How to Choose a Candle Wick

Updated 28 August 2026 · 14 min read

The wick is the engine of a candle. Get it right and the melt pool reaches the glass, the wax burns cleanly, and the fragrance carries. Get it wrong and you get a tunnel, a drowned flame, or soot on the jar.

This guide gives you the practical path first, then explains why the wick charts you'll consult disagree with each other — because they do, and knowing why is what lets you use any of them properly.

The short version

1. Measure the vessel's inside diameter. At the widest point of the eventual melt pool, usually near the top. Not the outside, and not the volume on the label — an eight-ounce jar from one supplier can be three inches (76 mm) across, the same volume from another 2.75 inches (70 mm).[1] That quarter inch is the difference between two wick sizes.
2. Pick a series that suits your wax. Broadly: ECO for soy and natural waxes, CD or HTP for harder and more viscous blends, LX for paraffin, zinc-core where a wick must stand rigid, wooden for a crackle. Each series has its own scale — an ECO 10 is not an LX 10 is not a CD 10.[1]
3. Read your diameter across that series' chart. Note what the chart assumes. Most are built around a base paraffin; some are calibrated for a named soy at a stated fragrance load; a few don't say.
4. Adjust for what the chart doesn't cover. Soy and single-pour paraffin usually need sizing up — one to three sizes is the common guidance.[2] Heavy fragrance loads and dark dyes push the same direction.
5. Burn test before you commit. The melt pool should reach the container wall within two to four hours of the first burn.[2] A ring of unmelted wax means the wick is too small. A deep pool with a dancing or smoking flame means it's too large.

That's the whole method. Everything below is why each step is harder than it looks.

There is no correct wick size for a vessel by itself. There is a wick that works in a particular combination of vessel, wax, fragrance, load and burn conditions. "CD 12 for a three-inch jar" is shorthand for a tested starting point, not a physical property of three-inch jars — which is why the charts below disagree without any of them being wrong.

Why the charts disagree

Take a 2.75-inch (70 mm) container in Golden Brands 464 soy. CandleScience recommend CD 12.[3] The Flaming Candle's chart puts CD 14 in the 2–3 inch (51–76 mm) container column.[4] Lone Star rate CD-8 for 2.5–3.0 inches (64–76 mm) in base paraffin, then tell you to size up one to three for soy — which lands between CD-10 and CD-14.[2]

None of them is wrong. They're answering slightly different questions, and there are three reasons why.

The base wax

Lone Star state plainly that their charts are built on paraffin and give you an adjustment rule for everything else.[2] CandleScience skip the adjustment by publishing a row per named wax — Cargill NatureWax C-3, Golden Brands 464, IGI 6006 and more than a dozen others.[3] The Flaming Candle organise purely by container size and note only that more viscous waxes "may require up to several sizes larger."[4]

So the same chart position means three different things depending on whose chart you're reading.

The fragrance load

CandleScience disclose theirs: burn tests run with no colour and 1.2 ounces of fragrance per pound — 7.5% — and they say dyed or heavily fragranced candles may need a larger size.[3] Most suppliers don't state a figure at all.

This isn't a failing peculiar to retailers. A 2023 study in Combustion and Flame examining twenty wicks across eight series noted that wick manufacturers themselves don't indicate which wax or fragrance loading their published data assumes.[5]

Slab versus container

The third reason is the one least often mentioned, and it explains a gap the first two can't.

Wick manufacturers publish a wax pool diameter for each size. Atkins & Pearce list HTP-62 at 4.07 cm (1.60 in), alongside its flame height and rate of consumption.[6] Those figures come from burning the wick in a large open slab of unscented wax.

You are not burning in an open slab. A container changes the thermal environment around the pool — heat reflects back off the wall rather than escaping — so a published pool diameter is not automatically what you will get in your vessel. Fragrance also changes the wax's burning behaviour, which is part of why a wick tested in unfragranced wax may behave differently in yours. NI Candle Supplies report the difference at roughly 10–12 mm (about 0.4–0.5 in), a wick's stated burn diameter running that much larger in a container than the chart says.[7] That is one supplier's reported observation rather than a physical constant — but the direction is what matters: a wick chosen strictly by datasheet pool diameter tends to over-perform in a jar.

This resolves a discrepancy that otherwise looks like an error. Blaze & Foam in Australia pair a 39–40 mm (1.54–1.57 in) vessel with HTP 62 — essentially matching the vessel to the manufacturer's stated pool diameter.[8] Lone Star recommend the same wick for 2.5–2.75 inches, or 64–70 mm.[2] One chart follows the datasheet; the other follows container testing.

