My Soy Candle Flame Looks Normal Immediately After Lighting but Keeps Getting Larger as the Melt Pool Deepens — What Should I Investigate?

Why a deeper pool makes a bigger flame

The wick is a pump. It draws melted wax up by capillary action, and the flame burns the vapour at the top. At lighting there is no pool: the flame melts its own fuel from the solid wax around the wick, a little at a time, so it is small. Once the pool reaches the glass it can only go down, and that is where your candle changes character.

Three things happen together as the pool deepens, and each one makes the flame taller. More hot liquid fuel. A deep pool is a reservoir of wax already at melting temperature; the wick never runs short and draws freely. A longer effective wick. If the wax level falls faster than the wick burns away, more wick stands above the pool every half hour. You trimmed it to 5–6 mm at lighting, but by hour three it may be standing much taller than that without your touching it. More heat around the flame. The glass above the pool and the wax below it warm up, so the fuel arrives hotter. These are common observations, not CSI measurements.

A matched candle reaches a balance: the pool reaches the glass, settles at a steady depth, and the flame stops climbing. The common rules of thumb are a full pool in about 3–4 hours on the first burn and a depth of about 0.5–1 cm. Many makers treat anything past about 1–1.5 cm as too deep. Those are rules of thumb, not standards, and they shift with jar width (about one hour per 2.5 cm of diameter for the pool to reach the glass: roughly 2.8 h for a 7 cm jar, 3.1 h for 8 cm, 3.5 h for 9 cm). Your candle has no balance point. The pool keeps deepening and the flame keeps climbing with it.

Soy matters here in one practical way. A softer, lower-melting blend commonly pools deeper on the same wick than a harder wax (a tendency). Premium Coconut Soy Wax CSI 468's page states "Melting Point 50°C – 53°C". So if you moved to a coconut-soy blend, or your supplier changed lot, the same wick can now overpower the same jar. That is a formula change, and it gets a wick retest, not a guess. If the deep pool only appears in the lower half of the jar after several good burns, that is late-burn heat, a different problem: very deep pool near the bottom and the deep-near-the-bottom post cover it.

The skewer log: measuring depth every 30 minutes

A labelled method, ours and many makers': simple and good enough to separate a pool that levels off from one that does not. You need a thin, straight stick, a millimetre ruler, a pen and a log sheet.

Here is what a log looks like when the pool is the problem. The numbers are a worked example, not CSI data, and they are the same numbers drawn in the diagram above.

Read the last hour, not the first. Between hour three and hour four the current wick added 0.6 cm of depth and the flame kept climbing. Over the whole session it grew 1.7 cm. One rung down added 0.1 cm in that hour and its flame grew 0.4 cm all session. That is the plateau you are looking for. A useful in-house rule (labelled, yours to set) is that a matched wick gains no more than about 0.2–0.3 cm of depth between hour three and hour four. Weighing agreed: about 9 g an hour against 6 g, roughly 22.2 against 33.3 hours for a 200 g candle if all of it burned.

Pool, mushroom, draught or room: the cause ranking

The ranking is our judgement for exactly your pattern (normal at lighting, growing as the pool deepens), and it is also roughly the order in which the log rules things in or out. Signs are common observations, not CSI burn data.

Cause 2 deserves a moment, because it is where most people stop too early. A wick that stands taller at hour three than at lighting has not "grown". The wax around it has gone down faster than the flame has consumed the wick. Trimming it mid-burn will shrink the flame for a while, and that is exactly why trimming can't be the fix: it hides the mismatch rather than ending it. The smokes-when-long post weighs whether that is over-wicking or customer maintenance. For this pattern, the skewer log usually answers it: a wick that creeps taller while the pool deepens is the same story as cause 1.

Cause 5 is easy to miss in India. A candle tested on a cool December evening and burned on a hot May afternoon can behave like two different candles. Ceiling fans belong to cause 4, and the flicker-or-airflow post has the free check. Fragrance can also shift the depth curve, and the effect is a tendency, not CSI data: heavier oils and higher loads commonly change how a wick burns, in either direction (bigger flame after lowering the load). Run the log first; the oil comes after.

