This is sourcing advice for the seven entries in the uploaded shortfall list, not an updated BOM or fabrication approval. No schematic, footprint, routing, or BOM properties were changed. Stock observations are point-in-time public catalog/distributor results, not a reservation in your JLCPCB order; checkout availability wins. LCSC stock is explicitly distinguished from JLCPCB assembly stock below.
What to select for each item
The requested purchase quantities come from the user's pasted order screen, not a new assembly quantity calculation. JLCPCB may require additional pieces for assembly attrition. Keep any required extra quantity when rematching.
I saved the seven-part replacement list—with exact LCSC numbers, stock checks, and unresolved qualifications—in your project files: JLCPCB Shortfall Replacement Guide; the immediately stocked recommendations are R2 → C346633 and R3 → C2684521.
The guide gives exact alternatives for the feedback resistor and charger pending checkout availability, a conditional capacitor candidate, and recommends sourcing the original inductor and modem rather than an unverified swap; I haven’t changed your BOM or board.
Before releasing the order, request a fresh DRC/connectivity check: your latest saved checkpoint still flags copper overlaps and unrouted USB connections.
My earlier pre-order suggestion wasn’t a confirmed purchasing route for your order, and I haven’t verified a safe, stocked drop-in for U6 or L2. JLCPCB distinguishes stock shortages, sourcing restrictions, and “Standard Only” assembly restrictions, so the exact message matters.
Can you share a screenshot of what appears when you try sourcing each part, and whether your order uses Economic or Standard assembly?
Scope: Read-only audit of the complete 60-component, 59-net schematic. No design changes were made.
Review basis: Current Flux schematic/netlist and manufacturer documentation for SIM7672X/SIM7670G-correlated pinout, BQ24074, TPS6282533, TXB0104, ESP32-S3-WROOM-1-N8R2, MAX98357A, USB Type-C sink requirements, and the installed support components.
Executive Summary
The schematic is not ready for fabrication. The main architecture is coherent, but two critical modem issues must be corrected before prototype release:
U7 VBAT is powered from VSYS_BAT, which reaches 4.3–4.5 V on USB power. The modem's recommended VBAT range is 3.4–4.2 V. The current design therefore operates U7 above its recommended maximum whenever the BQ24074 power-path output is regulated from USB.
PWRKEY is driven through U4 TXB0104, while U4 VCCA and OE are powered by U7 VDD_EXT. VDD_EXT is unavailable until the modem is already powered, so the MCU cannot reliably issue the initial PWRKEY low pulse. TXB0104 push-pull translation is also inappropriate for a pin intended to be pulled low against the modem's internal pull-up.
Additional high-risk findings include USB input-current noncompliance, incomplete modem VBAT/SIM protection, questionable reset driving through TXB0104, and direct exposure of USB and external connector signals without ESD protection.
Critical Findings
Required Modem Corrections — Locked
These are mandatory before fabrication release:
Correct U7 VBAT supply: U7 VBAT pins 55, 56, and 57 must be removed from VSYS_BAT/BQ24074 OUT. The replacement MODEM_VBAT source must remain within 3.4–4.2 V under all DC tolerance, startup, load-release, charging, and USB-present conditions, and must support the modem transmit burst with at least 100 µF local bulk capacitance plus the recommended 10 pF, 33 nF, 100 nF, and 1 µF ceramics.
Replace translated modem controls with open-collector drivers:
ESP32-S3 GPIO21 → base resistor → NPN Q_PWRKEY; emitter to GND; collector to U7 PWRKEY pin 1. Firmware drives GPIO21 high to pull PWRKEY low. PWRKEY must not pass through U4.
ESP32-S3 GPIO14 → base resistor → NPN Q_RESET; emitter to GND; collector to U7 RESET pin 16. Firmware drives GPIO14 high to pull RESET low. RESET must not pass through U4.
Use MMBT3904 or a similar small-signal NPN. A nominal 10 kΩ base resistor is appropriate; optional 100 kΩ base-emitter pulldowns keep both transistors off during ESP32 reset.
Preserve U4 TXB0104 only for the 1.8 V UART: U7 MAIN_UART_TXD pin 9 and RXD pin 10. U4 VCCA remains on U7 VDD_EXT pin 15, VCCB on 3V3, and OE must default low until both domains are valid.
