My Wick Keeps Drowning Only When the Candle Reaches the Bottom Third of the Jar — What Could Cause Late-Stage Wick Failure?

Late-drowning causes, ranked

"The bottom third" of a 150 g jar is a short distance. The jar's page lists a height of 8 cm, so its bottom third is at most about 27 mm, and the wax column is shorter than the glass. Within that span a few millimetres decide the diagnosis, which is why this page measures in millimetres. The ranking below is our judgement of how often each cause turns up, most common first; CSI publishes no failure statistics.

Rows 1 and 2 are the fixed-height causes that live at the tab; rows 4 to 6 are fixed-height causes that live above it; row 3 is the only one that moves with how you burn. A candle can have two at once: a lifted tab and a deepening pool will drown early on long burns and still drown on short ones, just later. Read the two measurements together, as the decision table below does. For the first-level answer to drowning in heavily fragranced soy, see why wicks drown in highly fragranced candles.

The tab-zone check, with a ruler

The tab zone is everything at the bottom of the wick that is not free wick: the metal tab disc, the crimped neck (often called the sustainer), any sticker or glue under it, and any gap if the tab lifted off the floor. CSI's wick pages state no tab or neck height in the text we captured, and the Braided Wick Roll describes its tabs only as "small metal tabs that are attached to one end of the wick", so measure your own.

In the diagram's worked example, candle 1 drowns at 8 mm with a tab-zone top of 7 mm, inside the zone and below the 1 cm stop: nothing to fix. Candle 2's tab lifted 7 mm during pouring, so its zone top sits at 14 mm and it drowns at 15 mm: a setting fault, fixed with adhesion and a holder, not a wick. Candle 3 drowns 17 mm above its tab: something else is going on. The figures are illustrative; your ruler supplies yours.

The burn-length pair: pool or fixed height?

When onset sits well above the tab zone, one comparison splits the cause table in two. Take two twins from the same pour to the start of the bottom third on your usual schedule. From there, twin L keeps your usual burn length (up to about four hours, a common-practice cap) and twin S switches to shorter burns (worked example: two hours), which tend to keep the pool shallower. Trim both before every burn, burn them in the same still room on alternate sessions, and record onset height for each. Why two-hour burns are a poor approval test is the long-burn post's argument; here the short burns are a probe, not a pass.

If you are pouring fresh twins, the worked example is two 150 g candles at 8% of total weight in CSI 464: 24 g of Roasted Coffee and 276 g of wax (12 g + 138 g each; 8% of total is 8.70% of wax weight), inside the oil page's 6–10% and CSI 464's "up to 10%" even on the conservative reading of an unstated basis (9.09% of total). Fix each tab with a sticker and hold the wick upright while the wax sets. Pour at 80–85°C in the 150 g jar (CSI 464 says 80–90°C and the jar page "Safe for hot pour up to 85°C"; the lower limit governs). The oil page says to cure soy candles at least 48–72 hours before burn testing; we would start at day 7 (judgement).

Costs exclude labour, freight, packaging and GST. With two unburned twins of your own, the pair costs nothing new. Reaching the bottom third takes most of each candle's burns, so start soon; the tab-zone check on the candle that already drowned can be done today.

From the two measurements to a decision

Only two rows can end in a purchase, and neither is a new oil. The tab-setting row costs a sheet of stickers. The pool row may end in a wick change, but only after the depth chart has said whether the lower section is over- or under-served, and after the candidate has been burned to about 1 cm; the depth-gates post sets how many candles and checkpoints that takes. If your jar has a thick base and you suspect it changes the heat at the bottom, the theory is the thick-base post.

What to buy for each answer

Four items, in the order the decision table might call for them: holding and fixing the tab, the wick rungs for the pool branch, and the worked-example oil. Chosen on page statements, price per unit and stock on 26 September 2026; CSI has no measured draw or burn data, so nothing is ranked on performance. The oil table shows every size, 150 g candles at 12 g, and 24 g pairs.

Frequently asked questions