Methods, Procedures, and Techniques
Methods, Procedures, and Techniques is about doing work the documented way and being able to prove it afterwards: reading the procedure before starting, meeting prerequisites, respecting hold points, measuring repeatably, and recording results at the moment they are taken.
Reading and running a procedure
- Confirm the revision status and that the document matches the equipment and product in front of you.
- Check prerequisites: training, permits, calibration status, consumables, and environmental conditions.
- Identify acceptance criteria and hold points. A hold point stops work until inspection is signed off; a witness point does not stop work.
- Sequence matters: clean before inspection, and tighten in the specified pattern and stages (star or criss-cross, stepped to final torque) to avoid distortion.
- Record readings as taken, in the units of the procedure, with the instrument identifier. Never reconstruct data later from memory.
Repeatability and sampling
Take more than one reading and report the spread. A simple repeatability check is the range of repeated readings on the same item under the same conditions. If the range exceeds the stated limit, suspect the method or the tool before judging the part.
Worked example
Five thickness readings on one sample: 10.02, 10.04, 10.01, 10.03, 10.05 mm. The drawing allows 10.00 ± 0.05 mm with a repeatability limit of 0.05 mm.
- Mean = (10.02 + 10.04 + 10.01 + 10.03 + 10.05) ÷ 5 = 50.15 ÷ 5 = 10.03 mm.
- Range = 10.05 − 10.01 = 0.04 mm, inside the 0.05 mm limit, so the method is stable.
- Limits are 9.95–10.05 mm. The mean is inside, and every reading is inside or exactly at a limit, so the item conforms. Reasoning: had the range been 0.08 mm, the correct response is to investigate technique and the instrument before judging the part.
Traps that cost marks
- Skipping a hold point because production is behind.
- Changing a step without change control.
- Writing results after the shift instead of at the time of measurement.
- Averaging readings in a way that hides an out-of-limit individual value.
- Following a procedure written for a different instrument model.