Neither is wrong. But a maker reading both without knowing which is which could land thirty millimetres apart on the same wick.

Two questions the industry hasn't settled

Most wick disagreements dissolve once you know each chart's baseline. Two don't.

Size up or size down at a borderline diameter?

If your container sits between two recommended sizes, which way do you go?

CandleScience say smaller. It's much safer and easier to correct a slightly under-wicked candle — which might leave a little wax on the sides during early burns — than an over-wicked one. Over-wicking can create an oversized flame, excessive soot, a dangerously hot container, and burn the candle down too fast.[3]

Candle Creations in New Zealand say larger. Where two or more wicks match your size, they recommend starting at the higher end of the scale and testing down.[9]

Both are experienced suppliers. The two errors aren't symmetrical, which is why we'd start smaller — an under-wicked candle wastes wax and disappoints, while an over-wicked one produces excessive soot, an oversized flame and a container that runs hot.[3] But you should know that reputable suppliers disagree, and that neither position is unreasonable.

Zinc core in soy?

Lone Star: they do not recommend zinc-core wicks in soy at all.[2]

The Flaming Candle: zinc core works well in paraffin but can be used in soy.[4]

There's a related factual dispute worth flagging. The Flaming Candle describe zinc core wicks as "coreless and self-trimming" in the same paragraph that describes their zinc core.[4] Atkins & Pearce and the Combustion and Flame study both classify cored wicks — zinc, paper and cotton core — as not self-trimming, in contrast to ECO, HTP, LX, RRD and Premier which are.[5][6] On that specific point the construction is not ambiguous — a zinc wire is physically present at the centre. Whether it self-trims is a separate question, and the two properties are worth keeping apart.

The wick series

Six families cover almost everything a container or pillar maker will use. What separates them is construction, and construction determines behaviour before any chart does.

Flat-braided coreless — CD, ECO, HTP, LX, RRD, Premier. Braided cotton, usually with a paper filament woven through for rigidity, engineered to curl at the tip so the wick trims itself as it burns. That curl is what limits carbon buildup and mushrooming. Best in containers.

Cored — zinc, paper and cotton core. A rigid centre keeps the wick standing straight while the candle is poured and while it burns, which suits votives, pillars and gel where a flat braid would slump. Cored wicks are generally not self-trimming.

⚠️ Construction and self-trimming are two different properties, and the industry does not divide neatly into cored-not-self-trimming and coreless-self-trimming. Treat the split above as the common case rather than a rule — the expander below covers where it breaks.

Wooden — two thin pieces of wood pressed together, producing a wide flat flame and a crackle. Containers only.

Multi-wicking

CandleScience put double wicking at wide containers, typically over 3.5 inches (89 mm), where a single wick can't safely create a full melt pool[3] — yet their own chart still lists a single wick up to 3.79 inches for most of the container waxes it covers, so where multi-wicking begins depends on the wax. More than one wick creates a larger melt pool and melts the wax evenly across a wide surface, where a single wick might not produce enough heat to do so and can tunnel.[7]

The sizing method is the same on both sides of the world: measure the container diameter, divide by the number of wicks, and select for that fraction. NI Candle Supplies give the worked example — a 30 cl jar at 77 mm (3.03 in), halved, means wicking for 38.5 mm (1.52 in).[7] Candle Creations in New Zealand describe the same method.[9]

For two evenly spaced wicks in a round vessel, halving the diameter is a useful first approximation. Beyond two, do not treat diameter ÷ number of wicks as a formula. The melt zones overlap, spacing and vessel geometry enter the problem, and the arithmetic stops being a straight division. Use the supplier's multi-wick guidance and test the whole arrangement rather than each wick's notional share.

Then space the wicks evenly from each other and from the edge. Spacing is a safety property, not an aesthetic one. In 2007 AtHome America recalled about 6,300 four-wick candles because the wicks sat too close together, causing the candle to burn unevenly and faster than expected.[10]

Three recalls, one failure mode

The 2007 case is about spacing. The more recent ones are about flame height, and the pattern across them is the argument for testing the whole combination.

All three are multi-wick. All three are excessive flame height.

