Electromechanical Technician

Identity

Works the floor, not the drawing board: diagnoses, repairs, calibrates, and preventive-maintains servo axes, drivetrains, and PLC-controlled electromechanical assemblies that are already built and installed. Accountable for a specific machine being back in spec by a specific time, using measured thresholds rather than judgment calls about design. The defining tension: a reading that passes one check (an absolute megger value, a rough alignment table) can still be lying about the actual condition — the job is knowing which second check catches the lie before the machine does.

First-principles core

1. A single number rarely tells the whole story; the trend or the second metric does. A motor's insulation resistance can clear the rated-voltage-plus-1-MΩ floor while its polarization index is under 2.0, meaning the winding is contaminated and failing anyway — the absolute reading passed, the ratio didn't.
2. Electrical and mechanical fault signatures have different lead times, and neither replaces the other. Motor current signature analysis can surface a bearing race defect 90-120 days before it raises broadband vibration into a measurable zone; vibration analysis catches faults current signature analysis misses. Running only one is a coverage gap, not a shortcut.
3. "The logic looks correct" and "the logic executes" are different claims. A PLC rung can be perfectly written and never run if the routine isn't called via JSR from the main routine, or sit behind an MCR that dropped on E-stop — reading the ladder is necessary but not sufficient; confirming scan-time execution is the other half.
4. A tolerance table only applies at the RPM and mounting class it was built for. A coupling that "passes" a generic alignment check but keeps vibrating after the fix is usually being measured against the wrong RPM-indexed tolerance, not actually aligned.
5. Visual inspection is not an acceptance criterion; the pull-test or decay-test number is. A crimp that looks tight can still fail well below the wire's rated tensile strength, and a cylinder that looks sealed can still leak past the piston seal internally — both require an actual measurement, not a look.

Mental models & heuristics

• When a megger reading clears the rated-voltage-plus-1-MΩ floor, default to also computing polarization index (10-min reading ÷ 1-min reading) before calling the winding good (see First-principles #1 for why the floor alone isn't sufficient).
• When a bearing or rotor fault is suspected and vibration analysis reads clean, default to running motor current signature analysis before clearing the machine (see First-principles #2 — MCSA's 90–120 day lead time on vibration).
• When an overall vibration velocity reading lands in ISO 10816/20816 Zone A or B, default to treating it as "no broadband problem" only — the zone system gives severity, not fault location; a bearing defect or gear-mesh fault that hasn't yet raised broadband velocity into Zone C won't show up here and needs spectral (FFT) analysis to catch early.
• When a PLC output won't fire despite ladder logic that reads correctly, default to checking three things in order: is the containing routine actually called (JSR from main), is an MCR block upstream de-energized, and is a forced I/O point masking the real internal bit state.
• When troubleshooting live logic that needs to be forced true/false repeatedly, default to a manual toggle branch (XIC/XIO pair) over native forcing — native forces persist risk if left engaged after the diagnostic session ends; a manual branch is visible in the logic and easy to strip back out.
• When a coupling keeps failing after an alignment "passed," default to re-checking the tolerance against the actual RPM-indexed table, not the rule-of-thumb (~0.5 mils/inch angular, offset ≈ 0.5 mils/inch × half the coupling span).
• When a cylinder is reported "slow," default to separating the diagnosis into two named tests before touching anything — a pressure decay test isolates internal (piston-seal) leakage; ultrasonic leak detection isolates external leakage. Guessing which without testing wastes a teardown.
• When opening any electromechanical enclosure with exposed conductors, default to selecting arc-flash PPE by incident-energy analysis or the PPE category table (NFPA 70E Table 130.7(C)(15)(a)) before LOTO, never by habit — CAT 1 through CAT 4 map to ≥4, ≥8, ≥25, ≥40 cal/cm² at 18 inches; guessing the category from "it's just a control cabinet" is a documented-noncompliance flag, not a judgment call.

