Arcbound FieldThe EVSE DirectorySubscribe
← Learn

Primer 05 · ~9 minFree

Site power vs charger fault

When undervoltage, phase loss, and LMS caps wear an EVSE mask — and how to prove it.

Many “all chargers died” tickets are feeders, transformers, LMS caps, or utility events. If you start inside a power module, you will ship crates and still be dark at 17:00.

Pattern recognition

Points to site / LMS

  • Multiple brands or cabinets fail in the same minute
  • Idle voltage looks fine; loaded voltage sags
  • Breaker handle mid-trip while HMI only says Faulted
  • Power returns when peak demand ends (LMS/SOC floor)
  • Recent utility, tap, or panel work

Points to EVSE

  • One gun or one cabinet only
  • Peer stalls healthy under the same feeder
  • Fault follows a cable swap or wash-down on that stall
  • Vendor string names isolation/cooling/module specifically

Measurements that matter

  1. Voltage at the cabinet under the failing load profile (idle lies).
  2. Which feeder/breaker owns the pad (one-line).
  3. LMS / dynamic load setpoints and whether they command 0 A — LOAD-01.
  4. Coincidence with BESS SOC floor or fire inhibit on depots.

Write the readings in the ticket. Electricians act on numbers; they ignore “chargers are broken.”

Soft vocabulary traps

Undervoltage may appear as OverCurrent, OtherError, or SuspendedEVSE depending on OEM. Cluster timing beats the English on the screen.

Field action

Use the free guide Site power quality basics. Then decide: electrician for feeder/phase, ops for LMS, or EVSE path for single-stall ISO/GF. Safety boundary: LOTO primer.

LMS and “0 A charging”

Sessions authorize then deliver nothing at peak. That can be legitimate curtailment. Confirm with ops before module RMA. Free sample LOAD-01.

Harmonics and PF trips

EVSE rectifier load can upset sensitive PQ relays. This is an engineering conversation, not a gun swap. Capture PQ events timed with session ramps.

BESS-backed pads

SOC floor and EMS inhibits starve chargers on a schedule. EVSE techs document coincidence; BESS ops own cells. Primer context: EVSE levels.

Evidence for the electrician

  • One-line snippet or panel schedule photo
  • Breaker position photo
  • Voltage table: idle vs loaded, all phases
  • Timestamp aligned to CSMS Faulted wave

Professionals respond to that pack. They ignore “come fix the Electrify stall.”

Practice drill

Next multi-stall outage: force yourself to write site vs EVSE in the first sentence before you open any OEM app.

Worked example: peak-hour cluster

Four stalls Faulted at 17:10; breaker warm; idle voltage 480V; loaded 442V. Site sag. Electrician + LMS ops. Not four cable RMAs.

Worked example: single stall undervoltage code

Peer stalls fine on same feeder under load. Now EVSE input stage or local connection is plausible—still measure at that cabinet before module crates.

Coordination with OEMs

Send site voltage tables with OEM tickets when codes smell like grid. It cuts “replace module” ping-pong.

Utility vs customer-owned gear

Know where the POCO meter and main switch sit relative to your cabinets. Sag on the utility side needs a different ticket path than a loose lug on the customer panel. Capture which side your meter clamp was on.

Soft power limits vs hard faults

LMS commanding 0 A can look like a dead charger in the app while local status is Available or SuspendedEVSE. That is not a Faulted protective event. Confirm LMS schedule and site demand before opening power modules.

Single-phase loss on three-phase DCFC

One lost phase can produce odd OEM codes that smell like “module failure.” Measure all three phases under load. If one is dead at the disconnect, stop OEM RMA theater.

Field action checklist (site power)

  1. Measure voltage at the cabinet under the same load that fails.
  2. Compare peer stalls on the same feeder before blaming one OEM.
  3. Separate LMS soft limits from protective Faulted codes.
  4. Capture phase-to-phase and phase-to-ground tables for the ticket.
  5. Open the free site-power guide when cluster timing matches demand peaks.

Cluster vs singleton decision tree

Cluster at the same minute across stalls → site or LMS first. Singleton with peers healthy under load → EVSE path. Mixed pattern (two of six) → check shared subfeed and LMS groups before shotgun module swaps.

Why voltage tables beat cabinet photos alone

A photo of a Faulted HMI without a loaded voltage table cannot separate site sag from module failure. Spend two minutes with a meter under the failing load. Attach numbers to the ticket. OEMs and your future self both need them. The free site-power guide exists so this step is not optional folklore.

Documenting for the next shift

Leave voltage tables, LMS group IDs, and which disconnect you opened in the ticket. The next tech should not re-discover site sag from scratch. That handoff is part of the field action—not optional admin work.