30 components and 16 nets built using verified component information from supplied exports. Source pin-to-net connectivity was absent; this is an engineered reconstruction, not an exact electrical import. Schematic prototype only: not ready to power or manufacture.
Verified topology
Dual-pot diode-steered 555 timing, charge and discharge paths with fixed limiting resistors; pins 2 and 6 share timing capacitor.
Timer output drives gate-driver channel B, then MOSFET gate through 47 ohms; gate pulldown fitted. Channel B noninverting polarity is documented, but fresh reviewer retrieval did not independently settle it.
Fuse and series input diode feed VLOAD and block intentional return to battery.
You were right—I hadn’t used your files yet; now I have built a 30-component, 16-net schematic using their verified component information, with adjustable timing and capacitor energy capture. The exports lacked electrical connectivity, so this is newly wired—not an exact import.
Do not power it yet: the controlled return converter remains unfitted, peak-current protection is incomplete, and a MOSFET pin-numbering defect blocks PCB use. The remaining fixes and verification results are in Schematic Review Results.
I’ll finish the controlled back-EMF return stage so captured energy feeds the circuit’s positive rail, with battery backfeed blocking and current/overvoltage protection. I’ll complete and check the schematic first; powered testing will still be needed before it can be called safe to use.
I haven’t finalised it in this turn—no circuit changes were made. The project remains at 30 components and 16 nets, with adjustable pulse timing and back-EMF capture into a capacitor, but the controlled return to the positive rail is still incomplete.
Still required:
Fit and wire the regulated return converter.
Add peak-current and recovery-capacitor overvoltage protection.
Correct the MOSFET pin mapping and verify the completed circuit.
Do not power it yet or connect the recovery capacitor directly to the positive rail. Execution stopped because your ACUs ran out.
Schematic build only for the 12V Pulse Energy Recovery Prototype. Experimental low-side inductive-load driver, conservative externally enforced 1A peak target. Not ready to power, fabricate, or order; no arbitrary-load compatibility or battery-charging claim.
Completed: direct source inspection; verified inventory; reuse U1 NE555P, U2 TC4428AEPA and Q1 IRFP250NPBF; actual protected input, adjustable timer, gate driver, load connector, capture rectifier, reservoir/bleeder/clamp and external return interfaces wired.
Completed: 30 components, 16 named nets, 144 persisted wire vertices and three intentional U2 no-connects verified by fresh full schematic read.
Completed with limitation: controlled-return interface is J3/J4 plus D3; the CC/CV regulator itself is NOT selected, populated, or implemented. This is an external module requirement, not an integrated regulator schematic.
Failed/blocked: Q1 library physical source/drain numbering is wrong; PCB integration blocked until library repair. Q1 is explicitly excluded from PCB and pick/place, retained for off-board wiring by verified lead function only. Quality report filed.
Unattempted by scope: PCB placement/routing/manufacturing, SPICE, measured recovery efficiency, independent safety review. Schematic uses automatically placed grouped parts and named-net stubs, not a polished hand-arranged circuit drawing.
Downloaded both directly and inspected actual text/archive. GenCAD identifies KiCad 9.0.9, units inch; outline bounding extent is 200.0 x 120.0mm. There are 18 electrical components, plus two blank root pseudo-components. SIGNALS lists +12V, CTRL_BYPASS, 555_OUT, SW_NODE, TIMER_DISCH, HV_PULSE_OUT, GATE_DRV, MOSFET_GATE, TIMER_RC, GND, but no NODE assignments. ZIP contains four copper Gerbers, masks/pastes/silks, edge cuts, drill file, pick/place, multiple BOM formats and .d356. The .d356 marks every component terminal N/C; only vias carry GND. Thus neither source provides a reliable component-pin netlist. All present wiring is newly engineered, not a recovered/imported original schematic. Copper forensic reconstruction was deliberately not pursued.