MODEM_VBAT architecture comparison
Option A — Dedicated regulated MODEM_VBAT rail
Add a synchronous buck-boost or suitable regulator between the system/battery source and U7 pins 55–57.
Recommended design target: approximately 3.8–3.9 V nominal. Select the regulator and feedback tolerances so the calculated static maximum is no more than 4.05 V, leaving at least 150 mV for startup/load-release overshoot; the validated worst-case at U7 must be ≤4.20 V.
Advantages: guarantees the modem remains inside its recommended range; stabilizes RF performance as the cell moves from full charge toward depletion; can provide an MCU-controlled enable for a true hard recovery cycle; isolates some charger/system-rail disturbances.
Disadvantages: additional cost, area, quiescent current, EMI, heat, and a demanding converter selection/layout for 2 A-class bursts. Efficiency is reduced when converting from a cell voltage already close to the modem voltage.
Worst-case voltage: must be demonstrated from the selected regulator datasheet and divider tolerance. The release requirement is ≤4.20 V including tolerance and overshoot; no exact value can be claimed until the regulator is selected.
Option B — Feed U7 directly from BQ24074 BAT / the fitted cell
This removes the 4.3–4.5 V OUT overvoltage and gives the modem a low-impedance path to the battery. BQ24074 permits 4.5 A recommended BAT discharge current, so a 2 A transmit burst is within the charger's BAT-path current rating if the cell, connector, protection PCB, copper, and capacitors also support it.
Advantages: simplest and most efficient architecture; no 2 A buck-boost; battery naturally supplies transmit peaks; fewer switching-noise and thermal risks.
Disadvantages: the modem voltage follows the cell and can approach the 3.4 V lower limit under low state-of-charge plus burst droop; modem load current flows directly at the battery node and can disturb charge termination/recharge detection; there is no regulator enable for a hard modem power cycle; battery wiring/protection resistance becomes critical; the modem remains battery-connected while logically off.
The BQ24074 regulates BAT to 4.20 V typical but 4.23 V maximum over its specified conditions. Therefore direct BAT feed does not strictly guarantee the modem's 4.20 V recommended maximum, even though a nominal lithium-ion cell is called “4.2 V.”
Worst-case voltage:4.23 V DC from charger regulation tolerance, before considering transient overshoot. This is below the modem's 4.5 V absolute maximum but 30 mV above its recommended operating maximum.
Decision
Because the stated requirement is that U7 must never exceed its recommended 4.2 V maximum, direct BQ24074 BAT feed is not fully compliant without an additional clamp/drop element and validation. Use the dedicated regulated MODEM_VBAT option unless the requirement is relaxed to allow operation up to the modem absolute maximum. Direct BAT remains the lower-complexity prototype alternative if the 4.23 V tolerance and low-cell burst droop are explicitly accepted and tested.
C1 — U7 modem VBAT exceeds its recommended operating voltage on USB power
Affected:VSYS_BAT, U1 OUT pins 10/11, U7 VBAT pins 55/56/57.
BQ24074 OUT regulation is 4.3 V minimum, 4.4 V typical, 4.5 V maximum when a valid input is present.
SIM767x VBAT recommended range is 3.4–4.2 V, typical 3.8 V; absolute maximum is 4.5 V.
Therefore the present USB-powered state is nominally 200 mV above the modem's recommended maximum and can reach the absolute maximum boundary.
This can increase stress, RF current, heating, and reset/shutdown risk. It is not a valid production operating point.
Sources: SIM7672X Hardware Design v1.01, Table 8 and Table 9, p.25; Table 51, p.61. BQ24074 datasheet, Power Path electrical characteristics and pin OUT description.
U4 VCCA and OE are powered from U7 VDD_EXT. VDD_EXT is generated by the modem and is low when the modem is off/sleeping.
Thus U4 is disabled before the modem starts, preventing the MCU from issuing the required initial active-low PWRKEY pulse.
U7 PWRKEY is an internally pulled-up active-low input. The reference circuit drives it with a low-side transistor/open-drain stage, not a push-pull automatic-direction translator.
PWRKEY requires approximately 50 ms low to power on and at least 2.5 s low to power off. Low must be ≤0.4 V.
The translated high level is only around 1.8 V, while the hardware manual lists a typical PWRKEY high level of 2.1 V. Although the internal pull-up normally establishes the idle state, actively driving the node high through TXB0104 is not robust.
