Spearhead Electrical Master Document

Doc ID: SPH-E-001 Revision: 0.19 Author: errrks Date: June 2026 Status: Phase 1 specified. Phases 2–4 are architecture stubs pending detailed design.

Purpose and Scope
System Architecture
Requirements Traceability
Phase 1: Electric VTOL (Detailed)
- 4.1 Battery System
- 4.2 Power Distribution
- 4.3 Kill Switches
- 4.4 Voltage Regulation
- 4.5 VTOL Motors and ESCs
- 4.6 Wiring Harness and Connection Plan
- 4.7 Flight Controller
- 4.8 GPS and Navigation Sensors
- 4.9 CAN Bus Architecture
- 4.10 Tail Electronics
- 4.11 Servo System
- 4.12 Telemetry and Communications
- 4.13 Health Monitoring
- 4.14 Heading LEDs
- 4.15 Component Locations
- 4.16 Phase 1 BOM and Cost Estimate
Phase 2: Hybrid Integration
Phase 3: BVLOS (Stubs)
Phase 4: Payload Bay (Stubs)
Work Packages
Open Questions
Revision History

1. Purpose and Scope

This document defines the electrical architecture for Project Spearhead across all four development phases. Phase 1 (electric VTOL validation) is specified to component level. Phases 2–4 contain architecture stubs and open questions to be resolved as the project progresses.

Scope: power system, propulsion electronics, avionics, sensor integration, communications, harness design, component layout, and per-phase work packages.

Out of scope: structural design, ArduPilot parameter tuning (see docs/reference-guide/ardupilot-quadplane.md), flight mechanics.

2. System Architecture

2.1 Power Topology

                     ┌──────────────────┐
                     │  12S LiPo Pack   │
                     └────────┬─────────┘
                              │  QS8-S (pack socket, anti-spark, ground safing disconnect)
                  [No hardware HV kill, PT1: software E-stop + QS8-S unplug, OQ-05 closed]
                              │  8 AWG
                        ┌─────┴──────┐
                        │ FCHUB-12S  │  44.4V nominal
                        │    PDB     │  battery pads = FC power tap point
                        └┬───┬───┬──┬┘
                        ESC1 ESC2 ESC3 ESC4  (all DroneCAN)
                          │   │   │   │
                          M1  M2  M3  M4   (FR / RL / FL / RR)

   FCHUB battery pads ─ 16 AWG XT60 pigtail ─ [PM02 V3]
                                                5.2V/3A analog → 6-pin → Pixhawk 6C POWER1
   FCHUB 440A sensor + VBat divider ─ analog ─→ Pixhawk 6C POWER2 (vehicle V/I, sense only)

                  FCHUB-12S PDB rails
                        │
          ┌─────────────┼──────────────────┐
          │             │                  │
   [5V/5A BEC]  [HV spur → Tail PM12S-3]  [12V/4A BEC]
  peripheral 5V   (Vx → ruddervators,     payload (Ph 4)
  (CAN/RC/telem)   5V → adapter+Here4)    + contactor coil
                                           if D2 picks one)

  Flaperon servo power: forward robocombo DC-DC buck @ 8.4V (HV input
  at the FCHUB pads), not shown above. See §4.4.
  Tail PM12S-3 HV input: 18 AWG branch off the ESC-lead extension
  joint at the boom root (not a FCHUB rail), behind the HV kill. See §4.10.
          │
    [LV Kill Switch]
   (FC + periph 5V)
          │
     ┌────┴──┬────────┐
  Pixhawk   GPS   Sensors/RC/
    6C    (shelf) Telemetry
 (POWER1 = PM02 V3 supply, POWER2 = FCHUB analog V/I sense)

2.2 Communications Topology

Pixhawk 6C
│
├── CAN1 ─── [120Ω] ── CAN splitter/hub
│        ├── ESC1..ESC4 (nodes 1–4, short boom stubs)
│        └── Here4 (tail, node 11) ── CAN-PWM adapter (node 12) ── [120Ω]
│
├── CAN2 ─── [reserved / Phase 3 expansion]
│
├── GPS1   ─── u-blox F9P primary GPS (serial, SERIALx_PROTOCOL = 5)
├── TELEM1 ─── Holybro SiK V3 telemetry radio (MAVLink 2.0; STORK-borrowed first-flight unit, long-term TBD)
├── RC IN  ─── Radiolink R9DS receiver (SBUS), paired with the AT9S Pro transmitter (D3 RC side closed June 18)
├── TELEM3 ─── [reserved / Phase 3 long-range link]
├── GPS2   ─── Ainstein US-D1 altimeter (UART; CAN alternate frees this port)
│
├── I2C   ─── [backup airspeed transducer if I2C DLVR]
├── ADC   ─── Pitot transducer (analog); POWER1/POWER2 carry battery V/I
│
├── PWM OUT (IO MAIN 8 ch + FMU AUX 8 ch):
│     Output 1: Aileron, right flaperon (physical PWM from the 6C IO)
│     Output 2: Aileron, left flaperon  (physical PWM from the 6C IO)
│     Output 3: RuddervatorLeft  (SERVO_FUNCTION 79, via DroneCAN to tail adapter, node 12)
│     Output 4: RuddervatorRight (SERVO_FUNCTION 80, via DroneCAN to tail adapter, node 12)
│     Outputs 5–8: VTOL motor commands via DroneCAN ESC nodes 1–4 (no physical PWM)
│
└── USB-C ─── Ground configuration / parameter upload

VTOL motor commands go over DroneCAN via ESC_BM bitmask. Physical PWM outputs 5–8 are unused in Phase 1 unless DroneCAN falls back to PWM.

CAN bus length nose to tail: approximately 2.5 m. No signal integrity issue at 1 Mbps for this length.

2.3 Physical Location Overview

Zone	Contents
Nose (slimmed, v5)	No avionics requirement. Pitot mast and external antenna placement only if needed
Forward battery bay	motorobit 12S 22Ah 15C solid-state pack (3,709g, 190×78×126mm), QS8-S socket (anti-spark, ground safing disconnect), HV kill (implementation TBD, OQ-05). Moved forward for CG per the v5 layout
Avionics shelf (above battery, top access)	Pixhawk 6C, PM02 V3 analog power module (FC POWER1 supply; POWER2 = FCHUB analog V/I sense), primary GPS (F9P, serial) on the shelf cover (sky view), RC receiver, telemetry radio, LV kill switch. Compartment ~180×250mm footprint × ~90mm tall between the longerons (June 12 estimate, pending CAD)
CG bay / boom roots	Matek FCHUB-12S PDB (HV distribution + 5V/12V BECs, battery pads carry the FC power tap), CAN splitter/hub, forward robocombo buck (flaperon 8.4V rail)
Boom roots / tips (×4)	ESC externally mounted on boom near motor (within 150mm of motor centerline)
Motors (×4)	MAD V8013 PRO IPE 150KV, 3–5° forward tilt relative to aircraft body (motors vertical relative to gravity when aircraft is in cruise attitude)
Tail group (boom-mounted)	Here4 (secondary GPS), Matek CAN-L4-PWM CAN-to-PWM adapter, PM12S-3 (Vx servo rail + fixed 5V for adapter and Here4), ruddervator servo connectors
Wing panels (×2)	Flaperon servos (Kingmax CLS3015S, 8.4V), wing panel connector (signal + 8.4V + GND, 8.4V from the forward robocombo buck)
Airframe exterior	WS2812 heading LED strips (wing tips, nose, tail)

Zones follow Alperen's fuselage conceptual v5 (June 11): longer thinner nose, forward battery for CG, avionics shelf above the battery with top access, pusher between the fuselage and the boom-carried tail. The tail harness (CAN trunk + HV spur) therefore routes along a tail boom, not through a fuselage run. Shelf and bay dimensions are open, see the open-question list in docs/pt1-electrical-information-note.md (SPH-E-002).

3. Requirements Traceability

Derived from the Project Spearhead DAO proposal (AIP-006, March 2026) and meeting decisions through May 2026. The formal project charter referenced in proposal §5 has not yet been published to the repo. Until then, the AIP-006 proposal text is the authoritative requirements source.

ID	Requirement	Phase	Source
E-REQ-01	VTOL thrust-to-weight ≥ 1.5:1 at MTOW 25 kg	1	AIP-006 §9.2
E-REQ-02	Independent hardware kill switches for HV and LV. PT1 deviation (June 17): LV hardware kill only; HV path uses software E-stop + QS8-S unplug, no hardware HV kill (recorded §9.3 deviation, see §4.3). Full HV kill returns for the final product	1	AIP-006 §9.3
E-REQ-03	Real-time telemetry: battery SoC, generator output (Phase 2+), fuel level (Phase 2+)	1/2	AIP-006 §9.3, §9.5
E-REQ-04	Redundant GPS	1	AIP-006 §9.4
E-REQ-05	Radar or laser altimeter	1	AIP-006 §9.4
E-REQ-06	Front-facing obstacle avoidance sensor	3	AIP-006 §9.4
E-REQ-07	Long-range telemetry > 200 km	3	AIP-006 §9.5, §9.6
E-REQ-08	Secondary short-range RF link for local ops	1	AIP-006 §9.5
E-REQ-09	Payload bay: 12V power, CAN or Ethernet, quick-release	4	AIP-006 §9.7
E-REQ-10	Pre-flight diagnostics: battery, GPS, altimeter, sensor health	1	AIP-006 §9.8
E-REQ-11	Heading indicator LEDs	1	AIP-006 §9.8
E-REQ-12	Link-loss and engine-out autonomous recovery	1/2	AIP-006 §9.8, §2.5
E-REQ-13	Pixhawk-standard FC with quadplane transition logic	1	AIP-006 §9.4
E-REQ-14	IC engine ignition and optional starter	2	AIP-006 §9.2
E-REQ-15	Generator: avionics power and battery recharge during cruise	2	AIP-006 §9.3

4. Phase 1: Electric VTOL (Detailed)

Phase 1 scope: all systems needed for VTOL hover, control surface authority, structural validation, and detachable wing testing. No IC engine. No generator.

4.1 Battery System

Sizing

Based on the MAD V8013 PRO 150KV datasheet (measured at 48V, 25°C, sea level with FLUXER PRO 26×7.8 MATT prop):

Datasheet voltage vs. 12S operating voltage:

A 12S LiPo charges to 50.4V (4.2V/cell) and the datasheet was measured at 48V — placing the test conditions squarely within the normal mid-flight range of a 12S pack. The datasheet numbers are therefore directly representative rather than requiring large corrections. As the battery discharges, bus voltage drops and the motor must spin faster (more throttle) to maintain the same thrust, increasing current draw. The motor's operating zone shifts progressively across a flight.

Required thrust per motor: 25,000 g / 4 = 6,250 gf

Hover operating point across battery discharge (sea level):

Battery state	Bus voltage	Throttle (approx.)	Current/motor	Total draw	Motor zone
Full charge	50.4V	~57%	~17A	~68A	Continuous
~75% SoC (matches datasheet)	48V	~60%	~19.8A	~79A	Continuous
~50% SoC	45V	~63%	~22A	~88A	Continuous
Nominal	44.4V	~65%	~24.4A	~98A	Continuous / edge
Low battery warning	42V	~68%	~28A	~112A	Short-term zone

Motor operating zones from datasheet footnote (motor thermal limits, not ESC limits — see §4.13):

Below 24A/motor: Sustainable continuous
24–78A/motor: Short-term only, ~10–30s
Above 78A/motor: Non-working

The motor operates in the continuous zone for most of the usable flight window. It only enters the short-term zone near the low-battery warning threshold (42V / 3.5V/cell) — which is also the point at which landing should already be initiated. At altitude (1,200m, ~12% lower air density), add approximately 2–3% throttle to each row above, nudging the nominal and low-battery points further into short-term territory.