And the Magnolia case is this page's own thesis in a federal recall notice: the stated cause was the wrong wax, not the wrong wick. Thirteen reports of excessive flames breaking the container. A wick that behaves in one wax and not another is exactly what the chart disagreements above are about — and here it reached 4,350 consumers.

The part nobody mentions: the tab

The metal disc at the base of the wick is called a sustainer, and it does more than hold the wick upright.

Sustainers are classified by three things: diameter, typically 15 or 20 mm; the hole size the wick feeds through; and neck length.[12] That third one matters. Most wick makers use a 3 mm neck as standard. Long-neck sustainers run around 6.4 mm, and Lone Star supply 10 mm as standard.[2][12]

A taller neck leaves a thin layer of wax unburned at the bottom of the container, so the flame goes out before it can heat the base of the glass.[2] That end-of-life behaviour is one of the things ASTM F2417 addresses, alongside flame height, container integrity, stability, secondary ignition and flame impingement.[13]

There is a widely repeated field observation here, and it is worth labelling as one. Candle Shack report that IKEA — Europe's largest candle retailer — moved to long-neck sustainers several years ago, and that the change was associated with a reduction in complaints and candle fires.[12] That is a supplier reporting another company's experience, not an independently verified causal finding.

If you're buying pre-tabbed assemblies, the neck length is worth asking about. It's rarely on the product page.

Testing

Every source on this page says the same thing: the chart is where you start, and the burn test is what decides. Here's how to make the test tell you something.

It helps to keep three different activities apart, because they have different stopping points. Screening narrows a shortlist of candidate wicks. Performance testing asks whether the winner burns well in your exact combination. Standards testing asks whether it passes ASTM F2417, which is a laboratory protocol on a minimum of 24 candles and is not something a kitchen bench reproduces. Most of what follows is the first two.

Burn in four-hour increments, checking each hour for melt pool, flame height and hot throw.[3] Test all the way to the bottom of the container over multiple burns — candles typically burn hotter as the flame descends and heat is trapped.[3]

Make three to five otherwise identical test candles, changing only the wick size, and burn them side by side.[6] One candle burned alone tells you far less than five burned together — and if more than the wick differs between them, the comparison tells you nothing about the wick.

Worth noticing which way that disagreement runs. Suppliers who cannot agree which wick fits a three-inch jar agree closely about what a correct melt pool looks like once you have one. The disagreement is about getting there, not about the destination.

What the standard requires, and what it does not

Everything above is a performance window — whether the candle burns well. ASTM F2417 asks a different question, and answers it with different numbers.

The standard sets a safety ceiling: flame height must not exceed 3.0 inches (76 mm) under the standard test, across a minimum of 24 identical candles, with religious and ceremonial candles allowed up to 3.75 inches (95 mm).[15][16] There is a floor as well: a patent describing testing under F2417 treats a reading above three inches as failure and a quarter inch as the desired minimum, so acceptable flame height sits above ¼ in and below 3 in.[17]

The 2023 revision also put two things this page describes into standard language. Wick migration: the wick shall not migrate or lean from its original position such that the flame contacts the container for longer than five seconds. Merged flames: for multiple-wick candles, two or more wicks shall not come close enough to form a single flame for longer than five seconds.[18] So flames merging into one is a defined failure mode rather than a matter of taste, and the spacing question in the multi-wicking section above has a threshold attached to it.

The same revision puts wick attachment on the manufacturer: either properly attach the wick, or clearly advise in the instructions how the consumer should position and adhere it.[18]

⚠️ These figures come from the 2023 and 2024 revisions, read through testing-laboratory summaries and a patent rather than from the standard itself, which is paywalled. The current edition is F2417-26.[13] Check the current text before relying on any of it commercially — the CPSC's own voluntary-standards page still names F2417-17, which is two revisions behind, and the making hub covers why a summary page tells you a standard exists but not which revision is current.

A record sheet

The method above is only worth running if you write the results down. One row per burn, per candle.

Record the wick series and size, the wax, the fragrance and its load, and the cure time at the top of the sheet — those are the variables that make one candle's results comparable to another's.

What to take away

Wick sizing has no single right answer, and the appearance of precision in a published chart is misleading. Every chart is calibrated on something — a particular wax, a particular fragrance load, an open slab or a real container — and most don't tell you which.

Find out what your chart assumes. Adjust for what it doesn't cover. Then burn a candle and watch it, because that's what everyone in the industry, including the people who manufacture the wicks, says you have to do.

About · Contact · Privacy Policy · Terms of Use · Editorial Standards · For brands