H1 — U7 RESET is also driven through an unsuitable auto-direction translator
Affected: U3 GPIO14 → U4 B4/A4 → U7 RESET.
RESET is an internally pulled-up, active-low control input with a recommended low-side transistor/reference circuit and a typical 0.5 s low pulse.
TXB0104 is intended for push-pull CMOS data lines and its automatic direction sensing can interact poorly with internal pull-ups and slow edges.
U4 is disabled when VDD_EXT is absent, so reset is unavailable during modem-off states. This is acceptable only if reset is intentionally usable after VDD_EXT becomes valid; it is not a reliable hard-recovery path.
H2 — Fixed 1.3 A USB input limit ignores Type-C source-current advertisement
Affected: J7, R18/R19, U1 EN1/EN2/ILIM, R2.
CC1 and CC2 each correctly use 5.1 kΩ Rd, making the board a USB-C sink.
U1 is permanently configured for the resistor-programmed input-current mode: EN1 is low and EN2 is tied high to VSYS_BAT; R2 programs approximately 1.3 A.
A USB-C sink must determine the advertised current from CC voltage before drawing above default USB current. The schematic has no CC-voltage detection or Type-C port controller.
A legacy/default-current source may therefore be overloaded before firmware can intervene; U1 has no MCU-controlled current-mode selection in the current schematic.
Sources: BQ24074 datasheet, Table 7-2 and ILIM equation. USB Type-C sink behavior; each CC pin requires independent Rd, but Rd alone does not grant 1.3 A.
H3 — Modem VBAT filtering is incomplete relative to the hardware manual
Missing from recommended network: 10 pF and 33 nF local ceramics.
A VBAT TVS is recommended but absent.
The manual recommends at least 100 µF total external capacity when the supply supports peak current and lists 10 pF/33 nF/100 nF/1 µF ceramics near VBAT.
C14 is described in component metadata as “non-polarized,” but its Rubycon 16SEV100M MPN is a polarized aluminum electrolytic. This should be corrected in documentation/BOM polarity handling even if the physical footprint is correct.
H4 — USIM protection/filtering is incomplete and SIM_VDD has a questionable series resistor
Affected: U7 SIM pins, R11–R14, C18/C19, J4.
R12/R13/R14 correctly provide 22 Ω series elements on DATA/CLK/RST.
R11 adds 22 Ω in series with SIM_VDD. The reference circuit shows the series resistors on signal lines, while SIM_VDD is locally bypassed; a supply resistor is not part of the standard shown circuit and can introduce supply droop during card transients.
Missing: 33 pF shunt capacitors on the USIM signal lines and low-capacitance ESD protection at the socket.
J4 VPP is tied to SIM_VDD. This is commonly acceptable for sockets/cards where VPP is unused, but should be verified against the exact socket/card convention.
SIM_DET is connected directly without the optional filtering/protection shown for hot-swap implementations.
C10 and C11 provide the required 100 nF bypass capacitors.
OE is tied directly to VDD_EXT and has R7=100 kΩ to GND. This keeps the translator disabled until VDD_EXT rises, which is good for power sequencing but causes the PWRKEY deadlock described above.
UART channels A1/B1 and A2/B2 are push-pull and suitable in principle.
Control channels PWRKEY and RESET are not good TXB0104 applications because they are pulled-up active-low controls rather than normal push-pull data paths.
TXB0104 must not be used with strong external pull resistors; TI recommends external pulls greater than 50 kΩ. The modem's internal control-pin networks and edge behavior require bench verification even after startup is solved.
Sources: TXB0104 datasheet, Sections 3, 8.1, 9, and 10.
M2 — Modem shutdown sequencing cannot be enforced in hardware
The modem manual prohibits simply cutting VBAT during normal operation and requires PWRKEY or AT+CPOF first.
Current U7 VBAT is permanently tied to the charger/system power-path output and has no MCU-controlled load switch/regulator enable.
This avoids accidental firmware-controlled hard cutoff, but it also prevents a controlled hard power cycle for recovery, which the manual recommends the host be able to perform.
Any battery removal or deep power collapse can still interrupt flash operations.
Sources: SIM7672X Hardware Design v1.01, Sections 3.1.2 and 3.2.2, pp.27 and 30–31.
