What it is

The flame sensor is a single metal rod sitting in the burner flame, wired back to the control board. Its only job is to prove to the board that there's actually a flame burning after the gas valve opens. No proof, the board closes the valve fast — usually within a few seconds — because the alternative is dumping unburned gas into a hot box. That's a safety, not a nuisance.

If a furnace lights and then drops the flame after about 4–10 seconds, over and over, flame sensing is the first suspect — not the gas valve, not the igniter.

How it works

This is flame rectification, and the physics matters because it explains every failure mode.

The board pushes an AC voltage onto the flame rod. A flame is electrically conductive (it's full of ions), so current flows from the rod, through the flame, to a ground reference — usually the grounded burner assembly. Here's the key part: the flame acts like a one-way valve (a rectifier). Because the rod is small and the burner ground is large, current flows much more easily in one direction. The AC gets "rectified" into a small pulsing DC current, and the board specifically looks for that tiny DC microamp signal.

Why it's done that way: a simple short or a wet sensor would pass AC. Only a real flame rectifies AC into DC. So the board can tell the difference between "there's a flame" and "something is leaking current to ground." That's why the signal is measured in microamps DC.

Two things have to be true for a good signal:

  1. The rod has to be clean enough to emit current (oxide buildup insulates it).
  2. The flame has to have a solid path to a good ground. A small rod and a big grounded burner = strong rectification.

In the field

To read it: Put your meter in microamps DC (µA). Break the flame-sensor circuit and put the meter in series — the simplest way is a flame-sense test lead/adapter, or pull the sensor wire and meter between the wire and the rod terminal. With the burners lit and flame on the rod, read the current.

Decision tree:

  • Reading is solid (well above the board's dropout) → sensing is fine, look elsewhere.
  • Reading is low but present → clean the rod first. Light abrasive — fine steel wool, a Scotch-Brite pad, or fine emery cloth. Do NOT use coarse sandpaper; you want to remove the oxide film, not gouge the metal (a rough surface actually collects more crud). Wipe it clean, reinstall, re-read.
  • Reading is still low after cleaning → suspect the ground path or the rod itself. Check: is the burner assembly grounded? Is the sensor's porcelain cracked (leaking current to the bracket before it ever reaches the flame)? Is the rod positioned properly in the flame envelope? A rod that's barely in the flame, or in a lifting/lazy flame, reads low even when clean.
  • Zero microamps with a visible flame on the rod → broken rod, broken wire, no ground, or cracked insulator shorting to the bracket.

Normal values & targets

  • Healthy flame-sense current: roughly 1–6 µA DC on most residential furnaces (some run a bit higher). What matters is headroom over the board's dropout, not a magic number.
  • Board dropout threshold: many boards drop the flame somewhere around 0.5–1.0 µA. If you're sitting at 1.2 µA, you're "working" but right on the edge — that furnace will nuisance-trip on a cold morning when conditions sag.
  • Rule of thumb: if you read under ~2 µA, clean it and aim to get well clear of dropout. A freshly cleaned good rod with a good ground usually jumps back up to several microamps.

Common faults & what they mean

  • Lights, runs 4–10 sec, drops out, retries, locks out after 3 — classic dirty flame sensor or weak ground. Clean first.
  • Reading collapses as the furnace warms up — could be a cracked insulator that conducts more when hot, or a marginal ground that changes with thermal expansion.
  • Cleaning helps for a few weeks, then back to dropping out — often the burner flame quality is the real problem (dirty burners, low gas pressure, lifting flame) so the rod fouls fast or sits in a poor flame. Fix combustion, not just the rod.
  • Brand-new sensor, still low — you replaced a part that wasn't the problem. Go back to the ground path and flame position.
  • Polarity / reversed line wiring — flame rectification needs a proper neutral/ground reference. Reversed hot-neutral at the furnace can kill or weaken the signal even with a perfect rod. Check incoming polarity if everything else looks right.

Tech tips & gotchas

  • Cleaning the flame sensor is the single most common gas-furnace repair there is. On any no-heat where the furnace lights and drops out, pull and clean the rod before condemning a board or valve.
  • Don't over-clean. Smooth, shiny, light passes. A gouged rod fouls faster.
  • The flame rod and the hot-surface igniter are two different parts even though both sit near the burners. Don't confuse them.
  • A flame sensor reads current through the flame to ground — so a bad chassis/burner ground mimics a bad sensor exactly. Verify ground before you keep swapping rods.
  • Always re-read microamps after the repair and write the number down. "Cleaned the sensor, came back at 4.8 µA from 0.9 µA" is a real before/after that proves the fix and protects you on a callback.

Safety / code notes

  • The flame-proving circuit is a primary safety. It exists to stop the furnace from flowing gas without a flame present. Never bypass it, jumper the sensor input, or trick the board into thinking it sees flame.
  • If the furnace keeps proving flame poorly because of bad combustion (sooting burners, lifting flame, low manifold pressure), that's a combustion problem — verify burner condition and check for CO with an analyzer rather than just chasing the µA reading.