Schematic Evidence and Limits
Latest ADC-held supply and ERC checkpoint — 2026-10-10
Actual schematic changes now supersede the earlier no-mutation continuation: D29/D30, C48/C49 and R153–R156 add independent diode-isolated ADC-only reservoirs; U12/U13 VDD and C40/C42 bypasses moved onto held rails, while TMUX/supervisors remain upstream. Net/polarity readbacks completed. Native1 µs collapse simulations with80 µF,1 mA ADC stress load and ideal active surrogates show no clamp injection for5 µs and100 µs switch-opening cases, versus1.48303 mA/channel without hold-up. This is bounded model evidence, NOT full-temperature/ramp/I2C-backfeed qualification; live-input supply-transition restrictions remain. Exact mutations, models, citations and itemized status are in Analog Collapse and ERC Repair, latest section.
Fresh ERC2026-10-10T17:08:44.901Z: 30 Errors/6 Warnings, unchanged. Individually verified passive/socket-drive and external-feed paths remain conditionally explained, not cleared; no false pin types or suppression. R63–R70 canonical15 kΩ normalization completed; explicit100 Ω value/unit follow-up7d21d030 completedR6–R9/R37–R40, all eight independently read back as value100/unitΩ with original nets retained. D25–D28 ineffective component Pin Type overrides removed. Terminal metadata API blocker and failed feedback remain. R153/R154 production MPNs remain unsourced.
Completed: ADC-held hardware implementation, readbacks and bounded simulations, per-path ERC audit/partial metadata cleanup. Failed/incomplete: unrestricted analog qualification, independent approved-SPI motor controls and hardware fail-safe reset/watchdog EN, exact module straps/accessibility and exact devboard reset/bridge evidence, clearing30/6 findings. Unattempted by instruction: VM regen protection/brake sizing and PCB/exports. Minimum VM envelope: returned energy per stop, peak regenerated power/current and repetition, source tolerance/OV and sink rating; alternatively an explicitly rated absorbing source/brake system. Keep motorR145–R150DNP, VM contacts open and modules uninserted/unpowered. Earlier historical counts, no-motor-interface and zero-ERC claims below are not current release status. No PCB/export mutation was made in this batch.
SPI approval and incomplete electrical completion — 2026-10-10
User now explicitly approves SPI for the two exact BIGTREETECH TMC5160T Pro V1.0 modules; prior wording that motor protocol architecture is pending is superseded. User-measured socket pitch 2.54 mm and row spacing 15.24 mm remain accepted nominal geometry, not manufacturer tolerance evidence.
Fresh read-only ERC at 2026-10-10T16:41:14.177Z reports 30 Errors and 6 Warnings: 8 VNQ command inputs, 6 SEL/SEn biased inputs, 2 FaultRST biased inputs and 14 externally sourced supply terminals; warnings are D5/D11/D25–D28 terminal metadata. Prior connectivity dispositions are not a claim the checker is cleared. No suppression or terminal-type falsification applied. No component, connectivity, PCB or export mutations occurred in this continuation; only this existing evidence file was updated.
Analog fast-loss protection remains FAILED/incomplete; U16/U17 slow-collapse improvement is not sufficient. ADG4612 was researched but not placed: its switch opens above VDD+VT and VT is 1.8 V typical, which does not establish ADS rail+0.3 V protection. Exact supply-collapse turn-off threshold remains unverified after incomplete retrieval, not proven absent. No new simulation was run.
Motor completion remains FAILED/incomplete: VM path/regenerative protection, hardware reset/watchdog EN interlock, exact mating orientation/mode configuration, socket ratings and sequencing are unresolved. Source/regeneration sizing specifically requires a maximum returned-energy/peak-power/repetition envelope or a specified supply/braking system that absorbs it; a VM feed fuse or source cutoff alone cannot bound locally regenerated voltage. No current ceiling is newly qualified by this continuation. Keep R145–R150 unpopulated, modules uninserted and VM unpowered. PCB placement/routing/export are UNATTEMPTED by scope. This is not a manufacturing release.
