I Have a Straight-Sided 3-Inch Candle Jar — How Do I Choose the First Three Wick Sizes to Test Rather Than Guessing Randomly?

Why a bracket beats a guess

Many makers test one wick, see a reasonable pool, and ship it. The trouble is that "reasonable" has no reference point. A wick one step bigger might have reached the glass an hour sooner; one step smaller might have burned cooler and cleaner through the last third. Without the neighbours on either side, you cannot tell whether your wick sits in the middle of the acceptable range or at its edge, and edges fail first: in a warm room, with a different oil, or with a customer who burns for five hours.

A bracket fixes that: put one test on each side of your best guess. If the lower rung leaves a ring of solid wax and the upper rung races to the glass with a tall, restless flame, you have boxed the answer in. The middle rung is no longer a guess; it is the only candidate the evidence leaves.

What "randomly" usually looks like

Three wicks chosen at random tend to cluster: three thin ones because the jar "looks small", or three from whatever pack arrived. If all three fail the same way, you learn only that the answer is somewhere else, and you have used nine candles to learn it. A bracket is deliberately spread so that at least one rung should fail each way. Two failures and a pass is the best result a first test can give.

Why three jars per rung

One candle per wick is an anecdote. A slightly off-centre wick, a draught from a window, a jar poured a few degrees cooler: any of these can make one candle misbehave. Three jars per rung let you see whether a result repeats. If two of three C2 candles reach the glass in about three and a half hours and the third lags, look at that third jar for a cause (centring, draught, fill) before you blame the wick. The grid is nine candles because three × three is the smallest layout that gives both a spread and a repeat.

Before you choose: is it 3 inches inside, and is it straight?

Three inches is 7.62 cm. If that is the outside diameter, as a supplier's figure usually is, the wall takes some of it away twice. With a wall of about 3.5 mm (the illustrative figure CSI uses across this series), the inside is about 6.92 cm. The pool area is π × r² on the inside radius: 37.61 cm², against 45.60 cm² if you took the outside figure at face value. Choose rungs for the outside and you are planning for 21.3% more pool than the jar holds. Thick decorative glass makes the gap larger. Why the inside diameter, not the ml, drives the wick is the subject of the diameter-and-wick hub; how to turn an external figure into an internal one is The Supplier Gives the External Diameter of the Candle Jar — Can I Use That for Wick Selection or Do I Need the Internal Diameter?.

Three quick measurements

Inside diameter at pool height. Callipers across the inside, a centimetre or two below the fill line, read twice at right angles. No callipers: lay a steel ruler across the mouth, measure the rim, then measure the wall at the lip and subtract it twice. Inside diameter lower down. Repeat near the base, or as low as your callipers reach. Depth. Rim to inside floor, which tells you whether a 10.16 cm Cotton Braided wick will need trimming before you pour and how many burns the jar holds.

If the top and bottom readings match within what your tool can resolve, the jar is straight-sided for wicking purposes, and one bracket covers the whole burn. If they differ, you are not wicking a 3-inch jar; you are wicking a taper, and the decision about which diameter to wick for belongs in the narrowing-jar post, with the burn-down consequences in the wider-at-the-top post.

Choosing the rungs: what CSI's wick pages do and do not tell you

A good middle rung comes from the best evidence you have, in this order: a wick you have already proven in a jar of similar inside diameter with the same wax and oil; a hint on the jar's own page; the positioning a wick page gives itself. Then the outer rungs are the nearest steps either side. With CSI's range, the pages say this much, and no more:

Read the ladder with its limits in view. C1 and C2 are the two sizes of one product, so the step between them is a real step in the same construction. Cotton Braided is a different construction whose page aims it at larger candles; the step from C2 to it is bigger and less tidy, which is fine for an upper rung whose job is to overshoot. The middle rung is C2 because it is the bigger of the two Eco sizes and a 3-inch jar sits below the 8 cm the braided page mentions: a reasoned starting point, nothing more.

When your evidence points somewhere else

You already have a winner in a narrower jar. Make that winner the lower rung, not the middle: a wider pool usually needs at least as much wick. How far to step up when only the width changes is the subject of the wider-jar step-up post. Your oil is heavy (dense vanilla, resin or gourmand notes) or your load is high: the pool may need more wick than an unscented guess suggests, so keep Cotton Braided in the bracket rather than dropping it. The reasoning is in our post on heavy fragrance oils and wick size. You want a wooden wick: run a cotton bracket anyway if you can spare the jars; it tells you whether the jar and wax behave, so that a wood result is about the wood.

