My First Candles from a Batch Look Normal but the Last Few Are Oily — Could the Fragrance Be Distributed Unevenly in the Melting Pot?

Two stories behind oily last candles

A pour session is a small production line, and the pot changes along it. Two things drift as you work down it, and both land on the last candles.

Temperature. A pot off the heat, or on a melter whose display is not what the wax is doing, loses heat for the whole session. Each pitcher loses heat faster than the pot, and the last candle of a fill is poured cooler than the first. Wax nearer its setting point is thicker and sets sooner in the jar, and oil that was taken in well at the start tends to separate when the wax is poured too cool: an oily top, a greasy film, oil at the glass (tendencies, not certainties). The worked example on the chart starts at 84°C and, never reheated, finishes at 72°C, well under CSI 464's stated pour temperature of 80–90°C.

Position. If the oil was not fully taken into the wax at the start, or the pot sits unstirred for a long time, the oil may not stay evenly through the pot. Depending on the oil it can drift towards the base or the surface; CSI does not publish oil densities, so we can't say which way yours goes. If it gathers low, the last fills, drawn from the bottom, carry more of it. In the worked example, a last candle that received 15 g of oil instead of 12 g would be at 10.0% of total weight rather than 8%, and a higher load than the wax holds for that oil tends to sweat. That extra 3 g is ₹11.34 of Roasted Coffee: nothing in money, a lot in how the candle behaves.

There is a third, smaller story that deserves its own name: the dregs. The very last candle is often poured from what you scrape out of the pot: the coolest wax, whatever gathered at the base, and anything half-set from the walls. If only the last one or two candles are oily, suspect the dregs before anything else. Whether the scent of first and last candles differs, rather than their surface, is a related question with its own blind test: the first-weaker-than-last diagnosis.

End-of-batch pattern → likely cause → ruled out by

Ranked most common first as editorial judgement; CSI publishes no failure statistics. Read the rule-in and rule-out columns: each is a line you can write on the log.

Rows 1 and 4 are the two halves of your question, and they are the ones a single "usual" session cannot separate, because in a pot that is never reheated, later and cooler are the same thing. That is the whole design problem the log solves: you need at least one session where time keeps passing but temperature doesn't fall. Rows 5 and 6 are cheap to rule out on the same sheet; row 5 looks like the jar-base post's cooling row, seen across a batch rather than in one jar.

The pour-session log, fill by fill

The log costs no extra materials: it is a sheet you fill in while you pour a normal batch. The worked example on the chart is thirty 150 g candles in CSI 464 at 8% (360 g of Roasted Coffee and 4,140 g of wax), poured in 10 fills of three from a 600 ml pitcher. Scale it to your batch; keep the numbering.

Readings are a worked example, not CSI data. Notice what this sheet can and cannot say. It shows the oily candles all came from fills poured below the window, which is enough to change the process: those fills were out of the wax page's range whatever else happened. It cannot yet prove temperature was the cause, because candle 25 is also far down the queue. Oddly, fills 5–7 were below the window too and came out clean in this example, which is a hint that temperature alone may not be the whole story; the held session settles it.

Reading the log: temperature or position

The confirming run is your next ordinary batch, poured as a held session: before every fill, reheat the pot back into the window (80–85°C in the worked example) and log it, but do not restir. Time still passes and the queue still has a front and a back; only the temperature is held. Then read it:

The rule reads the pattern, not a single candle. One oily jar in the middle of a clean held session is noise; look at its jar and its shelf. And the rule never needs you to judge glass by touch: every reading is from the thermometer in the wax, never from how warm a jar feels in your hand.

If your pot is large enough that holding the window is itself the challenge, the cooling curve and pour-window plan for big pots belong to the 20 kg cooling post; this page stops at the diagnosis.

The two fixes, and the one for the dregs

If it follows temperature: reheat and restir between pitchers

Before every fill, read the pot. If it has fallen below the wax page's range, bring it back in, restir from the base, and only then fill. Read the pitcher at its last candle; if it has left the window, the fill was too big or you poured too slowly. For CSI 464 in the 150 g jar that window is 80–85°C: the wax page states 80–90°C and the jar page "Safe for hot pour up to 85°C", and the lower limit governs. Reheat in small steps; don't overshoot to buy time.

If it follows position: fix the mix

The oil needs to be taken in thoroughly at the start and kept evenly through the pot. Stir at the wax page's add temperature for at least the one stated figure CSI has (Eco Soy CSI 400: "stir gently for 2 minutes"), longer for big pots as judgement, sweeping the base and walls where oil and cooler wax collect. Restir the full depth before every fill. For the routine at 5–10 kg, the large-batch mixing post and the 10 kg uniform-batch plan go further; the how-long-to-mix post is the first-level version.

The dregs: plan for them

Weigh your batch so the last sellable candle is poured before you scrape. Pour the scrapings into a tester or a sample for your own burn tests. In the worked example that means planning 31 candles' worth of wax for 30 sellable candles, or accepting that candle 30 is a tester.

Per candle that is ₹45.36 of oil, ₹70.02 of wax, a ₹94.40 jar and a ₹5.90 wick. These are material costs only and exclude labour, freight, packaging and GST. What happens to candles already oily (grade, hold, rework or salvage) is the lot-disposition post's decision, not this page's.

Before you buy a mixer or a melter

If you own none of them, the thermometer, the Large pitcher and the mixer together are ₹1,411.00; the thermometer alone (₹354.00) answers the first and commonest branch in our judgement. The Steel Pouring Pitcher's Small size is out of stock on 26 September 2026, so the table above never offers it.

Five things that make the log work

In the order the log needs them: the thermometer that makes it a log, the numbered pitcher, the mixer for the position verdict only, a second vessel for fills that cool as you pour, and the wax whose page sets the window. Chosen on page statements, price and stock on 26 September 2026; CSI publishes no measured performance data, so nothing is ranked on it.

Frequently asked questions