Experiment Design

Why Experimental Methods:

• Prove design concept before product integration
• Prove/establish reliability
• Verify/create design specifications
• Prove/establish accuracy of math models
• Characterize product/environmental behavior when modeling is not useful/possible

Best Practices in Experiment Design:

• Conduct test case where result is well known
• Repeat experiment multiple times
• Collect many data points spanning the full range of test conditions
• Use multiple instruments
  • Different types of sensors (when available)
  • Multiple sensors of the same type
• Calibrate instruments immediately before experiment if possible
• Simplify experiment as much as possible
• Avoid interpolation beyond experimental range when analyzing data
• Keep detailed notes/data logs
  • Date
  • Ambient conditions (room temp/pressure/humidity etc…)
  • People present
  • Goal of experiment
  • Define independent and dependant variables
  • Sketches/pictures of apparatus
  • Description of instrument
    • Model number
    • Accuracy
    • Short description of instrument
  • Design equations with explanation of variables and units
  • Detailed description of the steps of the experiment
  • Summary of the expected outcome of the experiment
  • Tables of data
    • Columns and/or rows labeled clearly/neatly
    • Label units
  • Write down all observations during experiments
  • Plots of data (computer generated or hand drawn on graph paper with ruler)
  • Use ink (can’t be erased/rubbed off by you or others)