My Replacement Jar Has the Same External Diameter but Thicker Glass and a Smaller Internal Diameter — Should I Treat It as a New Candle System?

Why the outside measurement misleads

A replacement jar with the same external diameter looks like a like-for-like swap: it fits the same lid, the same box, the same shelf and the same label. That is exactly why this change slips through. If the glass is thicker, all of the extra thickness comes out of the inside, and the internal diameter — the one that sets the size of the melt pool — shrinks by twice the wall change, because the wall is on both sides. Your title gives no numbers, so this page uses a labelled illustration: both jars 8.4 cm outside, the old with a 3 mm wall and the new with a 5 mm wall.

Two things move at once, and they are the reason this is a bigger question than a simple width change. The melt pool gets 10.0% smaller, so the approved wick now has less surface to melt and sits 2 mm closer to the glass all round. And the wall itself contains 63% more glass per centimetre of height — the ring of glass in cross-section grows from 7.63 cm² to 12.41 cm². That is geometry, not a thermal measurement: CSI publishes no thermal data for any jar, and how much heat a particular glass holds depends on the glass itself. But in general practice a heavier wall is commonly described as warming up more slowly and holding heat for longer, which is why the jar-side lines of your test deserve more attention here than in a simple width change.

Each extra millimetre of wall takes about 5% off the melt-pool area in a jar this size and adds roughly a third more glass to the wall. So a supplier who says "same size, just a sturdier glass" may be describing a candle with a noticeably smaller pool. The mirror image — same inside, thinner glass — changes only the wall term. For a change that is purely internal diameter, the 7.6 → 7.9 cm hub post has the sensitivity table.

Is it a new candle system? The answer, split

It helps to separate two meanings of "system". The formulation — wax, oil, load, dye, cure — has not changed and does not need to be redeveloped. The burn system — the wick in that formula in that jar — has changed on two counts: a smaller pool and a heavier wall. So the honest answer is: treat the burn system as new, keep the formulation. In practice that means repeating every wick-and-jar test, including the full-life burn, rather than just confirming the early burns as you might for a 3 mm diameter change.

The NEW line is specific to this jar. A thicker wall is commonly described as holding heat after the flame is out, which matters for how the jar feels to a customer who picks it up after snuffing it. You never needed to measure that before because your old jar and your old test were matched. Add one reading five minutes after extinguishing, at the same point on the glass, with the same thermometer. Compare it with the control; there is no absolute number to hit, and CSI does not publish one.

The test: nine candles, full-life on every set

Because the pool is smaller, the sensible bracket goes down, not up. In the worked example the approved wick is Eco Thick C2 and the bracket is Eco Thin C1 — the only two Eco sizes CSI lists. We do not claim either is right for this jar or that they are one step apart for your wax; if your wick comes from a range with more sizes, bracket with the next size down in that range. All three sets go to the end of life, because the late burns are where this particular change is most likely to show.

The priced test kit

Built from listed, in-stock CSI sizes on the 24 September 2026 pull, using Luxury Soy Wax CSI 464 (rated up to 10%) and Sandalwood as the worked example, at our 8% working load with a 5% allowance.

Decision rule

CSI publishes no numeric flame or jar-temperature limit. The rule is relative: no worse than the approved control on every safety line, within your jar supplier's guidance and the candle fire-safety standard you work to (ASTM F2417 is a commonly cited US one). Set tolerances for the performance lines before you read the data.

A FAIL on the jar line is a legitimate result, not a failure of your testing. If both wicks leave the glass warmer than the control, a third wick is unlikely to fix a wall problem, and the right conversation is with the jar supplier. The alternative-container post shows how to score other candidates quickly so that conversation does not stall production.

Record the jar as a specification, not a name

This change happened because the jar was specified by its outside and its name. Close that gap now. Your jar record should carry the outside diameter, the inside diameter at the fill line, the wall thickness range, the base thickness, the internal height and the fill weight — all measured by you, with the date and supplier. The reproducibility-record post shows the format, and the approved-supplier list post explains how to stop purchasing from accepting a "same size" jar without those numbers.

Base thickness is on that list for a reason: jars made with heavier walls are often heavier at the base too. If yours is, the end-of-life part of the test carries even more weight — the thicker-base post covers that directly. For the general list of what to retest after any jar change, see the only-the-jar post.

Related jar changes

If your new jar is thicker-walled and tapers inward toward the base, the smaller pool gets smaller still as it burns down — the narrowing-jar post works that case. If the glass is the same thickness but frosted or coated, the geometry is unchanged and the test changes shape instead: the frosted-glass post. And if the new jar is simply narrower inside with the same wall, the maths on the same-weight post runs in reverse.

Frequently asked questions