# Power, charging and prototype measurements

J103 accepts the two solar-panel leads. U101 harvests energy into its own storage reservoir, and U103 selects either that reservoir or VBUS from J101. A single battery charger, U102, powers the `SYS_AON` bus and charges the pack connected to J102. The timer, RTC and wake logic stay on the always-on rail. When `MAIN_EN` is asserted, U104 brings up `VMAIN` at 3.3 V for the radio, sensor and display.

## Solar panel and cold start

The procurement specification is ABRA SP-5V-230MA-100X80, 100.8 × 82.5 × 3 mm, 5 V at maximum power point, 230 mA. The manufacturer specifies a power of 1.15 W, without a numerical limit for `Voc` at cold condition. The project's acceptance limit is 9.0 V at open circuit terminals under the most severe illumination and temperature conditions within the installation. At 25 °C, the diode post-voltage ratio `Vmp/Voc` is verified on the I–V curve to be 0.77…0.85. D102 is a Schottky PMEG2010ER of 20 V and isolates the reverse panel. The panel is connected to J103.1 positive and J103.2 return. Wires have mechanical anchoring within the enclosure. [ABRA Panel](https://abra-electronics.com/solar-products/solar-cells-chargers-modules/0-1-watt/sp-5v-230ma-100.8x82.5mm.html) · [Nexperia PMEG2010ER](https://assets.nexperia.com/documents/data-sheet/PMEG2010ER.pdf)

U101 SPV1050TTR operates with `CONF=0` and uses a buck-boost bridge. L101, 22 µH, is connected between `L_HV` and `IN_LV`, while `IN_HV` receives voltage through D102 and resistor R126. R126 is a Panasonic ERJ3GEY0R00V resistor of 0 Ω / 1 A. Terminal 2 receives `PV_IN` from the cathode of D102. Terminal 1 supplies `PV_IC_IN` to U101.20, C103, and R105. To measure solar current, disconnect both power sources, remove R126, and insert the ammeter in series. This allows measurement of the MPP current divider. Assuming 50 mΩ and 230 mA as conservative values, the additional voltage drop is 11.5 mV, and the power loss is 2.65 mW. For cold start, ST indicates 2.6 V and 5 µA typical, 2.8 V at maximum voltage threshold. R105–R107 have values of 6.8 MΩ, 360 kΩ, and 1.5 MΩ, respectively. The nominal tracking ratio is 1.5 / (1.5 + 0.36) = 80.65%. Assuming ±1.5% tolerance on each resistor, the range is 80.17% to 81.11%, and the `MPP` input reaches a maximum of 1.979 V for a 9 V reference, below the ST threshold of 2.1 V. The nominal current through the divider is 0.72 µA at 6.25 V. C104 stabilizes the reference between sampling operations. The device interrupts conversion approximately every 16 seconds for about 400 ms. In very low light, the verification must include the charging time of C103 within the acquisition window. [ST SPV1050, sections 5–6 and figures 12–19](https://www.st.com/resource/en/datasheet/spv1050.pdf) · [Bourns SRN3015](https://www.bourns.com/docs/product-datasheets/srn3015.pdf)

R108–R110 are 6.8 MΩ, 1.5 MΩ, and 2.7 MΩ, respectively, between STORE and GND. The total resistance of 11 MΩ falls within the 10…20 MΩ rule from Equation 1 of ST. The 1.23 V reference sets the `EOC = 5.011 V` and `UVP = 3.221 V` thresholds. With ±1% tolerance for the reference and ±1.5% for each resistor, the resulting ranges are 4.850…5.178 V and 3.131…3.315 V. The lower EOC threshold provides a margin against VIN-DPM of up to 4.63 V as provided by U102. The result assumes the solar panel can supply power to the load. [ST, section 6.1](https://www.st.com/resource/en/datasheet/spv1050.pdf) · [TI BQ24074, table 8.5](https://www.ti.com/lit/ds/symlink/bq24074.pdf)

