My Candle Burns Perfectly During the First Test but the Flame Becomes Too Large Halfway Down the Jar — Why Does Wick Behaviour Change as the Candle Burns Lower?
शेयर करना
Is it the wick, the jar or the way it was burned? In our judgement, most common first: an untrimmed wick and long sessions, then a wick at the hot edge for the lower section (a pool deeper than about 1 cm), then heat building in the glass in the final third, then draughts or a drifted wick. The oil is rarely the lead suspect for a growing flame, though heavy oils can add to carbon.
How do I tell which it is? Light an unburned twin beside the half-burned candle on the same evening, both trimmed to the same length. If the half-burned one still burns taller at the same trim, depth is doing it. If the flames match once trimmed, the wick length was.
What does it cost, and when do I stop? Nothing new if you kept a twin (its material value is ₹239.00 in the worked example). If depth is confirmed, the likely purchase is one wick step down: ten Eco C1 for ₹59.00. Stop every burn, test or customer, with about 1 cm of wax left, and never judge glass heat by touch.
- Straight answer: why the wick changes
- Same wick, different candle: what changes with depth
- The depth-stage map: top, middle, final third
- Late-flame patterns and what each points to
- The same-evening depth comparison
- Why about 1 cm is the stop line
- What not to buy for a flame that grows
- The four products this page uses
- Map of this cluster: 3241–3260
- Frequently asked questions
Same wick, different candle: what changes with depth
A container candle is not the same object from top to bottom. On the first burn the flame stands at the rim, in open air, over a cold, full jar. Halfway down it stands several centimetres inside a glass wall that has been warming for hours of cumulative burning, over wax that may already be warm from the session before. The wick is the same length of the same braid. The conditions it burns in are not, and a flame responds to conditions.
1 · The glass above the pool starts to matter
Once the flame drops below the rim, the glass above it does two things. It shelters the flame from room air, so the flame stands straighter and often taller than it did in the open. And it absorbs radiant heat and passes some back to the wax near the wall. In the middle third that tends to melt the pool faster and wider; in the final third, with the flame close to the base and a drifted wick close to a wall, it is where glass heat builds. None of this is a jar fault; it is why a wick is proven by a burn-down, not a first pool (why the final third decides a container).
2 · The pool tends to get deeper
A melt pool about 0.5–1 cm deep is a commonly cited target at the top (a rule of thumb, not CSI data). Lower down, with warmer glass and warmer wax underneath, the same wick commonly melts deeper. A deeper pool is more liquid fuel within reach of the wick, and a wick with more fuel available tends to carry a taller flame. That is the single most common reason, in our judgement, that a wick which was right at the top reads as over-wicked lower down. Tracking depth against remaining height, burn by burn, is its own method: the depth-versus-height chart.
3 · The wick stands taller unless you trim it
In many candles a session consumes wax faster than it consumes the wick, so an untrimmed wick tends to stand a little taller at each lighting, and carbon can collect on the tip (mushrooming). A taller wick carries a taller flame. Here the cause is the burn habit, not the wick size. The Candle Wick Trimmer page's tip is "Trim your wick to 1/4 inch before every burn", about 6 mm. If your first test was trimmed and your later burns were not, you have compared two different wicks. The trimmed-versus-untrimmed twin proves which one you have.
4 · The session gets longer and the jar never fully cools
Burns late in a candle's life are often longer, because the pool takes a while to spread and the owner is used to the candle by then. A longer burn in an already warm jar compounds the first two changes. Two burns a day with little cooling in between compounds them again. The wick-length and heat effects are also why a short test and a long one can disagree; the short-versus-long cycle comparison owns that.
None of these four changes needs a faulty wick, jar or oil. What decides whether a candle crosses the line is how close the wick was to its hot edge at the top. A wick chosen because it gave the fastest, widest first pool is often already near that edge. Which is why "it burned perfectly on the first test" is less reassuring than it sounds.
