← Back to: Battery Charging / Gauging / Protection / BMS
Scenario & problem statement: why wearables need a nano-leakage charger
Wearables run on very small cells (typically 50–200 mAh, sometimes even 30–80 mAh). Users also leave them idle for a whole weekend or keep them in warehouse/storage for 1–3 months. In these intervals, even a few microamps from the charger itself (not the MCU) can eat a big portion of the battery. That is why we need a nano-leakage charger that sleeps together with the system.
Typical field complaints come from three patterns: (1) the MCU already entered deep-sleep but the charger did not; (2) an NTC divider or TS pin is permanently enabled and keeps stealing current; (3) the wearable is not docked (VBUS = 0) but the front-end detect circuit is still waiting for a wake. Our target for this page is to make all three visible and to write current limits directly into the BOM.
Minimum BOM-level requirements we will enforce in later chapters:
- Shipping / storage current of the charger IC: ≤ nA–a few µA @ VBAT = 4.2 V, system off.
- Charger quiescent current must be stated with VBAT = 4.2 V, no system load, NTC gated.
- Wake-on-VBUS / wake-on-button must not leave the charger in a high-current steady state after charge completion.
We do not expand JEITA here (only one sentence: wearable chargers shall still expose JEITA zone). We do not talk about high-voltage / 6S / 12S / automotive packs. We do not cover USB-C PD power-path coordination because that belongs to the USB-C sink + charging page.
Takeaway for purchasing: “Charger must also sleep” is a BOM-level requirement, not only a firmware feature. Do not approve alternatives that lack a documented ship-mode / deep-sleep current.
Operating states timeline for nano-leakage wearable chargers
A wearable charger does not live in a simple ON/OFF world. It moves along a low-power timeline: Shipping → Storage → User attach / dock detect → Charging (pre / fast) → Back to deep-sleep. At each state, the allowed leakage is different and the ability to report telemetry is different. Only certain wake sources are allowed to pull the charger out of deep-sleep.
Shipping
Target: ≤1 µA total
May skip telemetry; wake by VBUS / button.
Storage
Target: ≤2–3 µA incl. NTC gating
Do not keep NTC divider on in this state.
Dock detected
Temporary active, <200 µA
Used to confirm real charge intention.
Charging
Normal charger mode
Must report charging state + JEITA zone.
Deep-sleep return
Back to µA range
After charge: do not stay mA-level.
Disallowed for this page:
- No long-idle power-path priority for large batteries.
- No multi-cell re-enumeration logic.
- No cloud-OTA wake-up flow (handled in cloud/coordination page).
BOM takeaway for purchasing: “Wake-on-VBUS / button shall not increase steady-state current above µA range after charge completion.” If a replacement charger does not document this behavior, it shall not be approved.
Ultra-low leakage strategy: the charger side must also sleep
Many wearable failures come from reading the charger datasheet only at Iq (EN = 1) and assuming it is the real shipping / storage current. That value is often measured in an enabled, normal-bias condition and does not reflect what the device will draw when the MCU has already entered deep-sleep and the user is not docking the device. In this chapter we pin down the leakage sources and the exact 4-point test that must be written into the BOM.
Where leakage comes from
1) Charger internal bias / monitors
Reference, safety and status circuits that stay alive even when the system is off. Must have a published shipping current number.
2) NTC / TS divider
Most wearables leave the temperature divider powered; this alone can burn 10–50 µA. In ship-mode the divider must be gated.
3) Dock / VBUS detect front end
Magnetic / pogo detect circuits waiting for a charger. If they are always on, they defeat the purpose of nano-leakage.
How to read charger leakage specs
If the datasheet only gives you Iq (EN = 1) or IBAT quiescent at 25°C, that is not enough for a wearable. We must force the vendor (or ourselves) to test under the real shipping condition and to check it again at elevated temperature.
- Look for a distinct shipping / storage current spec.
- If the TS pin / NTC is enabled in the example circuit, ask what the current becomes when TS is gated.
- If the device has a “VBUS present” low-power mode, specify that we need the numbers with VBUS = 0 and the device waiting for a dock.