C106 and C107 charge to `PV_STORE`. The `BATT` pin of U101 has its own C105 of 1 µF. When the reservoir first reaches EOC, `BATT_CONN` pulls down gate Q101 and connects `PV_VALID`. At UVP, the switch opens. R111 recharges the gate from STORE. Its value of 1 MΩ adds approximately 5 µA only in the active state. This arrangement allows startup to occur before U102 draws current and removes power from the STORE terminal as recommended by ST. The internal BATT transistor has 6…8 Ω. [ST, sections 5 and 6.1](https://www.st.com/resource/en/datasheet/spv1050.pdf) · [AO3401A](https://www.aosmd.com/sites/default/files/res/datasheets/AO3401A.pdf)

The Murata curve for each 47 µF capacitor gives 17.73 µF at 5.011 V and 26.86 µF at 3.221 V. Integration of the two curves provides 14.10 µC up to VIN-DPM maximum and 78.42 µC up to UVP. At 10 mA, the first duration is 1.41 ms, the second is 7.84 ms. The battery pack helps supply the load during the 400 ms sampling interval. ST provides the efficiency curve for a 6 V panel in Figure 15. A 230 mA panel can exceed the current drawn by the converter under strong solar conditions. Solar charging current is characterized by measurement, including restarts in low light. [Murata characteristic curves](https://ds.murata.com/simsurfing/mlcc.html) · [ST, Figure 15](https://www.st.com/resource/en/datasheet/spv1050.pdf)

## USB-C, input selection and the battery

J101 is a GCT USB4135-GF-A with six contacts and four shield terminals. Each CC contact has 5.1 kΩ to GND and is protected by USBLC6-2SC6. The two power contacts are connected together, as are the ground and shield. The port is a USB-C consumer operating at 5 V, exclusively for power supply. The mounting position allows both cable orientations. SWD pads are used for programming. [GCT Schematic](https://gct.co/files/drawings/usb4135.pdf) · [ST USBLC6-2](https://www.st.com/resource/en/datasheet/usblc6-2.pdf)

The TPS2116 has VIN1 and MODE connected to VBUS, VIN2 connected to PV_VALID. Resistors R103 (301 kΩ) and R104 (100 kΩ) provide a nominal 4.01 V for USB priority. The calculated range is 3.607 V to 4.430 V with a reference of 0.92 V to 1.08 V and ±1.5% tolerance on the resistors. When VBUS is absent, MODE and PR1 are low, and the device selects the higher source in diode mode. A degraded USB source that maintains priority at approximately 4.1 V to 4.5 V can reduce load even when the panel is available. BQ limits the input through VIN-DPM, and the battery can supplement the output. Both inputs of U103 allow a maximum of 5.5 V and include reverse current blocking. [TI TPS2116, sections 7.3–7.4](https://www.ti.com/lit/ds/symlink/tps2116.pdf)

U102 BQ24074RGTR operates in USB100 mode, with EN1 and EN2 tied to GND. The current limit at the IN pin is 90…100 mA, typical value 95 mA. R113, 2 kΩ, maintains the ILIM pin biased as specified in the datasheet. The selection and battery temperature monitoring circuits draw power independently from the input. The connector acceptance limit is 101 mA in steady state. Use a 5 V USB-C power supply with default USB current available. [TI BQ24074, Table 8.5 and Section 9.3](https://www.ti.com/lit/ds/symlink/bq24074.pdf)