Expansion electrical follow-up — 2026-10-10
Current repair adds U16/U17 brownout supervisors, R151/R152 pull-ups and C46/C47 bypasses to replace rail-tied TMUX enables. Actual topology, cited supply/logic limits, nominal SPICE comparison and residual rapid-collapse/backfeed blockers are in existing Analog Collapse and ERC Repair; no unrestricted live-input power-cycle or fabrication qualification is claimed. Fresh ERC 2026-10-10T16:29:48.951Z remains 30 U5/U6 Errors/6 diode terminal metadata Warnings, not the historical zero-error statement below. Motor remains isolated DNP and unpowered pending architecture; exact module is documented SPI, while silicon UART capability does not prove assembled-module UART access (see existing Stepstick Expansion Decision Gate). All original functions retained; no PCB/export changes.
Current requirement supersession — 2026-10-08
Current continuous ceilings are 1 A EACH /4 A load per bank /8 A aggregate load plus actual switch/support draw, replacing the former2 A/8 A/16 A requirement in this historical evidence file. They are not silicon limits or qualified ratings.100 mΩ conduction-only estimate is0.4 W/package/0.8 W total, not old1.6/3.2 W; actual thermal/control/switching burden remains to be qualified. Current schematic has199 components/122 nets;30 conditional source-model ERC Errors remain visible. Later connector/passive/source evidence in bench-assembly-evidence and electrical-release-follow-up supersedes outdated sourcing gaps below. Revised physical path evidence and individual qualification blockers are in revised-current-path-review. External branch protection, inductive suppression and vehicle conditioners remain mandatory.
Final Correction Pass
Historical schematic pass below: retained cited evidence and then-current counts/sourcing gaps are superseded by the current-requirement header and later bench/electrical files. Thermal2A/16A calculations, enclosure proposals and zero-ERC claims below are archival, not active requirements or current check results.
- Added R85–R92: eight 47 kΩ input-side AHCT pull-downs. With the removable MCU absent or unpowered, each A input has a defined LOW state; OE defaults and connector-side idle-low resistors remain.
- Replaced U2/U4 SN65HVD230 with TCAN1042HGVDRQ1 and explicitly remapped every pin. VCC is supplied from each domain's 5 V rail, VIO from its local 3.3 V rail, and GPIO18 now controls active-high STB.
- Reused C6/C15 as 5 V VCC bypass capacitors and added C23/C24 as 3.3 V VIO bypass capacitors.
- Changed VNQ FaultRST bias from LOW/autorestart to HIGH/latched thermal shutdown.
- Corrected sensor-current metadata to 90 mA continuous budget and 139 mA fault-sizing ceiling.
- Replaced obsolete relay/coil/gate/LOW signal terminology with positive high-side command/output names while preserving connectivity.
- Updated J3 to require an external protected regulated 5 V source rated at least 3 A.
- Replaced J18/J37 with Phoenix Contact 1729034 four-position screw terminals without changing their verified sensor-net connectivity, and added U9/C25 as a local MCU-A-only I2C board thermometer.
- Set R45–R48 to an explicit 0.25 W minimum, matching R14–R17. At 16 V continuously across 47 kΩ the ideal dissipation is only about 5.45 mW; however, the production resistor MPN must still have a documented working-voltage rating above the final conditioned-trigger residual envelope.
- Reconciled J1/J2/J21/J22 to the supplied identical-board pinout. Previous schematic numbering had incorrectly treated several socket positions as matching GPIO numbers. Power and all GPIO8–21 functional connections were moved to their actual physical contacts on both MCUs.
- Added J40/J41 spare interfaces: local 3.3 V, protected GPIO39, protected GPIO40 and GND. R93–R96 are 100 Ω series elements and D21/D22 are TPD2E2U06 dual-channel ESD arrays.
AHCT Input Bias Evidence
TI requires unused inputs to be terminated to VCC or ground and permits a pull resistor for intermittently connected inputs. SN74AHCT125 maximum input leakage is ±1 µA and VIL(max) is 0.8 V; a 47 kΩ pull-down's worst leakage-derived voltage is 47 mV.
This establishes an input default but does not provide powered-off isolation. LED strips or other external sources must not drive the carrier while its buffer supply is off.
