Judging Ceramic Glaze Defects on Chinese Sanitary Ware: Pinholes, Specking, Crazing and the AQL Argument

Glaze defects are the argument that never ends. A container of basins arrives, the buyer opens three cartons, finds a pinhole the size of a pin head near the drain, and sends a photo demanding a credit for the whole lot. The factory replies with a photo of a perfect basin and a note saying the defect is within "normal industry tolerance". Both parties are acting in good faith. They are just working to two different acceptance standards that were never written down.

This is the single most common cause of quality disputes I see on ceramic orders, more than chipping, more than dimensional problems, more than colour. And it is almost entirely avoidable, because glaze defects are visible, countable and classifiable. You just have to agree the classification before production rather than after it.

The defect names matter, because they age differently

Buyers tend to lump everything into "defect", which is fine for a photo caption and useless for a specification. The types that actually matter on sanitary ware and tiles behave quite differently over a product's life, and the acceptance threshold should reflect that.

Pinholes. Tiny craters where gas escaped through the glaze before it sealed. Glaze thickness, kiln atmosphere and the bisque surface all contribute. Cosmetically annoying, and on a food-contact or shower surface they collect dirt. They do not grow, but they will hold a stain.
Blistering or boiling. A raised bubble in the glaze, often with a broken crown. Usually more serious than a pinhole because the glaze shell is thin there, and a bump can expose the body. On a rim or a seat contact face, this is a reject regardless of size.
Specking or iron spots. Small dark dots from iron contamination in the body or the glaze. Common after a material change or when a factory runs faster and skips magnet cleaning on the slip lines. Cosmetic, but a speck on white ceramic is very noticeable, and on a large flat tile it repeats in a pattern that looks like a design error.
Pin-holing at the drain or around the overflow. This one is often a glaze drainage problem in the piece, not a kiln problem, and it is where water sits. Treat it as a functional zone rather than a cosmetic one.
Crazing. Fine surface cracking of the glaze, sometimes appearing months after installation. This is a materials and firing mismatch, glaze versus body expansion, and it is a genuine quality indicator. A factory producing crazed work has a process problem, not an occasional accident.
Running or pooling. Glaze that has flowed heavily in one direction, thinning elsewhere. On flat tiles it can cause a visible sheen difference across a batch.
Lime popping or body defects. Small chips of body erupting through the glaze later, sometimes after installation. These are the ones that generate warranty claims in year two. If you see them in a sample, walk away rather than negotiate a threshold.

The distinction that matters practically: pinholes and specking are cosmetic and can be tolerated in defined zones, at a defined level. Crazing, lime popping and blistering are process indicators and should be rejections at any count. Writing this distinction into the spec is far more useful than a single "no defects" clause that nobody can enforce.

Zones beat counts

The most practical thing I ever did on ceramics was introduce zones. A basin does not have one acceptance standard. It has at least three, and a factory will agree to a strict standard quickly once it understands you are not applying it to the entire surface.

Zone A, the front and top surfaces the guest sees. Inside rim, visible interior of the bowl, the tap deck, the front face. Strict. Typically no visible pinholes above a minimal size, no specking, no blisters.
Zone B, secondary faces. Outer sides, back of the basin, underside of the rim not normally visible. A higher AQL is reasonable here, and it is where you will actually find the pinholes anyway.
Zone C, non-visible surfaces. Base, unseen back, area hidden by the countertop. Defects here that do not compromise strength or hygiene should be tolerated, and a reasonable allowance saves you a lot of pointless rejections.

On tiles, the same idea works differently: define a viewing distance and a lighting condition. A defect that is invisible at 1.5m under normal room light and unacceptable under a raking spotlight are two different specifications, and designers and contractors will use whichever one suits their argument. Put the viewing condition in writing.

AQL is a sampling plan, not a quality level

Buyers routinely say "AQL 2.5" as if it were a description of quality. It is not. It is the percentage of defective units you are willing to accept in shipping inspection lots, given a particular sample size and acceptance number. Changing the lot size changes the practical meaning.

For ceramics on a project order, this is what I would put in the contract rather than a bare AQL figure:

A defined inspection per shipment, or per stated batch where the order is large enough to split. For most 20 or 40 foot container loads, one lot per container is realistic, but if a container is filled from more than one glaze batch, treat it as more than one lot.
An acceptance standard based on zones and defect classes, as above, with a clear "critical defects are zero tolerance" list. Cracks, chipping through the glaze at rim or seat contact surfaces, and exposed body go on that list.
A defined sample size. The general level default of a widely used standard is a useful starting point, but for project orders I would increase the sample rather than tighten the number, because with ceramics the issue is often consistency across the batch, not a random bad unit.
Who inspects. The factory's own QC is fine for routine screening but should not be the deciding party on a project shipment. Use a third party, and give them your zone drawings, not just your AQL number.

A word on critical defects: a cracked basin will eventually fail, and a chipped rim on a toilet seat contact face will chip the seat. Those are not negotiable at any AQL. Write them as zero allowed and inspect 100% for them if the order is going into a hotel or a retail unit that will generate returns.

Colour and glaze consistency between batches

White is a colour, and it varies. Two batches from the same factory can differ enough that a mixed wall looks wrong under a warm light even though every piece is nominally "gloss white".

The practical controls:

Ask for and retain a physical master sample of the glazed finish, sealed and dated, held by both parties. A photo is not a master sample.
Insist that a large order be produced from a single glaze batch where possible, and written into the order confirmation that the batch is reserved at deposit.
For tiles, get the batch or shade code on every carton, and buy 5 to 10% overage from the same batch for breakage and future repairs. A tile from a different batch will not match, and no amount of complaining changes that.
Check colour under the lighting the product will live in. A factory inspection under daylight neon and a bathroom under 3000K warm LED will show the piece differently. If you cannot travel, ask for photos at the specified colour temperature, and treat that as a weak control, not a strong one.

I have seen a project accept a delivery where the floor tiles in one room came from a different shade lot and, under the hotel's lighting, the room looked like it had a rug-shaped patch. That is a repairable problem, but only by re-tiling.

Where ceramic QC actually fails, and it is rarely the inspection

The inspection is usually competent. The failure is upstream, in three places.

The standard was agreed verbally. Someone in the factory's sales team said "our quality is export standard, no problem". That sentence has no content. The production team then makes the product they always make.

The first article was not inspected. The first pieces off a new mould or a new glaze batch should be inspected against the zone drawings and signed off before the run continues. Skipping this is how you discover a systematic pinhole problem in a defect concentrated in one area of the piece after 800 units.

Nobody looked at the packing. Glaze defects are often created in the carton. Poor separators, insufficient foam at the rim, cartons stacked too high, or a pallet that gets corner-dropped at a building site all produce chips that get reported as manufacturing defects. If you are getting an inexplicably high chip rate on arrival, ask for the loading photos and the carton drop test result before you blame the kiln.

A realistic conversation with the factory

What works, in my experience, is not a demand for perfection but a proposal. Something like: front and top surfaces strict, back and underside toleranced, critical defects zero, a master sample held by both sides, batch reserved at deposit, third-party inspection with our zone sheet, and you will pay for the re-inspection if the first one fails. Factories respond well to this because it is an actual workable standard, and because they would rather know up front than argue about a credit six weeks later.

What does not work is the same photo argument, repeated per shipment, in which the buyer escalates and the factory stonewalls, and the relationship slowly degrades until nobody wants to take the next order. Glaze defects are one of the few quality areas where the fix is almost entirely documentary. Write it down once, properly, and the argument goes away.