Original record-planning diagram: part state, measurement basis and the linked result are separate entries. Not a fixture design, measured customer result or equipment photograph.

In short: agree the acceptance state first

Ask which part state the drawing and acceptance plan require. Record whether the sample is clamped, released or supported in a specified way, and the agreed temperature or cooling condition. Do not treat a fixture check as a released-part result unless the agreed criterion explicitly calls for that check. This is a suggested buyer discussion, not a tolerance specification or an instruction to release clamps from a hot part.

TWI explains that weld-area contraction during cooling can cause shrinkage. TWI, “Distortion — Types and Causes”, publication and web revision dates not stated; accessed 9 October 2026.

Does clamping prove the finished dimensions?

A record should say what was actually checked. TWI notes that restraint can limit movement but increase residual stress. TWI, “What Are the Factors Affecting Distortion?”, publication and web revision dates not stated; accessed 9 October 2026. This does not establish a clamp-release method or acceptable distortion for your part. The responsible welding and inspection personnel must agree safe handling and the applicable acceptance conditions.

For an illustrative tube-frame trial, ask whether mounting-point alignment is assessed in the welding fixture or in the specified inspection support. This is an example question, not a CBOXTEC case or a claim that either state always governs acceptance.

Keep each result tied to its measurement basis

A proposed record should link these fields:

  • Sample ID, drawing revision, inspected feature and datum reference.
  • Required part state, actual state and recorded cooling condition.
  • Measurement method and instrument, result, criterion and reviewer.

Let the responsible inspection specialist define the method and any equipment suitability checks.

If a temperature, support arrangement or acceptance limit has not been agreed, mark it pending. Do not insert a convenient number or relabel an unrecorded observation as a pass. Keep differently conditioned readings as separate entries until the reviewer explains whether comparison is valid. Identify missing evidence before transferring a trial result to production.

FAQ

Should every sample be measured after unclamping?

Not automatically. Follow the applicable drawing and agreed inspection plan. Any release and handling sequence must be determined by responsible personnel, not this article.

How long should the welded part cool?

This article specifies no universal time or temperature. Agree the measurement condition for the actual part and record how it was confirmed.

Does robot repeatability guarantee the part tolerance?

Do not substitute a robot specification for measured part acceptance. Request evidence for the specified features under the agreed conditions; this article promises no CBOXTEC dimensional result.

Discuss the trial acceptance scope with your drawing, critical features and proposed inspection conditions.