Simba GaoMECHANICAL / MECHATRONICS
← All projects

UNIVERSITY OF MICHIGAN / PRECISION SYSTEMS DESIGN LAB

Diaphragm Flexure Fabrication and Testing

I helped manufacture and test a research flexure, with experimental results agreeing with the model within 9%.

My roleCAD refinement, manufacturing and assembly planning, alignment, and stiffness testing within a four-person team. The sponsor supplied the original architecture and FEA.

  • Precision assembly
  • Tolerance analysis
  • Wire EDM
  • Capacitive sensing
Manufactured diaphragm flexure on an optical breadboard with a capacitive displacement probe.
Built flexure and capacitive measurement setup
9%Experimental agreement with the model
7075 / 6061Aluminum alloys used in the assembly
±25 µmSelected drawing tolerance targets

01 / ENGINEERING PROBLEM

Build a precise, testable flexure

The sponsor’s sandwich flexure was intended to resist in-plane motion while allowing compliant out-of-plane travel. Our task was to make that architecture manufacturable, align its thin features, and measure its stiffness. Manufacturing lead times and tolerance limits shaped the final assembly.

02 / METHODS & ANALYSIS

From research geometry to a test assembly

Project figures and hardware
Click images to enlarge

01

Choose fabrication methods around the geometry

Exploded diaphragm flexure showing two folded-beam layers and a monolithic center interconnect.
  1. Thin folded-beam layersWaterjet-cut aluminum 7075
  2. Complex center interconnectMonolithic aluminum 7075 · outsourced wire EDM
  3. Frames and sensor mountsMilled parts; 6061 where strength and availability allowed

I contributed to the CAD and fabrication plan, reserving wire EDM for the complex center piece and using waterjet and milling for the remaining parts. The team balanced precision, cost, and lead time rather than assigning one process to every feature.

02

Align the layers and instrument the load path

Assembly CAD showing dowel pins and gaps for 1 mm alignment shims.
Physical flexure fixture with a hanging calibration weight.

Two dowel pins constrained relative layer motion; 1 mm shims aligned the center interconnect. Probe mounts and loading hooks made the assembly testable. Waterjet beam-gap variation reached about 100 µm, exceeding the intended 40 µm variation and motivating a more precise cutting process.

03 / RESULTS & OUTCOME

In-plane response measured

PHYSICAL PROTOTYPE · THREE-LOAD TEST
Original in-plane force-displacement comparison: measured displacement is greater than sponsor FEA at a given force.
Experiment and sponsor FEA both showed an approximately linear response; the hardware was more compliant.
Within 9%Experimental agreement with the model
3In-plane force–displacement measurements
Out-of-planeAnd rotational stiffness remained untested

The prototype was fabricated, aligned, and tested. Its measured in-plane response followed the FEA trend, with lower stiffness than the ideal model. Assembly compliance and manufacturing variation remained plausible contributors. Out-of-plane and rotational testing remained incomplete.