Skip to content

Rev B1 ECN-15 — Battery connector polarity (GND on pin 1) (draft)

Draft — EverTag Station battery variants — reverse-polarity risk

Reverse the battery header pin assignment on Connectivity Module PCBs so the JST PH connector matches the 232250 (EEMB LP472040F) cable as delivered: Pin 1 = GND, Pin 2 = BAT+. The released A-series footprint has Pin 1 = BAT+, which connects the cell backwards through the keyed connector. Not approved. Fix targets the B1 Altium release. Does not apply to Tag (230220).

Ships in the same B1 respin as Rev B1 ECN-01 — GPIO remap. Independent of, and to be coordinated with, Rev B1 ECN-12 — Battery connector rotation and Rev B1 ECN-14 — CHARGE_ENABLE pulldown.

Problem

The released A-series battery header (JST S2B-PH-SM4-TB, J3) is documented and laid out as Pin 1 = BAT+, Pin 2 = GND (232201 §3.2).

The mating battery is article 232250 (EEMB LP472040F-PCM-LD, JST PHR-02, 10 cm cable). The EEMB datasheet defines the cell's connector polarity by wire colour — red = + , black = − — and by its circuit diagram. It does not print pin numbers ("1"/"2") on the housing.

Two facts make the current assignment unsafe:

  1. JST PH 2-pin has no universal pin-1 polarity convention. Roughly half of vendors place + on one end and half on the other; the JST/Adafruit 1S-LiPo de-facto standard is Pin 1 = GND (B−), Pin 2 = B+. Wire colour is reliable; pin number is not.
  2. JST PH is keyed — the PHR-02 plug only mates one way, so battery housing pin 1 always lands on header pin 1. With the cell's red (+) wire in the housing position that mates the header pin currently assigned GND, the cell is connected reverse into the protected charger / power-path input.

The released A-series documentation asserts "Pin 1 = Battery +" without referencing the 232250 cable's wire-colour-to-pin mapping, so the mismatch was not caught on paper.

Primary cause of the B3 overheat

Reversed battery polarity is now believed to be the primary cause of the severe B3 overheating / damage seen on the A-series first article: a back-to-front cell forces destructive current through the charger / power-path. Bench-confirmed (2026-06-16) — with the cell connected the correct way, B3 stays cool. The CHARGE_ENABLE default-bias / B3-enable weakness (ECN-14) is a secondary robustness issue (mild reverse stress on AC loss) that made the polarity fault worse. Both must be fixed for B1, but this polarity fix is the primary one.

What changes

Adapt the PCB to the purchased EEMB 232250 cable rather than custom-crimping a vendor cell cable in volume. Swap the J3 net assignment to the JST 1S convention.

A-series (released) B1 (proposed)
J3 Pin 1 BAT+ GND (black wire)
J3 Pin 2 GND BAT+ (red wire)
Connector MPN S2B-PH-SM4-TB(LF)(SN) Same — net/pad assignment only, no connector change
Battery article 232250 (PHR-02, red +/black −) Unchanged — board adapts to the cable
Silk polarity marks 1 / 2 (no polarity) Add − at pin 1 / + at pin 2 (or GND / +)

This is a net-to-pad reassignment on the existing footprint — reroute BAT+ and GND copper from the swapped pads to the charger / power-path / sense divider. No charger, power-path, or sense-divider topology change.

Rotation does not fix polarity

ECN-12 rotates the J3 footprint 180° for cable routing. Rotation moves the physical position of the pads but does not change pin numbering or polarity (keying preserves pin1↔pin1). The polarity swap in this ECN is required in addition to the rotation. Do not "mirror" the footprint to fix polarity — mirroring breaks the land pattern. Coordinate both changes in the same J3 placement step.

