Tritium

Tritium — field brand hub

Tritium field work splits hard between PK350 (liquid-cooled ultra-fast) and RTM75 (air-cooled mid-power). Sites often mix them under one CPO; treat cooling architecture as a first-class branch in every ticket. Isolation, cable-leak, and thermal paths are not interchangeable — name the platform before you pull a rectifier.

When this OEM shows up / cabinet vs dispenser

PK350 deployments typically present as a power/rectifier section + dispenser UI with liquid-cooled cables. Isolation and leak-detect faults often track the cable assembly or quick-connects, not the whole power stage. RTM75 is air-cooled — fan path, inlet filters, and GF under weather dominate; do not apply PK350 bleed procedures to RTM.

Dispenser / UI ownership: no-boot HMI, auth timeouts / empty local list, cellular failover flapping.

Power / rectifier ownership: rectifier stage fault, isolation (PK), ground fault (RTM).

Cable / thermal ownership: liquid-cooled cable leak detect (PK), gun temperature derate (PK), inlet fan blocked (RTM).

Internal link: RTM internal CAN bus error — prove harness/termination before power-stage RMA.

If one dispenser faults and the peer stays Available, chase cable/gun/ISO first. If the whole unit Faulted with cooling codes, stay in the CDU / cabinet cooling path. Ultra-fast sites also expose feeder limits when multiple PK units ramp together — measure power before condemning modules. Dispenser UI dark with cellular flapping is a backhaul/UI ticket until proven otherwise.

How faults typically present

HMI / service LEDs

• PK dispenser UI: fault strings / service LEDs (TRI-PK-*).
• RTM: service LED / local codes (TRI-RTM-*).
• UI no-boot: blank or looping screen while power section may still be alive — check supply and link before HV.

OCPP / Veefil / CSMS

• Faulted / Unavailable with Tritium vendor codes in info.
• Cellular failover flapping: Heartbeat gaps, reconnect storms — use /guides/ocpp-faulted-triage.
• Auth timeout with empty local list: offline auth policy / CSMS sync, not a gun fault (/guides/fleet-cms-triage).
• Soft derate long before hard Faulted on thermal paths — pull MeterValues curve.

Physical cues

• Glycol smell, puddles at holster, hot cable near handle on PK350; blocked RTM inlets after storms; water at glands on outdoor RTM.

Ranked common failure patterns

PK350

1. Isolation fault — Cable/gun IR or wet path before rectifier condemnation. Prove with known-good cooled cable. Entry: /entries/tritium-tri-pk-iso-pk350-isolation-fault. CableCheck stage: /entries/protocol-ccs-h03-cable-check-isolation-fail, /guides/ccs-handshake-basics.

2. Liquid-cooled cable leak detect — Quick-connects, jacket damage, residual after swap. Stop sessions; do not “top off and ignore.” Entry: /entries/tritium-tri-pk-cable-liquid-cooled-cable-leak-detect. Loop procedure: /guides/liquid-cooling-loops. Follow OEM fill/bleed after any cable swap — airlocks fake flow faults (/entries/site-generic-cool-air-airlocked-loop-after-fill).

3. Dispenser UI no boot — Supply, board, or link; prove backhaul/power to UI. Entry: /entries/tritium-tri-pk-disp-dispenser-ui-no-boot.

4. Cellular failover flapping — SIM/APN/antenna/RSSI in metal cabinets; RSSI OK ≠ stable websocket. Entry: /entries/tritium-tri-pk-net-cellular-failover-flapping. Heartbeat path: /entries/protocol-ocpp-f02-heartbeat-timeout-to-csms, /entries/protocol-net-c01-dns-resolution-failure-to-csms, /entries/sierra-wireless-mdm-rssi-rssi-too-low-in-cabinet.

5. Gun temperature derate — Coolant flow, sensor, sun load, repeated high-power sessions. Entry: /entries/tritium-tri-pk-gun-gun-temperature-derate.

6. Rectifier stage fault — After isolation, cooling, and grid proven. Entry: /entries/tritium-tri-pk-pwr-rectifier-stage-fault. Grid check: /guides/site-power-quality-basics.

RTM75

1. Ground fault — Rain, wash-down, boot seals, or true PE open. Entry: /entries/tritium-tri-rtm-gf-rtm-ground-fault. PE: /entries/protocol-gnd-01-ground-continuity-fail-at-dispenser.

2. Inlet fan blocked — Filters/debris; clear inlet before sensor swaps. Entry: /entries/tritium-tri-rtm-fan-inlet-fan-blocked.

3. Auth timeout — local list empty — Offline list / CSMS sync after network outage. Entry: /entries/tritium-tri-rtm-auth-auth-timeout-local-list-empty. Related: /entries/protocol-ocpp-f03-authorize-rejected-invalid-idtag.

4. Internal CAN bus error — Harness, termination, moisture at connectors. Entry: /entries/tritium-tri-rtm-can-internal-can-bus-error.

Cross-platform

• Handshake / latch before current with healthy power stage: /entries/protocol-ccs-h01-ccs-slac-timeout, /entries/protocol-ccs-h02-sdp-failed-secc-not-responding, /entries/protocol-ccs-h04-precharge-voltage-mismatch.
• Stuck Faulted after reboot: /entries/protocol-ocpp-f01-statusnotification-stuck-faulted-after-reboot.
• LMS caps that look like hardware failure: /entries/protocol-load-01-dynamic-load-management-caps-to-0-a.

Log sources / tools

Source	What you pull
Dispenser UI / service LEDs	Live code, boot state
Tritium Veefil / CSMS	Event log, FW, SIM/modem stats
OCPP export	Heartbeat gaps, StatusNotification, Authorize, MeterValues
Coolant tools (PK)	Level, pressure/flow, leak dye/visual
IR / DMM / megger	Gun heat, PE, IR, phase
Known-good cooled cable (PK)	Leak/iso moves with cable

For cellular tickets: RSRP/RSRQ, APN, antenna connector seating, and whether Ethernet failover exists. Never run a coolant pump dry to “test.” Photos of leak residue and holster seals (scrub PII).

Field checklist (first 15 minutes)

1. PK350 or RTM75? (cooling branch)
2. One gun or whole unit?
3. Leak smell/puddle? If yes, stop and contain — do not keep testing.
4. Recent cable/coolant work?
5. OCPP snippet for the fault window
6. Known-good cable swap if ISO/CCS and dry enough
7. Feeder voltage under simultaneous ramp if multi-unit site

When to escalate

Electrician / site power — True PE open or feeder events correlating with GF / rectifier trips; supply issues to dispenser UI (separate from rectifier HV); utility/transformer limits under multi-PK ramp.

Network / CSMS ops — Failover flapping with modem stats pointing at APN/SIM/DNS/TLS; empty local auth list after prolonged offline — policy/sync, not hardware; wrong OCPP identity after board/SIM swap.

Tritium OEM — Rectifier stage with isolation + cooling + grid cleared; confirmed cooled-cable leak path needing OEM cable/RMA; UI board no-boot after supply and link proven; internal CAN after harness/termination documented; persistent cooling after bleed + filter + leak-free loop.

Do not escalate UI no-boot or cellular flap as a rectifier fault. Do not treat RTM GF like PK cable-leak coolant procedure. Payment-only failures belong to the billing stack (/entries/ingenico-pay-to-payment-terminal-timeout).