M3 — GNSS backup supply is absent
U7 GNSS_VBKP pin 116 is unconnected.
This is allowed for cold-start operation but prevents optimal hot-start retention. The manual recommends about 2.8 V and gives typical 16 µA at 2.8 V.
GNSS_VDD is correctly tied to VDD_EXT with C20=2.2 µF, and GPIO2→10 kΩ→GNSS_PWRCTL plus crossed AUX UART lines through 1 kΩ matches the non-standalone reference architecture.
M4 — ESP32 strapping GPIO45/GPIO46 are only relying on weak internal pulls
GPIO0 has the correct 10 kΩ pull-up and external boot-button connection.
GPIO45 and GPIO46 are unused and rely on their default weak pulldowns, which is valid if no layout leakage or attached circuitry disturbs them.
For maximum production robustness, explicit weak external pulldowns can be considered, especially GPIO45 because a high strap changes VDD_SPI behavior. They are not strictly mandatory for this module configuration.
GPIO3 is floating as defined by the default JTAG-source configuration and is unused.
Sources: ESP32-S3-WROOM-1 datasheet, Tables 4-1 and 4-3, pp.13–14.
M5 — Native USB lacks common recommended series/ESD provisions
U3 GPIO19 and GPIO20 are correctly used as USB D− and D+.
The connector's duplicated A/B data pins are logically joined correctly.
No USB ESD array is present, and there are no optional source-series resistors/test provisions. Espressif reference designs commonly route native USB directly, so absence of series resistors is not automatically wrong, but footprints or 0 Ω options improve tuning/debug flexibility.
M6 — BQ24074 thermal performance may limit programmed charging
R1=1.13 kΩ correctly programs approximately 0.8 A fast charge.
R2=1.18 kΩ correctly programs approximately 1.3 A input limit.
R3=4.12 kΩ correctly programs approximately 110 mA termination.
At high system load plus charging, the linear charger can enter its 125 °C thermal regulation loop. Copper area and exposed-pad thermal design determine whether 800 mA charging is sustained.
With the modem drawing transmit bursts, DPPM will reduce charging and the battery will supplement the system. This behavior is expected but should be included in firmware/UI expectations.
Sources: BQ24074 datasheet, Section 10.2.2.2 and thermal-regulation section.
Block Reviews
BQ24074 charger — mostly correct
IN bypass C1=4.7 µF: within required 1–10 µF.
BAT bypass C2=10 µF: within required 4.7–47 µF.
OUT bypass C3=10 µF: within required 4.7–47 µF.
ISET R1=1.13 kΩ: correct for ~800 mA.
ILIM R2=1.18 kΩ: correct for ~1.3 A.
ITERM R3=4.12 kΩ: correct for ~110 mA.
TS R4=10 kΩ to GND: explicitly permitted when no battery NTC is used, but removes battery-temperature protection.
CE low: charging enabled.
TMR low: safety timers disabled. This is intentional but risky for unattended charging because time-based fault protection is removed.
EN1 low, EN2 high: external ILIM mode.
PGOOD and CHG are unused: permitted, but valuable diagnostic signals are lost.
Thermal pad and VSS are both connected to GND: correct.
TPS6282533 buck — electrically correct with margin checks
Exact part is fixed 3.3 V, 2 A.
FB is correctly connected directly to the 3V3 output; no external divider should be used.
L1=0.47 µH is the standard recommended inductance.
C4=4.7 µF nominal input meets the 3 µF minimum effective requirement only if DC-bias derating leaves ≥3 µF at the actual input voltage. The selected 10 V 0603 part should be checked using its bias curve.
C5+C6=20 µF nominal output matches the standard recommended combination; verify effective capacitance remains ≥10 µF.
L1 is rated/saturates at 3 A. For 2 A output, this is plausibly sufficient and near the datasheet's 20–30% saturation-current margin target, but the exact inductance-vs-current curve should be checked.
Supply is 3.3 V with C7=100 nF and C8=10 µF local support. Datasheet requires a supply capable of at least 0.5 A.
EN has R5=10 kΩ to 3V3 and C9=1 µF to GND: matches Espressif's recommended RC delay.
GPIO0 has R6=10 kΩ pull-up and external active-low boot button: correct.
Native USB pins are correct: GPIO19=D−, GPIO20=D+.