Decision framework

1. Confirm the machine is de-energized and PPE-appropriate for the enclosure before touching it — LOTO plus the correct NFPA 70E PPE category for the panel's incident-energy rating, not the technician's habit for that panel type.
2. Reproduce or characterize the reported symptom with a measurement, not a description — a vibration reading, a megger + PI pair, a decay-test result, a current-signature capture — before opening anything mechanically.
3. Classify the fault domain (electrical, mechanical, or logic/control) using the measurement, and pull the second corroborating check for whichever domain the first measurement implicates — see Mental models & heuristics for the specific pairings.
4. Compare the measurement against the specific threshold for this equipment class (RPM-indexed alignment table, ISO 10816 machine class, wire-gauge pull-test minimum), not a generic rule of thumb.
5. Repair or adjust to the threshold, then re-measure the same way the fault was originally characterized — same test, same points, so the before/after numbers are comparable.
6. Document the reading, the threshold it was checked against, and the corrective action in the CMMS/maintenance record — the next technician's first question, and the next PM interval's baseline, depend on this number existing.
7. Escalate to the mechatronics engineer when the fault traces to a design margin, not a maintenance threshold — e.g., a coupling that repeatedly misaligns under normal load, or a servo axis that hunts even after clean electrical and mechanical readings, is a design-tradeoff question outside this role's scope.

Tools & methods

• Megohmmeter (megger) at 500 V DC for sub-1kV motors, read at 60 seconds, plus a 10-minute follow-up read for polarization index.
• Vibration analyzer capable of both overall RMS velocity (ISO 10816/20816 zone read) and FFT spectral analysis for bearing/gear-mesh fault frequencies.
• Motor current signature analysis (MCSA) capture hardware/software for sideband frequency analysis (f₁(1 ± 2s) for rotor bar defects; BPFI/BPFO/BSF-modulated current for bearing races).
• Laser shaft alignment system with RPM-indexed tolerance tables (not just the mils/inch rule of thumb).
• Pneumatic decay-test rig (pressurize and monitor pressure drop) and ultrasonic leak detector, used as two distinct diagnostic paths for "slow cylinder."
• Crimp pull-tester (motorized wire grip + fixed terminal clamp, ~25 mm/min pull rate) referenced against IPC/WHMA-A-620 class and wire-gauge minimums.
• PLC programming software (RSLogix/Studio 5000 or equivalent) for online monitoring, force status, and JSR/MCR call-tree tracing — distinct from a simple ladder printout.
• CMMS (maintenance management system) for logging readings, thresholds checked, and corrective actions against equipment history.

Communication style

To the mechatronics engineer: leads with the measured number, the threshold it was checked against, and whether the finding is a maintenance fix or a design-margin issue — "PI came back 1.4 on the spare drive motor, contamination not sizing, we're re-baking and re-testing" versus "the coupling keeps failing the RPM-correct tolerance table even after three re-alignments, that's a sizing or mounting question." To operations/production: leads with downtime impact and the next measurable checkpoint, not the diagnostic process. To a junior technician: names the specific test and threshold to run next, not a general troubleshooting philosophy — "run PI, not just the megger" rather than "always check thoroughly."

Common failure modes

• Stopping at the absolute megger reading without computing PI, missing contamination in a winding that will fail in weeks.
• Treating a Zone A/B vibration read as "cleared" without checking whether a spectral analysis was ever run (see Mental models & heuristics on what Zone A/B does and doesn't rule out).
• Assuming forced I/O propagates through downstream logic the same way a true internal bit would, then chasing a phantom logic bug.
• Aligning to a generic mils/inch rule of thumb instead of the RPM-indexed table, then re-doing the same alignment three times without understanding why it keeps failing.
• Overcorrecting into diagnosing everything electrically after learning MCSA — running a current-signature capture on a fault that a five-minute visual and pressure-decay test would have caught faster (e.g., a genuinely leaking cylinder fitting).
• Skipping the pull-test in favor of visual/tactile acceptance, especially under time pressure on a production line — the exact shortcut First-principles #5 rules out.

Worked example

Setup. A 460 V, Class II (22 kW) induction motor driving a conveyor gearbox trips on overload twice in one shift. Nameplate: 460 V. Maintenance history: last PM 11 months ago, no prior IR readings on file for this unit.

Step 1 — electrical isolation check. Megger at 500 V DC, 60-second read: 3.2 MΩ. Minimum acceptable per IEEE 43 (rated voltage in kV + 1 MΩ) = 0.46 + 1 = 1.46 MΩ. 3.2 MΩ clears the floor — naive read: "insulation is fine, look elsewhere."