The depth-stage map: top, middle, final third
The map at the top of the page is the whole idea on one drawing. Split the wax between your fill line and a stop line about 1 cm above the inside base into three equal bands, and mark them on the outside of a test jar with a pen or a strip of tape before the first burn. Then log the same few readings at each stage, so the lower sections are judged against the top rather than against memory. The table below turns the map into a log.
| Stage | What commonly changes | Record at this stage | Red flag (judgement) | Method post |
|---|---|---|---|---|
| Top third | First full pool forms; the flame is in open air; the jar starts cold | Time to a full pool (about 3–4 h is the rule of thumb; about 2.8 h by the rough 1 h per 2.5 cm of width for a 7 cm jar), depth at the end, flame height at a fixed trim on a cm card, exposed wick before lighting | Pool short of the edge (too little heat); or edge reached very fast with a deep pool (already near the hot edge) | 3258 |
| Middle third | Glass above the flame shelters and warms; the pool tends to deepen; an untrimmed wick lengthens; carbon can build | Flame height at the same trim, compared with the top-third baseline; pool depth; carbon cap; exposed wick before trimming | Flame clearly taller than baseline at the same trim; depth past about 1 cm | 3242, 3243, 3247 |
| Final third | The flame is nearest the base and walls; glass heat builds; a drifted wick nears a wall; the tab zone approaches | Glass reading with a non-contact thermometer at fixed spots and times (never touch); gap from flame to the nearest wall; soot band after wiping; flicker; any drowning near the tab | Flame touching or licking the glass; a wick leaning to a wall; persistent soot; a glass reading above the limit you set before testing | 3244, 3246, 3249 |
| Last ≈1 cm | Not burned: the stop line | The wax height at which you stopped; the state of the tab and wick | Any burn you had to stop early for safety fails the combination | 3260 |
Three measuring habits make the log worth keeping. A fixed trim. Trim to the same length before every burn and measure the exposed wick first, so a longer flame can't hide a longer wick. A fixed clock. Read the flame at the same times each burn (for example at 1 h and 3 h, a worked-example schedule) against a card marked in centimetres stood behind the jar, and photograph it. A fixed room. Same spot, level surface, door shut, fan and AC off. A ceiling fan that was off in the first test and on in the fifth is a different candle again.
On glass heat, the rule for this whole cluster is short: never judge glass heat by touch. A hand is not a thermometer, and the final third is exactly where a jar can be hot enough to hurt. A non-contact (infrared) thermometer read at the same spots and times is the tool. CSI does not list one; our Pen Thermometer is a probe for the wax pot, and its page states no range. Set your acceptance limit before the test, write it on the log, and compare wicks against it. The safety criteria that fail a wick in the final third belong to the hot-jar post, and the jar-or-wick question when the final burn runs hot to the very-hot final burn post.
Late-flame patterns and what each points to
The ranking below is editorial judgement, most common first; CSI publishes no failure statistics and no measured burn data. What makes it useful is the last two columns: the observation that rules each cause in, and the one that rules it out.
| # | Pattern observed | Likely cause | Rules it in | Rules it out |
|---|---|---|---|---|
| 1 | Flame taller in later burns; wick visibly longer before lighting; no trimming routine | Exposed wick length (burn habit) | Trimmed to the same length, the half-burned candle's flame matches a fresh twin | Still taller than the twin at the same trim |
| 2 | Flame taller at the same trim; pool well past about 1 cm deep in the middle third | Wick at the hot edge for the lower section | Depth on target at the top, rising lower down (log it burn by burn) | Depth stays about 0.5–1 cm all the way |
| 3 | Flame fine in the middle, taller in the final third; glass reads warmer than it did at the top | Heat building in the glass near the base | Non-contact readings rise stage by stage at the same spots and times | Readings flat while the flame grows (look at trim or draught) |
| 4 | Flame leans, flickers or dances; taller on one side; sometimes soot | Draught, fan or AC; or a wick that has drifted off-centre | Settles in a still room; the wick is visibly off-centre after cooling | Tall and steady in still air with a centred wick |
| 5 | A carbon cap that returns after trimming; soot; a heavy or dark oil at the top of its range | Carbon build-up adding to an already hot wick (wick or load) | An unscented control with the same wick carries less carbon | The unscented control mushrooms the same |
| 6 | Flame too large from the very first burn once you look closely | Over-wicked for this jar at every depth: not a depth problem | A fresh twin at the same trim is just as tall | The fresh twin is clearly smaller |
Patterns overlap: an untrimmed wick already near its hot edge produces patterns 1, 2 and 3 at once, tempting you to buy three fixes. Don't: remove the cheapest cause first (trim), then look again. Pattern 6 is the one to rule out early, because it changes the question from "what happens with depth" to "the bracket was wrong from the start", which calls for a fresh wick bracket from the top rather than anything in this cluster.