BOM 4-point test condition (write this exactly)
These 4 conditions must be in the BOM so purchasing, factory and FAE all measure the same thing:
- VBAT = 4.2 V (full cell condition)
- SYS = no load (system fully off / disconnected)
- NTC = not permanently driven (divider gated in ship-mode)
- Shipping pin / ship command = asserted (real low-leakage state)
Optional but recommended: repeat measurement at 40–50°C, because some parts meet the number only at 25°C.
Must-shut-down paths
1) Input / VBUS detect branch
In ship-mode this branch must be off. “Always-on detect” is not acceptable for sub-200 mAh wearables.
2) System-side LDO / bias
If SYS is off, bias must also be off. Do not accept parts that keep bias alive with no system load.
Don’t approve… alternatives that (a) only specify Iq with EN=1, (b) do not tell NTC / TS divider current in ship-mode, or (c) cannot disable dock-detect front ends in shipping state.
Ship-mode / deep-sleep wake path design
After we make the charger sleep, we must still be able to wake it. Wearables typically need exactly three wake sources: VBUS / pogo pad detect, user button / touch, and timed (RTC or gauge) wake. All of them must follow the same pattern: wake → charge / report → sleep again. A wake event must not leave the charger in a mA-level standby state.
Three wake paths & when to use them
VBUS / pogo pad detect
For magnetic / dock charging. Must detect attachment but return to µA after charge.
User button / touch
For user-driven wake. Short active period only, no permanent mA standby.
Timed (RTC / gauge event)
For periodic reporting to cloud / gateway. Must expose the wake reason.
Post-wake power limit
Every wake must converge back to the same low-leakage goal. We can write it as a BOM line:
“Wake-on-VBUS / button / timed shall return to µA-range within T = 1–3 s after charge completion or user detach.”
Wake reason telemetry (for cloud mapping)
Because we are in the BMS branch, cloud-side mapper must know which source woke the charger. Minimum fields to expose:
wake_source(VBUS / button / timed)charging_state(pre / fast / done / fault)JEITA_zone(to keep thermal control even in wearables)
If we swap across TI / ST / NXP / Renesas / onsemi / Microchip / Melexis and the wake reason is encoded differently, the cloud-side telemetry mapping must be updated before production.
Ship-mode ≠ standby ≠ disabled charger
Many datasheets use standby to mean “charger not running” but still powered; this is not the same as a wearables-grade ship-mode. A real ship/deep-sleep must:
- pull total current down to the nA–µA range,
- still be wake-able by one of the three sources,
- not require system power to wake,
- and expose the wake cause once the system is online.
Don’t approve… parts where “standby” is advertised as “shipping mode” but no current number is provided, or where standby cannot be woken by VBUS.
MCU handshake (push back to ship)
After a valid wake and charge, the system MCU should ACK the event and force the charger back into ship/deep-sleep. This guarantees that even user-aborted charge attempts do not leave the charger sitting at mA-level.
Coordination with gauge / NTC / JEITA for low-leakage profiles
Goal for this chapter: you want nano-leakage, but you cannot lose temperature protection. That means the NTC path must be switchable / gated, and when the NTC is shared with a gauge, the gauge must present a high-impedance input in ship-mode. We also keep a minimum JEITA reporting set: “JEITA zone must still be exposed even in ship mode.”
NTC sharing scenarios
NTC on charger side
Charger owns the NTC divider. In ship-mode the divider must be off, otherwise it will dominate the leakage budget.
NTC shared with gauge
Gauge reads the same sensor, but must be high-Z when the charger is in ship-mode. No always-on loading on the divider.
Wake → measure → sleep
On wake, enable NTC, read temperature, expose JEITA zone, then remove the divider again.
JEITA minimal reporting set
We do not expand full JEITA curve here. For this wearable low-leakage branch we only require:
- charging_state (pre / fast / done / fault)
- JEITA_zone (cold / cool / warm / hot)
- source_of_temp (charger-NTC / shared-NTC / gauge) — optional but recommended
This is enough for the BMS branch to keep thermal supervision even when the device is mostly sleeping.
Leakage budget when sharing NTC
Rule: if the device is expected to stay undocked for 2–3 days and the cell is 80–120 mAh, the temperature-sensing branch must not be always-on. Use a gated divider or an NTC power switch, measure in a short window, then go back to high-Z.