Battery capacity sizing: the QuadPlane hovers only during VTOL takeoff, landing, and transition and flies wing-borne the rest of the mission, so sustained hover is not the energy driver. At ~85A average hover draw a 22 Ah pack gives roughly 15 minutes of hover before reserve, well beyond the per-flight hover budget for Phase 1 testing and transitions. The pack selected June 17, 2026 is the 22,000 mAh (12S) motorobit solid-state LiPo (3,709g, 190×78×126mm, QS8-S), superseding the 16 Ah ProFuse picked June 11. This returns capacity to the earlier 22 Ah figure with no weight penalty: at 3,709g it is ~113g lighter than the 3,822g 16 Ah ProFuse, so the solid-state cell density offsets the extra capacity and the May mission-energy weight trade (which had dropped to 16 Ah to save ~1.7 kg) no longer applies.

Peak current (safety-margin thrust at 75% throttle, 36.6A/motor × 4):

4 × 36.6A = 146A (short-term zone, < 30s)
At 22,000 mAh: 146 / 22 = 6.6C. The 320A ESC-capped absolute peak is 14.5C, still inside the pack's 15C / 330A continuous rating

Thrust-to-weight at max thrust (100% throttle, 15,032 gf/motor × 4 = 60,128 gf = 61.3 kg):

T/W = 61.3 / 25 = 2.45:1 (E-REQ-01 of 1.5:1 exceeded)

Specification

Parameter	Target
Chemistry	LiPo, high-discharge
Cell count	12S
Nominal voltage	44.4V
Full charge	50.4V
Low-battery cutoff	3.5V/cell → 42V pack (ArduPilot warning), 3.3V/cell → 39.6V (critical)
Capacity	22,000 mAh (12S, motorobit solid-state pack, selected June 17, 2026)
C-rating	15C (330A continuous). Meets the ≥15C floor exactly. 330A continuous covers even the 320A ESC-capped absolute peak, but expect more voltage sag under the peak than the 60C ProFuse showed — verify on the first high-throttle hover
Connector	QS8-S socket on the pack (Amass QS8, anti-spark, ~120A continuous class). Harness side: QS8-S plug, 8 AWG to the FCHUB battery pads. Anti-spark, so there is no mating spark at 12S (see the §4.3 connector note). No Y-splitter: the FC power module taps the FCHUB pads
Form factor	Selected June 17: 12S 22,000 mAh 15C semi-solid LiPo (~290–300 Wh/kg; vendor lists it as "solid-state"; sourced from motorobit, TR), single pack, 190×78×126mm, QS8-S socket. Length 190mm drives the forward bay (34mm shorter than the ProFuse's 160mm height). Lock position with Alperen before the fuselage is built
Weight	3,709g, largest single CG impact item. ~113g lighter than the 16Ah ProFuse despite +6Ah, so the May weight trade is moot

Dumb LiPo pack (no BMS or CAN). A standard high-discharge 12S LiPo is used, not a smart battery. Alperen found the smart-battery option added roughly 1 kg, not worth it for the end product. Implications:

No CAN battery telemetry and no per-cell data. Vehicle current comes from the FCHUB-12S 440A sensor and pack voltage from the power module or a FC voltage sense (see §4.2 and §4.13)
Balance leads are used for charging only, not wired into the aircraft. In flight the pack connects to the HV path by main leads only
Set conservative pack-level low-voltage thresholds (no per-cell cutoff). Carry a plug-in cell checker for ground checks
Pack selected June 17 (WP-E01 reopened and reclosed): motorobit 12S 22Ah 15C solid-state LiPo, 3,709g, 190×78×126mm, QS8-S socket (supersedes the ProFuse 16Ah 60C selected June 11)

4.2 Power Distribution

Parameter	Value
Board	Matek FCHUB-12S V2, off-the-shelf PDB (no custom board for V1)
Architecture	Star: battery to the FCHUB-12S, individual leads to each ESC. The battery input pads also carry the FC power module tap (16 AWG XT60 pigtail), replacing the earlier battery Y-splitter
Input	8–60V (3–12S), battery main leads only (no balance harness)
Per-channel rating	70A continuous, 110A burst per ESC pad. Hover is ~24–31A per motor, peak ~80A per motor (capped by the 80A ESC), so the margin is ample
ESC connection	4× ESC power pads. Only the power pads are used. The signal and telemetry pads stay unused because the ESCs run DroneCAN (see §4.9)
On-board BECs	5V/5A (max 6A), 12V/4A (max 5A), 3.3V/0.5A. Rail assignment in §4.4
Current sensor	On-board 440A analog sensor for vehicle current (0–3.3V over 440A; ArduPilot `BATT_AMP_PERVLT 133.3`), wired to the 6C POWER2 analog CURRENT pin. Pack voltage from the V2 built-in 1K:20K divider (V pad set to VBat, `BATT_VOLT_MULT 21.0`) to the POWER2 VOLTAGE pin. FCHUB is the primary battery monitor (BATT1), PM02 V3 on POWER1 is BATT2. See §4.13
V2 breakout	JST-SH 8-pin: V (selectable 12V or VBat via the 1K:20K divider), G, Curr, TLM (BLHeli32, unused), S1–S4 (ESC signal, unused with DroneCAN). Ships with the SH1.0 8-pin cable
Switchable 12V	12VSW pad, on/off via FC PINIO (GPIO level), 2A constant. Candidate supply for the EV200 contactor coil (§4.3) or a switched LED feed
Per-branch fusing	100A fast-blow per ESC branch, optional for Prototype 1 (assess in WP-E02)
Physical	55 × 50 × 6 mm, 30.5 × 30.5 mm Φ3 mounting, 21g. Ships with a low-ESR bulk capacitor for ESC switching transients

The FCHUB-12S replaces the earlier custom PCB and bus bar option, in line with the V1 goal of minimizing custom boards. Holybro PM07 is not needed since the FCHUB current sensor handles vehicle-current monitoring. The 70A per-channel continuous rating comfortably exceeds the AMPX 80A ESC ceiling, and the 4×110A burst rating covers the 320A peak.

4.3 Kill Switches

Switch	Covers	Does NOT cover	Placement	Current rating
HV safing (no hardware kill, PT1)	Software E-stop (ArduPilot motor emergency stop on an RC switch) cuts CAN motor commands; QS8-S anti-spark unplug for ground safing	Does not physically interrupt the armed HV bus	RC switch + accessible QS8-S	software + connector
LV Kill	FC power (PM02 V3 5.2V to POWER1), peripheral 5V rail (GPS, RC, telemetry)	Both servo rails: forward robocombo (flaperons) + tail PM12S-3 (ruddervators + adapter + Here4), all HV-derived	Exterior, separate	≥ 10A

The tail HV→5V BEC powers the ruddervator servos, the CAN-PWM adapter, and the Here4, all from the HV bus spur. The whole tail control path therefore has no separate LV kill and de-energizes only when the pack is disconnected at the QS8-S (there is no hardware HV kill for PT1, see below), preserving ruddervator authority whenever the aircraft is armed. This is intentional, and it resolves the earlier concern about the adapter losing power on an LV kill, since it no longer draws from the CAN 5V rail.

PT1 implementation (resolved June 17: no hardware HV kill). AIP-006 §9.3 SHALL requires independent HV and LV kill switches. For PT1 the team accepts a recorded §9.3 deviation: no hardware HV kill. HV-path safing is the ArduPilot software E-stop (motor emergency stop on an RC switch, with the AMPX 2 s CAN watchdog as a backstop) plus unplugging the QS8-S anti-spark connector for ground safing. The LV kill is retained. Rationale (June 12): no clean contactor placement for PT1 without extra components and mass, and the QS8-S anti-spark already manages connection inrush. The full independent HV kill returns for the final product. The June 11 RC-switchable criterion and the candidates below are kept as record (superseded for PT1):

Candidate	How it switches	Mass	Notes
TE Kilovac EV200AAANA contactor	12V economizer coil (1.7W hold ≈ 0.14A) from the FCHUB V2 12VSW pad, switched directly by FC PINIO. No driver hardware. RC kill via RC_OPTION relay mapping, GCS kill via MAVLink relay	~450g	~€149. Proven interrupt rating (2000A at 320VDC). Closing onto the discharged ESC input caps at 50V is within rating, no precharge stage needed
MOSFET anti-spark switch (e-skate class, 200A cont / 400A peak, up to 20S)	Enable line interfaced to an RC switch or FC GPIO	~100–150g	~$80. Fails short (a failed kill no longer kills), no flight heritage, enable interface is an adaptation. Bench evaluation required before trusting it in the HV path
Software E-stop + ground safing only	ArduPilot motor emergency stop on an RC switch cuts CAN motor commands. AMPX throttle-loss protection backstops at 2 s. Pack connector unplug for ground safing	0g	Not a hardware kill. A recorded §9.3 deviation for PT1 only, full hardware kills return for the final product

The software E-stop layer costs nothing and gets configured regardless of which hardware path wins. The earlier WP-E02 options table (contactor + key switch, Anderson PP75 pull disconnect, E-stop pushbutton) is superseded: none of those are RC switchable, and the PP75 was undersized anyway at 75A-class continuous against a 98–124A hover draw.

Connector interaction (updated June 17): the selected motorobit pack carries a QS8-S socket, which is anti-spark. The connector itself limits the ESC cap inrush on mating, so there is no mating spark at 12S regardless of the kill choice. This removes the earlier "plain connector" argument for an inline hardware kill. With PT1 going software-only (no hardware HV kill, decided June 17), the QS8-S anti-spark is what manages mating inrush, and no AS150U adapter is needed.

LV kill switch: illuminated rocker or key switch, 10A @ 12V. Interrupts the FC power (PM02 V3 output) and the peripheral 5V rail (GPS, RC, telemetry). It does not cut either servo rail: the forward robocombo (flaperons) and the tail PM12S-3 (ruddervators + adapter + Here4) are both HV-derived and stay powered until the pack is unplugged at the QS8-S (no hardware HV kill for PT1).

4.4 Voltage Regulation

Rail	Voltage	Continuous Current	Loads	BEC type
FC power (POWER1)	5.2V	3A max	Pixhawk 6C (single supply on POWER1)	PM02 V3 (analog, bundled with the 6C), fed from a 16 AWG XT60 pigtail at the FCHUB battery pads. POWER2 carries the FCHUB analog V/I sense, not a power feed
Peripheral 5V	5.0–5.2V	4–6A	Shelf F9P GPS (serial), RC receiver, telemetry radio, sensors	FCHUB 5V, separate from the FC inputs. Tail devices (adapter, Here4) run on the tail PM12S-3 instead
LED 5V	5V	up to ~3A	WS2812B strips (wing tips, nose, tail)	Separate feed, high transient draw, kept off the FC rail
Flaperon servo	8.4V	~2–3A/servo, 15A available	2× Kingmax CLS3015S on wing panels	Forward robocombo DC-DC buck @ 8.4V (HV input at the FCHUB pads, 15A) via the wing connectors. See §4.11
Ruddervator servo	8V	~2–3A/servo, 15A available	2× Kingmax CLS3015S in tail bay	Tail PM12S-3 Vx rail (8V), fed from the HV bus spur. ~33 kg·cm at 8V vs 35 kg·cm at 8.4V, still ample. Its fixed 5V powers the adapter and Here4. No long LV run through the fuselage
Payload	12V	TBD	Phase 4 only — not needed for Phase 1	—

FC power: the Pixhawk runs from a single source, the PM02 V3 on POWER1. POWER2 is not a backup brick — it carries the FCHUB analog V/I sense (§4.13), so for V1 we trade FC power redundancy for vehicle-current sensing. This is a single point of failure on FC power, accepted to keep V1 simple. If FC power redundancy is wanted before extended flights, move the FCHUB sense to a spare ADC and free POWER2 for a backup BEC.