CAN Coordination Evidence
- TCAN1042HGV-Q1 is AEC-Q100 qualified and its H-variant CANH/CANL absolute maximum range is −70 to +70 V.
- D5/D11 PESD2CANFD24V have 24 V reverse standoff and 42 V maximum clamp at 1 A with an 8/20 µs waveform.
- U2/U4 therefore withstand the stated local TVS clamp level. This does not validate load dump, prolonged ground offset, harness energy or EMC for the installed vehicle.
- VCC and VIO each have the recommended 0.1 µF local bypass.
- STB is active high and TXD/STB bias is referenced to VIO on the V-suffix device.
VNQ Protection Limits
- FaultRST LOW selects autorestart; HIGH selects latch-off. R66/R70 now pull FaultRST HIGH to the applicable 3.3 V domain.
- A deliberate LOW reset pulse or appropriate power cycle is required to clear the latch; the specified latch-reset interval is 3 µs minimum, 10 µs typical and 20 µs maximum under the datasheet conditions.
- The VNQ high-current limiter is 21 A minimum to 38 A maximum at 13 V. It protects silicon/system components and must not be represented as 2 A branch protection.
- Each branch requires an external fuse and wire gauge selected from measured load inrush and inductive energy. A 2 A, 32 V MINI blade fuse such as Littelfuse 0297002.WXNV or Eaton ATM-2 is only a candidate pending time-current, ambient and harness validation.
- FaultRST pull-ups are powered from the removable MCU-domain 3.3 V rails. Latched behavior therefore depends on that rail being present and valid; removing or losing the MCU rail changes the fault-reset bias condition.
VNQ Thermal Limit
At four channels × 2 A × 100 mΩ, estimated package loss is 1.6 W. At 85 °C ambient:
- TJ <150 °C requires θJA <40.6 °C/W with no margin.
- The design goal TJ ≤125 °C requires θJA ≤25 °C/W before other losses.
The datasheet's 23.5 °C/W four-layer example and 57.3 °C/W two-layer example do not validate this PCB. Continuous 2 A/channel release is blocked by copper, via, airflow/enclosure and full-load temperature verification.
All eight channels at 2 A imply 16 A through the common system return and approximately 3.2 W total switch-package conduction loss under the same 100 mΩ assumption. Connector, copper, via and harness-return ratings remain layout/system validation items.
Sensor Current Evidence
TPS2553-Q1 equations for 232 kΩ give approximately 99.67 mA minimum, 117.01 mA nominal and 137.34 mA maximum at nominal resistance. Applying the resistor's ±1% tolerance expands the calculated bound to 98.67–138.64 mA.
Accordingly:
- Connector continuous budget: ≤90 mA, including I2C pull-ups and rail burden.
- Fault/current-path sizing ceiling: ≥139 mA.
- No externally powered sensor module and no SDA/SCL drive while module power is off. Reverse-voltage shutdown is delayed and is not galvanic or ideal-diode isolation.
External Power Limits
- J3 requires an external protected regulated 5 V source rated at least 3 A. No source or raw-vehicle front end is installed.
- J4/J39 require conditioned 9–16 V feeds with external bank fusing, surge/load-dump, reverse-polarity and backfeed control.
- D6/D20 SMBJ18A devices are local downstream clamps only.
- SMBJ18A has 18 V standoff and a listed 29.2 V clamp at 20.5 A. Because 29.2 V exceeds the VNQ's 28 V normal operating ceiling, the external conditioner must define a residual surge envelope and the design still requires an authoritative absolute/transient-margin coordination check. The local TVS alone is not qualification evidence.
- The carrier does not contain reverse protection or a separated protection-ground network and must not be described as raw-battery ready.
Connector Rating Evidence
- The KF301-2P source used for J3 and the other two-position field terminals lists 16 A rated current. Its 200 VAC/1 minute entry is a dielectric-withstand test, not a stated continuous operating-voltage rating; production operating voltage remains unresolved.
- Phoenix Contact 1729034 current and voltage ratings for J18/J37 remain evidence-pending in the datasheet service. Their actual sensor load is limited to 90 mA by the TPS2553 branch, but that circuit limit is not a substitute for connector documentation.