Rationale

Factor A-series B1
Polarity Cell connects reverse via keyed plug Cell connects correct — board matches 232250 cable
Source of truth Doc asserts pin 1 = + (no cable cross-ref) Pin assignment defined against 232250 red/black wire colour
Fix location — PCB (cannot re-crimp purchased EEMB cable safely at volume)
Convention Non-standard for 1S JST PH Aligns with JST/Adafruit 1S Pin 1 = GND
Scope — Pad/net assignment + silk only

Scope

Item B1 change
232201 Base Bat J3 polarity swapped — battery populated
232203 Base Bat+WiFi Same shared layout — battery populated
232204 Base Bat+Radar Same shared layout — battery populated
232200, 232202 J3 footprint on shared PCB — DNP on non-battery BOMs
232250 battery No change — existing PHR-02 red/black cable
230220 Tag No change

Altium implementation (after design approval)

  1. Swap net assignment on J3 — Pin 1 → GND, Pin 2 → BAT+ (battery positive into the protection / charger / power-path input as today, just on the other pad).
  2. Reroute BAT+ and GND copper from the reassigned pads; keep the BAT_VOLTAGE sense divider Kelvin tap on the BAT+ pad.
  3. Silk — add polarity marks at the connector (− / GND at pin 1, + at pin 2). Update any +/− legend.
  4. Coordinate with ECN-12 — apply polarity swap and 180° rotation together; do not mirror the footprint.
  5. Gerber lock — verify pin 1 net = GND, pin 2 net = BAT+.

Documentation updates (required — track to completion)

Done now (released docs carry the A2 as-built values plus an A2-defect callout and the B1 (ECN-15) corrected polarity, so shipping A2 hardware is not misdocumented). Flip to B1-only values at Gerber lock:

  • 232201 §3.2 Battery Connector — A2-defect !!! danger callout added; pin rows now show A2 (BAT+/GND) and B1 (GND/BAT+); 232250 wire-colour row added.
  • 232201 §3.1 — note added: polarity defined by EEMB 232250 wire colour (red/black), not an assumed pin-1 convention.
  • 232201 §7 Battery-Specific Tests — polarity line updated to Pin 1 = GND, Pin 2 = BAT+ with a cable-to-header multimeter check (232250 red → BAT+ net).
  • 232203 Base Bat+WiFi §4 — A2-defect callout added, pointing to 232201 §3.2 and this ECN.
  • At B1 lock: replace the dual A2/B1 pin rows with B1-only (Pin 1 = GND, Pin 2 = BAT+) once A2 is superseded.

Verification (assembly / QA)

  • Cable polarity: on the mated 232250 cable, red wire → J3 pin 2 (BAT+ net); black → pin 1 (GND) — confirm with multimeter against the board nets (do not rely on pin number alone)
  • BAT+ net reaches charger / power-path input with correct polarity (no reverse current into protection circuit)
  • TP10 reads ~+3.2 V (cell positive) with battery connected the correct way
  • Reverse-insertion is mechanically prevented by JST keying (sanity check both ends are PH series)
  • Combined with ECN-12 rotation: pin numbering/polarity still correct after 180° rotation
  • Non-battery variants (232200, 232202): DNP pads unaffected

Open items

# Question Owner Status
1 Confirm against a physical 232250 sample which PHR-02 housing pin carries the red (+) wire (multimeter) and attach the result Bench Pending
2 Confirm EEMB cannot/should not supply 232250 with reversed crimp (so the fix stays on the PCB, not the cable) Purchasing Pending
3 Confirm final silk polarity legend (+/− vs GND/+) and that it survives the ECN-12 rotation Altium Pending
4 Update 232201 / 232203 docs (§3.1, §3.2, §7) HW docs Done — A2 defect + B1 polarity documented; flip to B1-only at lock

Rev A interim (until B1)

Aspect Handling on rev A
HW patch verify the 232250 cable polarity against the board nets (red → BAT+ net) with a multimeter; rework cable/connector if reversed
Firmware none — firmware cannot detect reversed polarity
Verified B3 cool + charging once polarity is correct (2026-06-16)
Limitation do not connect a battery to an unverified A2 board — reversed polarity cooks B3

See the Rev A bringup reference for the cross-ECN table.