My Candle Lids Fit the First Batch of Jars but Are Loose on the Second Batch — How Do Commercial Brands Manage Dimensional Tolerances?

Why a lid that fitted in one batch is loose in the next

Glass jars are formed in moulds, and lids are pressed, turned or moulded somewhere else again. Every production lot carries small differences: moulds wear or are swapped, the glass is a little hotter or cooler when it is formed, lids come from a different tool or a different liner roll. General guidance on mass-produced glass is simple: it varies between batches, so you measure samples from each batch rather than assuming. The same is true of lids.

What makes lids unforgiving is that fit depends on a gap measured in fractions of a millimetre. In our worked example, a 0.5 mm smaller rim is only 0.7% of a 70 mm diameter, far too small to see, yet it roughly doubles the play between rim and lid. Your eye says "same jar"; the lid disagrees.

Four things that can move between batches

1. The rim outer diameter. For a lid that presses or sits over the rim, this is the dimension that grips. It is also the one most listings describe loosely: CSI's 150 g family says "outer rim diameter approximately 7cm", which is useful for choosing a jar and useless as a tolerance.

2. Rim roundness. A rim can measure 70.0 mm one way and 69.6 mm at right angles. A lid then grips on two sides and rocks on the other two. Measure twice at 90 degrees; the difference is your out-of-round figure.

3. The lid itself. A metal lid from a new lot can have a slightly larger skirt; a wooden or bamboo lid can move with humidity (the detail is in the wooden-lid tolerance guide); a glass dome can differ in its seating step.

4. The liner or seal. Many lids rely on a thin liner, a foam disc or a silicone ring. A thinner liner, or none, changes the feel of an otherwise identical lid.

Batch variation also affects inside diameter, fill line and wick behaviour, not just lids. That wider picture belongs to the batch-to-batch variation guide; this page stays with the lid.

The swap test: find out which part changed

Before you blame the glass, find out whether it is the glass. You need three jars and three lids from the old batch (this is why you keep golden samples) and the same from the new one. Try all four combinations and note which feel loose.

The last row matters more than it looks. If the loose jars are scattered through both batches, you do not have a batch problem; you have a design whose fit is too sensitive for mass-produced glass. That is a job for a closure with more give (a lug cap, a clip-top, a lid with a ring seal) rather than for tighter inspection.

Write a jar-and-lid spec card

Large brands write specifications into purchase orders. A small brand buying stock jars cannot dictate the glassworks' tolerances, but it can write down what its own approved jars measure, so that "the same jar" has a meaning. The card below is what we would record. The window column is our judgement for a small brand, not an industry standard: set yours from your own approved samples.

Two notes on the card. First, "opens without tools" matters as much as "holds": a CSI customer's four-star review of our tin containers says a lid was "difficult to open" once the candle was made. Too tight is a fit failure, just a quieter one. Second, the fill line row is on the card because a lid that clears the wax in one batch can touch a wick in a taller-filled batch; the headspace logic is in the headspace guide.

You will need a digital calliper. CSI does not sell one; any hardware or online store does. A steel rule cannot resolve a tenth of a millimetre, and a tenth is the scale this problem lives on.

Approved samples and your fit window

A fit window is only as good as the jars it was set from. The method we suggest for a small brand, as judgement rather than any standard:

Approve on at least ten. Buy one pack of 10 of the jar you want, measure all ten, and hand-test every lid on every jar you can. If all ten pass, the range they span is your approved range. In our worked example that is 69.8 to 70.2 mm.

Widen it a little, then prove it. Add about 0.1 mm each side (69.7 to 70.3 mm) only if you have jars at or near those edges that still pass the hand test. A window is a record of what worked, not a guess about what might.

Keep three golden samples. Three approved jars with their lids, sealed in a labelled box with the spec card. In our example that is ₹283.20 of glass you will never pour. They are what you swap-test against when the next batch arrives, and what you photograph and measure for a supplier conversation.

Re-approve when anything changes. A new colour of the same shape counts as a new jar for lid fit: CSI lists the 150 g lidded shape in clear, red, pink, orange and black as separate listings, and you should approve each colour you use. The same goes for a new lid colour or finish.

Incoming inspection, step by step

This is the routine for every delivery, before any jar is washed, warmed or poured. It is written for a small brand receiving tens to a few hundred jars; importers approving thousands need a formal sampling plan, covered in the import approval guide, and the full visual and dimensional checklist for glass is in the jar QC checklist.

