My First 50 Candles From a Large Batch Have Smooth Tops but the Last 50 Are Rough — Is the Wax Cooling Inside My Melter During Production?

What smooth-then-rough tops can and cannot tell you

A batch that gives you 50 glassy tops and then 50 rough, pitted or lumpy ones is telling you that something changed during the run, not between the raw materials and your test. The wax, oil and wick went into the melter as one mass. What differs between candle 1 and candle 100 is time: time in the melter, time in the pitcher, time the jar sat on the bench, and the temperature that went with each.

Cooling wax is the most likely cause, and the question in the title is the right one. But it is not the only cause, and a rough top is not proof of it. Three other things produce the same picture: the jars used later were colder (a second tray that came off a cold shelf), the room changed (a fan or air conditioner switched on mid-afternoon), or the pitcher was refilled from the melter surface where a thin skin had started to form. Each of these looks like 'the wax got cold', and each needs a different fix.

What rough tops usually do not mean is that the fragrance load is wrong or that the wax lot is faulty. If your first 50 candles are smooth, the formula can make a smooth candle. Changing the formula now means paying for a new test cycle to solve a problem the formula did not cause.

Log temperature against pour sequence before you change anything

The controlled way to find the cause is to make the next batch a measurement run. Keep the recipe, the jars and the people the same, and write down what the wax actually was at each point in the sequence. You need a separate thermometer for this. The imported melter's display carries a stated ±10°C deviation, so a display reading of 70°C could be anywhere from 60°C to 80°C in the wax itself. Read the melter display, then read the wax with your own thermometer, and log both.

Now lay the grades beside the log. If roughness begins where the pitcher reading drops below a certain figure, you have your cause and your limit. If roughness begins where the jar temperature or room changed while the pitcher reading stayed steady, cooling wax is not the driver.

The readings above are invented to show the shape of a useful log, not figures to copy. Notice the pattern they illustrate: the display never moved while the wax fell steadily. That gap is exactly why a melter display cannot be your process record.

Prove it with a small control poured at the late temperature

A log shows correlation. To prove cause, pour a 1 kg control from the same wax and oil lots and deliberately split it. Pour half the jars at the temperature your early candles saw, and hold the rest in a small vessel until they reach the temperature your late candles saw, then pour those. Use jars from the same carton, warmed identically, on the same bench.

If the deliberately cool pours come out rough and the warm ones smooth, cooling is confirmed and you can fix the process with confidence. If both halves are smooth, the late wax temperature alone is not the cause; look at jar temperature, the bench, or skinning in the melter. This control costs you about 1 kg of material instead of another 20 kg run, and it separates the process from the formula cleanly.

What to do with the 50 rough candles already poured

Take a worked example. The batch was 20 kg of scented wax at 8% fragrance: 18.4 kg of CSI 464 and 1.6 kg of Lavender, poured into 100 jars of 200 g. The wax cost is 18.4 kg × ₹507.40 per kg (the 10 kg pack at ₹5,074.00 divided by 10) = ₹9,336.16. The oil, bought as 1 kg at ₹5,221.00, 500 g at ₹2,620.00 and 100 g at ₹540.00, cost ₹8,381.00. Together that is ₹17,717.16 in wax and oil, so the rough half carries about ₹8,858.58 of material (₹17,717.16 ÷ 2) before jars, wicks and labour.

That is too much to scrap for a cosmetic fault. Treat the rough half in this order:

The process control that stops it recurring

Once the log shows the temperature at which roughness began, turn it into a rule. Set a minimum pitcher temperature for pouring, with a margin above the point where roughness started, and treat it as a suggested starting point to validate over two or three more batches. Then change the workflow so no pitcher can fall below it.

Practical controls that work without new equipment: draw smaller pitchers so each one is emptied quickly; keep the melter on at a hold setting instead of switching it off, and confirm that setting with your thermometer rather than the display; stir the melter gently before each draw so you are not skimming the coolest surface layer; warm the late trays of jars the same way as the first. Never run the melter at its maximum setting as a shortcut to keep wax hot, and remember a large melter pulls a heavy electrical load, so keep it on its own socket.

Keep the log going for the next two or three production batches, but thin it out to four checkpoints (a suggested starting point): candle 1, the candle after each refill, the halfway candle and the last candle. If every top grades smooth and every pitcher reading stays above your minimum, the control is working and the log becomes your routine batch record. If one batch goes rough again, the log will show you whether the rule was broken or whether the rule itself needs a higher margin. Either answer is useful, and neither needs a change of wax or oil. Keep the graded candles from the measurement run as reference pieces, labelled with their pour number, so new staff can see what a smooth, slightly rough and rough top looks like before they start judging their own.

Only when these controls fail across repeated logged batches should you consider splitting the 20 kg into two 10 kg scented halves. Do not buy a second melter until the log has shown you that heat holding, not workflow, is the limit.

Where this page fits

CSI already has guides that cover parts of this question. This page covers the failure investigation for the question above; these go deeper on the rest:

Three things to buy for this job

Frequently asked questions