I'm Using 10% Fragrance Oil but My Soy Candle Still Has Weak Hot Throw — What Am I Doing Wrong?

Before anything else: is your 10% actually 10%?

Every CSI wax page states its load the same way: as a percentage of total wax weight. On a 200 g jar, 10% means 20 g of fragrance oil poured into 200 g of wax, giving a 220 g pour. That is the only definition used anywhere on this page, and it is worth writing on the wall above your scales, because three of the four ways a 10% recipe quietly stops being 10% start with a different definition.

The first is the convention itself. If you took 10% of the finished candle, you worked out 10% of 222 g and weighed 22.2 g into 200 g of wax. Against wax weight that is 11.1%, which is above the figure the CSI 464 page states. Over a ceiling the surplus does not become extra throw; general practice is that oil the wax cannot bind separates — it beads on the top, wets the wick, and in a hot Indian room it sits as a film rather than evaporating from the pool. A candle can smell weaker at 11.1% than at 9%, which is the most counter-intuitive thing in this whole subject.

The second is resolution. A 1 g kitchen scale reading 20 g is honest to about ±1 g, which on this dose is ±5% of the oil. Two jars from the same batch can be 9.5% and 10.5% and you will never know which one you sniffed. Weigh to 0.1 g. CSI's Candle Making Weighing Scale is ₹354.00 and stock on it is thin at the moment; any 0.1 g jeweller's scale does the same job.

The third is transfer loss. Oil weighed into a beaker and then tipped into the melt leaves a film behind. If 1.5 g of a 20 g dose stays in the beaker, the wax received 18.5 g — 9.25% — and you recorded 10%. Weigh the oil directly into the melt on a tared scale, or scrape the beaker with a Silicone Spatula at ₹175.82.

The fourth is sequencing. Oil added to a large pot and then poured out over ten minutes does not stay evenly distributed if the stir was short: the first jars and the last jars differ. The Luxury Soy Wax Chunks page is specific about this — add fragrance at 80–90°C and stir slowly for two full minutes. Two minutes is longer than it feels.

Where 10% sits on your wax's page — and on your oil's page

The wax has a stated maximum and the fragrance oil has a stated range, and they are set by different people for different reasons. The wax figure describes how much oil the wax was formulated to hold in solution without seeping. The oil figure is the supplier's stated usage range for that fragrance in that format. Where the two disagree, the lower figure governs — always, and without exception.

On Luxury Soy Wax CSI 464 the page states up to 10%. That is why this question exists: you are on the ceiling, not below it, and the usual advice to add another two points is not available to you. On Luxury Soy Wax Chunks the page states up to 13% of total wax weight, with 8% as a starting point, 10% as the recommended figure and 13% as the maximum to be burn-tested thoroughly. Premium Coconut Soy CSI 468 also states up to 13%. Eco Soy CSI 400 does not publish a load figure at all — we will not guess one, so ask on WhatsApp before you dose it.

Now read your oil. Many CSI oil pages state 6–10% of total wax weight for soy candles — that is the figure on the Lavender, Roasted Coffee and Dark Vanilla pages. Jasmine states up to 10% for soy. Autumn states 7–11% for candles and describes itself as medium to strong. Cinnamon Vanilla states up to 13%. If you are pouring Lavender at 10% in CSI 464 you are simultaneously at the top of the wax's range and the top of the oil's range, which is a perfectly legitimate place to be and also a dead end for the dose lever.

At the ceiling, the wick is the biggest dial you have left

This is the mechanism that makes the dose lever feel so disappointing. Fragrance evaporates out of molten wax, so the rate at which a burning candle scents a room is governed by how much molten surface there is and how warm it runs. Doubling the dose in a jar with a small pool mostly buries more oil in wax that never melts.

The arithmetic is unforgiving because area goes with the square of the diameter. In a 70 mm jar, a full pool is 38.48 cm². A pool that stops 10 mm short of the glass on each side is 50 mm across and covers 19.63 cm² — about 51% of the surface. You did not halve the fragrance in the candle, but you roughly halved the surface it can leave from. That gap dwarfs the difference between 8% and 10% in almost every jar we have ever been sent a photograph of.

Measure it rather than eyeballing it. Burn one jar for three hours in a still room, then lay a steel ruler across the top and read the widest part of the liquid, edge to edge. Photograph it with the ruler in shot and write the number on the batch card. If the pool is short, move one step up the wick ladder — Eco Candle Wicks come in Thin (C1) at ₹5.90 each and Thick (C2) at ₹9.44 each in 100-packs, and CSI does not publish a diameter table per wick, so this has to be a test rather than a lookup. The live guide to choosing a wick by diameter and the piece on wrong wick size cover the ladder method in more detail.

There is a limit in the other direction, and it is worth stating plainly because it costs money: a wick that is too large does not keep improving throw. It runs the pool too hot, burns the candle down quickly, mushrooms, soots the glass and can scorch the fragrance it is meant to be carrying. A full pool by hour three or four is the target. A full pool in ninety minutes with a 15 mm-deep puddle is an over-wicked candle, not a well-scented one. Pool depth is worth watching too — 6–12 mm is the commonly cited working band in general candle-making practice.

Cure: the fortnight most 10% candles are never given

This is the most common single answer to the question in the title, and it is free. Soy container candles keep changing after the pour. The wax continues to set and rearrange; the fragrance distributes through it and binds; the surface stops giving off the loose oil that makes day-one cold throw so flattering. General practice for soy container candles is 14–21 days before you judge hot throw, with the longer end for heavy vanilla, amber and woody-oriental blends.