FC supply (Pixhawk 6C + PM02 V3, selected June 17): the FC runs from the analog PM02 V3 (2–12S, bundled with the 6C at rx-dynamic) on POWER1, fed by a 16 AWG XT60 pigtail soldered at the FCHUB-12S battery input pads. The pads are the same electrical node a battery Y-splitter would create, with one fewer connector pair in the 320A path, so the earlier AS150U/XT150 Y-splitter is dropped (it remains the fallback if the pad tap proves mechanically awkward). The module carries only the FC draw. The 6C's analog POWER ports are what enable the FCHUB vehicle V/I to feed POWER2 directly (§4.13); the digital PM02D used by the earlier 6X plan could not have done this, since its V/I ride I2C.

FCHUB-12S on-board BECs: the FCHUB provides 5V/5A (max 6A), 12V/4A, and 3.3V/0.5A (verified against the Matek page, June 2026). With the FC on its own power module, use the FCHUB 5V for the peripheral 5V rail (F9P GPS, RC, telemetry), and the FCHUB 12V for the Phase 4 payload and the HV kill contactor coil if an EV200 or GX11 is chosen. Keep the WS2812 LED strips on a separate 5V feed so their transients stay off the peripheral rail.

BEC / servo-rail selection (updated June 17): 1× Matek PM12S-3 (tail) + 1× robocombo DC-DC buck (wing).

Tail — Matek PM12S-3 (Aykut Havacılık TR, 6,900 TL, specs verified June 2026): 9–55V input with TVS protection, Vx selectable 5.25 / 6 / 8V at 15A continuous (25A peak), fixed 5.2V/4A, fixed 12V/4A, built-in 1K:20K divider, 53g. Vx set to 8V feeds the ruddervator servo rail (Kingmax CLS3015S), and its fixed 5V feeds the adapter logic and the Here4. Kept because it provides both the 8V servo rail and the 5V logic rail in one TVS-protected module.
Wing — robocombo DC-DC buck (8–60V in, 15A adjustable): HV input at the FCHUB pads, output set to 8.4V for the flaperon servos. Adjustability to 8.4V (which the PM12S-3 Vx cannot reach, capping at 8V) gets full Kingmax torque. Replaces the forward PM12S-3, which is dropped. Caveats: lock the output pot (threadlock), confirm it is a synchronous buck genuinely rated at 15A with heatsinking, and scope for output overshoot before a servo is connected.

This supersedes the late-June-11 "2× PM12S-3" plan and the earlier BEC12S-PRO / MBEC6S / Castle BEC Pro / PM07 candidates. PT1 runs the final 8.x V HV servos directly (no 5V test-servo stage, see §4.11).

4.5 VTOL Motors and ESCs

Status: ordered (6 of each, MAD Motor Poland, expected delivery 2–4 weeks from May 1, 2026; ATR certificate requested for Turkish customs — zero import tax if provided).

Motors

Parameter	Value
Model	MAD V8013 PRO IPE 150KV
KV	150
Weight	591 g
Battery	12S (44.4V nominal; datasheet tested at 48V)
Max temperature	87°C
Recommended hover thrust	4.5–6.5 kgf (Spearhead at 6.25 kgf/motor — top of range)
Maximum thrust	15 kgf (with FLUXER PRO 26×7.8 MATT at 48V)
Electrical efficiency	> 80% (peak ~82% at 65–75% throttle)
Max propeller	26" (boom spacing constraint)
Auto-center feature	Magnetic/mechanical indexing: props stop parallel to boom when power removed
Mount pattern	40mm bolt circle, M4×4
Bearings	Dual 6901 deep-groove
Quantity ordered	6 (4 installed, 2 spares)
Motor tilt	3–5° forward relative to aircraft body axis. The aircraft's natural cruise attitude (nose slightly elevated from wing incidence) places the motor thrust axes approximately vertical relative to gravity during hover — the structural tilt and aircraft attitude cancel. ArduPilot level calibration must be performed with the aircraft in cruise attitude (not flat on the ground) per `docs/reference-guide/ardupilot-quadplane.md §12`.

Performance table (MAD datasheet, 48V, 25°C, sea level, FLUXER PRO 26×7.8 MATT):

Throttle	Current (A)	Input Power (W)	Thrust (gf)	Efficiency (gf/W)	Elec. eff. (%)	Zone
30%	3.5	169	1,694	10.0	68.6	Continuous
40%	6.8	326	2,894	8.9	75.1	Continuous
50%	11.9	571	4,355	7.6	79.5	Continuous
55%	15.6	745	5,313	7.1	81.0	Continuous
60%	19.8	945	6,440	6.8	82.0	Continuous limit ≈ here
65%	24.4	1,163	7,318	6.3	82.2	Short-term (10–30s)
70%	30.0	1,420	8,476	6.0	82.3	Short-term
75%	36.6	1,726	9,759	5.7	82.1	Short-term
80%	41.7	1,961	10,389	5.3	82.0	Short-term
90%	58.5	2,718	12,985	4.8	80.4	Short-term
100%	77.8	3,569	15,032	4.2	78.0	Short-term / non-working

Datasheet operating zones:

Below 24A (< ~62% throttle at 48V): Sustainable continuous
24–78A (~62–99% throttle): Short-term only, ~10–30 s
Above 78A: Non-working zone

At Spearhead's MTOW and nominal battery voltage (44.4V), hover falls near 65% throttle at sea level and ~69% at 1,200m altitude — placing each motor at or just above the continuous zone boundary. Monitor ESC temperatures during all hover tests. Set Q_M_THST_HOVER accurately to minimize steady-state throttle.

Motor wire enamel risk: If motor phase leads are shortened, inspect strand insulation for enamel or clear varnish coating. Remove at the termination point before soldering. A coated termination creates a high-resistance joint that will overheat under load. (Ref: Feather PC3 incident — see Erick's failure notes.)

ESCs

Parameter	Value
Model	MAD AMPX 80A V2 (5–14S) DroneCAN
Protocol	DroneCAN V2.0 (also supports CyphalCAN; also accepts PWM 1050–1940µs)
Continuous current	80A (requires good heat dissipation)
Current limit (burst)	84A
Input voltage	5–14S, 16–64V
BEC output	5V / 200mA — insufficient for avionics; use only as CAN signal reference or auxiliary. Dedicated BECs required for avionics and servo rails.
CAN telemetry fields	Voltage, current, temperature, operation status (broadcast in real time over DroneCAN)
PWM input voltage	3.3V or 5V compatible
Protection level	IP67 (fully sealed, waterproof)
Operating temperature	-20°C to +65°C ambient
Thermal protection	>125°C: output reduces to 50% max. >140°C: full shutdown; will not restore until throttle zeroed AND ESC cools to <80°C
Short circuit protection	Shuts down output permanently — requires power cycle to recover. One motor loss per flight is unrecoverable in flight.
Throttle loss protection	Cuts power if CAN command absent for >2 seconds. ArduPilot CAN bus must maintain continuous ESC commands or motors cut.
Startup protection	Motor must start within 10s of throttle advance or ESC shuts down
Voltage protection	Below 16V or above 64V: alarm and no start (disabled during flight). 12S operational range 36–50.4V is within safe limits.
Sync freewheeling	Supported — active energy recovery on throttle reduction. Test interaction with IPE auto-center feature.
Weight	~90g (excluding wires)
Dimensions	79.6 × 36.0 × 23.5mm
Mounting	M2×4 screw holes (4 points)
Power wire	12AWG / 800mm
Phase wire	14AWG / 150mm
Signal wire	1000mm (bare leads: Black=GND, White=PWM, Red=CANH, Green=CANL + 5V BEC wire)
Quantity ordered	6 (4 installed, 2 spares)
Unit price	$60 (DroneCAN V2) + $30 shipping

DroneCAN node ID assignment:

Node ID	Motor position	Rotation	ArduPilot servo output
1	Front right (M1)	CCW	Servo 5
2	Rear left (M2)	CCW	Servo 6
3	Front left (M3)	CW	Servo 7
4	Rear right (M4)	CW	Servo 8

Key ArduPilot parameters:

CAN_P1_DRIVER = 1
CAN_D1_PROTOCOL = 1
CAN_D1_UC_ESC_BM = 0x000F   ; bits 0–3 enable ESC nodes 1–4
CAN_D1_UC_SRV_BM = 0x000C   ; servo outputs 3–4 (ruddervators) to the tail adapter, node 12
Q_FRAME_CLASS = 1             ; Quad
Q_FRAME_TYPE = 1              ; X

Full parameter list in docs/reference-guide/ardupilot-quadplane.md.

Propellers

Parameter	Value
Model	MAD FLUXER PRO 26×7.8 MATT — 2Pcs (SKU: 0103LX0013)
Diameter × pitch	660.4 × 198.1 mm (26 in × 7.8 in)
Material	Carbon fiber + resin
Finish	Matte (non-reflective)
Weight	69g per blade — 4 installed props = 276g total
Type	Fixed (not folding)
Hub bore	Ø10mm center shaft hole
Hub bolt pattern	Ø20mm bolt circle, 4× M3 screws
Accessories included	8× M3×14mm mounting screws, 2× cover plates per pair
Optimum RPM	1,900–4,000 RPM
Single thrust limit	18 kgf — not the binding constraint (motor limits at 15 kgf)
Working temperature	-40°C to +65°C
Quantity ordered	6 pairs (12 blades: 6 CW + 6 CCW); 4 installed, spares absorb early breakage
Unit price	$193.90/pair + $59.70 shipping
Reasoning (fixed not folding)	Folding props risk uncontrolled rotation in forward-flight airflow and uncertain locking behavior at speed

Hover operating point cross-check (prop test data vs motor datasheet):

From the standalone prop test table:

RPM	Thrust (gf)	Output Power (W)	Prop efficiency (gf/W)
3,226	4,110	444	9.3
3,619	5,120	625	8.2
3,971	6,160	834	7.4
4,232	6,940	1,008	6.9

At hover (~3,780 RPM, 44.4V nominal): interpolated ~5,700 gf, prop efficiency ~7.8 gf/W. Motor datasheet showed 6,440 gf at 3,811 RPM at 48V — consistent after accounting for voltage difference. Numbers agree across both datasheets.

Hover operating point (~3,780 RPM) falls within the 1,900–4,000 RPM optimum range.

Hub compatibility: The prop hub requires a 10mm shaft bore. Confirm V8013 PRO output shaft diameter is 10mm before installation. Torque plate bolt pattern (Ø20mm, M3×4) must match the motor's prop adapter plate — verify with Alperen's motor mount design.

