My humble Riello central heating system is malfunctioning. sometimes it locks down into a safe mode (i think) with a red light, turning it off and on again sometimes solves it (i.e. the boiler starts but it's a gamble if the flame is produced or not, the flame sensor shutting it down if there's no flame after 3 seconds of sparking) and while working, now it seems to lock itself after 20-30 minutes (it's very variable, sometimes it got to 1-2 hours of working) then it shuts itself off again. what might be the cause of this intermittent failure?

Nova Pro

Vote: control circuit

After reviewing the arguments from other models, I remain convinced that the control circuit is the most likely cause of the intermittent failure. The variability in the duration before the boiler locks down, along with the unpredictable ignition issues, strongly suggests a thermal-related failure within the control circuit. As the control circuit heats up, degraded components such as capacitors, relays, or solder joints could cause intermittent electrical failures, leading to the symptoms described. This aligns with the detailed explanations provided by GPT-5.4, Gemini 3.1 Pro, and Qwen 3.5 397B, which highlight the thermal instability and variable runtimes characteristic of a failing control circuit.

GPT-5.4

Vote: control circuit

I maintain vote D: control circuit. The strongest overall clue is the combination of two intermittent modes: unreliable ignition at startup, and separate lockout after 20–120 minutes of otherwise normal running. A single flame sensor fault explains failed flame proving, but it is a narrower fit for this whole pattern than a failing primary control/PCB. The other D arguments reinforced this, especially the thermal-failure point from Gemini 3.1 Pro and Qwen 3.5 397B. Heat-sensitive relays, solder joints, or board components can cause exactly this kind of variable runtime followed by lockout, then temporary recovery after reset/cooldown. Grok 4’s flame-sensor case is plausible, but less complete. A dirty/degrading sensor usually causes more consistent flame-proving trouble, especially at ignition. Here, the broad variability and the fact that different parts of the sequence seem affected still point more convincingly to the control electronics coordinating ignition and flame supervision. So my final case is that D best explains both the startup gamble and the later random shutdowns with one underlying intermittent fault.

Grok 4 (changed mind)

Vote: control circuit

The symptoms include both erratic ignition (gamble on flame after sparking, with flame sensor lockout) and successful runs followed by variable shutdowns after 20-120 minutes. This combination is not best explained by a consistently dirty flame sensor (A), which would typically cause repeatable startup failures rather than long successful operation followed by thermal-triggered lockouts. The control circuit (D) oversees ignition sequencing, flame proving, and safety shutdowns; a failing board with degraded capacitors, relays, or solder joints commonly exhibits exactly these heat-dependent intermittent faults—working when cold but failing as it warms. This matches the red-light safe mode and reset-by-power-cycle behavior. Exhaust (E), water/radiators (F), gas line (C), and spark (B) issues would produce more consistent or differently timed symptoms. The other models' arguments, especially on thermal expansion breaking connections in the PCB, were compelling enough to change my vote.

GLM 4.7 (changed mind)

Vote: control circuit

The symptoms described point strongly to a control circuit issue. The boiler exhibits two distinct problems: 1) intermittent ignition failures (sometimes flame is produced, sometimes not) and 2) variable runtime before shutdown (20-30 minutes to 1-2 hours). While a faulty flame sensor could explain the first symptom, it wouldn't adequately explain why the system can run successfully for extended periods before failing. The control circuit manages the entire operation sequence including ignition, flame detection, and safety limits. A failing PCB often exhibits thermal instability - working correctly when cold but malfunctioning as components heat up, which perfectly explains the variable runtime before lockout. The fact that the system can run for up to 2 hours before suddenly locking out is a classic sign of thermal failure within the control board, not just a sensor issue.

Gemini 3.1 Pro

Vote: control circuit

I am maintaining my vote for the control circuit (D). GPT-5.4 and Qwen 3.5 397B both provided excellent reasoning that reinforces this conclusion. Grok 4 argued for the flame sensor (A), but this overlooks a crucial detail in the prompt: the user notes it's a gamble 'if the flame is produced or not.' If the flame is genuinely not being produced, the flame sensor is doing exactly what it is designed to do—shutting the system down safely after 3 seconds. A faulty flame sensor would drop the signal even when a healthy flame is present, but it wouldn't cause the initial failure to produce a flame during the spark cycle. The control circuit (such as the primary control box, famously prone to failure on Riello burners) manages the entire sequence: the fuel solenoid, the ignition transformer, and the safety monitoring. A failing relay, degraded capacitor, or cracked solder joint on this board perfectly explains both the erratic ignition gamble at startup and the thermal cutouts after 20 to 120 minutes of operation. Heat builds up in the control box over time, breaking the electrical connection, which forces the hard lockout (red light). Therefore, the control circuit is the only option that accounts for all the intermittent symptoms described.

Qwen 3.5 397B

Vote: control circuit

While Grok 4 raises a valid point about the error message citing the flame sensor, that message indicates the symptom (loss of flame signal) rather than the root cause. A dirty sensor typically causes consistent startup failures, not variable runtimes of up to two hours. The thermal failure pattern described—working when cold but locking out after 20–120 minutes—is hallmark behavior of failing electronics within the control circuit, such as degrading capacitors or cracked solder joints expanding under heat. GPT-5.4 and Gemini 3.1 Pro reinforced this by highlighting how a single sensor fault rarely explains both erratic ignition and delayed thermal lockouts. The control board manages both phases, making it the only component capable of producing this specific combination of intermittent failures. Therefore, D remains the most comprehensive diagnosis.

Final Summary

The AI Roundtable reached a unanimous consensus that a faulty control circuit was the primary culprit behind the heating system's erratic behavior. While Grok 4 initially argued for the flame sensor, the collective logic of Gemini 3.1 Pro and GPT-5.4—highlighting how thermal expansion in the control board causes intermittent failures only after long runtimes—convinced the entire panel to settle on Option D.

All 6 models agreed on "control circuit" after discussion

Strongest Arguments

• flame sensor: The specific 3-second lockout after sparking directly indicates the system is failing to prove the presence of a flame, which is the primary job of the flame sensor.
• control circuit: Thermal failure within the control board (such as degraded capacitors or cracked solder joints) is the only fault that explains both the erratic startup 'gamble' and the variable shutdowns after 20-120 minutes of operation.

Key Turning Points

• Gemini 3.1 Pro and Qwen 3.5 397B introduced the 'thermal failure' theory, explaining how heat builds up over 20-120 minutes to break electrical connections.
• Grok 4 abandoned the flame sensor position after realizing a dirty sensor would cause consistent startup failures rather than successful two-hour runs.