- Samtec SSW-122-02-G-S per-contact current and working-voltage ratings were not recovered as citable evidence from the attached series drawing and remain unresolved. Do not infer them from distributor/BOM descriptions.
- The KF45C-9.5-2P bank connector's claimed 32 A capability is currently only a BOM-description assertion; authoritative current, conductor-size, temperature-rise and voltage ratings remain required before releasing an 8 A bank.
Bulk Capacitor Production Gap
- C7/C8 are generic 10 µF, 16 V, 0603 entries on protected 5 V; C9/C16 are generic 47 µF, 35 V, 1210 entries on the conditioned bank feeds.
- No production MPN or capacitor datasheet is attached. Effective capacitance under DC bias, ripple-current capability, ESR, temperature grade, lifetime and surge margin are therefore unverified. C9/C16 must not be considered validated merely because the nominal 35 V value exceeds the local TVS's listed clamp voltage.
Naming and Topology Corrections
The U5/U6 paths now use HS_OUTx_GPIO, HS_OUTx_CMD and BANK_x_HS_OUTx_POS naming. Output connector pin 2 is positive switched output, pin 3 is ground/reference and pin 1 remains intentionally unused.
Check Boundary
- Current schematic count: 195 components, 122 nets.
- The eight new resistors and two new CAN VIO capacitors are the exact added part count in this pass; U2/U4 were replacements, not count additions.
- Production MPN assignment for generic resistor/capacitor instances is a sourcing task, not a floating-pin ERC defect. R45–R48 now have explicit 0.25 W minima, but their production working voltage remains open.
- Fresh ERC: zero errors; two warnings remain because D5/D11 library symbols lack informative pin-type metadata. Connectivity is CANH to pin 1, CANL to pin 2 and GND to pin 3, but the metadata warning itself is not cleared.
Deferred Physical Work and Residual Blockers
- APPROVED current target: Polycase WC-35 + WX-34 aluminum backplate, factory clear polycarbonate lid. User fits standoffs to carrier holes; no existing enclosure boss match required. Prior Zulkit150 and all other targets below are withdrawn historical research.
- Current proposal130 ×130 mm,16,900 mm² gross,6,009 mm² aggregate component area (~35.6%),195components/122nets. Nominal254 ×174 ×88 mm cavity gives22 mm side/62 mm end margins; approximate254 ×173 ×1.524 mm plate gives21.5 mm side margins. Arithmetic only: CAD central-area/notch fit is unverified because direct drawing retrieval did not expose geometry. No sourced constraint currently requires changing130square, but actual fit is not claimed.
- See
pcb-placement-proposal for complete datum/connector plan, four nominal3.2 mm NPTH holes at5 mm edge offsets with8 mm hardware keep-outs and user-adapted backplate. Current lead-reported Flux SDI v3/JLCPCB SDI4Layer v1,35µm copper baseline remains unchanged;2oz is consideration only. Proposed4layers1.6 mm FR-4/through vias do not validate2 A/channel,8 A/bank or16 A return. No PCB resize/placement/routing/CAD/export performed; placement proposal awaits user approval.
- Current enclosure target supersedes all larger-box proposals below: user explicitly selected Zulkit 150 × 150 × 90 mm from the linked family, not a 150 mm square PCB. Confirmed gray two-pack G-3 / B0FY3DW2RW, source, lists internal 131 × 131 × 82 mm ±2 mm; general volume only, not certified clear floor. Clear-cover single-pack B07RNM73P3 separately lists the same general dimensions; pack/cover is unconfirmed and boss tooling must not be transferred between models.
- The previous 220 × 170 × 110 mm recommendation and 180 × 110 mm PCB envelope are withdrawn. A 120 × 120 mm candidate could be studied after user approval and usable-floor evidence: 14,400 mm² gross versus ~5,100 mm² footprints, nominal 5.5 mm per-side general-interior margin (4.5 mm at minus-2 mm bound). Area alone does not prove fit of 195 components, four ~57 mm socket rows at 25.4 mm pair spacing, terminal wiring/tools, USB plugs, antenna cables or VNQ thermal copper. Two domains side-by-side with opposite-edge terminal banks and remaining terminals on service/feed edges is provisional only; 8 A/bank and 16 A return still need dedicated copper and full-load verification.