Weight upgrade path: Ultra-light carbon variants (~20g savings per prop) deferred to ~August 2026 after initial testing (80g total saving across 4 props).

4.6 Wiring Harness and Connection Plan

Wire Sizing

Current ratings below assume silicone-insulated stranded copper in open-air UAV duty (not bundled in conduit).

Segment	AWG	Est. Length	Est. Current	Connector
Battery main leads	8 AWG	0.3–0.6 m (forward bay to FCHUB)	~98–124A hover, 320A peak	QS8-S plug at the pack (anti-spark; 8 AWG fits the QS8 solder cups, not 4 AWG), leads land on the FCHUB battery pads. FC power module taps the same pads via a 16 AWG XT60 pigtail
HV bus to each ESC (×4)	12 AWG	~1.4 m run (rear booms 20–30 cm longer)	~25–31A hover, 80A rated peak	XT90S or solder direct to ESC power leads
ESC power leads (factory, ESC to FCHUB)	12 AWG	800mm (factory)	80A rated	Bare tinned. Add XT90S or solder to FCHUB pads
ESC phase wires to motor (factory)	14 AWG	150mm (factory)	80A rated (short duty only)	Bare tinned — 5mm bullet at motor side
BEC input from HV bus	16 AWG	0.2–0.3 m	5A max	XT30

Wire notes:

The AMPX 80A ESC ships with 12AWG/800mm power leads and 14AWG/150mm phase leads. These are MAD's designed wire lengths for their thermal model at rated current.

Phase leads (14AWG/150mm): do not extend without upsizing to 12AWG or heavier. The 150mm is short by design — heat dissipates through the motor housing. A longer 14AWG run at 80A will overheat.
Power leads (12AWG/800mm): adequate for Spearhead's hover current (~25–31A/motor). At rated 80A the wires run warm but duty cycle is short (10–30s per motor zone limits). The boom run exceeds 800mm (confirmed June 11), so each ESC power lead is extended at the boom-fuselage junction with an XT90S splice. Extend with 12AWG throughout to match the factory lead. MAD rates that same 12AWG factory lead at the full 80A, and the extension carries the same current in series while Spearhead never sustains more than ~31A/motor in hover (the motor winding limit caps continuous draw at 24A, §4.5), so no upsize is warranted. Matching 12-to-12 also keeps the splice a clean same-gauge joint instead of a 12-to-10 step.
Main battery leads: 8AWG silicone on the pack's QS8-S (anti-spark). The earlier 4AWG spec does not fit the QS8 solder cups, and on this 0.3–0.6 m run the connector (QS8 ~120A continuous class), not the wire (~150A for 8AWG silicone in open air), is the thermal limit. Note the QS8's ~120A continuous rating sits at the top of the worst-case hover draw (124A at low-SoC nominal voltage), so monitor connector temperature during hover tests. Hover at 98–124A runs the lead warm, the 320A peak lasts seconds. Monitor connector and lead temperature during hover tests. The FC power module leg is small-gauge (FC supply only), an XT60 pigtail soldered at the FCHUB battery pads.

Current rating reference (silicone insulated, open air):

4 AWG: ~200A continuous
10 AWG: ~55–65A continuous
12 AWG: ~40–50A continuous
14 AWG: ~32–40A continuous

CAN Bus Wiring

Parameter	Value
Cable type	Twisted pair, 24 AWG, ≥ 10 twists/foot
Shielding	Single drain wire, grounded at Pixhawk end only
Connector	4-pin JST-GH (CAN H, CAN L, 5V, GND) per DroneCode standard
Total run	~2.5 m nose to tail — no signal integrity issue at 1 Mbps
Topology	Daisy-chain (node-to-node) with 120Ω terminators at each end of bus
Mid-bus nodes	No termination on intermediate nodes (ESCs, nose GPS)

Servo Wiring (Tail Harness)

Parameter	Value
Signal wire	26 AWG, < 0.5 m per servo within tail bay
HV spur to tail	18 AWG HV spur broken out at the tail-boom ESC-lead extension joint (boom root) to the tail BEC (5V is regulated at the tail, not run from the nose)
Signal logic level	5V (from the CAN-PWM adapter)
Servo power	8V (tail PM12S-3 Vx, shared ground)

Wing Panel Connector

The detachable wing panels each carry one flaperon servo. The connector at the wing root must break out signal, servo power, and ground.

Signal	Notes
Servo signal (×1 per wing)	5V logic PWM from the 6C IO output
8.4V servo power	From the forward robocombo buck via fuselage harness
Ground	Shared with avionics ground

Connector candidate: Molex Mini-Fit Jr 6-pin (known from Quiver, rated 9A per pin, latching). The Kingmax CLS3015S stall current (~2–3A) is well inside the per-pin rating. TBD in WP-E06.

Routing and EMI

Servo PWM wires run near ESC DC power in places. Alperen flagged a ~1 m parallel run between a main-wing servo PWM line and ESC power. Guidance for the discrete harness:

Separate signal and power runs where practical. A few centimeters of gap over a 1 m parallel run is usually enough at these currents.
Where a signal line must cross power, cross at right angles rather than running parallel.
Use twisted pair for PWM (signal with its own return) to reject common-mode pickup. CAN is already a twisted pair (see the CAN wiring spec above).
Keep PWM and CAN away from the motor phase wires, which carry the switching currents and radiate the most. Phase wires stay short and at the motor by design.
For a long run that must sit near power, use shielded cable with the shield grounded at one end only, the FC end. The same single-end grounding applies to the Phase 2 CDI shielding (§5.4).
Quiver avoids this class of noise mostly through multi-layer PCB routing with ground and power planes. Spearhead's discrete wiring does not have that, so the separation and twisting above carry the load.
The tail PWM is generated locally at the adapter (§4.10), so only short tail-bay servo leads run as PWM, not a 2.5 m run through the ESC field.
Flaperon PWM runs physically from the 6C IO outputs to the wing roots (~1–1.5 m per side). These are exactly the runs the separation, twisting, and crossing rules above exist for. A wing CAN-PWM node was considered on June 11 and dropped with the FC selection.

PT1 Harness Connection Table

Every PT1 cable end to end, per the June 11 request for the general layout. Lengths are estimates against the v5 fuselage and get confirmed once Alperen publishes shelf and bay dimensions. EMI rules from the previous subsection apply to H8, H10–H13.

#	From	To	Function	Wire	Connector(s)	Est. length	Notes
H1	Battery +/−	FCHUB-12S battery pads	12S HV main	8 AWG	QS8-S plug at the pack, solder at pads	0.3–0.6 m	No hardware HV kill in this run for PT1 (OQ-05 closed, software E-stop). QS8-S is anti-spark, so no mating spark; unplug it for ground safing
H2	FCHUB battery pads	PM02 V3 input	FC power leg	16 AWG	XT60 pigtail	0.2 m	Pad tap replaces the Y-splitter
H3	FCHUB ESC pads ×4	ESC power leads	HV per motor	12 AWG (factory 800mm + 12 AWG extension)	Solder at pads, XT90S splice at the boom-fuselage junction (boom exceeds the 800mm factory lead, confirmed June 11)	per boom geometry	Rear booms run longer, exact extension length pending Alperen's boom dimensions. On the tail-carrying boom this extension joint also breaks out the 18 AWG tail HV feed (H5)
H4	ESC ×4	Motor phase leads	3-phase	14 AWG (factory 150mm)	5mm bullets	150mm	Do not extend (wire notes above)
H5	ESC-lead extension joint (boom root)	Tail PM12S-3 input	Tail HV spur	18 AWG	Branch broken out at the ESC-lead extension joint (XT90S or soldered lap, per H3), XT30 at the PM12S-3	~1.5–2 m (boom-root joint to the tail group)	Breaks out at the same boom-fuselage junction where the tail-boom ESC lead is extended (H3), so no separate splice into a healthy conductor. Still behind the HV kill (downstream of the FCHUB ESC pad). Runs aft along the tail boom, path TBC with Alperen (inside-tube routing pending structural check)
H6	PM02 V3 6-pin	Pixhawk 6C POWER1	5.2V + analog V/I	Module cable	6-pin power	0.3 m	Stock cable. PM02 V3 = BATT2 (`BATT2_VOLT_PIN 8`, `BATT2_CURR_PIN 4`, MULT 18.18, PERVLT 36.36). Its current reads only the FC leg. See §4.13
H7	FCHUB 5V BEC	LV kill switch, then peripheral 5V rail	RC RX + telemetry power	20 AWG	Soldered bus / servo leads	Shelf-local	LED strips get their own feed (H14)
H8	Pixhawk 6C IO PWM 1–2	Wing root connectors	Flaperon PWM	26 AWG twisted with its GND	Molex Mini-Fit Jr 6-pin	~1–1.5 m per side	EMI rules apply: separation from ESC power, right-angle crossings, twisted with its own return
H9	Forward robocombo buck out	Wing root connectors	Flaperon servo power	18–20 AWG	Same Mini-Fit Jr	~1–1.5 m per side	8.4V (Kingmax CLS3015S). The robocombo HV input taps the FCHUB pads
H10	Pixhawk CAN1	CAN hub (CG bay)	DroneCAN trunk	24 AWG twisted pair	JST-GH 4-pin	0.3–0.5 m	120Ω at the FC end
H11	CAN hub	ESC1–4 CAN leads	ESC stubs	Factory 1 m signal leads	JST-GH / solder	Trim toward <0.3 m where layout allows	Daisy-chain along booms instead if stubs misbehave (§4.9)
H12	CAN hub	Primary GPS (shelf)	CAN branch	24 AWG twisted pair	JST-GH	0.3–0.5 m	Node 10
H13	CAN hub	Here4, then CAN-L4-PWM (tail)	CAN tail trunk	24 AWG twisted pair	JST-GH daisy-chain	~1.5–2 m	120Ω at the last tail node. Runs alongside H5 on the boom, stays twisted
H14	FCHUB 5V (separate feed)	WS2812 strips	LED power, data from FC GPIO	22 AWG	JST	Per airframe	Transients stay off the peripheral rail
H15	Tail PM12S-3 outputs	Vx → adapter servo V+ rail + ruddervators. Fixed 5V → adapter logic + Here4	Tail servo and logic rails	20 AWG	Servo leads / JST-GH	<0.3 m	Keep the two feeds separate. Vx set to 8V (PT1 = final HV servos, §4.10)
H16	CAN-L4-PWM PWM 1–2	Ruddervator servos	PWM, 5V logic	Servo leads	JR 3-pin	<0.3 m	Generated locally, no long PWM run
H17	Pixhawk TELEM1	Holybro SiK V3 radio	MAVLink UART	JST-GH 6-pin	JST-GH	0.2–0.3 m	SiK V3 (915 MHz, 100 mW) borrowed from STORK for first flight (D3, June 18). Long-term unit still TBD (range-limited)
H18	Pixhawk RC IN	Radiolink R9DS RX	SBUS	3-pin servo	servo	0.2 m	Radiolink AT9S Pro transmitter + R9DS receiver, SBUS on the dedicated RC-IN port (D3 RC side closed June 18)
H19	Pixhawk UART4	US-D1 altimeter	UART	JST-GH to US-D1 lead	JST-GH	0.3–0.5 m	US-D1 also offers CAN if the UART budget tightens
H20	FCHUB V/I sense	Pixhawk 6C POWER2 (Cur→pos 3, V→pos 4, G→pos 6; 4-wire, no 5V)	Vehicle current + voltage (FCHUB = BATT1, BATT_MONITOR 4)	26 AWG	6-pin power	Shelf-local	V-pad jumper = VBat (1K:20K, `BATT_VOLT_MULT 21.0`, ADC pin 5). Curr `BATT_AMP_PERVLT 133.3` (ADC pin 14). Leave POWER2 5V pins open. See §4.13
H21	Pitot transducer	Pixhawk I2C or ADC	Airspeed	Per model	Per model	Shelf-local	OQ-06