Reading the log: one rung down, or a burn-time limit?

Every soy pool deepens if you burn it long enough. What decides the fix is the time at which your current wick's pool crosses the depth limit you wrote down before the test. Many care labels advise burning for no more than about four hours at a time (common guidance, not a regulation), so four hours is the natural yardstick.

The third row is the honest version of "reduce burn session length". A four-hour limit is fair when the candle genuinely behaves at four hours from the first burn to the last, and you have tested it to that limit, including the bottom third, where heat builds. It is not fair when it's being used to hide a wick that misbehaves at three. The four-hours-or-longer post works through the label policy in detail. Don't set a two-hour limit to rescue a strong wick, either. Two hours is often too short for the pool to reach the glass, so you swap a deep pool for a tunnel.

If you are at the second row and the one-rung-down candle looks too small, check its pool width as well as its depth before you reject it. A smaller wick that reaches the glass by about hour three to four from burn two, with a calm flame, is usually the better commercial candle even if it feels less dramatic. If it never reaches the glass, see the tall-flame, short-pool post. And if your pool reaches the glass in well under an hour, read the under-an-hour post before you decide, because speed alone proves less than it seems. The full-life method behind all this is why the whole candle is tested.

The six-candle test and the approval rule

This test is deliberately small, because your question is narrow: does one rung down level the pool where the current wick does not? It uses the series' default test candle, labelled: 200 g at 8% = 16 g fragrance + 184 g wax (load as a percentage of total candle weight), poured in your jar. If your candle has another fill, keep your fill. For the wider three-candidate ladder (including a pair of smaller wicks), use the hub's nine-candle plan.

Worked example for the calendar: at 6 g an hour, a 200 g candle is about 33.3 hours, roughly 9 four-hour sessions. With a 3-day cure that is about 12 days before you know, burning daily. Don't stop at burn 2: a wick that levels the pool at the top can still over-deepen it at the bottom; the 1 cm-at-four-hours post shows how to read heat against depth.

What to buy, and what to skip

CSI's sized, tabbed wicks are three rungs: C1 → C2 → Cotton Braided. One rung down is Cotton Braided → C2, C2 → C1, or a C2 pair → a C1 pair. There is no sized, tabbed wick between C1 and C2 or between C2 and Cotton Braided, no smaller sized, tabbed wick than C1, and one wooden-wick listing with no widths. The untabbed options (Eco Braided Wicks by the metre, the Braided Wick Roll with 100 tabs, and the Eco Candle Wicks Loose, whose page says "designed specifically for 2 inch diameters") are cut-to-length constructions, not half-sizes. If C2 over-deepens your pool and C1 tunnels, we would rather tell you that plainly than sell you a third pack: ask on WhatsApp, try a pair of smaller wicks through the hub's ladder, or adjust the formula on the fragrance side.

Argued against our own interest: you may not need a wick at all. If your log shows the pool crossing your limit only after four hours, the fix is a printed burn limit and a full-life test to it, and the only thing you buy is nothing. If it shows a cap or a flicker, a new wick is the wrong purchase. When you do step down, the saving is small (C2 → C1 is ₹3.54 a candle), and it is never the reason to choose a wick.

After approval, buy by the month. Production wicks = candles × wicks per candle × (1 + a defect allowance). Worked example, labelled: 150 single-C1 candles a month plus 2% is 153 wicks; one 100-pack and three 20-packs give 160 for ₹944.00, the same ₹5.90 a wick as any bigger pack. The hub works the reorder point through (6901). A later change of wick lot or supplier is a formula change: compare wick lots before you trust a new delivery.

Five items, chosen on page statements, price and stock on 7 October 2026. CSI publishes no burn, depth or flame data, so nothing here is ranked on performance, and each card says when not to buy.

Skip for now: any pack over 20 before three candles pass to the end, and a new wax while the wick question is open.

Frequently asked questions