My 7 cm Jar Burns Perfectly with Unscented Soy Wax but Tunnels After I Add 10% Fragrance Oil — Why Did the Fragrance Change the Wick Requirement?

What the unscented candle did and did not prove

Burning an unscented candle first is a good habit, and yours has already told you useful things. Your jar holds the wax without trouble; your pour temperature and cooling give a surface the wick can start in; your wick sits centred; your room is still enough; and a wick of that size can pull a pool to the glass in plain CSI 464 in a 7 cm jar. Those facts carry over to the scented candle.

What does not carry over is the balance between the flame and the fuel. In an unscented candle the wick draws only melted wax. In a 10% candle it draws a mixture in which one part in ten, by weight, is fragrance oil: a blend of aromatic materials with its own viscosity, volatility and burning behaviour. A wick that sits comfortably in the middle of its range in plain wax can end up near the bottom of it once the fuel changes. That is the whole mechanism in one sentence, and it is why candle makers bracket in the finished formula.

So keep the unscented candle in your plan, just in the right job. In the re-bracket below it is the control: it has to reproduce "burns perfectly" in the same batch, same room, same week. If it doesn't, the problem is not the oil, and the whole test has told you something cheaper.

How fragrance oil changes a burning candle (tendencies)

Nobody can tell you exactly how your oil will change your wick: CSI publishes no wick-performance data for any oil, and fragrance oils vary widely in composition. What experienced makers report are patterns. Use them to form a hypothesis, then let the six candles decide.

Two cautions. First, these are directions, not sizes: none of it tells you whether C2 is the answer for your oil. Second, the effect of load on flame size is its own question. If your flame visibly shrank as you went from 6% to 10%, the load ladder in the 6%-to-10% flame post is the test for that; this page is about the step from zero to 10%. And if you are changing load across a whole range of scents, see requalifying a collection at 10%.

Two ceilings: is 10% inside CSI 464?

Two pages set the limit, and the lower one governs. The wax: Luxury Soy Wax CSI 464 states "Fragrance Load Up to 10%". The oil: our worked example, Jasmine Fragrance Oil, states "Recommended load up to 10% soy, up to 6% paraffin". Both say 10%, so on the face of it 10% is the ceiling and you are standing on it. These are formulation limits the pages state, not regulatory limits.

The catch is the basis. CSI's convention, and the one used in every figure on this page, is load as a percentage of the total candle weight: a 150 g candle at 10% is 15 g of oil and 135 g of wax. The CSI 464 page does not say whether its 10% means of total or of wax. Read it as of wax weight and your 10%-of-total candle is actually 11.11% of wax, over the line. The full method for converting between the two is the load-basis post; here are the three numbers you need.

Read the other way, 10% of wax weight is only 9.09% of the total. So the conservative top rung in CSI 464 is 9% of total, which satisfies both readings. Whether you stay at 10% or step to 9% is your product decision; the rule for this test is simply bracket at the load you will sell, because the wick answer at 9% and at 10% need not be the same.

If your wax is not CSI 464, use its own page. Luxury Soy Wax Chunks state "up to 13% fragrance load capacity", and its FAQ phrases it as "13% of total wax weight", which is ambiguous in the same way. A higher wax ceiling does not lift the oil's: Jasmine's page still says up to 10% in soy.

Rule out the cheap confounders first

Your unscented and scented candles were probably not twins. Before you buy a wick pack, check whether anything besides the oil differed. Each item below is free to check, and each one alone can produce a first-burn tunnel.

If any of these differed, fix it and burn a new pair (one unscented, one 10%, same wick) before anything else. That costs two candles, not six, and it is the right purchase more often than a new wick pack.

The scented re-bracket: six candles, three cells

The design is deliberately small. Cell 1 re-runs your unscented success as a control. Cell 2 re-runs the problem: same wick, 10% oil. Cell 3 changes one thing, the wick, by one rung. Two candles a cell is the minimum that separates a pattern from one odd candle; add a third per cell if you can.

Which wick is "one rung up" at CSI

CSI's ready-made rungs are Eco C1 and Eco C2 (both pre-tabbed) and Cotton Braided Wicks Ready Made. If your unscented wick was C1, wick B is C2, and that is the design this page costs. If it was another supplier's wick, make it wick A and ask us where C1 and C2 sit relative to it; we cannot map other brands' sizes for you with confidence. If it was already C2, the honest answer is awkward: the next CSI rung, Cotton Braided, is described by its own page as "engineered for large-diameter candles exceeding 8 cm", and a 7 cm jar (6.3 cm inside on our reading) is below that. You can still test it as a clearly labelled arm, but its page does not describe it for your jar. Before that, check the confounders and consider whether 9% is the product; or ask us on WhatsApp. Whether two rungs are enough to close a bracket is the adjacent-sizes post.

The bill of materials

Per candle: the unscented control is ₹176.41, a 10% candle on C1 is ₹233.30 and on C2 ₹236.84; the wick difference is ₹3.54. From nothing, the cash outlay is ₹2,093.80; if you own wax, oil and jars already, it is ₹212.40 for the two wick packs (or ₹94.40 if you already have C1). Costs are worked examples and exclude labour, freight, packaging and GST. The jar is the largest line: ₹94.40 each. Re-use nothing from a failed candle; a jar that has held a test burn is not a new jar.

Protocol

Pour all six on the same day from one melt: weigh the unscented wax for cell 1 first, then add oil to the rest at the same temperature and stir the same time. Label bases 1a/1b, 2a/2b, 3a/3b. Cure all six the same number of days (Jasmine's page: 48–72 hours for soy; we would wait about a week). Burn in one still room, sessions of about three hours with full cool-down, trim to about 5–6 mm each time, and log at 1, 2 and 3 hours: pool width (or the ring left on the glass in millimetres), depth near the wall, trimmed flame height, soot, mushrooming, and flame distance from the glass. Watch the glass; never touch it to judge heat. Run at least three burns before deciding, and take the survivor through its full life, stopping about 1 cm from the base.

Reading the six candles

The thresholds are ours, chosen because they can be read off a photo: the flame range many makers use is about 2.5–5 cm, and a pool about 0.5–1 cm deep is a common target. A clean pass in burns 1–3 clears the top of the candle only. In a jar about 8 cm tall the lower third arrives within a few sessions, and a wick chosen to fix an early tunnel is the kind that can run hot there. That is why cell 3 is a candidate, not an answer, until it passes the full life; the calendar for that is the full-life post, and the acceptance gates by depth are in how far down to burn a test candle.

Read cell 2 in millimetres, not adjectives. On a 6.3 cm inside, a ring of solid wax only 4 mm wide all round leaves 24% of the surface unmelted: a thin line on the glass, and nearly a quarter of the candle.

For the geometry of your jar, measure the inside at the rim, the middle and about 1 cm above the base; the 150 g jar's page gives a 7 cm outer rim and about 6.5 cm base, so expect it to narrow. On our 0.35 cm wall that is 6.3 cm inside at the rim (31.17 cm²) and 5.8 cm at the base (26.42 cm²). Why that taper matters to a wick is the jar diameter × wick hub.

What to buy, and what to skip

Five items, in the order the test needs them, chosen on page statements, price and stock on 29 September 2026. CSI publishes no wick-performance data, so nothing here is ranked on burn results. Oil candle counts are at 12 g (8%) and 15 g (10%) per 150 g candle.

Frequently asked questions