4.7 Flight Controller

Parameter	Value
Model	Pixhawk 6C + PM02 V3 (selected June 17, supersedes the June 11 6X). Full-size 6C (not the Mini), bundled with the analog PM02 V3 12S power module and a cable set. Selection rationale: the 6C IO PWM pins drive the wing flaperons directly (same as the 6X plan, so the wing CAN-PWM node stays dropped), and its two analog POWER ports let the FCHUB vehicle V/I feed POWER2 directly (§4.13). No GPS in this bundle (a +M9NGPS variant exists separately)
Firmware	ArduPlane 4.x with QuadPlane
IMU	Dual (ICM-42688-P + BMI088), on a vibration-isolated sensor board
Barometer	Internal MS5611 + external baro on the tail Here4 (CAN)
Magnetometer	External compass on the F9P + Here4, internal mag (IST8310) as backup
CAN ports	2× (CAN1 active, CAN2 reserved)
UART ports	5× (TELEM1, TELEM2, TELEM3, GPS1, GPS2)
PWM outputs	16× (8× IO MAIN + 8× FMU AUX), 3.3V/5V switchable
Power input	POWER1 from the analog PM02 V3 (single FC supply). POWER2 = FCHUB analog V/I sense (not a backup brick), see §4.13
Vibration isolation	Internal gel-foam isolators in chassis
Location	Avionics shelf, oriented with X-axis forward, mounted to isolator plate
USB-C	Ground access for configuration and parameter upload

Sourcing (June 17): ordered from rx-dynamic (TR) at ₺19,900 for the 6C + PM02 set (no GPS), kept TR-side to keep customs simple per the parts-list sourcing rule. The full 6C (vs the Mini) was chosen for the extra UARTs (5 vs 4) and the second analog POWER port. The analog PM02 V3 is now an advantage, not a constraint: it is what makes the FCHUB-into-POWER2 vehicle V/I sensing possible (the 6X's digital PM02D could not have done it). Earlier PM07/PM06 references are superseded (§4.2 FCHUB-12S handles distribution and vehicle current).

GPS (×2 redundant, AIP-006 §9.4)

Two independent GPS units on different interfaces (OQ-12 / parts-list D1 closed June 17). The primary is a salvaged u-blox F9P on a serial UART (GPS1), using the on-hand antennas. The secondary is a CubePilot Here4 on CAN (carried over from the tail electronics plan, now GPS only). This bundle has no M9N (a +M9NGPS 6C variant exists separately); the F9P is the primary, not a bench unit.

Parameter	Primary (shelf/nose)	Secondary (tail)
Model	u-blox F9P (RTK, salvaged)	CubePilot Here4
Interface	Serial UART (GPS1)	DroneCAN
Internal sensors	Baro + compass (module-dependent)	Barometer, IMU, magnetometer
RTK capable	Yes	Yes
GNSS constellations	GPS, GLONASS, Galileo, BeiDou	GPS, GLONASS, Galileo, BeiDou
Supply voltage	5V (peripheral rail, via the GPS port)	5V (CAN bus)
Node / port	SERIALx (GPS1)	CAN node 11

ArduPilot parameters:

GPS_TYPE = 2          ; u-blox serial (primary, F9P on GPS1)
GPS_TYPE2 = 9         ; DroneCAN (secondary, Here4)
SERIAL3_PROTOCOL = 5  ; GPS on the GPS1 UART (confirm the port's SERIALx index)
GPS_GNSS_MODE = 0     ; auto
GPS_GNSS_MODE2 = 0
GPS_AUTO_SWITCH = 1   ; auto-switch to best GPS

GPS yaw / heading: moving baseline GPS yaw (dual-antenna heading) requires two matched RTK units on a shared interface configured as base plus rover. A mixed pair (serial F9P plus CAN Here4) does not support moving baseline, so Phase 1 yaw comes from the compasses on the GPS units rather than from GPS. If GPS yaw is wanted later, move to two matched RTK units on one interface. The ~2 m nose-to-tail separation still gives spatial diversity for the redundant fix.

Airspeed Sensor

Parameter	Value
Hardware	Pitot tube from storage (model TBD — Alperen to confirm)
Interface	I2C or analog ADC depending on pressure transducer
Location	Nose or boom, away from prop wash and wing downwash
ArduPilot	ARSPD_TYPE = 1 (analog) or 7 (I2C DLVR) depending on transducer

If the stored pitot's transducer is unknown or incompatible, replacement option: Matek AP_ANALOG_AIRSPEED or Holybro airspeed sensor (I2C DLVR, ~$25).

Radar Altimeter — Ainstein US-D1 (selected June 11, OQ-03 closed)

Model	Type	Range	Update rate	Interface	Supply	Weight	Cost
Ainstein US-D1 (selected)	Radar	0.5–50 m	100 Hz	UART or CAN	5–13V (5.5V rec.)	110g	Quote (~$280 est.)
Benewake TF03-180 (not selected)	LiDAR	0.1–180 m	100 Hz	UART	5V	76g	~$250

Selection rationale: radar tolerates the prop wash dust kicked up during VTOL landing on dirt fields, where LiDAR returns degrade. Specs corrected against Ainstein's current page (June 2026): the US-D1 is 110g and 100 Hz, not the 56g / 12 Hz carried in earlier revisions, so the weight argument for LiDAR is gone but the robustness argument stands. Sourcing: a unit is believed to be on hand in Turkey already (June 11). Confirm its location and interface variant (UART vs CAN) before buying anything. Only if it does not turn up: no public price, quote via Ainstein or a distributor, and US sourcing plus Turkish customs would make it the longest lead item on the list (parts list line 5).

ArduPilot integration: UART on a spare serial port with the rangefinder protocol and RNGFND1_TYPE set to the USD1 serial driver, or the DroneCAN variant if the UART budget tightens.

First-flight note (June 12): the team treats the rangefinder as optional for first flight, not a blocker. Alperen may have an older Feather LiDAR unit available as a fallback — confirm its interface (CAN vs UART) if the US-D1 does not turn up.

4.9 CAN Bus Architecture

CAN1 bus node map:

Node ID	Device	Type	Notes
1	ESC1 — Motor 1 (FR)	AMPX 80A	DroneCAN ESCStatus + RawCommand
2	ESC2 — Motor 2 (RL)	AMPX 80A
3	ESC3 — Motor 3 (FL)	AMPX 80A
4	ESC4 — Motor 4 (RR)	AMPX 80A
11	Here4 (tail)	GPS + baro + compass	Secondary GPS (primary F9P is serial, not on CAN)
12	CAN-PWM adapter (tail)	Matek CAN-L4-PWM	DroneCAN→PWM for ruddervators

Topology: CAN1 leaves the Pixhawk and runs to a passive CAN splitter/hub near the CG, which fans out short stubs to the four boom ESCs. The trunk continues from the splitter to the tail (Here4 and adapter). The primary GPS is no longer on CAN (F9P moved to a serial UART, §4.8), so the nose GPS branch is dropped. This keeps the CAN fan-out off any custom board for V1. Keep each splitter stub short (ideally under 0.3 m at 1 Mbps), and confirm whether the splitter has a built-in 120Ω terminator. If the boom stubs show marginal signal integrity, daisy-chain the ESCs along the booms instead.

Bus termination: 120Ω at the Pixhawk CAN port (built-in or terminator plug) and 120Ω at the last physical node at the tail (whichever of the Here4 or CAN-PWM adapter sits at the bus end). The CAN splitter sits mid-bus and is not terminated. No termination on the other intermediate nodes.

CAN2 is reserved. Future uses: generator controller (Phase 2), companion computer CAN bridge, additional sensors.

ESC node IDs are assigned via DroneCAN UI Tool (pre-flight on bench, not in the field). Set before boom installation.

4.10 Tail Electronics

Problem: Running PWM servo signal cables 2.5 m from the nose avionics bay to the tail surfaces through an environment with 4× 80A ESC switching noise causes signal degradation and potential interference.

Solution: CAN to the tail, PWM generated locally at the tail. A standalone DroneCAN-to-PWM adapter at the tail converts the ruddervator servo commands (broadcast over CAN) into local PWM. The tail Here4 is a GPS node only. A local Matek PM12S-3 powers everything: its Vx rail feeds the ruddervator servos and its fixed 5V feeds the adapter and the Here4. No custom PCB, and no long LV run through the fuselage.

Supersedes: earlier revisions used the tail Here4 itself as the CAN-to-PWM converter (via its second breakout cable) plus a custom tail PCB for regulation. That plan is dropped. Here4 doing both secondary GPS and tail conversion would let a single Here4 failure remove tail control-surface outputs, not just GPS, and Here4/RTK failures have been seen on Quiver, so a dedicated adapter avoids that single point of failure. Decoupling the servo signal path from the GPS also removes the dependency on the Here4 breakout, and an off-the-shelf adapter removes the custom PCB spin. The former Here4 breakout PWM bench test is no longer needed.

CAN-to-PWM Adapter — Matek CAN-L4-PWM

Item	Detail
Function	DroneCAN node. Receives RuddervatorLeft/Right (SERVO_FUNCTION 79/80) over CAN, outputs PWM
CAN	4-pin JST-GH, daisy-chained from CAN1. Node ID 12, 120Ω term if at bus end
Logic power	4.5–5.5V from the tail PM12S-3 (not the CAN 5V rail), ~30mA
PWM outputs	9× PWM (8× DShot-capable), use 2 for ruddervator L/R at 5V logic. AP_Periph firmware (STM32L431), 3.5g
Servo rail	V+ rail fed from the tail PM12S-3 Vx output (8V, PT1 = final Kingmax HV). Adapter logic and Here4 stay on the PM12S-3 fixed 5V, deliberately separate so the 8V rail cannot reach the logic or the GPS
ArduPilot	Map the CAN node's PWM channels to SERVO_FUNCTION 79/80, confirm in WP-E04

Sourcing (June 11): in stock in Turkey at Aykut Havacılık, 2,060 TL. The Matek CAN-L431 from the same shop (2,500 TL, 5× PWM, 3× UART, I2C, dual parallel GH-4P CAN connectors) is an acceptable alternate running the same AP_Periph family on the same STM32L431. The L4-PWM stays primary: more PWM headroom, and it sits at the bus end where its single CAN connector plus the 120Ω terminator fits the topology. The L431's pass-through CAN connectors would matter only if the adapter had to sit mid-bus.