Table
Source designators
Verified source inventory / values
Present design decision
U1
NE555P
Reused existing U1
U2
IR2127PBF
Existing TC4428AEPA retained for low-side drive; IR2127 not imported
Q1
IRFP250NPBF
Reused logical pin names; PCB-numbering defect isolated by off-board exclusion
J1,J2
XY301V-A-5.0-2P
Same connector type; interfaces newly assigned
D2
CVFD20065A
Not imported; current D2 is battery block, D4 UF4007 is capture diode
BOM says ECA-1HM100 but 1000uF and 450V properties
Internally contradictory, NOT treated as real rating. Replaced with explicit off-board 1000uF/100V reservoir interface
New components are RV1/RV2 100k linear pots, R6 default-disable, J3/J4 external buck interfaces, J5 enable jumper, D1 timer steering, D2/D3 reverse-blocking Schottkys, D4 capture rectifier, D5 TVS, F1 1A fuse and C8 470uF/25V rail bulk.
Actual circuit and interface mapping
J1 pin1 battery positive -> F1 -> D2 anode; D2 cathode -> VLOAD. J1 pin2 GND. D2 blocks intentional recovered current into battery; semiconductor leakage is not zero. Bench input 12V nominal, maximum 14.4V, no charging source or automotive surge environment.
J2 pin1 VLOAD, pin2 SW_NODE. External inductive load lies between these pins. Q1 functional D to SW_NODE, S to GND, G to MOSFET_GATE. Resistive loads have no useful inductive recovery; motors/transformers require additional application-specific analysis.
U1 OUT -> U2 IN_B -> OUT_B -> R3 47ohm -> gate. R4 100k gate-source pulldown. U2 IN_A grounded, OUT_A/NC/NC_2 intentionally unconnected. Manufacturer page-1 polarity diagram shows A inverting and B noninverting;
inspected diagram. U2 verified operating supply 4.5–18V.
U1 VCC and U2 VDD use VLOAD with separate 100nF plus 10uF bypass/bulk. Recovery may sustain the controller with the battery removed: open J5 before shutdown.
Timer: VLOAD -> R1 1k -> RV1 rheostat -> DISCH; D1 anode DISCH/cathode TIMER_RC bypasses discharge branch during charge. TIMER_RC joins U1 TRIG/THRES, C1 positive and RV2 rheostat; RV2/R2 return to DISCH. Both pots use library pin1 as one end and pins2+3 tied. Datasheet physical wiper is pin2, whereas library labels it pin3; tying both pins2+3 makes this particular rheostat electrically valid despite that labeling mismatch.
Inspected TT diagram. Confirm knob direction by ohmmeter.
R6 10k pulls RESET to GND; J5 short connects RESET to VLOAD. Open J5 is default disabled, not a safety-rated interlock. U1 CONT -> C2 10nF -> GND.
D4 anode SW_NODE/cathode VREC. C5 off-board P1 is explicitly positive lead, P2 negative/GND. R5 10k/1W and D5 cathode VREC/anode GND. Candidate ECA-2AHG102 is verified 1000uF/100V, +/-20%; negative sleeve marking identifies negative lead. Its library part had two unlabeled '~' pins and no confirmed polarity-to-pad mapping, so it was removed rather than guessing; C5 remains an explicit generic external interface with no assumed footprint.
IRFP250N physical leads are G1, D2, S3 (tab drain), not library G1/S2/D3; VGS maximum +/-20V and drain breakdown minimum 200V under datasheet test conditions.
D4 UF4007 is 1000V repetitive reverse, 1A average under stated thermal/lead conditions.
D2/D3 1N5822 are 40V blocking and 3A average under specified board/temperature conditions, not unconditional current capability.
SMBJ33A: standoff 33V, breakdown 36.7–40.6V at 1mA, specified maximum clamp 53.3V at 11.3A. Its 600W 10/1000us rating is nonrepetitive, temperature-derated, with specified mounting. It is not a continuous dump load and does not guarantee safe repetitive operation.
Computed at 53.3V: R5 dissipation 0.284W; reservoir stored energy 1.420J; ideal discharge to 5V approximately 23.7s. Verify with meter; dielectric tolerance and external module alter this. Stored energy remains hazardous to parts and can spark.
R1 worst-case 14.4V/1k dissipation 0.207W, so 0.25W minimum modeled; increase margin in physical part selection. Generic packages are schematic intentions, not confirmed purchasing selections.
TI conventional astable equation is tL approximately 0.693 RB C. With 1k–101k and 100nF, ideal off-time is 69.3us–6.999ms. Diode steering changes charging equation; on-time is approximate, not the unmodified TI astable equation. Real frequency/duty require scope measurement; controls adjust charge and discharge durations, not independent frequency and duty knobs. Pot value comes from exact B100K library identity; physical pin diagram was directly inspected, datasheet semantic query did not resolve its bundled question.