The same-evening depth comparison
You need to know whether the growth comes from depth or from something that also varies between evenings: trim, room, draught, session length. The cleanest control is a candle identical in every way except depth. An unburned twin from the same pour is exactly that. Light both on the same evening, trimmed to the same length, in the same still room, and read them against the same clock.
The set
Candle A: the candle whose flame grew, now at or below halfway. Candle B: an unburned twin from the same pour (same wax, oil, load, wick, jar and age). Worked example: 150 g in the 150 g jar at 8% of total weight, 12 g of Sandalwood and 138 g of CSI 464, on an Eco C2 wick, poured at 80–85°C (the jar page's 85°C governs). If you kept no twin, pour one with the same everything and let it cure at least as long as the oil page's stated cure (Sandalwood's page states 48-72 hours for soy); A will then be older than B, which is acceptable for a flame comparison, though not for judging scent.
| What you see | What it points to | Next step | Purchase |
|---|---|---|---|
| A's wick was long before trimming; once trimmed, A's flame matches B's | Exposed wick length: a burn habit | Confirm with the trimmed twin test; trim before every burn; tell customers | Nothing (clean scissors on a cooled wick) |
| A still taller than B at the same trim; A's pool well past about 1 cm | Wick at the hot edge for the lower section | Chart depth (3242), then run the current wick against one step down (3244) | Ten Eco C1 at ₹59.00 |
| A taller than B at the same trim; pool on target; A's glass reads warmer | Heat in the lower section of the jar | Apply the final-third criteria (3244); jar or wick (2744) | Nothing until the criteria have run |
| A leans or flickers, B steady | A drifted wick or a draught at A's spot | Swap positions on evening two; check centring (3250) | Nothing; centre better on the next pour |
| Both flames equally tall, and both taller than your top-third baseline | Not depth: over-wicked from the start, or the room | Re-bracket the jar from the top; check for draughts | Depends on the bracket |
| Both on target on both evenings | The growth came from that night: an untrimmed wick, a long session, a fan | Log sessions; write burn instructions (3288) | Nothing |
What it costs: if you kept a twin, nothing new. The two candles in the comparison carry ₹477.99 of material value in the worked example (one twin is ₹239.00: ₹65.14 of Sandalwood at the 50 g price, ₹70.02 of CSI 464, ₹94.40 of glass and ₹9.44 of wick). If the rule points to a wick step, two confirmation candles on C1 add ₹470.91 of materials (₹235.46 each), and C1 is ₹3.54 a wick cheaper than C2. Costs are worked examples and exclude labour, freight, packaging and GST.
Why about 1 cm is the stop line
Many makers stop every burn with about 1 cm of wax left. That is common practice, not a CSI figure and not a regulation, and it exists for reasons this cluster keeps meeting. Near the base the flame is at its closest to the glass floor, heat has had the whole candle's life to build, the wick is approaching its tab, and a thick glass base holds heat under the pool (thick bases and the bottom of the burn). Any drift of the wick is now close to a wall. It is the zone where the problems in this cluster show up, and one where a candle has little left to give.
For testing, the stop line is where the burn-down ends, not where the test gets skipped. A combination passes only if it reaches the stop line without failing any final-third check on the way down. If you had to stop early because the flame was touching the glass or the jar read above your limit, the combination failed, however good the first pool and the throw were. How many candles and which gates to use is the how-far-down post.
For customers, the stop line is an instruction you give, not one they know. A short care line ("burn within sight; trim the wick before each burn; stop when about 1 cm of wax remains") on the jar or in the box does two jobs: it sets the habit that prevents pattern 1, and it keeps the last centimetre out of use. Whether you rely on care instructions to accept a slower first pool is a business decision with its own risk weighting: the slow-first-pool decision.