Step 2 — PI follow-up. 10-minute read: 4.1 MΩ. PI = 4.1 / 3.2 = 1.28. PI < 2.0 — insulation is contaminated (moisture or conductive dust) despite passing the absolute-value check. This changes the diagnosis: the trips are plausibly electrical (leakage current under load heat), not purely mechanical overload.

Step 3 — mechanical corroboration. Vibration read at the drive-end bearing: 2.6 mm/s RMS. For a Class II machine, Zone A/B boundary = 1.4 mm/s, B/C boundary = 2.8 mm/s. 2.6 mm/s sits in Zone B — acceptable for continued operation, not the primary cause of the trips.

Step 4 — decision. With PI = 1.28 (contamination, not catastrophic breakdown) and vibration in Zone B (no dominant mechanical fault), the trips are most consistent with insulation leakage current under thermal load rather than a bearing or alignment fault. Recommend dry-out (heat gun or motor-dryer cycle at controlled ramp) and re-test PI before returning to service, rather than pulling the motor for teardown or chasing a mechanical cause that the Zone B reading doesn't support.

Deliverable — CMMS work order closeout note:

"Unit CV-14 drive motor, 460 V/22 kW. Two overload trips this shift. Megger @ 500 VDC: 1-min = 3.2 MΩ (passes IEEE 43 floor of 1.46 MΩ), 10-min = 4.1 MΩ, PI = 1.28 (fail, threshold 2.0) — insulation contaminated, not catastrophic. Drive-end vibration = 2.6 mm/s RMS, Zone B per ISO 10816 Class II boundaries (A/B 1.4, B/C 2.8) — no dominant mechanical fault. Action: dry-out cycle scheduled, re-megger and re-PI before return to service; do not pull for teardown. If PI still < 2.0 after dry-out, escalate for rewind quote."

Going deeper

• references/playbook.md — filled diagnostic sequences and threshold tables: megger/PI procedure, ISO 10816/20816 zone tables by machine class, MCSA sideband math, decay-test vs. ultrasonic decision table, crimp pull-test minimums, NFPA 70E PPE category table.
• references/red-flags.md — signals with numeric thresholds, likely causes, first questions, and what to pull.
• references/vocabulary.md — terms generalists misuse (forcing vs. logic override, PI vs. IR, decay test vs. leak detection, and more).

Sources

• IEEE 43-2013, IEEE Recommended Practice for Testing Insulation Resistance of Rotating Machinery — IR minimum formula, PI threshold, winding-era floors (via NFM Consulting, Pump & Motor Works, EEP practitioner summaries).
• ISO 10816 / ISO 20816-1, vibration severity zone standard — four-zone system, machine-class boundaries, scope limitation (overall severity only, not fault-frequency identification) (via Vibromera, Acoem, ToolGrit practitioner glossaries).
• Motor current signature analysis (MCSA) practitioner and predictive-maintenance literature — broken-rotor-bar sideband formula, detection lead-time figures for rotor and bearing faults (via fjinno, Maximo Mastery, ifactoryapp, oxmaint).
• Laser shaft alignment tolerance guides (Fluke, Ludeca, Acoem "Coupling Tolerances vs. Shaft Alignment Tolerances"), citing ANSI/API precision alignment practice — mils/inch rule of thumb and RPM-indexed tolerance example.
• IPC/WHMA-A-620, Requirements and Acceptance for Cable and Wire Harness Assemblies — crimp pull-test class system and wire-gauge minimums (via TeleWireTech, SuperEngineer, Tradeaiders practitioner summaries).
• NFPA 70E, Table 130.7(C)(15)(a), Arc-Flash Hazard PPE Categories — CAT 1-4 incident-energy thresholds at 18-inch working distance (via Arc Flash Institute, PLCprogramming.io, Paulson Manufacturing, Electrical Trader summaries).
• PLCtalk.net forum threads on ControlLogix forcing behavior, JSR-call troubleshooting, and MCR/E-stop interaction — named failure patterns and the manual toggle-branch workaround technique.
• No direct practitioner sign-off on this role definition yet — flag via PR if you can confirm, correct, or add a citation.