Power Electronics

Power Electronics — field brand hub

Power Electronics tickets here center on NB 360-class multi-module DC cabinets. Isolation, module-communication loss, cooling-pump contactor chatter, and input phase-loss dominate. Input phase-loss often originates upstream—prove the feeder before a module-bank swap. Name NB360 and whether one module, the cooling plant, or the whole cabinet shares the event.

When this OEM shows up / cabinet vs dispenser

NB multi-module cabinets share conversion stages and often a cooling plant. Ownership:

Cable / ISO side: CableCheck isolation—known-good cable before module condemnation.

Module / share side: Module communication loss / current-share fail (early-access depth)—prove which module and link before full bank RMA.

Cooling side: Pump contactor chatter, flow/pressure, airlock after service (/guides/liquid-cooling-loops).

Grid side: Input phase loss—measure all three phases under load at the cabinet disconnect.

Do not shotgun modules because one phase is dead at the disconnect. Do not ignore contactor chatter as “nuisance noise” if the pump is cycling into a dry or airlocked loop.

How faults typically present

HMI / keypad

• NB series codes (PE-NB-*).
• Module map with one stage offline or not sharing.
• Coolant / pump alarms with audible contactor chatter.
• Phase-loss / undervoltage banners during peaks.

OCPP / service export

• Faulted at CableCheck—/guides/ccs-handshake-basics, /guides/ocpp-faulted-triage.
• Soft derate then Faulted on thermal/coolant paths.
• Pad-wide Faulted at the same minute → feeder/PQ first (/guides/site-power-quality-basics).

Physical cues

• Chatter at pump contactor, glycol weep, warm modules next to a silent peer, single-phase dead at disconnect, recent civil work on the feeder.

Ranked common failure patterns

1. NB360 isolation — Cable/moisture before module bank. Free sample: /entries/power-electronics-pe-nb-iso-nb360-isolation. /entries/protocol-ccs-h03-cable-check-isolation-fail.

2. Module communication loss (early-access / Pro depth) — Backplane/link/share; confirm which module before full bank swap. Entry: /entries/power-electronics-pe-nb-mod-module-communication-loss. Not a free sample unlock.

3. Cooling pump contactor chatter — Control voltage, coil, airlock, low level, or failsafe cycling. Entry: /entries/power-electronics-pe-nb-cool-cooling-pump-contactor-chatter. Airlock: /entries/site-generic-cool-air-airlocked-loop-after-fill. Contactor cousins: /entries/site-generic-ctr-coil-contactor-coil-open.

4. Input phase loss — Upstream breaker, lug, or utility; prove at cabinet under load. Entry: /entries/power-electronics-pe-nb-grid-input-phase-loss. /guides/site-power-quality-basics.

Log sources / tools

Source	What you pull
NB HMI / keypad	Module map, coolant, phase alarms
Service log export / OCPP	StatusNotification, MeterValues, fault history
Clamp / DMM	Phase-to-phase and phase-to-ground under load
Known-good cable	Isolation prove

Field checklist (first 15 minutes)

1. Photo HMI + module map + connector ID.
2. One module, cooling plant, or whole cabinet?
3. Measure three-phase voltage under the failing load.
4. If ISO: dry + known-good cable.
5. If chatter: level/flow/airlock before replacing modules.
6. If phase loss: stop module RMA; call electrician with voltage table.
7. If module-comm: identify which stage; check link before bank crates.

When to escalate

Electrician — Phase loss, lug heat, feeder trips, undervoltage clusters.

OEM (Power Electronics) — Confirmed module-comm after link/grid proven; pump/contactor after coolant procedure followed; ISO after known-good cable fails dry.

Do not escalate phase loss as module communication loss, or isolation after rain as a full bank failure. LMS soft caps can look like dead power—confirm site load management before crates (/entries/protocol-load-01-dynamic-load-management-caps-to-0-a).

Worked examples

Example A — Isolation one gun after cable swap. Known-good cable; dry connectors. Do not pull the module bank for CableCheck.

Example B — One module offline on map; siblings sharing. Module-comm / share path (Pro depth). Identify stage and link before ordering a full set.

Example C — Pump contactor chatters; flow alarm after fill. Airlock or low level. Bleed per OEM; do not run dry “to test.”

Example D — Phase-loss banner; A-B healthy, C dead at disconnect under load. Electrician/feeder. Stop module crates immediately.

False friends

Looks like	Often actually
All modules bad	Single phase loss upstream
Sensor flaky	Pump contactor / airlock cycling
SECC failure	Cable ISO after service
Random Faulted	LMS soft 0 A at peak

Related learning

/learn/site-power-vs-charger-fault · /guides/liquid-cooling-loops · /help

Ticket hygiene

Capture: HMI photo, connector/port ID, peer status, OCPP errorCode + vendor string verbatim, weather/service notes, and (for thermal/grid) a loaded measurement. Scrub customer PII before /request or Pro Q&A. Prefer one branch per ticket—do not merge money-path and HV into one RMA narrative.

Coordination notes

When opening OEM or electrician tickets, attach: HMI photos, connector IDs, OCPP vendor strings, peer-stall status, and any loaded voltage or RF measurements. That evidence pack cuts “replace module” ping-pong. Keep one fault branch per work order.

Platform notes (NB360)

Multi-module current share means one sick module can look like “the cabinet is weak” without a hard Faulted on every stage. Contactor chatter on the pump circuit is audible evidence—record a short video for the OEM ticket. Phase-loss banners deserve a three-phase table under load before any crate is opened. Isolation after a cable swap is almost never a full bank failure on day one—prove the gun first. Early-access module-comm rows stay Pro-gated; use free ISO samples for the public branch.