Procurement prohibition list
- Do not approve chargers that only support fixed-temperature charging.
- Do not approve chargers that require NTC divider to stay powered in ship-mode.
- Do not approve chargers that cannot expose JEITA zone while in wearable low-leakage profile.
- Cross-brand alternatives must still support NTC gating and JEITA reporting (TI / ST / NXP / Renesas / onsemi / Microchip / Melexis).
Safety & BMS linkage for 30–200 mAh wearable cells
Very small batteries are not afraid of one big charge cycle; they are afraid of “looks like off but actually trickling” and “looks like storage but actually leaking through sensing branches.” To prevent this, the wearable charger must stay aligned with the BMS branch and always be able to upload: charging state, JEITA zone, and wake reason. As soon as we swap TI ↔ ST ↔ onsemi or any of the seven brands, the cloud-side telemetry mapping must be updated.
Typical risks for small cells
Looks off but is slowly charging
Dock detect or charger bias never returned to µA → weekend drain or storage drain.
Looks like storage but NTC is on
TS divider is permanently enabled. After 1–3 months the cell is flat.
Cross-brand payload mismatch
Charger changed → field names changed → cloud did not → data is wrong.
Mandatory fields to upload
For this nano-leakage wearable charger page, we keep exactly 3 fields aligned with the BMS mainline:
- charging_state — so the cloud knows the device is actually charging, not passively leaking.
- JEITA_zone — so thermal derating is still visible even in low-leakage mode.
- wake_reason — VBUS / button / timed; without it cloud cannot classify the event.
This also makes it possible to detect “suspicious” events such as a device waking up from storage too often.
What the cloud mapper does here
Different vendors encode charger telemetry differently. TI might expose charging_state + temp_code, ST might expose chg_status + bat_temp, and onsemi might expose chrg_evt + jeita_zone. The cloud-side mapper takes these different payloads and normalizes them into a single wearable charging schema.
Cloud-side telemetry mapping must be updated before cross-brand alternatives. Otherwise the cloud will misread charging-state / JEITA / wake-origin and the safety logic will be ineffective.
BOM takeaway: “Before approving TI / ST / NXP / Renesas / onsemi / Microchip / Melexis charger alternatives, synchronize the cloud-side telemetry mapper with the new payload.”
Small-batch procurement & cross-brand alternatives
For wearable nano-leakage chargers we only allow replacement inside the seven brands: TI / ST / NXP / Renesas / onsemi / Microchip / Melexis. Every replacement must keep: (1) real ship/deep-sleep, (2) µA-level leakage with NTC gated, and (3) telemetry fields (charging_state, JEITA_zone, wake_reason). Cloud-side telemetry mapping must be updated before using cross-brand alternatives.
Mistake 1: no ship-mode
Picking a classic linear charger (e.g. MCP73831) to replace TI BQ25180 without adding ship-mode gating.
Mistake 2: standby ≠ ship
Approving parts where “standby = 20–50 µA” and calling it “shipping current”.
Mistake 3: 25 °C only
Checking leakage only at 25 °C; warehouse 40–45 °C will drain small cells.
Mistake 4: NTC always powered
Not verifying if NTC/TS divider can be gated in ship-mode.
Preferred parts by brand (nano-leakage / wearables)
TI → BQ25180, BQ25185, BQ25120A
ship-mode, ultra-low Iq, wearable-grade
ST → STBC15, STBC08, STBC28
single-cell, low-power, NTC-capable
NXP → PF1550 / MC34PF1550Axx (charger part)
needs low-power profile enabled
Renesas → ISL9205A, ISL9205D, ISL9220A
single-cell charger, add NTC gating
onsemi → NCP1852, NCP1855 + LC709204F (gauge)
USB/dock wake, set post-charge low-power
Microchip → MCP73831, MCP73871
must add NTC/power gating for nano-leakage
Melexis → MLX91216 / temp-sensing nodes
keep telemetry sensor when charger swapped
Three replacement paths
A → A (same brand, same series)
TI BQ25180 → TI BQ25185
ST STBC15 → ST STBC08
Cloud: no change (fields compatible).