R115 (8.66 kΩ) sets a nominal charge current of 102.8 mA before input current limiting. The factor of 797...975 AΩ and the ±1% tolerance yield a current of 91.12...113.72 mA. These values remain below 125 mA, the maximum specified for LP223439. The standard charge current in the battery pack datasheet is 50 mA. Input limit and consumer load reduce the actual current. Precharge current ranges from 8.00...12.36 mA during the short detection condition phase, approximately at 3.0 V. U102 regulates the output voltage between 4.16...4.23 V. R123, 4.7 kΩ, at ITERM sets a typical termination current of 5.43 mA, with a range of 4.26...6.64 mA under USB100 conditions. R114, 68 kΩ, controls timing, providing 0.67...1.14 h for precharge and 6.73...11.45 h for fast charging. Thermal regulation and DPPM modify the effective time. The battery pack’s PCM provides secondary protection. [TI BQ24074, sections 9.3.5–9.3.7](https://www.ti.com/lit/ds/symlink/bq24074.pdf) · [LiPol LP223439](https://www.lipobattery.us/wp-content/uploads/2022/11/LiPo-Battery-Datasheet-LP223439-250mAh-with-NTC-and-JST-ACHR-03V-S.pdf)

J102 is JST BM03B-ACHSS-GAN-ETF(LF)(SN), 1.2 mm pitch. The ordered configuration is pin 1 black/negative, pin 2 yellow/NTC, pin 3 red/positive. The battery pack drawing identifies the wire colors but does not number the cavities. At incoming inspection, check the unmated connector for positive voltage between 3 and 1, and approximately 10 kΩ between 2 and 1 at 25 °C, prior to insertion. [JST ACH](https://www.jst-mfg.com/product/index.php?lang=2&series=9) · [Drawing LP223439](https://www.lipobattery.us/wp-content/uploads/2022/11/LiPo-Battery-Datasheet-LP223439-250mAh-with-NTC-and-JST-ACHR-03V-S.pdf)

## Charging temperature

The battery pack includes a 10 kΩ ±1% NTC thermistor B3435 and an allowed charge temperature window of +10…+45 °C. The generic TS profile of U102 does not match this lower limit. Resistor R112 (10 kΩ) maintains TS within the charger's allowed temperature window in external monitoring configuration. U105 MCP6542 compares the NTC physically against two hardware references independent of firmware from `CHG_IN`. Resistor R116 (22 kΩ) biases the NTC. Resistors R117/R118/R119 (62/22/22 kΩ) produce 44/106 and 22/106 of the supply voltage. At low temperature or in open circuit, channel A conducts. At high temperature or short, channel B conducts. Diode D103 (BAT54C) combines the signals through separate anodes. The common cathode raises CE to U102. Resistor R120 (22 kΩ) returns CE to ground when both outputs are low. The monitor and NTC are powered from the charger input and do not draw from the battery in darkness. [Microchip MCP6542](https://www.microchip.com/en-us/product/MCP6542) · [Nexperia BAT54C](https://assets.nexperia.com/documents/data-sheet/BAT54C.pdf) · [TI BQ24074, TS and CE](https://www.ti.com/lit/ds/symlink/bq24074.pdf)

Across 256 combinations per threshold, the cold turn-on threshold ranges from 11.75…15.96 °C, and the hot cutoff threshold ranges from 37.26…42.98 °C. The calculation includes resistors ±1.5%, R25 ±1%, a sensitivity B ±3%, an equivalent input ±15 mV, and 3.2…5.5 V. The 15 mV includes offset, hysteresis, and a margin for temperature and bias. The battery pack datasheet does not guarantee the B sensitivity tolerance, and the comparator's offset has specific test conditions. Both thresholds, hysteresis, cell temperature and thermal lag must be measured with the enclosure closed. The calculated margins up to cell limits are 1.75 and 2.02 °C.

Ngspice 47 resolved six defect cases and three normal transfers at 3.2/5.0/5.5 V, plus a finer sweep at a step size of 0.025 °C. The DC model uses a finite-gain comparator with 100 Ω output impedance, a Schottky junction, and a CE input load. At 3.2 V, an open or shorted NTC produces approximately 2.90 V at CE, compared to a minimum VIH of 1.4 V. The nominal charging window was 14.0…39.9 °C at a step size of 0.1 °C. These results verify the topology and polarity. Physical measurement of the delay upon sudden source application and at minimum voltage level remains necessary.