Tail Servo Power — Matek PM12S-3

Item	Detail
Input	HV bus spur (36–50V), an 18 AWG branch broken out at the tail-boom ESC-lead extension joint (not a dedicated FCHUB run), via the tail harness
Outputs	Vx at 8V, 15A cont, for the 2 Kingmax CLS3015S ruddervator servos. Fixed 5.2V/4A for the adapter logic and the Here4. Fixed 12V unused at the tail
Type	Matek PM12S-3 (9–55V input with TVS, 53g). Vx at 8V drives the Kingmax CLS3015S (6.0–8.4V range) at ~33 kg·cm vs 35 kg·cm at their full 8.4V — still well above target. 8.4V would need the robocombo buck (used on the wing); the tail keeps the PM12S-3 because it also supplies the 5V logic rail
Wiring	Vx lands on the adapter's servo V+ rail, servos plug into the adapter PWM headers. The fixed 5V feeds the adapter logic pad and the Here4 on a separate feed
Kill scope	On the HV spur, de-energized only by the HV kill, not the LV kill (see §4.3)

Common ground: all PM12S-3 rails share ground, so the Vx servo rail, the adapter logic, the Here4, and the PWM signals reference one ground.

Kill behavior: the tail PM12S-3 feeds the servos, the adapter, and the Here4 from the HV spur, so the entire tail control path is on the HV kill only and is unaffected by the LV kill. There is no longer a dependency on the CAN 5V rail for tail control.

Fallback for first hover: for the first VTOL-only bench/hover test, the adapter and servos can be powered from any bench 5V source before the tail harness and PM12S-3 are finalized.

Wing Flaperon Drive — 6C IO PWM (wing CAN node considered and dropped)

A second CAN-L4-PWM at the wing center was considered on June 11 and dropped: the 6C IO PWM outputs 1–2 drive the flaperons directly (the full 6C has the same IO PWM the 6X did), saving a node, its wiring, and shelf space. The flaperon PWM runs (~1–1.5 m to the wing roots) follow the §4.6 EMI rules. Servo power comes from the forward robocombo DC-DC buck @ 8.4V (HV input at the FCHUB pads) through the wing connectors, feeding the Kingmax CLS3015S V+. Only the tail keeps a CAN-PWM adapter (node 12), so CAN_D1_UC_SRV_BM = 0x000C (servo outputs 3–4).

4.11 Servo System

PT1 servo selection (June 17, 2026): PT1 runs the final-class HV servos directly, dropping the earlier 5V test-servo stage. All four control surfaces use the Kingmax CLS3015S (6.0–8.4V HV, 35 kg·cm at 8.4V / 25.7 kg·cm at 6V, 0.15 s/60° at 8.4V, metal gear, dual ball bearing, waterproof, 80g, 25T spline). Flaperons run at 8.4V off the forward robocombo buck, ruddervators at 8V off the tail PM12S-3 Vx (§4.4). The 35 kg·cm rating is far above the preliminary torque targets below, so the surfaces are not torque-limited even pending Zeynep's hinge-moment study (WP-E06). CG note: at 80g each the two ruddervators add ~160g at the tail vs the earlier ≤30g/servo target (~100g more aft), feeding the active CG/stability investigation — confirm with Alperen and Zeynep.

Ruddervator Servos (×2)

V-tail airfoil: NACA 0015, 280mm chord, control surface 30–35% chord (~84–98 mm deep surface). Each surface is one ruddervator providing mixed pitch + yaw authority (ArduPilot SERVO_FUNCTION 79/80).

Parameter	Requirement
Type	Kingmax CLS3015S (digital HV, 6.0–8.4V), run at 8V on the tail PM12S-3 Vx
Torque	~33 kg·cm at 8V (35 kg·cm at 8.4V) vs ≥ 8 kg·cm target
Speed	~0.16 s / 60° at 8V vs ≤ 0.12 s / 60° target
Weight	80g each (tail weight is CG-critical — see the CG note above)
Connector	Standard 3-pin JR/Futaba (2.54mm pitch), 25T spline, 333mm lead
Control	Via tail CAN-PWM adapter PWM output (see §4.10)

Torque is far above the preliminary target. Re-check against Zeynep's hinge-moment study (WP-E06 input); the binding constraint here is tail mass, not torque.

Flaperon Servos (×2)

Main wing: Clark Z, 20–25% chord flaperons.

Parameter	Requirement
Type	Kingmax CLS3015S (digital HV, 6.0–8.4V), run at 8.4V on the robocombo buck
Torque	35 kg·cm at 8.4V vs ≥ 6 kg·cm target
Speed	0.15 s / 60° at 8.4V
Weight	80g each
Connector	Standard 3-pin + wing panel connector (see §4.6)
Control	Via Pixhawk 6C IO PWM outputs 1 and 2 (§4.6 EMI rules on the wing runs)

Selected June 17: the Kingmax CLS3015S replaces the earlier 5V-test / HV-candidate split (Savöx SH-0255MG, KST DS135MG, MKS HV6130). One servo SKU now covers both flaperons and ruddervators. Confirm linkage geometry and endpoint travel against the Kingmax 25T horn in WP-E06.

4.12 Telemetry and Communications

RC Link

Parameter	Value
Range requirement (Phase 1)	VLOS, ≤ 5 km
Selected (June 18)	Radiolink AT9S Pro transmitter + bundled R9DS receiver (10-ch, 2.4 GHz)
Protocol	SBUS (R9DS output)
FC interface	Dedicated RC-IN port (SBUS)
Receiver location	Avionics shelf or fuselage side, antenna external
Prior candidates (superseded for PT1)	ExpressLRS (900 MHz, CRSF, 1W option), TBS Crossfire

Selected June 18: the team's RC direction is the Radiolink AT9S Pro transmitter with its bundled R9DS receiver, on SBUS into the 6C RC-IN port (D3 RC side closed). This is the transmitter the pilots will use, which is why the earlier ExpressLRS / Crossfire CRSF preference is dropped for PT1. SBUS is unidirectional, so RC and ground telemetry stay on separate links (the SiK radio below carries MAVLink) rather than the CRSF MAVLink passthrough that would have folded them onto one link. Confirm the AT9S Pro channel map and failsafe in WP-E08.

Telemetry Radio

Parameter	Value
First-flight unit (June 18)	Holybro SiK Radio V3 (915 MHz, 100 mW, MAVLink 2.0), borrowed from STORK
Protocol	MAVLink 2.0
FC interface	TELEM1
Data	Attitude, GPS, battery SoC, ESC temps, link quality
Long-term	Still open — SiK V3 is range-limited; a better unit (RFD900x-class, 1W) to be researched by Zeynep + Alperen

Telemetry is split off from RC: the Radiolink SBUS link carries no MAVLink, so a separate MAVLink radio is needed. For the first flight the team borrows STORK's Holybro SiK V3 (915 MHz, 100 mW) on TELEM1, which is adequate for VLOS but range-limited. The long-term Spearhead telemetry unit stays open (D3, telemetry side), with an RFD900x-class 1W radio the likely upgrade before the Phase 3 extended-range runs (E-REQ-07 / WP-E08).

Phase 3 requirement (> 200 km) likely requires satellite. See §6.1.

4.13 Health Monitoring

Battery Monitoring

The pack is a dumb LiPo, so monitoring is sensor-based rather than from a BMS. Vehicle current comes from the FCHUB-12S 440A analog sensor on the 6C POWER2 port, pack voltage from the FCHUB VBat divider on the same port, and SoC from ArduPilot coulomb counting. There is no per-cell data.

Parameter	Method
Pack current	FCHUB-12S 440A analog sensor (0–3.3V over 440A; ArduPilot `BATT_AMP_PERVLT 133.3`) → 6C POWER2 CURRENT pin (BATT1)
Pack voltage	FCHUB VBat divider → 6C POWER2 VOLTAGE pin (BATT1). PM02 V3 on POWER1 gives a redundant read (BATT2)
Cell voltages	Not available (no BMS). Use a plug-in cell checker/alarm for ground checks
SoC / consumed	ArduPilot coulomb counting from the FCHUB current sensor
Temperature	Pack not instrumented (no BMS). Optional external thermistor if needed
Low voltage warning	BATT_LOW_VOLT = 42V (3.5V/cell), GCS alert + buzzer
Critical voltage	BATT_CRT_VOLT = 39.6V (3.3V/cell), RTL trigger
Low capacity	BATT_LOW_MAH = 2000 mAh (reserve threshold)

ArduPilot parameters (Setup B — FCHUB on POWER2 as BATT1, PM02 V3 on POWER1 as BATT2).

The 6C analog POWER ports map to fixed ADC pin numbers (ArduPilot Pixhawk6C hwdef). Note the difference between the connector position (where the wire physically lands) and the ArduPilot pin number (the ADC channel that position routes to):

Port	Module here	Instance	Voltage: conn. pos → ADC pin	Current: conn. pos → ADC pin
POWER1	PM02 V3 (FC 5V supply)	BATT2	pos 4 → pin 8 (PC5)	pos 3 → pin 4 (PC4)
POWER2	FCHUB-12S V2 sense	BATT1	pos 4 → pin 5 (PB1)	pos 3 → pin 14 (PA2)

The instances are deliberately swapped from the board default (which assigns BATT1 to POWER1): the FCHUB is the primary (BATT1) because it carries the real vehicle current, so SoC, consumed-mAh, and the BATT_LOW_* / BATT_CRT_* failsafes ride on it. The PM02 V3 is BATT2, a redundant pack-voltage read only — its current reads just the FC leg off the pad tap, not vehicle current.

# BATT1 = FCHUB-12S V2 on POWER2 (primary: vehicle current + voltage + SoC)
BATT_MONITOR     = 4        ; Analog Voltage and Current
BATT_VOLT_PIN    = 5        ; 6C POWER2 voltage (PB1)
BATT_CURR_PIN    = 14       ; 6C POWER2 current (PA2)
BATT_VOLT_MULT   = 21.0     ; FCHUB 1K:20K divider (Matek-specified)
BATT_AMP_PERVLT  = 133.3    ; FCHUB 440A sensor (Matek-specified)
BATT_VOLT_OFFSET = 0
BATT_AMP_OFFSET  = 0
BATT_LOW_VOLT    = 42.0     ; 3.5V/cell
BATT_CRT_VOLT    = 39.6     ; 3.3V/cell
BATT_LOW_MAH     = 2000

# BATT2 = PM02 V3 on POWER1 (redundant voltage; current is FC-leg only)
BATT2_MONITOR    = 4        ; Analog Voltage and Current
BATT2_VOLT_PIN   = 8        ; 6C POWER1 voltage (PC5)
BATT2_CURR_PIN   = 4        ; 6C POWER1 current (PC4)
BATT2_VOLT_MULT  = 18.18    ; PM02 V3 (6C board default scale)
BATT2_AMP_PERVLT = 36.36    ; PM02 V3 (6C board default scale)

FCHUB-to-POWER2 cable: 4-wire from the FCHUB-12S V2 JST-SH breakout (silk ... TLM Cur G V) to the 6C POWER2 plug — Cur → POWER2 position 3 (CURRENT2), V → POWER2 position 4 (VOLTAGE2), G → POWER2 position 6 (GND). Leave POWER2 positions 1/2 (5V) open, POWER2 is sense only. Set the FCHUB V-pad jumper to V = VBat so that pin outputs the 1/21 divided voltage (0–2.857V, ≈2.4V at 50.4V). If it is left on 12V it will destroy the 3.3V ADC, so confirm the V wire reads ≈ pack-voltage ÷ 21 on a multimeter before plugging in. The PM02 V3 keeps its stock 6-pin cable into POWER1.

The 6C's two analog POWER ports are what make this split possible — no cutting the PM02 harness, unlike the single-port splice the digital PM02D would have forced. After install, calibrate BATT_VOLT_MULT against a multimeter and trim BATT_AMP_PERVLT against a clamp meter or a known load.

