How Do I Evaluate Container Temperature Consistently During Development Rather Than Relying on Whether the Jar “Feels Hot”?

Your question, made precise

You're developing a candle, or choosing between two wicks, and somewhere in the burn the jar seems hot. Maybe hotter than last week's version, maybe hotter in the bottom third than the top. Right now the evidence is a hand near the glass, a customer's comment, or a feeling. You want a method that gives the same answer when you or anyone on your team repeats it, so you can compare wicks, jars and fragrances properly and decide with numbers.

What "consistently" has to mean: the same quantity (the outside surface of the glass, not the wax, not the air), at the same place, at the same moment in the burn, under the same conditions, recorded the same way, and enough times that one odd candle doesn't decide. Each of those has a step below. What it does not mean is a universal safe temperature; that number is yours to set, and the method for setting it is in its own section.

The instrument: what a non-contact thermometer actually reads

CSI doesn't sell a non-contact thermometer; you'll buy one elsewhere. A handheld infrared unit reads the infrared emitted by the surface in front of it, averaged over a circular spot. Three things decide whether its number means anything here.

Spot size. Units are sold with a distance-to-spot ratio printed on them or in the manual; 12:1 is common on handheld units, but check yours. At 12:1 the spot is about one-twelfth of the distance across: at 12 cm about 1.0 cm; at 30 cm it is about 2.5 cm, 36% of a 7 cm jar's width, and part of it may be reading the air beside a curved jar or the table behind. Pick one distance that puts the whole spot on the glass, and use it every time. A short stick taped to the unit as a distance gauge is a cheap way to keep it constant without touching the jar.

What it sees. Glass is generally opaque to the infrared wavelengths most handheld units read, so the number is the outside surface of the glass, not the wax behind it, which is what you want. Shiny metal is the opposite problem: polished or metallic surfaces (a golden lid, a printed metallic label, a tin) commonly read far lower than they really are. Take the lid off and aim at bare glass, away from labels.

Emissivity setting. Many units have a fixed or default emissivity setting; glass is commonly listed near the high end of the scale. Whatever yours is set to, write it on the sheet and never change it mid-study, because changing it shifts every reading. Absolute accuracy matters less than consistency: you are comparing candles with each other and with a reference, measured the same way.

Points, times and the room: the protocol

Worked example, labelled: CSI's Clear Glass Jar with Golden Lid (150 gram), whose page gives "150g fill capacity", outer diameter 7 cm and height 8 cm (about 6.3 cm inside with 0.35 cm walls assumed; the page calls it "thick-walled", so measure). The candle is 150 g at 8% = 12 g fragrance + 138 g wax (load as a percentage of total weight).

That is 4 points × 5 times = 20 records per candle per session, each the highest of 3 shots. With six candles and 6 sessions each (worked example: 125 g above a 25 g floor at 5.5 g an hour is about 22.7 hours), the test fills 720 cells from 2,160 trigger pulls. That sounds like a lot; it's about a minute per candle per hour. Pre-print the sheet: one row per point, one column per time, and a column for the session's grams lost.

Point P is the one people skip and the one that matters most late. As the wax level drops, the flame sits lower and the glass just above the pool takes the most heat, which is why a jar can read cool at M and very warm at P in the same minute. Why the lower jar behaves this way is the mechanism post, why a wick correct at the top becomes too aggressive lower down; how to read the readings section by section is evaluating top, middle and bottom separately.

Why two readings disagree: causes, ranked

The ranking is our judgement for the most common reason two readings of "the same" candle disagree. The first four are measurement and environment (how you read and where you burn); the last three are the candle.

Causes 1 to 4 cost nothing to fix and account for most "this batch runs hotter" alarms in a first study, so fix them before you change a wick. Cause 6 is the one this protocol exists to catch, and it almost always shows in the bottom third first; the published 7 cm jar post on hot glass after several hours, strong hot throw but very hot glass after several hours, covers the wick decision. Cause 7: heavier oils and higher loads commonly change wick behaviour (a tendency, not CSI data), and glass varies between batches, so record the jar batch on the sheet.

Setting your own limit without inventing a number

We won't give you a glass-temperature number. None exists in CSI data, jar pages state pour figures, not burn ratings, and a number copied from the internet tells you nothing about your jar, table or customer. What we can give you is a way to set one that is honest, repeatable and written down.

Two cautions. A reference limit is only as good as the reference: if your "trusted" candle runs hotter than it should, the method inherits that, which is why step 1 says to make the reference conservative when in doubt. And "uncomfortable to touch" is not a limit and never becomes one; what to do when a candle with a perfect pool still runs warm is a perfect pool but a jar too hot to touch, and why a hot-jar candle's excellent throw doesn't earn it a pass is rejecting a wick with excellent throw.

Using the protocol: reference against candidate

Six candles in the 150 g jar, poured the same day from one wax lot, one oil lot and one jar batch: three reference candles on your current wick (C2 in this example) and three candidates one CSI rung down (C1). 72 g of your oil and 828 g of wax. CSI's tabbed rungs are C1 → C2 → Cotton Braided, with no sized, tabbed wick between them; from Cotton Braided the candidate is C2.

With your own jars: ₹625.40. From a Cotton Braided reference, the candidate is an Eco C2 10-pack instead: ₹1,545.80 with jars and wax. Why a two-hour test can't replace this is safe in the first half, extremely hot near the bottom; why a new jar's final third is its own glass test is testing the final third of a new glass container.

The CSI buy list and the step to production

CSI's sized, tabbed wicks are three rungs, C1 → C2 → Cotton Braided, with no sized, tabbed wick between C1 and C2 or between C2 and Cotton Braided, none smaller than C1, and one wooden-wick listing with no widths. If C1 passes the temperature gate but tunnels, and C2 runs over your limit at the bottom, you're between rungs we don't sell: the fragrance load (within both ceilings), a jar change, or a question on WhatsApp, where we'll say plainly what we have.

Argued against our own interest: the most important item in this post is a non-contact thermometer, and we don't sell one; buy a decent one elsewhere before you buy wicks from us. And if the protocol shows your current candle stays inside your limit in every third, you need no new wick at all.

The step to production quantity

Only after three candidate candles pass every gate to the floor. Wicks = candles × wicks per candle × (1 + a defect allowance). Worked example, labelled: 250 candles a month plus 2% is 255 C1 wicks: 2 × 100 + 50 + 20 = 270 for ₹1,593.00. A 500-pack is the same ₹5.90 a wick and ties up ₹1,357.00 more. Reorder point = daily use × your lead time + safety stock; CSI publishes no lead times, so ask on WhatsApp. Then keep the protocol alive in production: burn 1 retained candle per batch (our judgement) to the floor with the same sheet, so a drift in a wick, oil or jar lot shows on paper before it shows in a customer's hand. If a customer does complain, the records to check are the production-records post.

What to buy, and what to skip

Five items, chosen on page statements, price and stock on 7 October 2026. CSI publishes no flame, soot or heat data and sells no glass thermometer, so nothing is ranked on performance, and the first card tells you what not to use it for.

Skip for now: any 100-plus wick pack before the candidate passes; a second jar design in the same study (one jar batch, or the comparison is muddied); and any thermometer, ours included, that you'd be tempted to press against hot glass.

Frequently asked questions