What not to buy for a flame that grows
A candle that misbehaves late makes people shop for the wrong end of the problem. We would rather you ran the comparison first, even though several of these are our own products.
What each branch does need is short. Burn habit: a trimming routine and a care line. Hot edge in the lower section: ten wicks one step down, and two candles to confirm. Heat in the jar: the final-third criteria first, then a decision between a smaller wick and other options, which is the no-perfect-wick post. Drifted wick: better centring on the next pour, which the Wick Centering Device (Pack of 5) (₹118.00) helps with while the wax sets. Its page says it holds "wicks straight in the container while the wax sets"; it does nothing once the candle is burning.
The four products this page uses
Four items: the wick step the lower section most often needs, the jar the stage map is drawn from, the wax and the oil in the worked example. Chosen on page statements, per-unit price and stock on 26 September 2026; CSI has no measured burn data, so nothing is ranked on performance. The oil table shows every size, candles at 12 g and step-down pairs at 24 g.
| Size | Price | Per unit | What it covers here |
|---|---|---|---|
| Thin (C1) / 10 | ₹59.00 | ₹5.90 | the one-step-down candidate with duplicates, and eight spare |
| Thin (C1) / 100 | ₹590.00 | ₹5.90 | production, only after C1 has passed all three stages |
| Thick (C2) / 10 | ₹94.40 | ₹9.44 | the worked-example current wick; a step up if your test says too little heat |
| Thick (C2) / 100 | ₹944.00 | ₹9.44 | production, only after C2 has passed all three stages |
| Size | Price | Per unit | What it covers here |
|---|---|---|---|
| Pack of 10 | ₹944.00 | ₹94.40 | a fresh twin pair and a step-down pair, with spares |
| Pack of 20 | ₹1,888.00 | ₹94.40 | the same per jar; a range test across several oils |
| Size | Price | Per unit | What it covers here |
|---|---|---|---|
| 500 g | ₹260.00 | ₹0.52 per g | three 150 g candles at 8% (414 g of wax) |
| 1 kg | ₹507.40 | ₹0.51 per g | seven 150 g candles at 8% |
| 5 kg | ₹2,537.00 | ₹0.51 per g | the re-pour once the wick is chosen |
| 10 kg | ₹5,074.00 | ₹0.51 per g | production after a full burn-down pass |
| Size | Price | Per gram | 150 g candles at 8% | Step-down pairs (24 g) |
|---|---|---|---|---|
| 15 g | ₹93.22 | ₹6.21 | 1 | none (15 g of 24 g) |
| 50 g | ₹271.40 | ₹5.43 | 4 | 2 (+2 g) |
| 100 g | ₹502.68 | ₹5.03 | 8 | 4 (+4 g) |
| 500 g | ₹2,513.40 | ₹5.03 | 41 | 20 (+20 g) |
| 1 kg | ₹5,026.80 | ₹5.03 | 83 | 41 (+16 g) |
Map of this cluster: 3241–3260
This page owns the short answer, the stage map, the same-evening comparison and the stop line. Each of the other nineteen posts owns one test for one late-burn situation; pick the line that matches what you saw.
If your candle's problem is scent rather than flame (strong at the top, faint lower down), start at the cold-strong, hot-faint hub instead; if it is several problems at once, the multi-symptom hub later in this series orders them.
CandleMakingSuppliesIndia sells the wicks, wax, jars and oils named here, and does not sell finished candles. This page still tells you that the commonest cause of a growing flame costs nothing, that our trimmer is optional, that the Sandalwood 1 kg is no cheaper per gram than its 100 g, and that the jar's hot-pour rating says nothing about the final third.
Every price was verified on CSI's live store on 26 September 2026; every size in the pick tables is in stock on that date, and the tables flag any that are not. Jar dimensions, the hot-pour line, wax temperatures and load wording are quoted from the product pages; no CSI listing states an internal diameter, and where a page is silent in the text we captured, we say so.
The cause ranking, the stage red flags and the reading schedule are editorial judgement; the pool times and depths and the 1 cm stop line are commonly cited rules of thumb or common practice, not CSI data; CSI publishes no measured burn data. For bulk pricing, GST invoicing, and MSDS and IFRA documentation on request, message us on WhatsApp at +91 7397976926.