ESC Telemetry and Thermal

AMPX 80A V2 DroneCAN ESCs broadcast ESCStatus messages in real time: voltage, current, temperature, and operation status per ESC. ArduPilot logs these automatically and streams over MAVLink telemetry. No additional wiring required.

ESC thermal thresholds (source: AMPX 80A ESC datasheet — these are ESC junction temperature limits, independent of motor winding temperature):

ESC junction temp	Action
> 125°C	ESC reduces output to 50% max power; fault signal broadcast over DroneCAN
> 140°C	Full motor shutdown — will not restore until throttle zeroed AND ESC cools below 80°C

ESC ambient operating limit is 65°C. In warm weather (Ankara summer) with ESCs mounted externally on booms, ensure the heatsink fins have unobstructed airflow during hover. Log ESC temperature via DroneCAN telemetry and configure GCS alerts well below the 125°C threshold (suggest 90°C warning).

CAN watchdog — critical: The AMPX 80A cuts motor power if CAN commands are absent for more than 2 seconds. A CAN bus fault mid-hover causes all four motors to cut simultaneously after 2s. Verify CAN_D1_UC_ESC_BM is correct and confirm CAN bus health before first hover.

Motor Thermal

The V8013 PRO motor has no onboard temperature sensor and no direct telemetry output for winding temperature. The operating zone limits below come from the motor datasheet footnote (not the ESC) and describe the motor's sustainable current based on winding copper losses and prop-wash cooling at 25°C ambient:

Motor current	Zone	Duration
< 24A/motor	Sustainable continuous	Indefinite
24–78A/motor	Short-term	~10–30s
> 78A/motor	Non-working	Avoid

These limits are independent of the ESC. The ESC will run at 80A continuous without complaint; it is the motor windings that constrain sustained current.

As shown in §4.1, the motor runs in the continuous zone (< 24A) for most of the useful flight window and only approaches short-term territory near the low-battery warning voltage. The zones are more of a concern at altitude, where throttle must increase to compensate for lower air density.

Monitoring approach (no direct motor temp sensor):

Proxy	How	Limitation
Current per motor	ESC DroneCAN telemetry (ESCStatus.current)	Best real-time indicator of which zone you're in
Motor case temperature	Hand check after landing (bench tests only)	Rudimentary but effective for initial characterization
RPM	ESC DroneCAN telemetry	Confirms expected hover RPM vs. throttle; anomaly = mechanical issue

For Phase 2 characterization: consider adding NTC thermistors to the motor stator cans, routed along the boom wiring to the FC ADC, to establish actual winding temperature margins before committing to operational flight profiles.

Pre-flight Diagnostics

ArduPilot built-in pre-arm checks cover: GPS lock and HDOP threshold, compass calibration, RC calibration and failsafe, ESC arming, battery voltage above BATT_ARM_VOLT, CAN node health, and altimeter sanity check. No additional software needed for Phase 1.

Phase 2 Additions (stub)

Engine temp: type-K thermocouple at cylinder head and exhaust, via MAX6675 or AD8495 amplifier to FC ADC. Fuel level: ultrasonic or float sensor to ADC or UART.

4.14 Heading LEDs

Parameter	Value
Type	WS2812B (NeoPixel) addressable RGB
Control	ArduPilot NTF (Notify) subsystem
FC interface	GPIO pin configured as NeoPixel output
ArduPilot parameter	NTF_LED_TYPES = 1 (NeoPixel)
Color scheme	Front/port: red, front/starboard: green, rear: white/flashing red
Location	Wing tips (port/starboard), nose, tail
Power	5V avionics rail; WS2812B draws ~60 mA per LED at full white
Estimated LEDs	4–6 per wing tip, 2 at nose, 2 at tail
Total power budget	~16 LEDs × 20 mA average = ~320 mA at 5V

ArduPilot NTF patterns: arming flash, mode color (VTOL = yellow, FW = blue, RTL = red), GPS quality indicator.

4.15 Component Locations

Updated June 11 for Alperen's fuselage conceptual v5: longer thinner nose, battery moved forward for CG, avionics on a large shelf above the battery with top access, pusher between the fuselage and the boom-carried tail. Detailed layout remains WP-E07, pending shelf and bay dimensions.

Nose (slimmed, v5):
  - No avionics requirement. Pitot mast and antenna placement only if needed

Forward battery bay:
  - motorobit 12S 22Ah 15C solid-state pack (3,709g, 190×78×126mm): strapped, fore-aft position sets CG
  - QS8-S socket (anti-spark): reachable for ground safing
  - HV kill (OQ-05): adjacent, actuation per the chosen implementation

Avionics shelf (above battery, top access):
  - Pixhawk 6C: vibration-isolated, X-axis forward
  - PM02 V3 power module: inline on the pad-tap leg (POWER1); FCHUB sense to POWER2
  - Primary GPS (F9P, serial): on the shelf cover, unobstructed sky view
  - RC receiver (Radiolink R9DS) + telemetry radio (Holybro SiK V3): shelf edges, antennas external
  - LV kill switch: reachable through the shelf hatch
  - Shelf footprint reserves rectangular space for a future integration PCB (per Alperen)
  - Compartment ~180×250mm × ~90mm tall between the longerons (June 12 estimate, pending CAD). Fit check (WP-E07): 6C (84.8×44), PM02 V3, FCHUB-12S (55×50), CAN hub, F9P, RC RX, telemetry radio must all fit this volume

CG bay / boom roots:
  - FCHUB-12S PDB: near the boom-root convergence, short ESC lead runs
  - CAN splitter/hub: adjacent to the FCHUB
  - Forward robocombo buck (flaperon 8.4V rail): adjacent to the FCHUB

Booms (×4 motor arms, 2 carrying the tail):
  - ESCs: externally mounted near motors (within 150mm of motor centerline)
  - Phase wire exits sealed where they enter the boom
  - Tail harness (CAN trunk + 18 AWG HV spur off the ESC-lead extension joint): along one tail boom
  - Inside-tube routing is attractive but waits on the structural check (June 5)

Tail group (boom-mounted):
  - Here4 (secondary GPS): top surface, unobstructed sky view
  - CAN-PWM adapter + PM12S-3 (Vx @ 8V + fixed 5V): bulkhead-mounted, connectors accessible
  - Ruddervator servos (Kingmax CLS3015S, 80g each): local, short leads

Wings (detachable panels):
  - Flaperon servo: inline with control horn, servo output axis on hinge line
  - Wing panel connector: at wing root, mates to fuselage connector on insertion

4.16 Phase 1 BOM and Cost Estimate

Procurement view with priorities, candidate vendors, and lead flags: docs/pt1-parts-list.md (SPH-E-003). The tables below stay the cost baseline.

Already ordered (status: in transit, MAD Motor Poland, ~May 1, 2026):

Item	Qty	Est. Unit Cost	Est. Total
MAD V8013 PRO IPE 150KV motor	6	~$185	~$1,110
MAD AMPX 80A DroneCAN ESC (V2)	6	$60	$360
MAD FLUXER PRO 26×7.8 MATT (per pair)	6 pairs	$193.90	$1,163 + $60 shipping
Subtotal (ordered)			~$2,633

To procure:

Item	Qty	Est. Unit Cost	Est. Total
Pixhawk 6C + PM02 V3 set (FC + analog power module + cables, no GPS, selected June 17, rx-dynamic TR)	1	~$600	~$600
motorobit 12S 22Ah 15C solid-state LiPo (3,709g, 190×78×126mm, QS8-S, selected June 17)	1	TBD	TBD
Matek FCHUB-12S V2 PDB (HV distribution + 5V/12V BECs), incl. 1 spare	2	$30	$60
XT60 pigtail + 16 AWG for the FCHUB pad tap (Y-splitter dropped)	1 lot	$10	$10
u-blox F9P primary GPS, serial (salvaged unit on hand; antennas on hand, D1 closed)	1	$0	$0
Here4 GPS, tail (secondary GPS, GPS only)	1	$300	$300
Matek PM12S-3 (tail servo + logic rail, Aykut Havacılık TR, 6,900 TL)	1	$170	$170
robocombo DC-DC buck 8–60V 15A (wing flaperon 8.4V rail)	1	~$15	~$15
HV kill: none for PT1 (software E-stop, §9.3 deviation, decided June 17)	—	$0	$0
LV kill switch	1	$15	$15
QS8-S plug (mates the pack socket) ×3, anti-spark so no AS150U needed	1 lot	$15	$15
XT90S connector pairs (ESC boom junctions + spares)	8	$6	$48
5mm bullet connectors (×12 sets)	12	$2	$24
8 AWG silicone wire (2 m, battery harness)	1 lot	$15	$15
8 AWG silicone wire (10 m)	1 lot	$20	$20
12 AWG silicone wire (extension leads)	1 lot	$12	$12
LV wiring assortment (18–26 AWG)	1 lot	$30	$30
CAN bus twisted-pair wire (5 m)	1 lot	$10	$10
JST-GH crimp kit (CAN, UART, assorted)	1	$25	$25
CAN splitter/hub (DroneCAN, JST-GH, model TBD)	1	$15	$15
Kingmax CLS3015S HV servos (4 install: 2 flaperon + 2 ruddervator, + 2 spare; thkmodelucak TR, 3,950 TL)	6	~$110	~$660
Wing panel servo connectors (Molex Mini-Fit Jr)	2 sets	$10	$20
Radiolink AT9S Pro TX + R9DS RX (RC link, SBUS, selected June 18)	1 set	~$130	~$130
Holybro SiK Radio V3 telemetry (first flight, borrowed from STORK; long-term unit TBD)	1	$0 (borrowed)	$0
Ainstein US-D1 radar altimeter (selected June 11, believed on hand in TR, confirm)	1	$0–280	~$0
Pitot tube transducer (if needed beyond existing)	1	$30	$30
WS2812B LED strip + wiring	1 lot	$20	$20
Matek CAN-L4-PWM adapter (tail), incl. 1 spare (Aykut Havacılık TR, 2,060 TL each)	2	$50	$100
(tail servo rail = PM12S-3 line above; wing servo rail = robocombo buck line above)	—	—	—
Heat-shrink, zip ties, mounting hardware	1 lot	$40	$40
Subtotal (to procure, est.)			~$2,550 + TBD items
Phase 1 Grand Total (est., excl. labor)			~$5,185 + TBD items

Estimates only. Verify against actual quotes before procurement. Some items may be on hand from Quiver or other projects.

5. Phase 2: Hybrid Integration

Target: July–October 2026. Objectives: IC engine integration, transition flights, vibration characterization and notch filter tuning. Engine model for current layout work is a DLE55cc-class gasoline engine (June 11). The generator path moved out of Phase 2: PT2 may simply fly with a gasoline engine, and starter-generator work is now a PT3 / end-of-year sub-project candidate. Three starter-generator vendors responded to Alperen (Czech-linked, German, US), with a US vendor call pending. The §5.8 VESC architecture below predates this and reads as background for that research.

⚠️ STATUS: This section is preliminary. Significant research is still required before any of the Phase 2 content below should be treated as a design baseline.

