My New Silicone Mould Batch Produces Slightly Different Finished Dimensions — Should I Check Mould Consistency Before Accepting a Large Order?

When a few millimetres actually matter

Silicone moulds are cast, and castings vary. Small differences between moulds of the same design are normal, and they are usually invisible in a single candle. They stop being harmless in four situations: when finished candles go into fixed packaging, boxes or sleeves sized to the old dimensions; when several pieces are sold together and must match; when weight drives a burn-time claim on your label; and when a slightly wider pillar needs a different wick to melt properly. If none of those apply to your product, a small dimensional spread may be a record item rather than a rejection.

So the question at goods-in is not whether the moulds are identical. They will not be. It is whether the spread in finished pieces stays inside the limits your product needs, and whether the moulds you receive in the large order will look like the ones you tested.

CSI's pillar mould listings give a nominal size, for example 11.2 × 6.3 cm for the tall Pillar Candle Mould, but publish no dimensional tolerance, wall thickness or shore hardness. Treat any listed size as nominal and measure what you receive.

Separate the mould from the process first

A finished candle's size depends on the mould, but also on how full you pour it, the wax's shrinkage as it cools, pour temperature, and whether a thin wall bulged under the weight of hot wax. If you pour different moulds on different days from different pots, you cannot tell which of those caused the difference. Fix everything except the mould.

What to measure on every finished piece

Keep the list short and relevant to your product. Diameter at two heights shows taper and bulge. Height shows fill and cavity depth. Weight shows the total wax volume and is the most sensitive single number you have. A quick look at the base tells you whether the piece stands flat, which matters for pillar plates and for safe burning.

In this illustration, mould 2 is wider near the base and heavier, and the piece rocks. That pattern points to a bulging lower wall or a distorted cavity, not to your pour, because the other two moulds poured from the same pot came out close to the reference.

Wick hole and centre

On pillar moulds with a wick hole in the base, check where the hole sits on each unit. A hole cast slightly off-centre gives a piece whose wick burns closer to one side, which can tunnel on one side and thin the wall on the other. Push a straight wick through each mould, pour, and after demoulding measure the distance from the wick to the edge at four points around the top. Large differences between those four readings are a mould fault that no amount of careful pouring will correct.

Setting limits you can defend

There is no published tolerance to borrow, so set your own from what the product needs. A suggested starting point: diameter within ±1 mm of the approved reference where the piece must fit fixed packaging, weight within ±2% where your label states a weight or burn time, and a flat base on every piece. Validate those limits against your boxes and your wick tests, then keep them fixed for every future mould delivery.

Worked example: three moulds before thirty

You want 30 tall Pillar Candle Moulds at ₹942.82 each for a regular product line, which comes to 30 × ₹942.82 = ₹28,284.60. Before committing, you buy three: 3 × ₹942.82 = ₹2,828.46, and run the measurement plan above.

Two moulds land inside your limits; one produces a heavier, wider piece that rocks. That tells you something important about the large order: roughly one mould in three might need to be set aside. You now have three sensible choices. Ask the supplier whether the order can be checked or picked for consistency. Place the order and budget time to measure every mould on arrival and hold back outliers. Or, where the product must be rigidly consistent, trial the Polycarbonate Pillar Candle Mould, also ₹942.82, which is rigid and cannot bulge, though it releases and handles differently and needs its own tests.

Whichever you choose, keep the three pilot moulds and their measured pieces as references. When the thirty arrive, the new pieces are compared with those, under the same pour-by-weight routine, which turns the large delivery into a quick, repeatable check rather than a new investigation.

What you should not do is place the full order on the strength of one good mould. The pilot costs ₹2,828.46; discovering after production that a third of the moulds make pieces that do not fit the packaging costs the packaging, the wax and the rework.

Be realistic about what the pilot can and cannot predict. Three moulds are a small sample, so they tell you roughly how much units vary, not exactly how many in thirty will be outliers. Treat the pilot as a reason to plan an arrival check, not as a guarantee about the rest of the order.

Accepting a mould delivery

Keep decisions per mould, not per order. One outlier does not make the delivery bad, and one good unit does not make it good. If the moulds are from CSI, send your measurement log and photos of the outlier pieces on WhatsApp (+91 7397976926), ask whether units can be checked, and file the reply. A mould that fails your limits stays out of production even when it came from us.

Keeping dimensions consistent after approval

Moulds change as they age. Walls can soften and bulge more with repeated heat and fragrance exposure, so a mould that measured well at goods-in can drift. Add a simple routine: every so often, as a suggested starting point once a month or every few dozen pours, measure one finished piece from each numbered mould against your limits. Retire or reassign moulds that drift, and note the pour count when they do. That record becomes your best evidence the next time you judge a mould delivery.

Where this page fits

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

Three things to buy for this job

Frequently asked questions