The accept, sort and stop thresholds in step 5 are our judgement for a small brand. The logic behind them is simple: one outlier in ten is ordinary within-batch variation; three in ten means the whole lot has moved, as in the worked example, where 70% of the sample sat below the window.

The batch is already here: liners, gaskets and sorting

You have measured, the swap test says the jars moved, and you have a production date. Five options, roughly in the order we would try them.

1. Sort and pair

Measure every jar in the delivery and put them into two or three bands by rim diameter. Then try lids against each band. Lids vary too, and a lid from the larger end of its own spread may be snug on a small-band jar. Manufacturers call this selective assembly. It is slow, but it rescues a batch without buying anything.

2. Take up the clearance with a liner or gasket

A thin foam or silicone liner, or a silicone ring on a plug-style lid, fills a fraction of a millimetre of play. CSI does not sell liners or gaskets on their own, so source them locally and test them: a liner must not touch the wax or the wick, must not trap heat, and must go on only once the candle is fully cool. Approve the liner as part of the spec card, with its thickness, so the next batch gets the same one.

3. Redirect the out-of-window jars

A jar whose lid is slightly loose is still a perfectly good candle jar. Put those candles into a boxed gift set where the box holds the lid, sell them unlidded at markets, or use them as testers and samples. This is often cheaper than any fix.

4. Change the closure for the next order

If the swap test shows scattered failures across batches, the design is too sensitive. Threaded lug caps and clip-top seals tend to absorb rim variation better than press-on lids. The trade-offs in cost, weight and look between metal, wooden and glass lids are laid out in the lid-material comparison.

5. Talk to the supplier, with numbers

"The lids are loose" gets you sympathy. "Ten jars from the 3 October delivery measure 69.2 to 69.9 mm at the rim; our approved jars measure 69.8 to 70.2 mm; photos attached" gets you a conversation about what can be done. From us, that conversation is on WhatsApp: we can check stock, tell you what we can about the lot on the shelf, and discuss spare lids. We cannot promise that every future lot of a stock jar will match your first one, and no stock-jar supplier honestly can.

CSI lidded jars and what their pages say about fit

On 26 September 2026 CSI's catalogue sells lids only as part of a jar. Below are the in-stock lidded jars most relevant to this problem, with every pack size and what each page does and does not say about the lid. None states a tolerance. For a lidless design, the 200 ml Matte and Frosted family (the 200 ml jar hub) mentions no lid at all.

Three bamboo-lid listings (Glass Jar With Bamboo Lid, Amber Glass Jar with Bamboo Lid, White Iridescent jar with bamboo lid) are out of stock on 26 September 2026; the Green Iridescent jar with bamboo lid (₹150.00) is in stock, and its page gives only "Diameter - 7cm Height - 10cm". Wooden-lid fit is its own subject in I Want to Use a Wooden Lid on My Candle Jar — How Much Tolerance Should There Be Between Jar and Lid Dimensions for Bulk Production?.

What a loose batch costs, and what prevention costs

A worked example using the Clear Glass Jar with Golden Lid (150 gram) at ₹94.40 per jar. The loose share is illustrative, not a CSI figure: plug in your own count.

Per jar, a poured candle has ₹173.86 of jar, wax and wick in it, 1.84 times the empty jar, before the fragrance oil, which is usually the dearest ingredient per gram. That multiple is the whole argument for inspecting before pouring. Costs exclude labour, freight, packaging and GST.

Prevention is cheap by comparison. A sample-approval kit of one pack of 10 and a weighing scale for the fill-line checks is ₹1,298.00; the calliper is a one-off purchase elsewhere; the inspection itself is twenty minutes per delivery.

Fewer, larger orders mean fewer lots

Every separate delivery is a chance for a new lot. The Clear 150 g jar is priced flat per jar: 100 jars cost the same whether you buy them as two orders of 50 or one of 100: ₹9,440.00 for ten packs of 10 over two deliveries, ₹9,440.00 for five packs of 20 in one. There is no discount for buying more at once, but there is a consistency case for it when you know you will use them. Ask on WhatsApp whether a larger order can come from one lot and what stock depth there is; we will tell you what the shelf holds rather than promise a match.

Argued against our own interest, again: do not reorder 200 because 20 fitted. Buy the next 20, measure them against your golden samples, and only then commit. The fit record from two deliveries is worth more than a bigger first order.

Frequently asked questions