CSI's own pages sit either side of that. The Luxury Soy Wax Chunks page asks for a minimum of 48 hours and says ideally 1–2 weeks before burn testing. Several fragrance oil pages give shorter figures — 48 to 72 hours is common — because that is genuinely enough to test a candle and know it works. The two are answering different questions. Forty-eight hours tells you whether the candle burns and smells of the right thing. Fourteen days tells you how loud it will be on your customer's shelf, which is the number you are actually arguing about.

Cure conditions matter as much as cure days in India. Lids on during cure keeps dust out and slows the loss of top notes; a cupboard at 25–30°C is better than a shelf in direct sun at 38°C; and in Mumbai, Kochi or Kolkata in monsoon a sealed carton with the lids on is the difference between a candle that reads well in October and one that has been quietly breathing out its top notes for a fortnight. Cure indoors, away from windows, and do not stack hot jars — the live guide on why a candle needs a 14-day cure goes through the conditions in detail.

Where the heat took it: add temperature, stirring and the pour

Fragrance oil has to go into solution in the wax, and that happens at temperature and with agitation. Add it too cool and it does not disperse properly — it stays in richer and poorer regions of the melt, and the poorer regions are the ones that end up in some of your jars. Add it too hot, or hold the melt hot for a long time afterwards, and you lose more of the light volatile material than you need to.

Use the grade's own figures. The Luxury Soy Wax Chunks page asks you to melt to approximately 80–85°C, add dye at full melt, add fragrance at 80–90°C and stir slowly for two full minutes, then pour at 75–80°C. The Eco Soy CSI 400 page asks for fragrance at 85°C within a range of 80–90°C, a gentle two-minute stir, then a direct pour. The Luxury Soy Wax CSI 464 page gives 80–90°C as the working range for melting and pouring. Premium Coconut Soy CSI 468 is different again, with a melting point of 50–53°C and a pour temperature of 58–60°C — a grade you cannot run on soy habits.

A point worth retiring: flashpoint is not the temperature at which your fragrance burns off. It is a transport and storage safety property. British Rose states a flashpoint of 192°C, Jasmine states 185°C, and Roasted Coffee states approximately 191–193°C. Adding at 80–90°C is a long way below all three. If your throw is weak, the flashpoint is not the reason, and dosing at 60°C to protect the oil is a good way to get a badly dispersed candle instead.

Two process habits are worth building. First, stir with intent but without whipping: slow figure-of-eight passes that reach the bottom corners of the pitcher, for the full two minutes by a clock, not by feel. A CSI Pro Mixer at ₹349.00 is useful for larger melts, at low speed and below the surface. Second, pour promptly once the oil is in. A pitcher standing on the bench for twelve minutes while you fuss with wick stickers is losing top notes and dropping through its pour window at the same time.

When the oil is simply a quiet oil at 10%

This is the least popular answer, so it belongs on a page written by a supplier who sells fragrance oil: after the dose, the cure, the pool and the heat, a fraction of weak-throw cases are simply an oil that does not project strongly in soy at any allowed load. That is not a defect. Projection depends on the volatility of the materials in the blend, and a composition built around soft musks, heavy resins and creamy base notes is chemically committed to being close-range.

CSI's own pages tell you a surprising amount if you read them as specifications rather than as descriptions. Pages that describe an oil as strong-rated or as delivering strong cold and hot throw — Lavender, Jasmine, Roasted Coffee — are telling you the blend was built for it. A page describing an oil as medium to strong, as the Autumn page does, is telling you something different. And an oil written up for perfumery — Solar Bloom's page positions it as best formulated as an eau de parfum at 18–22% with moderate projection — is a fine oil that was never composed to fill a drawing room from a 200 g jar. One of the reviews on this page is a maker reporting exactly that, in a diffuser rather than a candle. We would rather point at it than pretend it does not happen.

The way to settle it in your own workshop is a reference oil. Pick one oil whose page states strong throw, buy it in a size you will finish, and pour one jar of it in every diagnostic round — same wax, same wick, same jar, same cure. British Rose works well as a reference because it is inexpensive at ₹3,068.00 for 1 kg, states a flashpoint of 192°C and states 0% vanillin, so it will not yellow a white wax and confuse a second variable into your test. If your problem oil is quiet and the reference is loud in the same jar on the same day, you have your answer and it is not the wax.

The four-jar round that settles this in sixteen days

Do not change four things and hope. Pour four jars on the same afternoon, from the same wax, into the same containers, and change exactly one thing in each. Materials for the round — 800 g of Luxury Soy Wax CSI 464, 80 g of British Rose at 10% and four Eco Candle Wicks — come to ₹674.96, or ₹168.74 a jar. That is less than most makers spend on a single panicked wax change.

What each answer costs — and what a wrong guess costs

The dose lever feels like the expensive one because oil is the expensive ingredient, and it is: on a 200 g jar of CSI 464 the wax is ₹101.48 and 20 g of Lavender at 10% is ₹104.42. The oil costs slightly more than the wax it is dissolved in. Which is exactly why it is worth knowing whether the extra two points are buying anything: dropping from 10% to 8% on Lavender saves ₹20.88 a candle, or ₹2,088.40 across 100 candles, and in a jar whose pool never reaches the glass it will cost you almost nothing in throw.

Set that against the cost of guessing. A 100-jar festive batch poured at a dose you have not verified, cured for three days, wicked by habit and then remelted because it smelled weak, costs the full wax again at ₹101.48 a jar, a new wick at ₹5.90, whatever fragrance is lost in the remelt, and two working days you did not have in the last week of October. The four-jar round is ₹674.96 and sixteen days. Run it in September and you never have that conversation with yourself in November.

Frequently asked questions