My 1 Kg Fragrance Formula Works Perfectly but the 25 Kg Batch Smells Inconsistent From the First Candle to the Last — What Is Going Wrong During Scale-Up?

What a 25 kg batch actually is on CSI's ladder

Start here, because most makers never check it. CSI's published wax ladder is 500 g, 1 kg, 5 kg and 10 kg, and there is no 25 kg pack of any wax. The one exception is Natural Soy Pearl Wax, which also comes as 50 kg at ₹21,240.00 (₹424.80/kg) — the only wax pack above 10 kg we publish. On everything else, 25 kg = two 10 kg packs + one 5 kg pack, and that has a consequence beyond price.

Now the part that explains the smell. Three bags are three fill events, and nothing on a bag guarantees one production lot. If one behaved slightly differently, a melt containing all three is a blend you have never tested, poured into jars in a sequence — and your 1 kg formula was validated on one bag. Before blaming the fragrance, read whether two lots of the same soy wax can go into one batch and decide whether you qualify lots or not. Either answer is defensible; not knowing which you did is not.

The 20 kg wall — your batch is bigger than any melter we publish

This is the uncomfortable centre of the answer. The largest melting vessel CSI publishes is 20 kg / 20 litres. There is no 25 kg melter, no holding tank and no automated filler, and we will not invent one. If you ran 25 kg in one go, you ran it in something not on this list — usually a pot on a hob or a bucket in a water bath. Neither holds a set temperature and neither lets you stir the bottom properly. That is the biggest change between your two batches, and it is almost never written down.

So the honest design answer is that a 25 kg day is not a 25 kg batch. It is two controlled cycles of 12.5 kg through one 20 kg melter, or two melters in parallel — 20 kg plus 10 kg is ₹25,960.00 + ₹21,240.00 = ₹47,200.00, two 20 kg units ₹51,920.00. Whether that spend is justified is your call, on volumes we will not forecast. But notice what two cycles buy: two shorter fills, each from a thermostatically controlled vessel. Two half-size batches that are each uniform beat one 25 kg batch that is not.

Temperature drift across a long fill

Pour temperature changes how wax crystallises and how well the fragrance stays bound in it. At 1 kg it never shows: the whole kilogram is in jars a minute after the reading you took. At 25 kg the first jar and the last are separated by a long fill, and unless the vessel holds temperature, they are different products made from the same pot.

Here is how narrow the published windows are. Premium Coconut Soy Wax CSI 468 publishes a pour temperature of 58–60°C and a melting point of 50–53°C — a two-degree window, with setting beginning a few degrees below. CSI 464 publishes a pour temperature of 80–90°C, ten degrees of room. Luxury Soy Wax Chunks publishes a recommended pour of 75–80°C. The other thirteen waxes publish no pour temperature at all — so there the window is one you set yourself at 1 kg, and if you never wrote it down you cannot tell whether the 25 kg batch stayed inside it.

The pitcher gets ignored. A 600 ml Steel Pouring Pitcher at ₹708.00 or a Candle Wax Pouring Jug at ₹531.00 holds a small fraction of a 25 kg batch, so you refill it many times across one fill — and every decant is a cooling event. The wax in the pitcher is colder than the wax in the melter, and the gap grows as the room pulls heat out of thin steel. Measure both: pen thermometers are ₹354.00, so two cost ₹708.00 and remove the largest blind spot in the process.

What you must not do is take a cooling rate from a supplier, us included. CSI publishes no cooling rate, no holding-temperature figure and no heat-loss data for any wax or melter — which is why the curve above carries no numbers. The number you need is on your own thermometer: note the melter temperature at the first jar and again every twenty jars, and plot it. If it leaves your window, you have your answer without changing a gram of anything.

Mixing, stratification and the minute count nobody publishes

Two listings in our own catalogue give a stirring instruction. Eco Soy CSI 400 says to "add your fragrance or essential oils at 85°C (185°F), stir gently for 2 minutes, and pour directly". Luxury Soy Wax Chunks says to "add fragrance at 80-90°C and stir for 2 minutes for full incorporation". Read those carefully: neither states a batch size. They are written for the scale a person works at with a jug and a spoon, and nothing in this catalogue states a mixing time for a 25 kg batch — not for 25 kg, not for 10 kg, not for any mass. Anyone saying two minutes is enough for 25 kg because the bag says two minutes is multiplying something that does not multiply.

It does not multiply because of geometry. More mass in a wider, deeper vessel multiplies the distance a stirrer must move material, and hand stirring fails first at the bottom and corners where flow is slowest. Fragrance oil is also a different liquid from molten wax and need not stay evenly distributed while the melter waits through a long fill. CSI publishes no stratification or settling rate for any oil in any wax, so whether yours separates in ten minutes or never is not a question any listing can answer. Your own samples answer it.

Weighing deserves the same scrutiny. CSI publishes no capacity and no graduation for the Candle Making Weighing Scale at ₹354.00, and there is no second scale in the catalogue — so read the graduation off the scale you own rather than trusting any supplier's description, ours included. Then run the arithmetic: a scale resolving to the nearest 10 g, on a 16 g single-candle dose, is ±62% and unusable; the same scale on 2.0 kg of fragrance for a 125-candle batch is ±0.5% and fine. That is the case for weighing fragrance at batch level — and against using a coarse batch scale for dye.

The top / middle / bottom test — stop guessing, take three samples

Everything above is a hypothesis. This is the measurement that separates them. It costs five jars and five wicks, and it is the only thing here that tells you which fault you actually have. Run it on the next batch, before you change anything.

Two disciplines make this worth repeating. Change one variable per batch — lengthen the stir, shorten the fill and switch melters at once and you will know something worked without knowing what. And keep retained samples: one jar from the start of every batch, one from the end, labelled with the batch number. Next time a customer says the scent is weak, a retained sample is the difference between an investigation and an argument.

The batch record that makes the next batch diagnosable

The reason this page cannot simply tell you what went wrong is that nobody wrote down what happened. A 1 kg test batch holds in your head; a 25 kg batch does not. CSI sells no batch sheet, SOP template or traceability software — this is general practice, not a product. Here is the minimum that makes a batch investigable.

Four temperatures belong on the sheet as separate lines, because they are separate decisions: melt, fragrance addition, first-jar pour, last-jar pour. The ones our listings state are CSI 464 pour 80–90°C; CSI 468 pour 58–60°C with melting point 50–53°C; Eco Soy CSI 400 fragrance addition at 85°C (185°F); Luxury Soy Wax Chunks recommended pour 75–80°C. That is the complete list. Every other temperature in your process is yours to establish and record — including on the thirteen waxes that publish nothing.

One commercial note, because it sets what the sheet is worth. A 125-candle batch on CSI 464 with 2.0 kg of Lavender Fragrance Oil at ₹5,221.00/kg is ₹12,685.00 of wax and ₹10,442.00 of fragrance before jars, wicks or labour — over twenty-three thousand rupees you can neither confidently sell nor discard. Build the sheet from a repeatable mixing and curing protocol, and set the batch size deliberately with the optimal batch size calculation.

Frequently asked questions