N8R2 does not use the octal 8/16 MB PSRAM restriction called out for IO35–IO37; the schematic's use of IO35–IO37 for RGB outputs is therefore not automatically prohibited. Confirm the exact module ordering code at procurement because substituting an octal-PSRAM variant would break these outputs.
GPIO45/46 are not externally biased; see M4.
Two I2S peripherals are available; allocating microphone RX and speaker TX to separate controllers is supported.
Both VDD pins are on VSYS_BAT, within the amplifier's allowed supply range for a single-cell/USB power-path rail.
C12=100 nF and C13=10 µF provide local bypass/bulk capacitance.
Thermal pad and all GND pins are grounded.
R8=100 kΩ from GAIN_SLOT to GND implements the requested 15 dB configuration.
R9=100 kΩ to 3V3 and R10=18 kΩ to GND bias SD_MODE into the mono-average region when the internal pulldown is included; the node is not floating.
OUTP/OUTN correctly feed a two-wire differential speaker connector; neither output is grounded.
Risk: amplifier VDD shares the modem's burst rail, so modem transmit droop/noise can modulate or reset the amplifier. Local bulk helps, but rail transient testing is required.
The stored LCSC PDF was unavailable during this review; configuration was cross-checked against the previously locked MAX98357A design requirements and schematic values.
USB-C — CC implementation correct, protection/current policy incomplete
R18 and R19 are independent 5.1 kΩ pull-downs to GND: correct UFP/sink declaration.
D+/D− mapping to ESP32 native USB is correct.
All four VBUS contacts are on 5V_IN and all ground contacts/shield are grounded.
SBU pins are unused, appropriate for USB 2.0-only operation.
Missing ESD and source-current interpretation; see H2/H5.
Decoupling Audit
Table
Device/rail
Present
Assessment
U1 BQ24074 IN
C1 4.7 µF
Correct nominal range
U1 BAT
C2 10 µF
Correct nominal range
U1 OUT
C3 10 µF
Correct nominal range
U2 TPS6282533 VIN
C4 4.7 µF
Verify ≥3 µF effective under bias
U2 3V3 output
C5+C6 20 µF
Correct nominal; verify ≥10 µF effective
U3 ESP32 3V3
C7 100 nF + C8 10 µF
Good
U4 VCCA/VCCB
C10+C11, each 100 nF
Correct
U5 MAX98357A
C12 100 nF + C13 10 µF
Good
U7 VBAT
100 µF + 22 µF + 1 µF + 100 nF
Bulk adequate; missing 10 pF/33 nF and TVS
U7 SIM_VDD
C18 100 nF + C19 1 µF
Good nominal; supply series resistor questionable
U7 GNSS_VDD
C20 2.2 µF
Matches reference
USB-C/external connectors
None
ESD protection missing
ERC and Connectivity Notes
Current schematic ERC reports no open errors or warnings.
ERC cleanliness does not validate voltage ranges, sequencing, current compliance, component equations, or suitability of TXB0104 for active-low modem controls.
The logical netlist contains all expected major endpoints. Physical PCB routing is outside this schematic review and remains a separate release gate.
Recommended Correction Order
Create a modem VBAT rail guaranteed within 3.4–4.2 V under USB and battery operation.
Replace TXB0104 PWRKEY/RESET channels with appropriate low-side/open-drain control stages; ensure PWRKEY works while the modem is off.
Implement USB Type-C advertised-current handling or constrain the standalone input limit to a universally safe value.
Add modem VBAT TVS and complete the recommended local capacitor spectrum.
Correct the USIM network: review SIM_VDD resistor, add signal capacitors/ESD, and validate detect polarity.
Add USB and external-connector ESD protection.
Decide whether disabled charger safety timers and fixed TS bias are acceptable for the intended battery/product safety case.
Re-run ERC and perform a power/transient simulation or bench test after corrections.
Sources
SIMCom, SIM7672X Series Hardware Design V1.01, used for the Table 4-correlated SIM7670G symbol and electrical reference: pp.19, 25–32, 39–41, 45–52, 61.
Texas Instruments, BQ2407x Standalone 1-Cell 1.5-A Linear Battery Chargers with Power Path, SLUS810N.
Texas Instruments, TPS62824/25/26/27 Family Datasheet, SLVSEF9I, revised March 2024.