Intermittent Flame-Proving Lockout on a Two-Wire Circuit
A 90% AFUE two-stage furnace with a hot surface igniter and a remote flame sensor has locked out three nights in a row. The technician watches the sequence: the inducer runs, the pressure switch proves, the igniter glows orange for 17 seconds, the gas valve opens, the flame lights and looks strong and blue across all burners, and then, three to six seconds later, the gas valve drops out before the blower ever starts. After three attempts the control enters a one-hour soft lockout and the LED flashes the failed-to-prove-flame code. The technician measures 0.4 microamps in series at the flame sensor with the burner box assembled. The sensor is clean. The furnace is fed by a 25-foot, 2-wire cable with no equipment ground to a junction box in the attic, then by 14/2 with ground to the panel. The junction box is plastic, the cabinet grounding lug has no conductor attached, and the circuit shares a neutral with a nearby attic light.
1. Explain how flame rectification works and why the microamp value matters more than the appearance of the flame.
Read the worked answer
When flame is present, the ionized gas between the sensor and the grounded burner conducts current asymmetrically, acting somewhat like a rectifier, so an AC voltage applied by the control produces a small DC current. The control interprets that DC microamp signal as proof of flame. Typical controls need at least about 0.5 microamps to hold, and healthy readings run roughly 1 to 6 microamps. A strong-looking blue flame can still fail, because the signal depends on how much of the sensor's surface sits in the primary combustion zone, the sensor's insulation, the flame's conductivity, and, critically, the integrity of the ground return path through the burner, the cabinet, and the equipment grounding conductor. At 0.4 microamps the reading is below the proving threshold and is marginal enough that it may prove briefly when the flame first engulfs the sensor and then drop out as the burner heats and the flame pattern shifts.
2. Based on the measurements, list the likely causes in the order you would investigate them and how you would prove each.
Read the worked answer
First, re-check the microamp reading with the burner box fully assembled and the unit in steady state, since flame current changes when the door is off. Second, verify the ground path, because the missing equipment grounding conductor is the standout defect: measure continuity from the furnace cabinet to the panel's ground bar and expect near zero ohms, then repair the circuit to provide a proper equipment ground. Third, check polarity at the junction box and the receptacle, since reversed line and neutral defeats rectification on many controls. Fourth, address the shared neutral, which can float the furnace's reference and produce erratic microamp readings. Fifth, verify the sensor is positioned in the flame and check its insulator for cracks and carbon tracking. Sixth, inspect the sensor terminal and harness at the control for corrosion or a loose pin. Any of these can drag a good flame signal below the proving threshold.
3. What repairs and documentation would you require, and would you consider bypassing the sensor to get heat on tonight?
Read the worked answer
No. Bypassing a flame sensor, jumping it, or substituting a simulator removes the device that shuts the gas valve off when there is no proven flame, and a delayed ignition or a failed ignition with the valve open is exactly the event that sensor exists to prevent. The correct repair is to restore a dedicated circuit with a proper equipment grounding conductor bonded to the furnace cabinet, correct any polarity or shared-neutral problem, and re-measure the flame current with the unit reassembled. Expect several microamps after the ground is repaired; if the reading stays marginal, replace the sensor with the OEM part and recheck. If heat cannot be restored safely the same night, say so and provide temporary heat rather than defeating a safety. Document the initial 0.4 microamp reading, the missing ground, the corrected wiring, and the final value, and note that the lockout was caused by the wiring defect, not by the control board.