The Invisible Input: Making Thermal Conditions a Purchasing Requirement for Precision Parts

A precision component can pass every dimensional check on a Friday afternoon and fail the same checks on Monday morning without a single cut being taken. The difference is not the machine, the operator or the drawing. It is the temperature of the part and the room in which it was measured.

In the UK, that is not a theoretical concern. A workshop in the Midlands may open its doors to a cold autumn morning, warm through the day as the sun reaches the bay, and cool again after the shift. A heatwave can push an unairconditioned inspection area well above the reference condition. A cold snap can leave a part below it. For work measured in microns, these are purchasing conditions, not background detail.

The drawing sets the size; the room sets the reading

A drawing defines nominal dimensions and tolerances. It may state a general tolerance, a surface texture and a material. It rarely states the thermal condition at which conformity is judged. The assumption is that measurement happens at the standard reference temperature of 20°C, with the part and instrument in equilibrium. That assumption is reasonable only if someone has made it happen and can show evidence.

Calibration supports the instrument, not the part. A CMM certificate confirms that the machine was assessed under defined conditions. It does not confirm that the component on the table has soaked long enough, or that the floor temperature has remained within an agreed band. A calibrated CMM can measure a cold part very accurately and still produce a misleading result.

Thermal expansion makes the scale of the issue visible. Aluminium expands by roughly 23 micrometres per metre per kelvin; steel by about 12; cast iron by around 10.5. A 100 mm aluminium part measured 8 K away from reference conditions can differ by approximately 18 micrometres before any other error is considered. Where a tolerance is tight, that is not a rounding matter. It is a large share of the available band.

Why a calibration certificate is not a climate certificate

Procurement teams are accustomed to asking for calibration certificates, material certificates and inspection reports. Those documents answer specific questions. They do not answer the question that decides many disputes: was the part in thermal equilibrium with the measuring equipment when the final result was recorded?

A supplier may have a temperature-controlled inspection room and still schedule final measurement immediately after a washing operation, a finishing process or a cold journey across the yard. A part can look stable and still be changing size. Soak time is not a formality. It depends on material, mass and section thickness, and it should be agreed rather than assumed from a generic note on a quality manual.

The practical question for a buyer is not whether the whole factory stays at 20°C all day. It will not. The question is whether final measurement is made under an agreed condition, with a defined tolerance around that condition, and with evidence that the condition was met. That distinction belongs in the purchase order.

Where the purchase order can carry the clause

A drawing is a technical document. A purchase order is a commercial instruction. The thermal condition of inspection sits between the two, which is why it is often missed. The buyer can close that gap with a short inspection-condition schedule attached to the order.

The schedule need not be elaborate. It should name the reference temperature, the permitted deviation, the soak requirement for the material and section concerned, the method used to check part temperature, and the records that will accompany the inspection pack. It should also state what happens if the floor drifts outside the agreed band: measurement stops, the part is re-soaked, and the result is recorded again. For micron-level work, a dispute procedure should say that a contested measurement is repeated after an agreed stabilisation period, not argued from two readings taken under different conditions.

Proportionality matters. A general machined bracket does not need the same thermal evidence as a precision spindle component. The clause should follow the tolerance, the material and the consequence of failure. Over-specifying every part adds cost without adding control. Leaving every part silent adds risk without adding clarity.

The seasonal problem is a scheduling problem

UK workshops do not enjoy a stable climate. Doors open, forklifts move between bays, heating comes on late, and summer afternoons can lift an inspection area beyond its normal range. The part itself may arrive from a cold store, a van or an outside inspection queue. Thermal soak is therefore a scheduling issue as much as a metrology one.

Buyers can ask a simple question during supplier assessment: where in the route does final inspection occur, and what happens to the part immediately before it? If the answer is that the part is measured straight after a wash, a coat of preservative or a move between buildings, the thermal state deserves scrutiny. If the answer is that the part is allowed to stabilise in the inspection area and the temperature is logged, the buyer has something to work with.

Six questions to ask before award

1. Where will final dimensional inspection take place, and at what point in the manufacturing route?
2. What reference temperature and permitted deviation apply, and how is the floor temperature recorded?
3. What soak time is used for this material and section thickness, and how is part temperature checked?
4. Does the measurement uncertainty statement account for thermal effects and environmental variation?
5. What happens if the inspection area moves outside the agreed band during measurement?
6. Will the inspection pack include a temperature and soak record tied to the part or batch identity?

These questions are not a substitute for a metrology plan. They are a procurement check that the plan will be executed under conditions the buyer has agreed. They also give the supplier a fair basis for quoting: a requirement that is known before award is easier to price than a expectation discovered at the point of dispute.

What good evidence looks like

The inspection pack does not need to be a climate report. It needs to be credible. A record of the inspection start and finish temperature, the soak start time, the part temperature check and the instrument used, signed and linked to the serial number, is enough to show that the condition was controlled. If a sub-tier supplier performs final measurement, the same requirement should flow down. A certificate from a distant factory does not describe the room where the part was measured.

None of this replaces a good drawing. It protects the drawing from being judged in conditions the drawing never intended. A calibrated CMM and a precise tolerance are necessary, but they are not sufficient. The buyer who specifies thermal conditions before the first cut is less likely to find the argument after the last measurement.