Lots of emphasis has been made in conveying the tolerance specification on a part. However, little has been done in terms of procedures to arrive at tolerances on a part. More time spent in arriving at tolerances lessens the risk of re-work and revision of tolerances and their associated costs.
Industries, in general, follow the design-build-test-redesign process to arrive at optimal tolerances for desired quality objectives. This extends the product development time and is fraught with undesirable cost escalation.
Simple steps to arrive at part tolerances are provided as under:
- Define Quality Objectives (DFQ) in measurable terms
- Classify Quality Objectives into two categories, namely, Performance and Assembly
- Perform Design Failure Modes and Effects Analysis (DFMEA) and classify the Quality Objectives as Critical, Essential and Desirable, based on the Risk Priority Number (RPN)
- Perform Design for Assembly assessment in alignment with DFQ and arrive at assembly sequence
- Select functional features involved in assembly with upstream part/assembly, as datum along with required datum precedence – reflecting assembly sequence
- Identify features of the part that influence the DFQ objective and specify tolerances based on Fit and Process Capability. Rest of the part features would be defaulting to general tolerances provided as company standard tolerance table
- Develop the cause-effect vector loop for the DFQ objective and see if the specified tolerances meet the same
- If the Tolerances meet the DFQ Objective, then increase the tolerances until Cost of Precision over-rides cost of Poor Quality
- If the Tolerances do not meet the DFQ Objective then reduce the same until compliance to DFQ objective is achieved.
- Perform Cost of Precision Vs Cost of Poor Quality analysis and justify the tolerances provided.
The 10-step process enumerated above seems too long and arduous to comply with, especially in the amount of time and/or effort required. In consideration to the time/effort/costs involved in Re-Work, Scrap, Recall, and the painful process of revisiting the tolerances after product release, the effort is worth the time spent. When the products do not meet the quality standards either on the shop floor or at the customer place, the rehashing of tolerances and re-work goes through the exact same steps as given above, albeit painfully.
DFQ Process is shown as under:
Define Quality Objectives (DFQ) in measurable terms
If a toleranced dimension cannot be measured, then the same should not find place on the drawing. Assembly tolerances are best examples of DFQ Objectives. Gap, Clearance, Assembly Run out are some examples of DFQ Objectives. An example of a Gap Measurement is shown below
Classify Quality Objectives into two categories, namely, Performance and Assembly
In the figure shown below, the pulley is mounted on a shaft supported on bearings. The vibrations at the bearing support need to be under control for the system to have a trouble-free life.
In this example, location of the centre of mass of the assembly that is rotating, about the instantaneous axis of the assembly, is important to control vibrations. This is a classic example of Performance Objective.
DFMEA & RPN Classification
Design Failure Modes & Effects Analysis forms the heart of the Design for Quality process. Unfortunately, this is considered as a non-value added practice, leading to either out-sourcing of the activity or adopting a cut-and-paste approach to completing this document. The RPN (Risk Priority Number) determines the critical nature of the failure mode in classifying the influencing toleranced dimensions as Critical, Essential or Desirable. A Sample example of DFMEA is shown as under wherein one can see the influence of tolerances for the failure mode enumerated therein.
Design for Assembly – Criteria for Tolerance Selection
Normally most of the industries use Cycle time reduction as the criteria for determining Assembly process. If the assembly process is not in alignment with the DFQ objective, disaster is waiting to happen. Assembly sequence directly influences the product quality. It is common knowledge that simplified assemblies outweigh simplified part designs. It is important to note that a part cost may be higher, but if the assembly cost is lower, with fewer things to go wrong in assembly, it is bound to succeed. This has been missed out by many industries leading to disastrous product launches, assembly-line ‘experts’ who perform selective assembly based on experience, untimely product phase out and significantly reduced profitability. ‘Final Assembly is the moment of truth’ – by Charles M. Fine (MIT) comes to mind while espousing the cause of Design for Assembly.
More to come in the next article in this series…