The protocol
Updated
The method is the same everywhere and the discipline is in doing it the whole way down rather than for one evening.
Set up
The exact production candle: same wax, same fragrance at the same load, same dye, same wick, same vessel, cured for the same time.
A draft-free room at normal room temperature, on a heat-resistant surface, away from anything flammable.
Wick trimmed to about 6 mm before the first light, and to the same length before each subsequent burn.
Never unattended. This is the part that makes testing slow, and there is no way around it.
The burn cycle
Burn in cycles of roughly four hours, extinguish, let it cool fully, trim, relight.
Four hours matters: it is about how long a customer burns a candle, and it is long enough for a full melt pool to form in most container sizes.
Repeat until the candle is finished. A candle that behaves perfectly for two cycles can misbehave at the bottom, where the wick is closest to the base and the glass is hottest.
What to record each hour
Melt pool diameter, and whether it has reached the vessel wall.
Melt pool depth.
Flame height, and whether it is steady or flickering.
Any mushrooming, soot, or carbon on the wick.
The temperature of the vessel wall and base, by touch or by probe, which is what tells you whether the glass is being asked for more than it can take.
Anything unusual: a flame that dances with no draft, a wick that leans, smoke.
Through to the end
The last centimeter is the whole reason for the test. End of useful life behavior is one of the things F2417 is about, and it is where a wick that was fine all the way down gets too close to the base.
Note where you stopped and why. "Self-extinguished with 10 mm of wax remaining" is a result. "Burned fine" is not a record.