## Main power domain and regulator stability

R124 and R125 are Panasonic 0 Ω jumpers in 0603, 1 A, placed at the input and output of U104. For servicing, remove the corresponding jumper to insert a current meter in either rail. The switching capacitors remain permanently connected to the U104 side. After the package fan-out, the bus traces reduce to 0.30 mm. VSEL is at GND and EN has R122 of 1 MΩ to GND. RT6160A starts with VMAIN at 3.3 V by default, at I²C address 0x75. SCL/SDA have pull-ups within the VMAIN domain. When EN is low, the regulator discharges the output through approximately 5 Ω typical. For a 100 µF capacitance in the domain, the lumped RC model predicts 1.20 ms for 3.3→0.3 V, while 44 µF provides 0.53 ms. Under Richtek’s test conditions, the enable delay is typically 220 µs and at most 300 µs. The output rise time to 95% is typically 300 µs and at most 1 ms. Check at the load terminals for display load. [Richtek RT6160A, sections 10 and 16](https://www.richtek.com/assets/product_file/RT6160A/DS6160A-05.pdf)

L102 is a 0.47 µH Coilcraft XGL4015-471MEC inductor. Its inductance drops by 10% at 4.4 A. C113 (22 µF/16 V) and C114 (100 nF) provide local decoupling for the bridge. At 4.5 V, C113 typically retains 8.07 µF compared to approximately 2.42 µF for the input component of 10 µF/0402 from Richtek's assembly. C115 and C116 are exactly the two GRM188R61A226ME15 components recommended by the manufacturer. At 3.3 V, they typically provide 18.18 µF together, respectively 12.36 µF in combined tolerance and temperature sensitivity. C109 from the battery provides 8.66 µF typically at 4.23 V and 5.89 µF at −20% × −15%. C110 from OUT provides the same 8.07/5.49 µF as C113 at 4.50 V, exceeding the recommendation of 4.7 µF at U102. Check the transient response on the prototype with an oscilloscope, taking MLCC aging into account. [Richtek, Table 1 and sections 16.15–16.17](https://www.richtek.com/assets/product_file/RT6160A/DS6160A-05.pdf) · [Coilcraft XGL4015](https://www.coilcraft.com/getmedia/c5d04454-742d-473e-8b80-06c97338c587/xgl4015.pdf) · [Murata SimSurfing](https://ds.murata.com/simsurfing/mlcc.html)

For a 100 mA operating scenario, with 5.5 V input and 2.9 V battery output, the estimated temperature rise at 1.5 mA of active current and TI’s junction-to-ambient thermal resistance of 44.5 °C/W on the reference PCB is 11.94 °C. The two-layer geometry and enclosure have inherent thermal resistance. Measure at 40 °C ambient temperature, under system electronic load and with a battery simulator. In darkness, the BQ24074 consumes typically 4.3 µA from the battery, with a maximum of 6.5 µA as per TI conditions. The full budget includes the LDO, RTC, timer, PCM, and event energy. [TI BQ24074, sections 8.4–8.5 and 11](https://www.ti.com/lit/ds/symlink/bq24074.pdf)

## Power-source combinations

| Panel | USB | Battery pack | Expected state |
|---|---|---|---|
| Darkness | Absent | Absent | After discharging the capacitors, SYS_AON and VMAIN are turned off. |
| Darkness | Absent | Present | The battery powers SYS_AON, while VMAIN is enabled only upon wake-up. |
| Darkness | 5 V | Absent | USB powers SYS_AON. An open NTC disables charging. The source capacity and the U102 limit determine the available load. |
| Darkness | 5 V | Present | USB powers the system and charges the battery within the NTC temperature window. The pack supplies additional current when the load exceeds the input capacity. |
| Active | Absent | Absent | After EOC, solar can power SYS_AON. During MPPT sampling or low light conditions, power interruptions/restarts may occur. Open NTC disables charging. |
| Active | Absent | Present | The system initially runs on the battery. After EOC, U102 charges from the solar source. The pack takes over the load in shade and during the MPPT sampling window. |
| Active | 5 V | Absent | A valid USB source takes priority. An open NTC disables charging. Unplugging the cable may restart the system if the solar reservoir and panel cannot sustain the load. |
| Active | 5 V | Present | USB takes priority. When unplugged, U103 selects solar if available, and the battery pack bridges the switchover. The NTC measures cell temperature independently of the copper temperature. |