- Target-specific usable-floor/boss geometry and threads were not recovered. One dimensioned usable-floor/mounting drawing for the chosen 150 mm variant is the remaining mechanical input to derive a real outline; no board resize, holes, placement, routing or CAD was performed. Older enclosure research below is historical, not an active recommendation.
- PCB remains unfinished on a provisional 160 × 100 mm outline.
- Current socket placement is still approximately 7.05 mm row spacing; 25.4 mm has not been applied or fit-checked.
- U2/U4 replacements and R85–R92/C23/C24 require board placement during the deferred layout pass.
- Actual enclosure mounts, smaller envelope, cable access and antenna keep-outs are unknown.
- Qualified power conditioners, real loads, harness gauges, branch fuse coordination, behavioral models, full-load tests and automotive EMC/transient tests remain unresolved.
- J19/J23 are manual feed links, not automatic power ORing; each must be removed before powering the corresponding removable board by USB.
- The supplied seller map is now implemented identically for both boards and recorded in
esp32-socket-pin-map. A final continuity check remains required because the evidence is a seller graphic, not a controlled manufacturer drawing. GPIO35–37, USB GPIO19/20, straps GPIO0/3/45/46 and onboard-RGB GPIO48 remain excluded.
- Socket geometry is user-verified as 2.54 mm pitch and 25.4 mm row spacing, but the unplaced PCB still shows approximately 7.05 mm spacing. Mount-boss positions are unknown. These block placement/fabrication, not the local schematic/BOM fixes above.
- Correction / rejected target: the user confirms the former LMioEtool enclosure's 176 mm span and 173 mm clear length are both accurate; lid threaded inserts and gasket roundover intrude past the screw centers. This is not internally conflicting evidence. The old enclosure, 184 × 70 mm outline suggestion and derived hole coordinates are withdrawn.
- Replacement link resolves to Zulkit T-4 / B0FY3GXLRW, 200 × 100 × 70 mm two-pack, listed internal 180 × 80 × 63 mm ±2 mm. Its narrow selected default is not the recommendation. Research recommendation for approval: 220 × 170 × 110 mm clear-cover Zulkit, B07RVN91WB, model 1, single-pack, listed general interior 204 × 152 × 103 mm ±2 mm. No verified floor/mount drawing or thread data was recovered.
- Linked-family 220 × 170 × 110 mm two-pack G-5 / B0FY3H5MLT prints internal 8.0 × 6.0 × 4.1 inches / 254 × 152 × 103 mm. The length is physically inconsistent; do not transfer another variant's internal dimensions or mounting geometry to it. The coherent clear-cover ASIN is the preferred researched option, not a fit release.
- Propose 180 × 110 mm planning envelope, centered datum, long axis Y, top toward lid; no mounting coordinates specified. The nominal clear-cover general-interior margins are 12 mm at ends and 21 mm along sides, not verified clearances. Minimum remaining mechanical input: one dimensioned base/boss-top outline and mounting pattern with datum, boss diameter/height, insert thread/safe engagement, and closed-lid height clearance. Exact machining/tolerance refinements may follow documented nominal geometry, but unavailable geometry must not be guessed.
- Domain-separated long-edge screw-terminal banks, adjacent high-side thermal copper, feed end opposite CAN/USB service end, top active/connector/socket population and allowed bottom low-profile passives are provisional. External antenna cable loops, USB plugs/bends, terminal wire/tool access and boss keep-outs need explicit room. Proposed 4-layer 1.6 mm FR-4 with through vias and consideration of 2 oz outer copper is subject to fab feature/clearance limits and a higher-than-standard four-layer quote; no cost quote or thermal fit validation has been performed. Board resize, holes, placement, routing and CAD remain unattempted by instruction.
- Planned physical direction remains 4-layer 1.6 mm FR-4, through vias, 2 oz outer copper under consideration, long-edge screw terminals, separated bank power paths, USB service ends and external IPEX antenna cables. Thermal and current-density confirmation is still required for every 2 A channel and the 16 A common return.