The Minimum Thickness on a Certificate Is the Lowest of Several Means

The source is ISO 2064:1996, Metallic and other inorganic coatings, definitions and conventions concerning the measurement of thickness, third edition, confirmed as current in 2024. ISO lists it at two pages of content, and its own foreword attributes it to ISO/TC 107, subcommittee SC 2, methods of inspection and coordination of test methods. It is the vocabulary that the coating standards on this site borrow.

Why the minimum cannot be a point

The introduction states the problem it is solving and then gives the reasoning, which is the most useful part of the document.

With some methods, the microscopical one for instance, it is possible to detect appreciable variations in thickness across extremely small areas, for example pits or cracks, which might be considered as places where the specified minimum thickness has not been achieved. With other methods, coulometric or various non-destructive ones, such minute local variations in thickness cannot be detected.

So:

Read that carefully, because it inverts the usual assumption. The definition was not chosen to describe the coating accurately. It was chosen so that a microscope and a magnetic probe would give the same verdict on the same part. Our own way of putting it, which is not the standard’s wording, is that the definition is set by the least sensitive instrument that anyone is allowed to use.

The consequence is stated in the next sentence. The mean of the measurements on the reference area should be taken as the local thickness, especially with non-destructive tests. A pit narrower than the reference area disappears into an average by design, not by oversight.

The square centimetre

The area itself is specified, and it is small enough to picture. For an article whose significant surface is greater than one square centimetre, the local thickness shall be determined within a reference area of approximately 1 cm², where possible a square of one centimetre side. Within it, up to five distributed measurements may be made, depending on the method, and the number shall be agreed between the parties concerned.

For an article smaller than that, the rule changes. The normal reference area shall be the whole of the significant surface of the article, and again the number of single measurements shall be agreed between the parties concerned. Smaller reference areas are allowed in special cases, but their sizes, number and location must also be agreed.

That phrase appears twice in a two page standard. The document that exists to define these words hands two of the decisions straight back to the buyer and the supplier.

The microscopical method gets its own rule: at least five distributed measurements along a length of a specified microsection. A footnote suggests the length may conveniently be 5 mm for electroplated coatings, and 20 mm for coatings that tend to be more uniform, giving anodic oxide as its example.

The measuring area is the instrument’s footprint

Definition 3.2 is the one that explains why two laboratories can measure the same part honestly and disagree. A measuring area is the area of the significant surface over which a single measurement is made, and the standard then defines it separately for each family of method.

Every one of those is a property of the equipment rather than of the part. Change the probe and you change the question being asked. That is also why the standard is careful to say the number of readings inside a reference area depends on the measuring method used.

When local and average become the same number

Clause 5 covers average thickness, and on small parts the two terms collapse into one. If the significant surface does not greatly exceed the reference area, the value of the local thickness shall be taken as the average thickness. If it markedly exceeds it, the average is the mean of three to five local thickness determinations distributed over the significant surface.

For the mass-loss route, the rule is the one a fastener buyer will recognise. If the significant surface of a coated article is less than the required minimum measuring area, a number of individual articles shall be selected to provide the measuring area required, and the result shall be regarded as the average thickness. If it does not greatly exceed that area, a single determination is the average for the article, and at least two articles shall be measured in order to verify the accuracy of the measurement.

So a thickness figure for a small fastener is very often an average over several parts, and the standard says so at the root rather than leaving each product standard to invent it. Which is exactly what the sherardizing standard does with its ten square centimetres.

The note about batches

One note in the definitions is a statistics requirement in disguise. For components coated in bulk, a product specification may require the average thickness of a bath to be determined. In such cases, the standard deviation must be known to be able to estimate the proportion of the batch that is below the thickness in question.

An average alone cannot answer the only question a buyer of barrel-plated parts actually has, which is how many of them are thin. The standard says as much in a single sentence and does not pretend otherwise.

The clause that explains why the others differ

The introduction ends with an escape hatch, and it accounts for most of the variation this site has run into.

So the base vocabulary tells you in advance that the hot dip family will not necessarily use it as written. That is worth holding next to the four answers the coating standards on this site actually give.

Reading them against ISO 2064, the differences stop looking like inconsistency and start looking like each committee deciding how far to depart from a common definition that openly invites departure.