A → B (cross-brand, same wake)
TI BQ25180 → onsemi NCP1852
ST STBC15 → Microchip MCP73831 (+ gated NTC)
Cloud: must update mapper.
A → A’ (same brand, new payload)
NXP PF1550 option → newer NXP PF1550 code
Renesas ISL9205A → ISL9205D
Cloud: change field names.
BOM-style examples
- “Use TI BQ25180 or BQ25185. Ship-mode leakage ≤ 2 µA @ VBAT = 4.2 V, NTC gated.”
- “If TI BQ25180 is not available, use ST STBC15 or STBC08 with external NTC gating. Update cloud-side telemetry mapping before use.”
- “Do NOT approve chargers that cannot expose both
charging_stateandJEITA_zoneto gateway.”
BOM remarks & telemetry hooks (wearable edition)
This section is the copy-ready version for small-batch orders. All lines below assume a single-cell wearable battery (30–200 mAh) and a nano-leakage charger that must sleep together with the MCU.
This page specific BOM lines
1. Charger / gauge must expose charging_state and JEITA_zone to edge gateway.
2. Ship-mode leakage ≤ 2 µA @ VBAT = 4.2 V, NTC gated, VBUS=0.
3. Cloud-side telemetry mapping must be updated before using cross-brand alternatives.
4. Devices that only support fixed-temperature charging must NOT replace NTC / JEITA-enabled parts.
5. If using Microchip MCP73831 / MCP73871, add NTC/power gating to keep nano-leakage profile.
6. If using onsemi NCP1852 / NCP1855, enable post-charge low-power and report wake source.
7. If using NXP PF1550, configure wearable low-power profile and map telemetry fields.
Telemetries to bind in cloud
Minimum telemetry set for this page:
charging_stateJEITA_zonewake_reason(VBUS / button / timed)hw_brand(TI / ST / NXP / Renesas / onsemi / Microchip / Melexis)fw_mapper_ver(so we know mapper was updated after replacement)
When to reject supplier alternatives
Reject if: (1) charger has no ship/deep-sleep mode; (2) NTC/TS divider cannot be gated; (3) wake reason cannot be reported; (4) supplier says “standby 20 µA” but no 4.2 V ship-mode number; (5) cloud mapper not updated after TI ↔ ST ↔ NXP ↔ Renesas ↔ onsemi ↔ Microchip ↔ Melexis swap.
Request a Quote
Testing & validation: how to prove “it really does not leak” on a bench
This procedure is for small-batch buyers, FAE teams and factories that need to verify leakage on a table-top setup without a full validation lab. We assume a wearable charger built from the seven allowed brands (TI BQ25180/BQ25185/BQ25120A, ST STBC15/STBC08, NXP PF1550, Renesas ISL9205, onsemi NCP1852, Microchip MCP73831/MCP73871, Melexis for sensing retention). The goal: show that in ship/deep-sleep, at VBAT=4.2V, with NTC gated, the current is in the µA (or sub-µA) range — and that it still wakes on VBUS and then goes back to µA.
Test supply
Bench supply → set VBAT = 4.2 V (repeat at 3.8 V). 4-wire or low-R wires preferred.
No system load
SYS / VSYS left open or high-Z. We only measure what the charger consumes.
Ship / deep-sleep asserted
Pull the ship/deep-sleep pin (TI BQ25180/BQ25185), or use the lowest-power enable (STBC15), then measure.
Step-by-step procedure
- Set VBAT. Configure bench supply to 4.2 V (full cell). Repeat later at 3.8 V to emulate a partially charged wearable device.
- Remove system load. Disconnect VSYS / system rail / LED drivers. We must measure charger-only leakage.
- Assert ship/deep-sleep. For TI BQ25180/BQ25185/BQ25120A pull the ship line; for STBC15/STBC08 put the device into its storage/low-power mode; for Microchip MCP73831/MCP73871 make sure TS/NTC is gated externally.
- Insert a µA meter in series with the battery side. Place the meter between bench supply and BAT to read actual shipping current. Zero/relative the meter if you work in nA range.
- Apply VBUS / dock. Connect a 5 V source to the VBUS/pogo pad to wake the charger. Watch the current spike and, after 1–3 s, confirm that current returns to the µA range.
