You should revisit design tolerances when the evidence shows the process is being forced to chase a requirement that is tighter than the product function, risk profile, and manufacturing reality justify.
This does not mean tolerance relaxation is always appropriate. Sometimes the right answer is to improve the process, tooling, fixturing, measurement system, or environmental control. But if a process is already reasonably controlled and the organization is still relying on sorting, repeated adjustment, high inspection effort, rework, or operator heroics to meet the print, then the problem may be the tolerance, not just the process.
The characteristic is functionally non-critical, but scrap, rework, or cycle time is disproportionately high because the tolerance is very tight.
The process is statistically stable, yet capability remains marginal even after realistic process improvements have been tried.
Meeting the tolerance depends on special handling, low throughput settings, excessive inspection, or highly experienced operators rather than normal production conditions.
Measurement uncertainty is a meaningful fraction of the tolerance, making accept or reject decisions noisy or contentious.
Different plants, machines, or suppliers show the same difficulty, which suggests a design requirement issue rather than a single local execution issue.
The tolerance appears inherited from legacy drawings, copy-forward practice, or conservative assumptions rather than current functional analysis.
The downstream assembly, performance, or field data do not show sensitivity at the current limit, even though manufacturing cost and disruption are high.
The requirement is safety-critical, interface-critical, or otherwise strongly tied to fit, function, reliability, or certification basis.
The process is not yet stable, and the current pain is mainly due to assignable causes, poor maintenance, weak standard work, or inadequate tooling.
The measurement system is not trustworthy enough to conclude the tolerance is the issue.
There is a clear and feasible path to improve capability without creating major downtime, validation burden, or disproportionate cost.
The decision should be based on more than operator feedback or one bad lot. At minimum, you need a credible view of process stability, capability, measurement adequacy, cost of poor quality, and the functional importance of the characteristic. In regulated and long lifecycle environments, that review should also consider drawing history, risk analysis, verification logic, and change control impact.
A practical decision sequence is usually:
Confirm the measurement system is adequate for the characteristic and tolerance.
Verify the process is stable enough that capability data mean something.
Separate special-cause issues from structural capability limits.
Check whether the tolerance is truly linked to product function, interchangeability, reliability, or downstream process needs.
Quantify the cost and risk of continuing to manufacture to the current tolerance versus changing it.
Evaluate the change under formal engineering, quality, and configuration control.
Revisiting tolerances can reduce scrap, lead time, inspection load, and supplier friction. It can also introduce real risk if it weakens fit, fatigue life, aerodynamic performance, sealing, repairability, or interchangeability. Those risks are highly part- and program-specific.
Pushing the process harder can preserve design intent, but it often raises hidden costs: slower feeds and speeds, more downtime, extra inspection, narrower approved equipment windows, more deviations, and greater dependence on tribal knowledge. In brownfield plants, those costs are often amplified by legacy machines, mixed controls, disconnected MES and QMS records, and limited opportunity for long shutdowns.
In mature regulated operations, full replacement of equipment or core systems is often not the practical answer to a tolerance problem. Requalifying equipment, updating interfaces across MES, ERP, PLM, and QMS, revising work instructions, and validating data flows can be more disruptive than the original issue. That is why organizations often need to compare three options honestly: improve the process, revise the tolerance, or accept the ongoing cost through formal controls. None is automatically correct.
If tolerances are changed, the change should be traceable through approved engineering change processes, affected manufacturing instructions, inspection plans, and supplier requirements. If they are not changed, the business should still be explicit about the recurring cost and operational risk of holding the current requirement.
So the short answer is: revisit design tolerances when the process is no longer the main source of variation, the requirement is driving disproportionate operational pain, and there is credible evidence that function and risk would remain acceptable under a revised limit. If those conditions are not met, pushing the process harder may still be necessary, but it should be done with a clear view of cost, control burden, and sustainability.
Whether you're managing 1 site or 100, Connect 981 adapts to your environment and scales with your needs—without the complexity of traditional systems.
Whether you're managing 1 site or 100, C-981 adapts to your environment and scales with your needs—without the complexity of traditional systems.