What to do with it

• Read a minimum as the lowest of several means. Local thickness is defined as a mean inside a reference area, and the minimum local thickness is the lowest of those means on one article
• Ask how many readings, and where. The standard leaves the number inside a reference area to agreement between the parties, twice
• Expect a small part to be measured as a group. Below the required measuring area, articles are combined and the result is an average
• Ask for the standard deviation on bulk-coated work, because the standard says it is what you need to estimate how much of the batch is under thickness
• Check whether the product standard departed from these definitions, since the introduction explicitly says some coatings do

The thing to take from two pages is the shape of the compromise. A specification has to be checkable by more than one instrument, so the quantity it names is built to survive being measured different ways. Everything else about a coating thickness follows from that, and it is stated once, at the front, in a document most people never open.

This page covers step 5, the finish. The whole order is substrate, thread, head, drive, finish, documentation, and why doing it out of order is rework rather than a tweak is in specifying a screw.

Common questions

What does minimum thickness mean on a coating certificate?

The lowest local thickness found on the significant surface of one article, where local thickness is itself defined as the mean of a specified number of measurements made inside a reference area. So the minimum is the lowest of several averages rather than the thinnest point on the part.

How large is a reference area?

For an article whose significant surface is greater than one square centimetre, approximately one square centimetre, and where possible a square of one centimetre side. Up to five distributed measurements may be made within it, depending on the method, and the number is to be agreed between the parties concerned. Where the significant surface is smaller than a square centimetre, the reference area is the whole of the significant surface.

Why is minimum thickness defined over an area rather than at a point?

So that different test methods agree. ISO 2064 notes that a microscopical method can detect variations across extremely small areas such as pits or cracks, while coulometric and non-destructive methods cannot, and concludes that the only practicable definition of minimum thickness is one that allows comparable results to be obtained by any of the approved test methods.

Why do two laboratories measure the same part differently?

Partly because the measuring area is a property of the instrument. ISO 2064 defines it separately for each method: for analytical methods the area over which the coating is removed, for anodic dissolution the area enclosed by the sealing ring of the cell, for microscopical methods the field of view at a specified magnification, and for non-destructive methods the probe area or the area influencing the reading.

Is the thickness on a small fastener measured on one part?

Often not. Where the significant surface of a coated article is less than the required minimum measuring area, ISO 2064 says a number of individual articles shall be selected to provide the area required and the result shall be regarded as the average thickness. Where the surface does not greatly exceed that area, a single determination counts as the average for that article and at least two articles shall be measured to verify the accuracy.

What does the standard say about barrel or bulk coated parts?

That an average is not enough on its own. A note states that for components coated in bulk a product specification may require the average thickness of a bath to be determined, and that in such cases the standard deviation must be known in order to estimate the proportion of the batch that is below the thickness in question.

Why do coating standards define thickness differently from each other?

Partly because ISO 2064 allows it. Its introduction states that for some coatings, such as hot-dipped and sprayed metal coatings, the specifications may call for compliance with a minimum local or an average thickness or both, that these may differ from the parameters defined in ISO 2064, and that the relevant product specifications should be consulted.

References

• ISO 2064:1996, Metallic and other inorganic coatings, definitions and conventions concerning the measurement of thickness. Introduction, clause 1, clause 3 definitions, clause 4 and clause 5

The text used here is the publicly available sample of SIST ISO 2064:1999, the Slovenian adoption, which reproduces the English text of ISO 2064:1996(E); it is not a preview published by ISO itself. ISO lists the standard at two pages of content and the sample carries both, covering clauses 1 to 5, and this page relies on it as printed there. The document’s own foreword attributes it to ISO/TC 107, subcommittee SC 2, methods of inspection and coordination of test methods, while the current ISO catalogue entry lists ISO/TC 107; the document is quoted as printed. The standard is the third edition, at stage 90.93, confirmed, last reviewed and confirmed in 2024, and it replaced the withdrawn ISO 2064:1980, which was not read. The phrase about the least sensitive instrument is this page’s own way of putting the argument and is not the standard’s wording; the standard’s own sentence is quoted in full above. Nothing from ISO 1463 is quoted beyond the length suggestions carried in a footnote of ISO 2064 itself. No relationship is claimed between this standard and any modern measurement method standard, none of which was read for this page. The four coating standards compared in the table are covered on their own pages and nothing is restated from them here beyond the single line each.