My Soy Candle Has Excellent Hot Throw for the First Hour but the Fragrance Seems to Disappear Later in the Burn — What Could Cause This?

What the first hour is, and what the third is

When you light a candle that has had its lid on, several things flatter the first hour at once. The headspace inside the jar has collected vapour since the last burn and it leaves in the first minutes. The first liquid wax is the surface layer, which still holds the oil's full top notes. You have just walked into the room, so your nose is fresh. And the door is often shut because you have only just settled. By the third hour, every one of those has changed: the headspace is empty, the pool has been hot for two hours, you have been breathing the scent continuously, and someone has opened a window or switched on the fan.

None of this means the candle got weaker in any physical sense, and some of it means the candle is doing what it should. A pool that reaches the jar edge and stays there is a working pool. What changes is how much of the oil's lightest material is still in it, how the flame is heating it, how fast the room exchanges air, and how sensitive the person judging it still is. The fade-point test below separates those four without changing the formula.

Four suspects, and what separates them

The table ranks the causes of an in-burn fade in the order we would suspect them. The ranking is editorial judgement: CSI publishes no failure statistics and has no measured throw data. What makes the table useful is the last two columns: for each cause, the observation that rules it in and the one that rules it out.

Causes combine. A maker who sits beside a candle for three hours with the fan on is adapted and ventilated, and the wick may have capped as well. That is why the test gives each twin one intervention only, and why the control twin matters: it shows what the fade looks like when you do nothing, so you can see which intervention moved the score.

Why top notes are last on the list

Top-note depletion is real, and oils built on a bright citrus top are the likeliest to show it: the citrus fade log measures exactly that with a timed sniff series. It sits last here because it is the only cause you cannot see or reverse at the fade point, so you reach it by ruling out the other three, and because it is the only one about the oil's own structure. Even then it is not a defect. An oil described as "sparkling bergamot and citrus top notes, warm amber and soft floral heart, woody musk and smooth vanilla base" is built, by its own description, to move from its top to its base; in the third hour you are smelling the amber and musk it promised.

The fade-point test, step by step

The test uses four twins: candles from the same pour, same wax, oil, load, wick and jar, and the same cure age give or take a day. Each is burned alone, on its own session, in the same room at the same time of day. Nothing about the candles differs; only what you do at the fade point does. If your batch has fewer than four unburned twins, run A and B first (adaptation is the commonest cause and costs nothing to test), then C and D with the next two.

Bill of materials, if you need to pour the twins

Using twins from the batch that fades is best: they are the candle you are diagnosing. If none are left, pour four in one session from one pot. At 8% of total weight each 150 g candle takes 12 g of oil and 138 g of wax; the set takes 48 g of oil and 552 g of wax. With CSI 464 in the 150 g jar, pour at 80–85°C (the jar page's 85°C governs), and let them cure at least as long as the batch you are comparing them with; the Bergamot Amber page states a 48–72 hour cure for soy, and many makers wait a week or more before judging.

Costs are worked examples and exclude labour, freight, packaging and GST. Starting from nothing, the cash outlay is ₹1,790.40, most of it jars and wax you keep. Running the set twice needs 96 g of oil: the 100 g bottle (₹560.00, ₹537.60 used) leaves 4 g, at the same ₹5.60 a gram as the 50 g. Swap in your own oil at your own load if the candle that fades uses something else; the method does not depend on the oil.

Reading the result: the decision table

Four of the seven outcomes end with "nothing", two with "nothing yet", and only one with a ₹59.00 pack of wicks you may not need. That is the point of catching the fade in the act: the causes that cost nothing to fix are also the commonest, and you only learn that by testing them rather than guessing.

What a clean result looks like, as a worked example

Say twin A peaks at 3 at 45 minutes and falls to 1 by 1 hour 30, and stays there. Twin B follows the same curve, but on the doorway breath after ten minutes outside it scores 3, and 2 again five minutes later. Twins C and D stay at 1. That is adaptation, cleanly: the candle was throwing all along, your nose stopped reporting it, and the quick return to 2 is the adaptation resuming. Now say instead that none recovers, the doorway breath in B is a 1, and what you smell is warm and sweet rather than bright. That is top-note depletion, and the candle is showing you the bottom half of its note pyramid. These numbers are illustrative; your scores are the data.

The burn-length cap and what it protects

Many makers and candle-care labels give two burn instructions: burn until the melt pool reaches the edge of the jar, and do not burn for more than about four hours at a time. Both are common practice, not CSI figures, and neither is a law of physics, but together they bracket the window in which a container candle does its best work. The first protects the pool: the commonly cited rule of thumb is a full pool to the edge in about 3–4 hours on the first burn, and roughly an hour per 2.5 cm of jar width, about 2.8 hours for the 7 cm-wide 150 g jar. Measure your own. The second protects everything else.

Past about four hours, several things that feed an in-burn fade tend to get worse together. The wick has had longer to build carbon, which is commonly linked to long burns as well as to over-wicking and heavy loads. The pool has been hot for longer, so more of its lightest material has already left. The pool itself tends to deepen, and heat builds in the glass above it as the level falls through later burns. And the people in the room have been adapting for four hours. None of that says a four-hour session will fail; it says the fourth hour is the least informative hour to judge a candle in, and the least kind to it.

For your own product, put the cap on the label: "Burn until the wax melts to the edge; trim to about 1/4 inch before lighting; do not burn for more than about four hours." That one line covers two of the four causes on this page for every customer you will never meet. For the customer who writes to say it "stops smelling after an hour", the full intake belongs to the customer-complaint post.

What not to buy for an in-burn fade

We sell every product on this page, and most of them are the wrong answer to "it fades after an hour". Here is what we would talk you out of, in order.

If your candle also shows a second symptom, a short pool, sweating or soot, do not try to fix them all from this page. The hub for this cluster, cold strong, hot faint, orders the checks, and the more-oil-worse-throw autopsy covers a wick that clogs at higher loads.

Products, in the order the test needs them

Five items in test order: the tool for twin C, the worked-example oil, the wick step if C points at the flame, the label that carries the burn cap, and the wax for a fresh set. Chosen on what each page states, price per unit and stock on 26 September 2026; CSI has no measured throw data, so nothing here is ranked on throw. The oil table shows every size, candles at 12 g each, and four-twin sets at 48 g.

Frequently asked questions