Archive for the ‘Tolerance Analysis’ Category

Tolerance Analysis: Evaluating a Complex Mechanical Assembly to Minimize Tolerance Risk

By Paul Thomas, M.E.

Thirty-nine dimensions were involved in calculating the vertical offset distance for this project.

[Download a PDF of this white paper.)

Andrews-Cooper was engaged by a medical device manufacturer to evaluate the design of a complex mechanical assembly. The client’s R&D reliability expert was extremely concerned that an important vertical travel distance between an upper working pad and its base might have a large, and possibly unacceptable, manufacturing variation. This variation could potentially cause improper operation during use.

The client’s manufacturing team was concerned that the component variations would create an overly difficult assembly procedure and require a highly-trained assembly staff specific to this product line. The end result would be a device-specific calibration requiring a high level of technical aptitude on both the assembly line as well as during the course of preventive maintenance or service in the field.

First Step: Understand the 3D CAD Model and Proof-of-Concept Prototype

The first step of the analysis was to understand how the device worked from its 3D CAD model and the proof-of-concept prototype provided – put simply, “what drove what?” in the assembly. Once the contributing components were isolated, the drawings were pulled and evaluated in-depth to determine the tolerance stack up for all contributing features.

The investigation showed that 39 dimensions were involved in calculating the vertical offset distance. We understand the method of adding and subtracting nominal dimensions and adding plus/minus tolerances. However, this mechanism was highly complex and included some interesting and unique challenges.

Challenges: Dimensions, Directions, Orientations and Mechanisms

These challenges are outlined below and the steps taken to address these aspects during the analysis are explained in detail.

  • Some dimensions had no tolerances documented on the component drawings.
  • The directions of forces effected clearances in the assembly.
  • A toothed coupling had “n” teeth giving it “n” discrete angular orientations.
  • Part of the assembly included a long ball screw drive mechanism.

For the purpose of this analysis, the dimensions without tolerances were assigned industry standard tolerances. This provided a starting point for a preliminary answer. The missing tolerances would need to be applied to the specifications of the various features going forward.

Click to enlarge image. 


Further Evaluations

Clearances were evaluated between mating parts, depending on the directions of forces applied to the mechanism. During the phase of the cycle under analysis, certain portions of the mechanism were in compression while others were in tension. The clearance moved from one extreme to the other depending on whether a particular joint is in tension or compression.

The mating halves of the toothed coupling were then evaluated. The coupling had teeth with sharp edged tips. You can imagine that the teeth could theoretically mate on their infinitely small sharp tips. However, if this were the case the teeth would then either slip clockwise or counterclockwise depending on the cycle direction. The worst-case tolerance was analyzed and this included, plus or minus half the tooth pitch in angular tolerance.

Worst-Case Analysis

Although the ball screw was relatively long, it was tightly toleranced. The ball screw’s effect on the overall assembly tolerance was minimal in both cycle directions.

A worst-case analysis revealed that the vertical distance between the upper working pad and its base was 167.2mm ±6.1mm, the RSS value had an even tighter tolerance range. 6.1mm was determined to be an acceptable tolerance range for operation. This analysis also provided the client with confidence that the assembly procedure could be completed by trained technicians and did not require extensive or excessive constraints.

Click to enlarge image.