner nearest it. A survey taken with a steady breeze carrying exhaust away from the wall proves only that the wind was favorable today. Record it as such: "wind from house toward unit, survey does not represent adverse wind", and schedule a repeat on a day with wind toward the house, or calm air, which lets the plume pool.

Step 4: Run the unit at load

Start the generator and put it under a real load, either by transferring the house and running large electric loads (an electric range, dryer, water heater, air conditioner; a gas appliance fired indoors adds its own CO to the readings) or by connecting a load bank per How to Connect a Portable Load Bank to a Residential Standby. Why load: exhaust volume rises with load, and an idling engine at light load is not the condition the house will see in an outage. Run at least 30 minutes, which is this card's default; take readings at 10, 20 and 30 minutes so a slow build indoors shows as a trend.

Step 5: Read every outdoor point

At each round, hold the meter at the opening itself, at the height of the opening, for long enough for the reading to settle (most meters take under a minute). Do not stand in the exhaust plume to reach a point; approach from upwind, and if your personal monitor alarms, back out. Outdoor readings near the unit will be well above indoor ones; their job is to show you which openings the plume reaches, so you know which rooms to watch.

Step 6: Read every indoor point

After the outdoor round, go inside and read each indoor point, walking the same order each round. Ask the occupants whether any alarm has sounded. Indoor readings are the ones that decide the outcome.

Step 7: Apply the stop rule

This card's default stop rule: any indoor reading 2 ppm or more above that point's own baseline on two consecutive rounds, or any single indoor reading of 9 ppm or more, or any CO alarm sounding. The 9 ppm figure is borrowed from the EPA's 8-hour outdoor ambient air standard as a conservative trigger, not as an exposure limit for this situation. Tune it to your shop's policy, but write the number down before you start.

When it triggers: shut the unit down at once, ventilate as above, and keep reading indoors until each point is back at its baseline. Then the unit stays off in the controller's off position until the exhaust path is fixed, which means relocating the unit or changing its orientation per the maker's manual and the siting rules in How to Lay Out a Standby Generator Site for Clearances. Tell the customer in writing that the unit is disabled and why. A unit left in auto after a failed survey will start itself on the next outage or exercise.

Step 8: Record and restore

Write every reading against its point, round time and wind. Restore the house: retransfer to utility, return any loads you switched, disconnect any load bank, return the controller to auto only if the survey passed, and confirm the house CO alarms still read normal.

Worked example: the fresh-air intake nobody looked at

The site. A new 20 kW natural gas standby on the side yard of a two-story house, set at the maker's setback from the wall with the exhaust end facing away. The house has an energy recovery ventilator whose fresh-air intake hood (P2) sits on the same wall, a few feet higher than the unit. CO alarms on both floors tested good.

Zero and baselines. Meter zeroed upwind at P0 after a same-day bump test. Indoor baselines: bedroom P5 at 0 ppm, laundry P6 at 1 ppm. Furnace off; water heater idle.

Wind. From the street side, roughly toward the house wall, light and gusting. A fair test.

Load. House transferred; electric range, dryer and water heater on.

Readings are ppm. At 10 minutes laundry P6 read 2 ppm, 1 ppm over its 1 ppm baseline, under the 2 ppm rise. At 20 minutes P6 read 5 ppm, 4 ppm over baseline (5 minus 1). One round over the 2 ppm rise is not yet the two-round trigger, but the trend and the intake reading told the story: P2, the ERV intake, read the highest of any outdoor point and was feeding the laundry, where the ERV supply duct lands.

The tech did not wait for a third round. A rising indoor reading with the source identified is reason enough to act, and the card's rule sets a floor for stopping, not a ceiling. Generator shut down at 21 minutes, laundry window opened, ERV switched off. P6 read 3 ppm at 30 minutes and 1 ppm at 45 minutes, back to baseline.

Outcome. Controller left in off, customer told in writing. The shop relocated the unit to the opposite side of the house, away from the intake. Re-survey on a day with wind blowing from the new location toward the house, ERV switched back on and running, 30 minutes at load: no indoor point rose above baseline on any round, and the highest opening reading was 2 ppm. Controller returned to auto.

What the record now holds: two sketches, two tables, wind at each round, alarm locations, and the decision. If a CO alarm sounds in that house in five years, the first question will be what was measured at install, and the answer is on file.

References

• See related: Exhaust Clearance and CO Safety Standard
• See related: How to Lay Out a Standby Generator Site for Clearances
• See related: How to Connect a Portable Load Bank to a Residential Standby
• OSHA 29 CFR 1910.1000, Table Z-1, carbon monoxide permissible exposure limit
• The generator maker's installation manual (exhaust orientation and setback)