Frequently asked questions
- My Candle Has a Good Melt Pool for the First Few Burns but Becomes Extremely Deep Near the Bottom — Does This Mean I’m Over-Wicked?
- My Candle Flame Gets Larger as the Wax Level Falls — Is Heat Building Up Inside the Container and Changing the Burn?
- My Candle Jar Is Safe During the First Half but Becomes Extremely Hot Near the Bottom — Should I Use a Smaller Wick Even if That Causes a Slower Initial Melt Pool?
- My Candle Tunnels During Early Burns but Becomes Too Hot During Later Burns — How Do I Choose a Wick When One Size Doesn’t Seem Perfect Throughout the Jar?
- My Wick Keeps Drowning Only When the Candle Reaches the Bottom Third of the Jar — What Could Cause Late-Stage Wick Failure?
- My Candle Passed a Two-Hour Burn Test but Failed After Several Longer Burns — Why Isn’t a Single Short Burn Test Enough for Commercial Candles?
- How Far Down Should I Burn a Test Candle Before Deciding the Wax, Wick and Jar Combination Is Commercially Acceptable?
- What Problems Only Appear During the Second Half of a Candle’s Life and Would Be Missed if I Test Only the First Melt Pool?
Customer reviews: The six reviews at the top of this page are genuine, published Judge.me reviews left by CSI customers on the product pages named on each card — Clear Glass Jar with Golden Lid (150 gram), Dark Vanilla Fragrance, Eco Candle Wicks, Luxury Soy Wax CSI 464 (Flakes Form) For Container Candles, Sandalwood Fragrance Oil, Woody Oud Fragrance Oil. They are general product reviews, not assessments of the diagnostic advice on this page. Names, star ratings, dates and products are as recorded by Judge.me and wording is unedited. "Verified buyer" appears only on reviews Judge.me recorded as verified — Monika Deep, Soumili Gope, Thenirvaofficial, Reena Bagal, Anonymous, Payal Patel. Reviews below five stars are included as published, not filtered out.
Figures verified 26 September 2026, CSI live pricing: Eco Candle Wicks Thin (C1) 10 ₹59.00, 100 ₹590.00; Thick (C2) 10 ₹94.40, 100 ₹944.00; Clear Glass Jar with Golden Lid (150 gram) Pack of 10 ₹944.00, Pack of 20 ₹1,888.00; Luxury Soy Wax CSI 464 500 g ₹260.00, 1 kg ₹507.40, 5 kg ₹2,537.00, 10 kg ₹5,074.00; Sandalwood Fragrance Oil 15 g ₹93.22 (₹6.21/g), 50 g ₹271.40 (₹5.43/g), 100 g ₹502.68 (₹5.03/g), 500 g ₹2,513.40 (₹5.03/g), 1 kg ₹5,026.80 (₹5.03/g); Candle Wick Trimmer ₹588.82; Candle Wick Cutter ₹295.00 and ₹177.00; Wick Centering Device (Pack of 5) ₹118.00; Pen Thermometer For Candle Making ₹354.00. All in stock on 26 September 2026. Per gram = price ÷ grams; per wick or jar = pack price ÷ units; two decimals. Worked costs use the exact per-gram price, then round.
Non-price figures: load is a percentage of total candle weight (150 g at 8% = 12 g oil + 138 g wax), inside CSI 464's stated "up to 10%" and the Sandalwood page's "up to 10% for soy". Jar figures (height 8 cm, outer diameter 7 cm, base about 6.5 cm, hot pour up to 85°C) are page wording; no CSI listing states an internal diameter, and the circle areas are illustrations. Melt-pool times (about 3–4 hours; about an hour per 2.5 cm of width) and depth (about 0.5–1 cm), the 1 cm stop line and 3–4 hour sessions are commonly cited rules of thumb or common practice. The 1/4 inch trim is the trimmer page's tip. The cause ranking, red flags and reading times are editorial judgement. Costs exclude labour, freight, packaging and GST. Diwali 8 November 2026 is a calendar fact.