First emulate a cell disconnected by its PCM. U102 provides detection of an absent battery and can supply power to the system from an external source at the OUT pin. The PCM-specific disconnection threshold is 2.75 V ±50 mV after 14…26 ms, with reconnection at 3.05 V ±50 mV. To reactivate the cell, the authorized PCM method is followed and current is monitored. During discharge tests, disconnect the load when the cell reaches 2.75 V. If the NTC is open, charging remains inhibited.

## Implementation and measurement access

The RT6160 uses a 15-ball array with 0.40 mm pitch. The masked variant features copper lands Ø0.250 mm, mask and paste Ø0.220 mm. These dimensions fall within Richtek's SMD range and maintain a 0.150 mm gap between adjacent copper lands. ENIG and local stencil of 0.08 mm provide an area ratio of 0.6875. QFN packages have paste windows on the thermal pad. GND, PGND, and EP at SPV are connected via short traces on the component side. IN_LV/L_HV and SW1/SW2 have local loops, and return vias are located outside the MLCC pad areas. J101 has a maximum height modeled at 3.35 mm. The geometry for J102 is verified by pad number correspondence and pin 1 orientation. [Richtek, WL-CSP geometry](https://www.richtek.com/assets/product_file/RT6160A/DS6160A-05.pdf) · [ST, Section 7](https://www.st.com/resource/en/datasheet/spv1050.pdf)

| Pad | Measurement | Pad | Measurement |
|---|---|---|---|
| TP101 | PV_STORE | TP104 | SYS_AON |
| TP102 | CHG_IN | TP105 | GND |
| TP103 | VBAT | TP106 | VMAIN |

The file `verification/power/connector-models.json` contains local numeric transformations. For J101, the six contacts and four mounting tabs correspond to the KiCad model within a tolerance of 0.003 mm. For J102, five contacts and mounting tabs from the CAD sample JLCEDA/EasyEDA were used, compared against the footprint derived from the JST drawing. A -90° rotation and an X-axis correction of +0.07657 mm yield a maximum residual of 0.0301 mm. The model origin statement is located in `hardware/models/JST-model-source.md`. [KiCad – USB4135 model](https://gitlab.com/kicad/libraries/kicad-packages3D/-/tree/master/Connector_USB.3dshapes) · [JLCEDA/EasyEDA Official Library](https://lceda.cn/) · [EasyEDA](https://easyeda.com/)

## Prototype tests and supporting records

Physical tests have not been run. The procedure in `docs/qualification.md` covers daily energy targets, the radio/EPD profile, P04 transitions and the temperature test. For initial power-up, use a current-limited USB supply and a battery simulator that can sink current. Connect the actual pack after verifying pin assignments, polarity and thresholds with the simulator.

