I’m Producing 1,000 Candles and Small Weighing Errors Become Kilograms of Material — How Do I Establish Scale Checks Before Each Production Run?

Why a tiny scale error becomes kilograms at 1,000 candles

When you made twenty candles, a scale that read a gram or two off did not matter much. Each candle was weighed separately, the errors wandered up and down, and the order was too small for the total to add up to anything. At 1,000 candles the arithmetic changes because a scale that is slightly biased repeats the same error in the same direction on every weighing.

Start with the order itself. A thousand 200 g candles is 200 kg of scented wax. At 8% of total weight that is 184 kg of wax and 16 kg of oil. If you run it as 25 kg batches, that is eight batches of 23 kg wax plus 2 kg oil. Each batch involves at least one oil weighing, one or more wax weighings, and 125 fill weighings before losses. Every one of those is an opportunity for the same bias to repeat.

None of these is dramatic on its own. Together they explain why a formula that was perfect in a 1 kg test can arrive at the end of a large order with the wrong yield, a slightly different load, and a costing that does not reconcile.

Worked example: one biased oil scale across eight batches

Suppose the order uses CSI 464 scented with Woody Oud. The oil is weighed on a bench scale that, unknown to the team, reads 1% high at 2 kg. When the display shows 2,000 g, only 1,980 g is in the jug.

Each 25 kg batch then contains 23,000 g of wax and 1,980 g of oil. The true total is 24,980 g, and the effective load is 1,980 ÷ 24,980 = about 7.93% rather than 8%. Across eight batches, 8 × 20 g = 160 g of oil that the batch sheets claim went in never did.

Flip the bias the other way. If the scale reads 1% low, each batch gets 2,020 g and the order uses 160 g of extra oil. Woody Oud at ₹4,531.20 per kg is ₹4.5312 per gram (₹4,531.20 ÷ 1,000), so 160 g costs 160 × ₹4.5312 = ₹724.99 that nobody budgeted. The fill-scale case is larger still: 2 kg of scented wax at 184 g of CSI 464 and 16 g of Woody Oud per 200 g candle is 10 × (184 × ₹0.5074 + 16 × ₹4.5312) = 10 × ₹165.86 = ₹1,658.61 of material.

The pre-run scale check, step by step

This check takes a few minutes and should happen at the start of every production run, before the first weighing. The limits below are suggested starting points to validate against your own scales and tolerances, not fixed rules.

What counts as a check object

A check object is anything stable that you keep only for this purpose: a sealed container of sand, a steel pitcher filled and taped shut, a metal block. Weigh it on a scale you trust on the day that scale was last verified, write the weight on the object, and store it where it will not be chipped, wetted or used for anything else. It will not tell you whether your scale is absolutely correct, but it will tell you immediately if today's reading differs from the last verified one. Certified calibration weights answer the absolute question; CSI does not sell them, and whether you need them is a separate decision.

When a check fails, and what to do with a run already weighed

A failed check is useful information only if it stops the run. The instinct to carry on and adjust later is what turns a scale fault into a formula fault.

Keep the scale log beside the batch sheets. When a batch later behaves differently, the first question an investigation asks is whether the scale was checked that day. A signed pass for that run eliminates one variable in seconds.

Make sure the scale is the right scale before checking it

A daily check cannot rescue a scale that is wrong for the job. Weighing 23 kg of wax needs a platform scale with a capacity comfortably above that load, which CSI does not sell. Weighing 2 kg of oil needs a scale whose reading steps are fine enough that one step is a small fraction of your tolerance. The CSI scale's capacity and resolution are not published, so verify both on your own check objects before assigning it a job.

Location matters as much as the instrument. A scale placed beside a running melter sits in warm, moving air and on a bench that may vibrate when people lean on it. Give each production scale a fixed spot away from heat sources, fans and doorways, on a rigid surface, and do the pre-run check in that spot. A scale checked in one place and used in another has not really been checked.

Assign each scale a role and a label: wax, oil or fill. Checking three scales takes a few extra minutes, but it means every critical weighing in the order has a logged pass behind it.

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