High CO at Low Fire After a Burner Linkage Replacement
A 700 hp, 150 psig gas-fired D-type boiler was retuned last week after the burner linkage was replaced. It now passes its annual efficiency test at high fire, but at low fire a portable analyzer reads 7.5% O2 and 1,100 ppm CO (dry), and the flame scanner signal swings between 4 and 9 VDC. At high fire the same analyzer reads 3.5% O2 and 18 ppm CO with a steady scanner. The combustion technician wants to "open the air damper at low fire" to bring O2 down to 3.5% everywhere. The boiler serves a process that spends about 60% of its operating hours at low fire, and the boiler-room CO monitor has alarmed twice in the past week. Stack temperature is 375°F at low fire and 455°F at high fire. There is no O2 trim; the flue-gas recirculation (FGR) damper is linked mechanically to the firing-rate cam. Production will not accept a shutdown longer than one shift.
1. Calculate the excess air at both firing rates and explain why the technician's instruction will make the problem worse.
Read the worked answer
Using the dry O2 approximation, EA% = O2 ÷ (20.9 − O2) × 100:
- Low fire: 7.5 ÷ 13.4 = 56% excess air
- High fire: 3.5 ÷ 17.4 = 20% excess air
So low fire already has nearly three times the theoretical air, and more than twice what high fire uses. At 56% excess air, a 1,100 ppm CO reading cannot be an air-deficiency problem: there is ample oxygen in the flue. The CO is being formed because the fuel and air are not mixing in the right place at the right temperature - the flame is quenching, cooling below the CO-to-CO2 completion temperature before the CO can burn out.
The technician's instruction is the reverse of what is needed. Opening the air damper further at low fire will:
- increase excess air toward 70–80%, lowering flame temperature further and driving CO up, not down;
- increase gas velocity through the register, which can lift the flame off the diffuser and worsen the scanner signal;
- raise NOx control risk in the other direction - FGR is doing the NOx work, and more air partially defeats it;
- increase stack loss at the rate where the boiler spends most of its hours.
The 4–9 VDC scanner swing confirms an unstable flame rather than a sensor fault - a healthy flame holds a steady signal. Note also that a 1,100 ppm CO reading makes the O2 reading suspect; a poisoned or water-loaded O2 cell should be checked with calibration gas before you tune to it.
2. List the most probable causes and the order in which you would check them during a one-shift shutdown.
Read the worked answer
Order by probability after a linkage replacement, then by cost to check:
- Verify the analyzer first (10 minutes). Zero and span with calibration gas, check the water trap and filter, confirm the probe is far enough into the flue. Tuning to a bad analyzer is the most common way to make a real problem worse.
- Firing-rate cam position at low fire. New linkage geometry changes the mechanical advantage. Confirm the low-fire position of every cam (gas valve, air damper, FGR damper) against the burner manufacturer's curve, and check whether the low-fire switch actually lands on the cam's low-fire notch or 10–15% above it.
- FGR position at low fire - most likely single cause. FGR linked directly to the firing cam often opens proportionally; at low fire the gas rate is small, so a proportional FGR flow is a much larger fraction of the combustion air and quenches the flame. Check whether the FGR has its own low-fire limit or a separate drive; the manufacturer's curve normally cuts FGR back near low fire.
- Burner gas pressure at low fire. Compare manifold pressure at low fire to the data plate. A regulator that overshoots at low flow gives a rich, lazy flame that quenches; an undersized low-fire orifice gives lift-off.
- Air damper/register and diffuser position. Confirm the low-fire air damper is set where the curve says, and that the gas gun/diffuser was not moved during the linkage work (a few millimetres of diffuser position changes low-fire stability dramatically).
- Furnace and flue condition. Check for a partially collapsed baffle or a leaking FGR duct joint that dumps flue gas into the burner inlet only when the wind direction changes - consistent with intermittent CO alarms.
3. Describe the test plan and the acceptance criteria you would require before the burner goes back to automatic.
Read the worked answer
Plan (one shift):
- Work in manual firing rate only, with a second person at the burner and a third watching the analyzer and CO monitor.
- Establish each firing rate in steps (low fire, 25%, 50%, 75%, 100%) and record O2, CO, CO2, stack temperature, scanner signal, manifold gas pressure, and FGR damper position at each point. Hold each point for at least 5 minutes so the readings stabilize.
- Adjust in this order: FGR at low fire first, then air damper/register, then gas cam, then re-verify high fire. Never adjust two variables at once.
- After tuning, run a full modulation sweep up and down and confirm no point goes out of limits and no point hunts.
Acceptance criteria (typical, confirm against the burner data sheet and local rules):
- CO below 100 ppm across the whole range, and no point where CO rises as excess air rises.
- O2 at high fire 3–5%; O2 at low fire deliberately higher (5–7%) - low-fire excess air is your flame-stability margin, not waste.
- Scanner signal steady, with no swing greater than roughly 1 VDC at any fixed rate.
- Stack temperature at high fire no higher than the pre-work baseline plus 20°F.
- Burner proven through a full trial-for-ignition, flame-failure trip, and low-water trip before being left in automatic.
- Log every reading, and leave the analyzer printout with the boiler log. Notify the insurer if the burner's rated input or the combustion settings changed.