| Test | Assembly and capture | Acceptance threshold |
|---|---|---|
| PV Thresholds | With the sources disconnected, discharge the reservoir to GND before assembly. Connect the panel emulator and a controlled downstream load. Record TP101, the Q101 gate, TP102, TP103 and TP104, then gradually reduce the light level. | First connection to 4.85…5.18 V at TP101, disconnection at 3.13…3.32 V. Measure `VMPP` with an active probe of at least 1 GΩ or buffer to maintain the divider impedance. |
| MPP and solar energy | Measure the I–V curve and `Voc` in full sun and at the temperature limit. Compare the energy delivered to the battery over 1 h at 25 °C with the pack partially discharged and the load idle. Time at least three consecutive samples. | `Voc ≤ 9.0 V`. At acceptance, the post-diode `Vmp/Voc` ratio must be 0.77…0.85. When the converter can draw the available current, the tracked voltage must be within ±5% of the sampled reference. Use the net solar current from `docs/qualification.md` for operational sizing. |
| USB in both orientations | Connect a C–C cable to 4.75…5.25 V using a compliant power supply, in both orientations. Measure separately the current at the connector and the current through U102 via the temporary instrumentation footprint. | Both power orientations supply the board. At 25 °C, steady state, without wake-up events, the input current of U102 is at most 100 mA, and the port current is at most 101 mA. |
| CCCV and termination | At 25 °C replace the NTC with a 10 kΩ resistor on the battery-pack simulator. Slowly raise the simulator voltage from 2.5 to 4.25 V. Repeat with each source. | With sufficient input, between cell detection and the low-voltage region, expect a precharge current of approximately 8…12.4 mA. Under all conditions pack current must be ≤125 mA and voltage ≤4.25 V. With USB input and no DPPM or thermal limiting, expect termination at 4.26…6.64 mA and regulation at 4.16…4.23 V. |
| Charging interruption | Using a calibrated resistance decade box, switch the NTC input to 0 Ω, high resistance, the values corresponding to 10 °C/45 °C, and intermediate resistances. Repeat on the bench at 3.2/4.35/5.0/5.5 V. Apply 5.5 V only at the selected input and stay within each source rating. Capture CE and current on source connection, in both ramp directions. Then repeat in a thermal chamber with the thermocouple attached to the pack. | CE ≥ 1.4 V in defect and at 10/45 °C, CE ≤ 0.4 V at 25 °C. With valid input, stable charge current ≤ 100 µA in inhibited state. With normal request, the NTC and the reference thermocouple must confirm the allowed cell temperature. |
| Main power supply | For each combination in the table, test both with VMAIN off and during a full refresh. Use a short probe ground lead at TP106. Superimpose a 0→100 mA load step with a 10 µs edge on the idle load. Repeat with the battery simulator at 3.0/3.7/4.2 V and USB at 4.75/5.25 V. | Static VMAIN: 3.20…3.40 V. With a functioning battery and load within the pack’s 250 mA limit, VMAIN must stay at 3.0…3.55 V during the event, without an MCU reset. For the complete assembly, target ≤2 ms from EN to 95% of final voltage. With EN low, target VMAIN <0.3 V within 5 ms. Record the actual time separately and show every startup and shutdown. |
| Losses and temperature | Run the load test at 40 °C ambient in the closed enclosure with a battery emulator. Record the temperatures of U102, BME280 and the battery pack together with the current in each branch. Disconnect USB, solar and SWD, then measure the total current drawn at the pack. | The U102 case must stay below 85 °C, with no thermal limiting in nominal operation. Charging must stop if the cell reaches 45 °C. For RTC traffic, deep sleep and the daily budget, use the full acceptance limits in `docs/qualification.md`. |

File `verification/power/check.py` reads values and nets from `hardware/power.json`, generates `results.json`, netlists, numerical results, and ngspice logs from `verification/power/spice`. Compressed Murata curves and SHA-256 hashes are located in `verification/power/sources/`. The `solar_reservoir` section integrates differential capacitance with respect to voltage. Independent simulation of U103 using the TI macromodel and the three convergence tests is found in `verification/independent/mux/`. TI’s version 1.10 model omits diode mode, so the solar-only case with VBUS absent still requires a bench test. The ST archive for SPV1050 has been downloaded and inspected. The 84,000-byte library `SPV1050_macromodel.lib` is encrypted for licensed OrCAD PSpice. `verification/power/spv-model-inspection.json` retains identifiers and hashes. For U101, threshold, capacitance, and energy calculations are available, while MPPT acquisition and harvested current still require measurements on the prototype. [ST, platform requirements for model](https://community.st.com/power-management-53/i-would-like-to-use-this-ic-spv1050-may-i-know-what-simulation-software-can-i-use-to-simulate-this-circuit-3380) · [TI TPS2116, models](https://www.ti.com/product/TPS2116)