The nine-candle grid, step by step

The grid only works if the three rungs are the only difference between candles. Wax, oil, load, fill line, pour temperature, cure, room and burn length are held constant. We use a 150 g candle at 8% (12 g oil + 138 g wax) as a worked example, the planning figure CSI uses for jars of about this size; weigh your own fill before you pour, as the 200 ml fill-weight post shows.

Reading the bracket: six outcomes and your next move

Judge each rung on the three-candle pattern, not the best candle. The pass criteria are common rules of thumb, and they are what we would use: the pool reaches the glass all round within about 3 to 4 hours of the first burn, sits about 0.5 to 1 cm deep, the flame is steady without soot, and the candle behaves the same way in the final third.

The unfair tests that fake a result

Before you move a bracket, rule out the usual culprits. A first burn stopped at two hours sets a memory ring that later burns tend to follow; tunnelling after a short first burn is a burn-length result, not a wick result. A cure of a day or two can make a whole grid burn differently than it will in a fortnight. A load above the wax page's ceiling (10% for CSI 464) changes everything. A fan, an open window or an air-conditioner vent will pull pools one way. And a wick that leans after pouring melts one side first. If any of these apply, fix it and rerun the same bracket before stepping up.

When the answer changes with the oil

A bracket is valid for the jar, wax, oil and load you tested. Change the oil and the wick may need to move; when to retest and how is covered in our post on choosing a wick after changing fragrance oil. Change the jar width and see the 7 cm to 8 cm retest post. Change nothing and your bracket result stands.

What to buy for the bracket

Five items, in the order the method uses them. We chose them on what their pages state, price and stock on 26 September 2026; CSI has no burn data for any jar-and-wick combination, which is why every card ends in a test rather than a promise.

Candle counts and cost: the 12-candle programme

The programme is nine grid candles plus three confirmation candles. Worked example: a 150 g candle at 8%, so 138 g of CSI 464 and 12 g of oil per candle. Lavender is our example oil because its page states "Recommended fragrance load 6-10% for candles" and every size is in stock; test with the oil you will sell, because the wick answer belongs to that oil. Prices are CSI live pricing on 26 September 2026; jars are not included because you already have them.

The oil is tight on purpose: 150 g bought, 144 g used, 6 g spare. Weigh it straight into the melted wax on the scale rather than via a cup, and if a confirmation candle needs pouring twice, buy another 50 g (₹270.00; Lavender is ₹5.40 per gram at both 50 g and 100 g). Wick stickers (₹100.00 for 40) are extra if you use them. If you only run the nine-candle grid, the wax and oil used (1,242 g and 108 g) plus the wicks and holders come to about ₹1,722.79.

Why the wick's price should not choose the wick

Worked example, one 150 g candle at 8%: 138 g of CSI 464 at the 10 kg price is ₹70.02, 12 g of Lavender at the 1 kg price is ₹62.65, and a C2 wick is ₹9.44, so ₹142.11 before the jar, labour, freight, packaging and GST. The wick is 6.6% of that. Moving from C1 to C2 adds ₹3.54 per candle; from C2 to Cotton Braided, ₹2.56. Those differences are too small to matter next to a candle that tunnels or soots. Choose on the burn; the price follows.

What skipping the bracket puts at risk

Pour 100 candles on an untested wick and the wax, oil and wicks in them are worth ₹14,211.32 in the same worked example, before jars, labour and the customers who received them. That is 6.1 times the whole 12-candle programme. A mis-wicked batch can sometimes be saved by melting out, re-wicking and re-pouring, but that costs labour and a second set of wicks, and the batch is no longer the candle you tested. The pre-order version of this argument, for a 500-jar line, is the 500-jar post.

Fitting the bracket in before Diwali, 8 November 2026

Diwali falls on 8 November 2026, 43 days after these prices were checked. The bracket itself needs one afternoon to pour, a cure, then four burn sessions; the confirmation adds another pour, cure and burn. Dispatch is from Pune and delivery time depends on your location, so confirm it on WhatsApp before you plan backwards from a date.

If the first bracket misses, the second bracket from the spare wicks costs about a week; decide by 12 October whether you ship this Diwali with a proven wick or hold the jar for Christmas (25 December 2026). Shipping on an unproven wick to make a date is the one move we would argue against.

Frequently asked questions