External regulator requirements — not populated
Use a documented nonisolated buck CC/CV module, common input/output ground, continuous input rating at least 80V, with supported pulsed-source/capacitive-load stability and adjustable current limit. Do NOT use an ordinary LM2596/XL4015 module on VREC. Target CV 10.5V before D3, CC <=0.25A initially; load rail contribution begins below approximately target minus diode drop, intentionally lower than normal battery-fed VLOAD. Require UVLO/shutdown near 16V VREC or higher, controlled startup, harmless no-load behavior and no output overshoot above 11V. Confirm input absolute maximum exceeds worst measured clamp/overshoot, not just 33V nominal TVS rating. Output diode is present on this schematic even if module lacks reverse blocking.
No selected module is asserted to meet these requirements. Returning stored energy is blocked until module documentation and bench verification exist. Capture testing without a module is short, low-energy only; otherwise capacitor reaches clamp and TVS overheats. Inductive flyback can include continued battery energy during turn-off, so capacitor charging alone is not proof of battery savings or recovery efficiency.
Fresh verification and unresolved issues
Fresh full read confirmed every intended net and no unexpected floating functional pins; three unused U2 pins are NC. Fresh floating-pin/driver-conflict/shorted-component checks reported no violations, but Missing Power ERC has two Errors: U1 VCC and U2 VDD on VLOAD are not recognized as driven. External battery through passive F1/D2 is the intended source; no artificial source/flag was added merely to silence this. Resolve source semantics in a later review; errors do not validate power safety.
Fresh airwire check has 54 Errors, expected because PCB is wholly unrouted and layout work was explicitly excluded. These are real PCB incompleteness, not schematic disconnections. No claim of overall ERC/DRC pass or PCB readiness.
Remaining blockers/checks: repair Q1 library numbers before PCB; source actual generic R/C parts and validate footprint/polarity/ripple ratings; confirm F1 DC interrupt rating and fuse coordination (1A fuse is NOT a 1A instantaneous limiter); independent electrical review and transient simulation; regulator selection/documentation; input/inductor peak-current enforcement; repetitive TVS/diode thermal energy and wiring overshoot; EMI and layout return paths. Limit scope to measured <=1A peak with lab current limit and short initial pulses, not arbitrary load operation.
Bring-up plan (future, after blockers are cleared)
Keep J5 open, no battery, no module, no inductive load. Verify all actual lead functions and polarized parts with meter/datasheets; Q1 uses physical G1/D2/S3. Verify no VREC-to-VLOAD short and D2/D3/D4 direction.
Use isolated current-limited bench 12V supply, initially <=50mA, with load/module absent. Expect VLOAD approximately input minus D2; Q1 gate should stay below 0.5V when J5 open. Stop on heat, unexpected current or oscillation.
Close J5 only for unloaded timer/driver checks; scope TIMER_RC and gate. Gate must switch cleanly and stay below 15V; open J5 must force gate low. Characterize full pot extremes and shutdown/battery-removal behavior. Do not assume startup pulse absent without checking.
Use a known small inductor with series resistance and external current measurement/limit, begin far below 1A and minimum on-time. Check load current ramp and drain/VREC waveforms before increasing repetition. Require measured VREC <60V and drain <65V in initial test setup; these are engineering test limits, not guarantees from the TVS rating. Stop if reservoir rises toward clamp, rail exceeds intended range, or parts heat. Ordinary current-limited supplies may not limit short pulses because C8 supplies them.
Separately verify candidate module with controllable source: UVLO, CC <=0.25A, CV 10.5V, no overshoot, stability and reverse behavior. Only then connect J3/J4 correctly, characterize handover as VLOAD sags, and prove no intentional battery charge current.
Measure battery input energy, load output energy and reservoir energy change over identical load cycles; account for losses and no-load capacitor precharge. Open J5, disconnect battery/module, wait and measure reservoir below 5V before handling. Never short-discharge capacitor.
Scope and status
Source evidence
Actual circuit and interface mapping
Rating evidence and calculations
External regulator requirements — not populated
Fresh verification and unresolved issues
Bring-up plan (future, after blockers are cleared)