- Repeat at 40–50 °C. Put the board inside a desktop thermal box, re-run steps 3–5. Record if shipping current increases at high temperature.
Test points to record
1) BAT / cell side — VBAT=4.2 V, SYS=NC, ship asserted → target ≤ 2–5 µA (adjust to your BOM value).
2) VIN / VBUS side — VBUS=0, dock absent → detect branch should also be in µA range.
3) NTC / TS line — ship-mode must gate the divider. If you see permanent current here, the part is not suitable for weekend-not-worn wearables.
Record everything at 25 °C and 40–50 °C. Some chargers only specify low Iq at room temperature. For small cells (30–200 mAh) this is not enough; high-temperature leakage must also be known and written into the BOM / cloud telemetry.
Test results must be written into the BOM / cloud record, e.g.: “Ship-mode leakage ≤ 2 µA @ VBAT=4.2 V, TS gated, 25/50 °C, charger=TI BQ25180 → if replaced by ST STBC15 or onsemi NCP1852, update cloud-side telemetry mapping first.”
Frequently Asked Questions — Nano-leakage wearable charger
All questions below belong to this page. They focus on low-leakage, ship-mode, wearables (30–200 mAh) and on the seven-brand-only replacement rule (TI, ST, NXP, Renesas, onsemi, Microchip, Melexis). Use them directly in WordPress without extra styling.
Why do I still need ship-mode if the MCU already sleeps?
Because the charger itself can draw several µA even when the MCU is in deep-sleep. For weekend-not-worn wearables you must make both MCU and charger sleep, otherwise the cell will drain in 2–3 days.
What shipping current is acceptable for 80–150 mAh wearables?
Use ≤ 2–5 µA @ VBAT = 4.2 V, VBUS=0, SYS=NC, NTC gated. If the part cannot meet this, it is not for the nano-leakage branch.
How do I gate the NTC / TS divider but still report JEITA?
Enable the divider only on wake or on charge, read temperature, map to JEITA zone, then disable it again. The field JEITA_zone must still be sent to cloud even in ship-mode.
Can I keep dock/VBUS detect always-on for user convenience?
Yes, but its current must be counted in the leakage budget. If always-on detect exceeds the budget, make the charger path sleep as well.
What if the replacement charger cannot report JEITA zone?
Then it cannot be approved for this page. Either pick an alternative inside TI / ST / NXP / Renesas / onsemi / Microchip / Melexis that reports JEITA, or update the cloud mapper to synthesize the zone.
How do I test wake → charge → sleep again on a bench?
Supply 4.2 V, assert ship, apply VBUS and log the spike, confirm that within 1–3 s the current drops back to µA. Repeat at 40–50 °C like in Chapter 9.
Can I use a consumer USB charger IC here?
Usually no. Many consumer parts only have standby, not true ship. They may also keep TS divider powered. That breaks the nano-leakage requirement.
Which parts can I swap inside the same brand without cloud changes?
TI BQ25180 ↔ BQ25185 ↔ BQ25120A, ST STBC15 ↔ STBC08, onsemi NCP1852 ↔ NCP1855. Still re-measure leakage like in Chapter 9.
What should I write in BOM when I approve cross-brand alternatives?
Write: “Charger/gauge must expose charging_state and JEITA_zone; cloud-side telemetry mapping must be updated before using cross-brand alternatives; ship-mode leakage ≤ 2 µA @ VBAT=4.2 V, NTC gated.”
Why update cloud-side telemetry mapping before swapping TI ↔ ST ↔ onsemi?
Because each brand reports slightly different field names / payloads. Without mapping, the cloud cannot see the real JEITA zone or wake source and safety logic will fail.
What is the difference between standby and real ship/deep-sleep?
Standby can still draw tens of µA. Ship/deep-sleep must go to a few µA or lower and still allow VBUS/button/timed wake. For wearables only the second is acceptable.
How do I measure nano-leakage current in a desktop thermal box?
Same as Chapter 9: VBAT=4.2 V, SYS=NC, assert ship, insert µA meter in series, then raise temperature to 40–50 °C and confirm current stays within your BOM limit.
