The Quiet Cost of Just-in-Case Precision: Rethinking Tolerance Specification on UK Machined Parts

Ask ten UK machinists what constrains their capacity and most will eventually mention the same quiet culprit: drawings that demand more precision than the part will ever use. A tolerance of plus or minus 0.01 mm on a bracket that sits behind a cover panel does not make the product better. It adds grinding time, inspection time, scrap risk and a supplier risk premium that ultimately lands on the buyer's invoice. The pattern is common enough to deserve a name. Call it the tolerance tax.

Where the tolerance tax is collected

The cost of an over-specified feature rarely appears as a single line item. It is spread across the manufacturing chain. A tighter tolerance may force a slower feed rate, a second operation, a different machine, a stress-relief step or a specialised fixture. It may require a temperature-controlled inspection room, a CMM programme and a full dimensional report on every part rather than a first article plus sampling. It may also narrow the pool of suppliers willing to quote, which reduces competition and raises the price of the remaining bids.

There is a secondary cost that buyers seldom see: the risk premium. When a supplier is asked to hold a tolerance that the process cannot comfortably achieve, they price the job as though something will go wrong, because eventually it will. Scrap, rework and concession requests follow. Over a twelve-month programme, the accumulated cost can far exceed the difference between a standard machining rate and a precision one.

Fit, function and safety: three questions worth asking

Not every tight tolerance is wasteful. Some features genuinely determine whether an assembly works, whether a seal holds, whether a bearing runs true or whether a safety case stands up to scrutiny. The discipline lies in telling the two categories apart before the drawing is released.

A useful test is to interrogate every dimensional callout against three questions. First, does this feature interface with another component in a way that affects assembly or alignment? Second, does it carry load, seal a fluid, locate a shaft, transmit motion or influence vibration? Third, does it appear in a regulated submission, a safety assessment or a customer-approved specification? If the answer to all three is no, the tolerance deserves a challenge.

That challenge does not require a buyer to become a metrologist. It requires a short conversation with the designer or the customer and a note on the drawing explaining why the value was chosen. Where no rationale exists, a general tolerance is usually sufficient.

Reading the drawing as a cost model

Experienced buyers learn to scan a drawing for the features that drive price before they look at the quantity. Tight positional tolerances on hole patterns, fine surface finishes on non-sealing faces, flatness calls on thin sections and concentricity requirements on parts that are not rotating are all familiar cost multipliers.

Geometric dimensioning and tolerancing is a powerful language, but it is also easy to misuse. A true position callout that is tighter than the mating part requires adds inspection complexity without adding function. Datum schemes that do not reflect how the part is actually located in assembly can force the supplier to hold a dimension that has no bearing on performance. A tolerance that is tight relative to the process capability of the chosen manufacturing route is a specification problem, not a machining problem.

Surface finish is another frequent offender. Ra 0.4 on a cosmetic cover may be justified. Ra 0.4 on an internal mounting face that is never seen and never slides against anything is usually not. Similarly, a full material certificate to EN 10204 3.1 on a commercial-grade bracket may add paperwork and cost without adding traceability value.

A fit-for-purpose specification checklist

Before releasing a drawing package to tender, it is worth working through a short discipline:

• Classify every dimension as critical, functional or general, and apply tolerances accordingly.
• Use a stated general tolerance block for non-critical features rather than individual tight callouts.
• Tighten only the interfaces that affect assembly, alignment, sealing or safety.
• Specify surface finish by function, not by habit, and state the measurement method.
• Request inspection by exception: first article plus agreed sampling, with full reports only where the risk justifies them.
• Ask for material certification only where traceability, regulatory or safety requirements demand it.
• Reference the applicable standard rather than describing the report format in prose.
• Flag cosmetic requirements explicitly so the supplier can price finishing realistically.
• Avoid vague notes such as supplier to advise, which invite conservative pricing.

The early-stage conversation

Most of the tolerance tax is avoidable at the quotation stage. A supplier who receives a drawing with a one-page brief explaining the part's function, the assembly context and the consequence of failure can often suggest where precision matters and where it does not. Good subcontract machinists will flag cost drivers without being asked, but they need permission to do so.

Buyers should also ask what a supplier would change if the volume doubled or if the programme ran for five years. The answer often reveals features that could be relaxed after first article approval, or inspection requirements that could move from every part to a sampled basis. Both are legitimate ways to reduce cost without reducing quality.

What good looks like

A well-specified drawing is not a loose one. It is a document that records where precision matters and why, and that gives the supplier enough freedom to choose an efficient process for everything else. Reviewed annually, with tolerance decisions documented and challenged, it becomes a commercial asset rather than a source of silent expenditure. Precision should be spent where it protects the product, not where it merely protects the drawing.