My Candles Have Better Adhesion When They Cool Slowly — How Do I Reproduce Slow Cooling at Commercial Scale?

What your observation means, and the first thing to measure

Many makers notice that the candles which cooled slowly — the batch left in a closed room overnight, the jars pushed together in the middle of the table, the winter evening batch covered with a cloth — have cleaner glass than the batch that cooled fast. That fits how wet spots are commonly described: places where wax has separated from the glass as it contracts while cooling and setting, with fast or uneven cooling one of the most commonly cited contributors. Slowing the cooling down gives the wax and glass a gentler ride. So your observation is plausible. But it is still an observation, and there are two problems with scaling it directly.

The first is that "slow" is not yet defined. You do not know how slow the good batches cooled, so you cannot tell whether a cabinet, a cover or a warm room reproduces it. The second is that the good batches may have differed in something else — the jars sat in the same warm room before the pour, the pour was hotter, the batch was bigger. The outer-versus-centre post shows how easily position, fill order and cartons hide inside a cooling pattern. So the first step is free apart from a thermometer: in your next normal batch, put a probe jar in the open and another in whatever slowed the good batch, and log the centre-wax reading every fifteen minutes until both reach room temperature. That gives you two curves — the shape of "fast" and the shape of "slow" in your own room.

Once slow cooling has a number — for example, the reading at 30, 60 and 120 minutes, or the minutes until the top turns fully opaque — you can test whether it really improves adhesion, against a fast control in the same melt. Only then is it worth designing a way to reproduce it for every batch. The rest of this page is that sequence: measure, prove, then scale, with the capacity cost of slow cooling counted honestly.

Step 2: measuring a cooling curve

A probe jar is a sacrificial candle filled from the same melt, with a thermometer tip held in the middle of the wax at mid-height. The Pen Thermometer measures wax, not glass; that is what you want here, because the wax is what sets and contracts. Fix the probe in place with a clip or a wick holder so it does not drift, and do not count the probe jar as a test candle — the probe changes it. Read and write the temperature at the pour and every fifteen minutes after (a labelled working interval), until it reaches room temperature or stops changing. Write the room temperature at the start and end, the load (how many hot jars were cooling in the same space) and whether anything was covered.

The curve is yours; nobody publishes one for any CSI wax, and we will not draw you a "correct" one. What matters is the comparison: the open-bench curve against the slowed curve in the same room on the same day. From the two curves pick a simple house measure you can check every day without a probe — for example, the reading at 60 minutes, or the time until the top is fully opaque — and confirm that it separates the two conditions clearly. That becomes the number your scaled-up method must hit.

In this worked example the candle is CSI's Yankee Glass Jar 130 ml with clear cap. Its 130 ml is brim volume and its fill weight is not published: a labelled approximate 0.86–0.90 g of wax per ml would put a brim-full jar at about 111.8 g–117.0 g, so this page uses 100 g as an example working fill — weigh one test pour to your fill line and use your own figure. The wax is Natural Soy Pearl Wax, which publishes no melt point, pour temperature, fragrance ceiling or adhesion statement, so the 8% load (8 g of oil, 92 g of wax) is a labelled test setting inside the 6–10% range commonly used for tests, and the pour temperature is whatever the wax maker's instructions say — ask CSI on WhatsApp if you do not have them.

Step 3: proving that slow cooling is the cause

Pour 14 jars from one melt: 6 on the open bench (arm A, control), 6 under the cover that gave you the slow curve (arm B), and one probe jar in each arm. The cover must be the only difference — same bench, same spacing, same fill order rotated between arms, same room and fan setting. A cardboard hood over a tray, a lidded plastic crate turned upside down, or an unheated cupboard are all labelled non-CSI examples; whatever you use, it must not touch the jars, and it should be dry and clean so dust and condensation do not land on the tops.

