TechMedixby BCR
TechniciansCertificationsSign In
L3

Curriculum · Level 3 of 5

Specialist Curriculum

~14 hours study + 45-minute exam 5 modules L2 Technician certification + 6 months field experience

L3 Specialist is the multi-platform tier. You diagnose across robot families, work in the frequency domain instead of the time domain, and lead fleet-level investigations. The math gets more serious here, but every concept ties directly to field decisions.

Modules

  1. 1

    FFT & Vibration Analysis

    2.5 hours

  2. 2

    MTBF & Fleet-Level Diagnostics

    2 hours

  3. 3

    FMEA & RPN Scoring

    2.5 hours

  4. 4

    Multi-Platform Diagnostics

    3 hours

  5. 5

    Fleet-Level Correlation

    2 hours

01

Module 1 · 2.5 hours

FFT & Vibration Analysis

A vibration signature tells you which bearing is failing, which gear is damaged, and whether a joint is in resonance. This module teaches you to read a spectrum and calculate bearing defect frequencies.

Learning Objectives

  • Explain the FFT transformation and its limitations (windowing, leakage)
  • Calculate BPFO, BPFI, BSF, and FTF from bearing geometry
  • Identify gear mesh frequencies and their sidebands
  • Distinguish resonance from forced vibration

From time-series to spectrum

The FFT takes a time window and tells you how much energy is at each frequency. Short windows trade frequency resolution for time resolution.

Use a Hann window for most bearing work. Rectangular windows cause leakage that smears peaks across bins.

Sampling rate sets your upper frequency limit (Nyquist = Fs/2). For bearing defect work, sample at 10–20× the highest defect frequency you expect.

Bearing defect frequencies

BPFO (outer race): (n/2)(1 − (d/D)cosφ) × shaft speed

BPFI (inner race): (n/2)(1 + (d/D)cosφ) × shaft speed

BSF (ball): (D/2d)(1 − ((d/D)cosφ)²) × shaft speed

FTF (cage): (1/2)(1 − (d/D)cosφ) × shaft speed

Where n = number of rolling elements, d = ball diameter, D = pitch diameter, φ = contact angle. You will memorize these for the exam.

Field shortcut

Most modern BLDC actuators use a known bearing family. The platform definition in TechMedix contains the pre-computed BPFO/BPFI values for each joint — check there before calculating from scratch.

Resonance vs forced vibration

A resonant peak shifts frequency when you change the operating speed; a forced vibration peak stays at the driving frequency.

Treat resonance by modifying the structure (stiffening, damping). Treat forced vibration by fixing the source (balance, alignment, bearing).

02

Module 2 · 2 hours

MTBF & Fleet-Level Diagnostics

Moving from single-robot to fleet-level changes the question you are asking. Instead of 'what is wrong with this robot', you ask 'why is this model family failing at this rate'.

Learning Objectives

  • Calculate MTBF from a service history and understand its assumptions
  • Build a fleet survival curve from censored and uncensored data
  • Identify failure clusters: same model, same component, same window
  • Know when MTBF is the wrong metric (non-constant hazard rate)

MTBF defined and misused

MTBF = total operating time ÷ number of failures, assuming a constant failure rate. That assumption is wrong for most mechanical components — they fail at a rising rate as they age.

For a young fleet with few failures, MTBF will over-estimate reliability. Always look at the failure-rate curve, not just the headline number.

Building a survival curve

Every robot in the fleet contributes operating time. If it hasn't failed yet, that time is 'censored' — we know it survived at least this long.

Plot cumulative failures vs. operating hours. The shape of the curve tells you whether you are in infant-mortality, random-failure, or wear-out territory.

03

Module 3 · 2.5 hours

FMEA & RPN Scoring

Failure Mode and Effects Analysis is the framework for deciding which risks matter. This module teaches the RPN scoring system and how to use it to prioritize field work and escalation.

Learning Objectives

  • Decompose a system into components, failure modes, and effects
  • Score Severity, Occurrence, and Detectability on a 1–10 scale
  • Compute RPN and set action thresholds
  • Escalate correctly when RPN exceeds the platform threshold

The RPN formula

RPN = Severity × Occurrence × Detectability. Each factor scored 1–10.

Severity: how bad is the outcome if this failure happens?

Occurrence: how often do we expect it to happen?

Detectability: how likely are we to catch it before it causes damage? (10 = invisible, 1 = immediately obvious)

RPN maxes at 1000. Any RPN > 200 triggers mandatory action.

04

Module 4 · 3 hours

Multi-Platform Diagnostics

L3 certification requires fluency across at least four platform families. This module covers what transfers, what doesn't, and how to build a mental model for a new platform quickly.

Learning Objectives

  • Compare actuation strategies across Unitree, Boston Dynamics, Agility, DJI
  • Identify where safety envelopes and failure modes differ
  • Use the BCR platform definition framework to learn a new robot in hours
  • Know when platform-specific training is non-optional

What transfers, what doesn't

Transfers: safety fundamentals, electrical diagnostics, sensor calibration principles, CAN bus reading, FFT analysis.

Does not transfer: joint naming conventions, torque specs, thermal baselines, firmware update procedures, specific failure-mode catalogs.

The BCR platform definition framework exists specifically to make the non-transferring knowledge structured and discoverable.

05

Module 5 · 2 hours

Fleet-Level Correlation

Patterns that are invisible at the single-robot level become obvious when you correlate across the fleet. This module teaches the correlation techniques that surface firmware bugs and design defects.

Learning Objectives

  • Correlate anomaly timing across units of the same model
  • Identify shared environmental stressors (heat, dust, duty cycle)
  • Distinguish random failure from systemic defect
  • File a platform-level escalation with supporting evidence

Three questions that find systemic issues

Did the failures cluster in time? A 48-hour window across multiple units usually points to a firmware rollout or an environmental event.

Did the failures cluster in component? Same joint, same bearing, same sensor across units = design or supply batch issue.

Did the failures cluster in customer? Same site across different models = environmental or operator issue, not a platform defect.

References

  • ISO 10816 — Mechanical vibration evaluation

    International reference on vibration severity zones.

  • BCR Platform Definition Framework

    Taxonomy for mapping a new platform into the BCR diagnostic system.

  • FMEA Handbook, SAE J1739

    Industry-standard reference for failure mode analysis.

Ready to test?

Take the L3 Exam

5 questions · Passing score varies by level · Results emailed instantly

Back to L3Start Exam