My 2 Kg Test Batch Gives Great Cold and Hot Throw but My 20 Kg Production Batch Smells Noticeably Weaker — Could Mixing Be the Problem?

The mass-balance rule: mixing moves oil, it does not remove it

Your 2 kg test was 1.84 kg of Eco Soy CSI 400 and 160 g of Roasted Coffee, 8% of total weight, poured into ten 200 g jars. Cold throw and hot throw were both excellent. The 20 kg batch was 18.4 kg of the same wax and 1.6 kg of the same oil, a hundred jars, and it smells noticeably weaker. Mixing is the obvious suspect, because the batch is ten times larger and you stirred it by hand.

Start with one piece of logic that settles more than people expect. Stirring does not create or destroy fragrance. It only shares out the 1.6 kg across the hundred jars. If mixing was poor, some jars received less than 16 g of oil and others received more. The average is unchanged. So if every jar you sample is weaker than the test, poor mixing alone cannot explain it — unless oil that never mixed in stayed behind in the melter or vessel instead of reaching the jars. And if you can find some jars that are stronger than the test, mixing moves right up the list.

That is why the answer to “could mixing be the problem?” begins with looking at the spread between jars, not at the average.

Read the pattern before you blame the stir

Pull jars from the first, middle and last part of the pour — three from each position is a suggested starting point to validate — and add three jars from the original 2 kg test or, better, a fresh 2 kg control at matched cure age. Compare cold throw first, then burn coded.

For the cold comparison, keep it simple and fair. Uncap coded jars one at a time, hold each the same distance from the nose, and rest between jars so the nose does not tire. Write a rank, not a description. Two people ranking separately and agreeing on a first-to-last trend is far stronger evidence than one person's impression.

The last row matters. Oil that sinks or clings as a separate layer and stays in the vessel's heel is oil the jars never received. That is the one way poor mixing can make the whole batch weaker, and it leaves a trace: the scrapings and heel from the melter or vessel smell far stronger than any jar. If you have already cleaned the vessel, note that the evidence is gone and rely on the jar pattern.

The composite test: is the oil there or not?

There is a direct way to separate “oil unevenly shared” from “oil missing or changed”. You recombine a sample of the batch and see whether the average is right.

If the production composite matches the control composite, the oil is present in the batch and the weakness you smelt was uneven distribution — the jars you tested happened to be the lean ones. That is a mixing problem. If the production composite is still weaker, the batch average is low: oil was short, left behind in the vessel, or lost intensity to heat. Then the next checks are the leftover oil weight and the temperature log, not the stirring stick.

What to do with the hundred jars

Mixing is the most rescuable of all scale-up failures, because the oil is still there. If the composite test shows the batch average is correct, the jars can be remelted in manageable portions, each portion thoroughly mixed, and repoured. Burn-test the rework at full cure before releasing it, because the second heat cycle is a change of its own. Wicks usually need replacing on remelted jars, so allow for that cost: if your jars use Eco Candle Wicks in the Thin (C1) size at ₹5.90 each, a hundred replacements come to 100 × ₹5.90 = ₹590.00, before the labour of cleaning every jar. Compare that with the value of the batch below before deciding that a write-off is simpler.

If the composite shows the batch average itself is low, remixing will only spread a lean batch evenly. Rework then needs a calculated addition of oil per melt — only possible if you know how much was missing from the leftover bottle weight — or the batch should be downgraded. Guessing an extra percentage into wax that has already lost intensity is the most expensive route of all.

For scale, the materials here are worth about ₹13,602.00: the wax bought as three 5 kg packs at ₹1,996.00 and four 1 kg packs at ₹399.00 (₹5,988.00 + ₹1,596.00 = ₹7,584.00 for 19 kg), plus 1.6 kg of Roasted Coffee as 1 kg ₹3,738.00 + 500 g ₹1,890.00 + 100 g ₹390.00 = ₹6,018.00.

Validating mixing for the next 20 kg run

Write the mixing step as a procedure rather than a duration: how the oil is poured in (slowly, across the surface rather than in one spot), what reaches the bottom and the far wall, and how you know it is done. Then prove it once by sampling. After mixing, draw small samples from near the top, the middle depth and the bottom of the vessel, pour each into a coded jar, and burn them with a control. Whatever stirring time you then adopt is a suggested starting point to validate, not a rule, and it should be re-proved if the vessel, the fill level or the oil changes.

Keep the procedure physical and checkable. A long enough stirrer to touch the vessel floor, a route that includes the base and the wall opposite where the oil went in, and a final lift from the bottom to the surface are all things a helper can be shown and a batch sheet can tick. “Stir well” cannot be ticked. When a helper runs the batch instead of you, the written procedure is what keeps the next hundred jars like this validated set, and the sampling result is what tells you it still works.

Honestly, do not buy a mixer yet. CSI does not sell mechanical mixers, and more importantly, a mixer is only the answer if the composite test and the position pattern both say distribution is the problem. If the batch average is low, a mixer would have mixed a lean batch very evenly. When sampling does point to mixing, look for equipment that reaches the full depth of your vessel and can be run without whipping air into the wax. And if you want a second pair of eyes on your pattern results, send them to CSI on WhatsApp at +91 7397976926.

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