The scoring sheet

Score four marked sides per jar through the glass against a plain light background, 0–4 per side, 0–16 per jar, with your house pass line written down beforehand (for example, "no side at 2 or more"). No industry figure exists for acceptable spotting. Record the tops as well, because slower cooling can change how the top sets — a cleaner wall with a worse top is a trade-off, not a win.

Your rule for a real result, written first: for example, arm B has fewer failing jars than arm A at both checkpoints, its tops are no worse, and its probe curve is clearly slower. If B wins on glass but its probe curve was not slower, something other than speed changed under the cover — humidity, dust or light — and that needs its own test. If B's curve was slower and its glass no better, slow cooling is not your lever, whatever the earlier batches suggested.

Step 4: making slow cooling work at batch size

Here is the part most scale-up attempts miss. Under a cover, the jars warm each other and the air around them. The more hot wax under the cover, the more heat it holds — so the same cover on a quarter-full day and a full day is not the same condition. The table below is not heat data; it is simply how much hot wax sits under one cover at different loads of the example 100 g candle.

So qualify the method at the loads you actually run. Put a probe jar under the cover on a quarter-load day, a half-load day and a full-load day, and compare each curve with your target. If a quarter load cools too fast to hit it, you have three choices: always fill the cover (consolidate small batches), use a smaller cover for small batches, or add a labelled non-CSI heat source and test that as its own variable. Whatever you choose, write the load into the batch record — jars and kilograms — so a bad batch can be traced to a light-load day.

Three other ways to slow cooling at scale, each a separate test against your measured curve: keep the batch as a tight block with a sheltered edge (7351), cool on an insulating layer rather than a cold top (7353 and 7354), and cool in a room held warmer by a labelled example amount (7357 shows how to record room temperature as a parameter). Warm phases in any test stay well below the point where wax would soften again; CSI publishes no softening point for Natural Soy Pearl, so keep any heated enclosure modest and logged.

Step 5: what slow cooling costs in capacity

Slow cooling is not free: jars that sit longer under a cover hold their slots longer. If you need to move a batch before the next one can be poured, the hold time decides how many slots you need. The example below uses 200 candles a day and 8 working hours for cooling turns — both examples you replace — and three example hold times you will measure from your own curves.

Doubling the hold time doubles the slots, and tripling it triples them. That cost is mostly bench or shelf area and covers — the cheapest possible equipment — but it is real. Against it sits the value of the candles you stop rejecting. The example 100 g candle costs ₹120.50 in CSI materials at Natural Soy Pearl's 5 kg rate (₹0.42 a gram — the only CSI soy wax with a real bulk break, 5.6% below the 1 kg rate), Fresh Strawberries at the 1 kg rate (₹3.54 a gram), an Eco Thin C1 at ₹5.90 as the example wick line and the Yankee jar at ₹47.20, before labour, packaging, freight and GST. If covering saved an example 5 jars per 100 from seconds, a cardboard hood at an example ₹250.00 pays for itself in materials in 0.41 batches of 100 and a closed cabinet at an example ₹6,000.00 in 9.96. Replace every example with your own figures; the reject-cost post adds selling price, labels and complaints.

The priced screen

If you already make this candle, run steps 2 and 3 inside a production pour and buy only the thermometer if you lack one. If you need a dedicated screen, here it is in listed, in-stock CSI packs, with a 5% pour-loss working allowance on wax and oil only. Natural Soy Pearl starts at a 1 kg bag, so there is wax left over for production.

The 647.6 g of wax and 32.4 g of oil left over are enough for a few more scored jars if you want seven or eight per arm instead of six — each extra jar costs about ₹124.71 in CSI materials at small-pack prices — or they go into your next production pour.

The decision rule

Stop at the first line that is true. Thresholds are your house choices, written before the pour.

When the first line is true, write the method down as numbers: cover type, load range, the probe reading at your checkpoint, and the hold time. That is what makes it reproducible on a busy day, by someone else, in a different season — and pilot batches in different seasons is how you check it survives the year.

The CSI products in this test

Frequently asked questions