The IC engine, fuel system, starter, generator, ignition, RPM sensor, throttle servo, and vibration isolation subsections in §5 are working notes based on general 2-stroke RC engine practice and a candidate engine class (35–55cc Stinger / DLE). The actual engine has not been selected or purchased, no bench testing has occurred, and most parameters (fuel consumption rate, electrical load profile, vibration spectrum, alternator output, starter inrush current) will only be known after empirical measurement on the chosen unit. Treat all numbers, parts, and architectures in §5 as candidate values to be validated, not commitments. Specific items requiring further research before commitment:

Engine selection (Stinger vs DLE, 35cc vs 55cc) — buy both and test empirically per current plan

Starter motor selection and current draw (Pilot RC: ~30 days lead after order, and the vendor says it cannot double as a generator)

VESC dual-role starter-generator viability (no off-the-shelf precedent at this scale)

Generator output topology (alternator vs starter-as-generator) and rectifier specifications

Ignition battery: shared 12S vs dedicated pack

CHT / EGT sensor selection and ArduPilot integration

Real vibration spectrum and final notch filter tuning (cannot be predicted, must be measured)

Throttle servo torque and stroke for the actual carburetor lever geometry

Phase 1 (§4) is the working design baseline. Phase 2 will be re-scoped after Phase 1 hover testing and once the engine candidates are in hand.

5.1 Fuel System and Mix

2-Stroke Fuel Mix

DLE and Stinger engines in the 35–55cc class are air-cooled single-cylinder 2-stroke gasoline engines. They require a premixed fuel-oil mixture — no separate oil reservoir.

Stage	Ratio (fuel:oil)	Notes
Break-in (first 2–3 tanks)	25:1	Richer oil mix lubricates and seats rings; run at varied throttle, avoid sustained WOT
Running	40:1 to 50:1	Manufacturer-specific; DLE 35cc manual specifies 40:1 typical

Oil type: semi-synthetic or full-synthetic 2-stroke oil designed for air-cooled engines. Do not use 4-stroke oil, marine 2-stroke (designed for water-cooled), or lawnmower 2-stroke oil. Brands: Motul 800 2T, Castrol Power 1 Racing 2T, or equivalent.

Fuel: 92+ octane unleaded gasoline, ethanol-free preferred. Ethanol is hygroscopic (absorbs atmospheric moisture), corrodes aluminum carburetors and fuel lines over time, and degrades stored fuel faster. Use ethanol-free pump gas (E0) or aviation mogas where available.

Altitude correction: 2-stroke carburetors are calibrated for sea level. At 1,200m operational altitude (air density ~88% of sea level), the mixture runs slightly rich at the same needle position. Most engines have a main jet needle clip with 3–5 positions for mixture adjustment. Research required: determine the DLE 35cc / Stinger jet needle adjustment procedure and characterize the correction needed at operational altitude. An engine running rich at altitude produces less power, runs hot in a different way, and fouls plugs.

Pre-flight mixing protocol: Mix fuel fresh before each test campaign. Fuel-oil premix has a shelf life of approximately 30 days before oil separation begins. Do not use old premix.

Fuel Storage and Delivery

Item	Notes
Tank material	Aluminum or HDPE (compatible with gasoline). Oratex and standard foam are NOT compatible with gasoline — design fuel bay accordingly.
Tank volume	Target ~6L (per Alperen's fuselage layout). At ~35cc/min consumption estimate for DLE 35 at cruise, 6L gives ~170 min endurance (needs empirical verification).
Fuel lines	Viton or Tygon fuel tubing (gasoline-compatible). Standard silicone tubing is NOT gasoline-compatible and will degrade.
Fuel filter	Inline filter between tank and carb. Replace every 5–10 flights.
Clunk pickup	Weighted clunk at tank inlet ensures fuel feed at any aircraft attitude.
Vent	One-way vented cap to prevent vapor lock.

5.2 Throttle Control

The throttle arm on the engine carburetor is driven by a servo, controlled via ArduPilot.

Parameter	Value
Servo type	Digital, metal gear, standard torque (≥ 3 kg-cm)
ArduPilot function	`SERVO9_FUNCTION = 70` (throttle) — maps ICE throttle to servo output 9
Servo travel	Must be calibrated to match carburetor idle stop (minimum) and wide-open throttle (maximum). Set `ICE_IDLE_RPM` and test with tachometer.
Idle RPM	Engine-specific; DLE 35cc target ~1,400–1,600 RPM at idle. Low enough to not produce significant thrust, high enough to stay running.
Throttle during VTOL hover	ArduPilot keeps IC engine at idle during all VTOL modes (`ICE_IDLE_RPM` via `ICE_ENABLE`). Engine is not shut down during hover — it idles at cruise throttle stop.
Throttle during transition	ArduPilot advances IC throttle as VTOL motors ramp down across `Q_TRANSITION_MS`.
Emergency cutoff	Ignition kill (see §5.3). Throttle servo drives to idle position as secondary measure.

ArduPilot IC engine parameters (research required — no direct experience with ICE ArduPilot integration):

ICE_ENABLE = 1
ICE_START_CHAN = <RC channel for starter relay>
ICE_IDLE_PCT = <idle throttle %, calibrate empirically>
ICE_IDLE_RPM = 1500   ; target idle RPM, uses RPM sensor feedback
ICE_IDLE_DB = 50      ; RPM deadband ±50 RPM
ICE_RPM_CHAN = 1      ; RPM sensor instance

Confirm ArduPilot ICE library supports DroneCAN ESC coexistence — the ICE library was designed for planes with a single IC engine and may need validation in a QuadPlane configuration.

5.3 Choke and Cold-Start Sequence

Most DLE and Stinger engines use a manual choke plate (butterfly valve) in the carburetor air intake. There is no electronic choke — it must be mechanically actuated.

Choke Actuation Options

Option	Description	Pros	Cons
Manual lever	Ground crew or pilot operates choke arm directly before flight	Zero weight/complexity	Requires physical access to engine; impractical after integration in fuselage
RC-servo actuated	Micro servo on choke arm, assigned to a transmitter switch	Remote operation, no hardware complexity	Adds 1 servo; needs physical routing inside engine bay
Pull cable	Bowden cable from cockpit hatch to choke arm	Simple, no electronics	Cable routing through fuselage, potential binding

Recommendation for Phase 2 prototype: servo-actuated choke on a dedicated RC channel (transmitter 2-position switch). Assign as a passthrough output, not an ArduPilot-controlled function. Pilot operates it manually during the start sequence.

Cold-Start Sequence (proposed, to be refined with engine testing)

Transmitter on, aircraft in QLOITER or QSTABILIZE, VTOL motors armed but at idle
Confirm throttle servo at idle position (minimum)
Close choke (RC switch → choke servo)
Engage electric starter (RC switch → starter relay) for 2–5s
Engine fires → immediately open choke partway
Run at fast idle (~2,000 RPM) for 30–60s to warm up
Open choke fully — engine now at normal idle
Confirm RPM sensor reading matches expected idle RPM
Proceed with VTOL takeoff as normal

Hot restart (engine already warm): skip choke, engage starter briefly.

Note: If the engine does not start within 3 attempts, investigate flooding (excess fuel in cylinder). Recovery: open choke fully, wide-open throttle, crank briefly to clear — a procedure that should only be done on the ground with props removed.

5.4 Ignition System

2-stroke engines at this scale use a CDI (Capacitor Discharge Ignition) module mounted on the engine. The CDI generates high-voltage spark from a low-voltage input and a trigger from a hall effect sensor or pickup coil on the flywheel.

Item	Notes
Power supply	CDI typically requires 5–6V regulated DC, < 1A. Use a dedicated low-noise 5V BEC (separate from the avionics BEC to avoid back-EMF transients on the avionics rail).
Kill wire	Most CDI modules have a "kill" wire: ground this wire to stop ignition immediately. Connect to an FC relay output (`SERVO10_FUNCTION = 28` relay) and an accessible hardware switch. Both the pilot RC switch and the hardware switch must independently kill the engine.
Spark plug	Replace per engine manual (typically every 10–20 flight hours). Inspect for fouling during initial testing.
RF interference	CDI spark generates broadband RF noise. Shield ignition wires with braided metal sleeve, ground the shield at one end. Route CDI wires away from GPS antennas and CAN bus cables. Confirm GPS fix quality doesn't degrade with engine running during ground test.

5.5 Electric Starter

Pilot RC Auto-Starter (primary candidate)

The Pilot RC DLE35RA electric auto-start unit (420g) is a gear-driven brushless motor that engages the engine crankshaft for starting. Vendor status (May 2026): not in stock, can be prepared roughly 30 days after order. The vendor stated the unit cannot be used as a generator (reason not given), so a generator path would need a separate VESC and brushless motor (§5.8), not this starter.

Starter power circuit:

[2S LiPo, 7.4V, ~1,500–2,000mAh]
        │
  [High-current relay, 30–50A]  ←── FC relay output (SERVO10 or AUX)
        │                              or dedicated RC transmitter switch
  [Starter motor]
        │
  [Engine crankshaft]

Parameter	Value
Supply voltage	7.4V (2S LiPo) or 11.1V (3S) — confirm with Pilot RC spec
Peak current	~20–40A for < 3s (estimated; measure at bench)
Relay	30–50A relay (Axiom, Tyco, or automotive relay). Switching inductive load — add flyback diode across relay coil.
FC control	ArduPilot `ICE_START_CHAN` assigns an RC channel. When triggered, the FC engages the relay, waits for RPM sensor to confirm start, then disengages.
Battery isolation	Dedicated 2S or 3S LiPo prevents starter inrush from sagging the main 12S VTOL battery. Also electrically isolates a potential starter motor failure from the avionics rail.
Safety interlock	Starter must only be able to engage when engine throttle servo is at idle and aircraft is on the ground (or explicitly commanded in-flight for in-flight restart). Use ArduPilot `ICE_STARTTIME` parameter to limit maximum starter engagement duration.

Fallback: Manual Pull-Start

If the Pilot RC unit is unavailable, the engine is pull-started on the ground before arming. The aircraft must not be armed (VTOL motors must not be able to spin) during pull-starting. This is standard practice for Phase 2 initial testing regardless of whether an electric starter is installed.

5.6 RPM Sensor

The RPM sensor is a Phase 2 day-one requirement, not optional. ArduPilot uses it for:

ICE_IDLE_RPM closed-loop idle control
INS_HNTCH_MODE = 3 — harmonic notch filter uses RPM to dynamically track engine vibration frequency
Engine health monitoring and telemetry
Fuel consumption estimation (if correlated with throttle position)

Sensor Options

Option	Interface	How it works	Pros	Cons
CDI tach wire	FC GPIO, interrupt	Most CDI modules output a 1-pulse/rev signal on a tach wire	No added hardware, directly from engine	Confirm DLE/Stinger CDI has tach output; may require pull-up resistor
Hall effect sensor	FC GPIO, interrupt	Magnet epoxied to flywheel; hall sensor mounted on crankcase	Very reliable, clean signal	Requires access to flywheel for magnet installation
Optical sensor	FC GPIO, interrupt	Reflective tape on prop shaft; IR emitter/detector pair	Easy to install	Sensitive to oil/dirt fouling

Recommendation: try CDI tach wire first (zero added hardware if it exists). If unavailable, use hall effect on the flywheel.

ArduPilot parameters:

RPM1_TYPE = 1         ; GPIO interrupt-based
RPM1_PIN = <GPIO>     ; pin number on Pixhawk 6C
RPM1_SCALING = 1.0    ; 1 pulse per revolution
RPM1_MAX = 10000      ; max expected RPM
RPM1_MIN = 500        ; min RPM to register as running

For harmonic notch filter:

INS_HNTCH_ENABLE = 1
INS_HNTCH_MODE = 3    ; tracks RPM sensor
INS_HNTCH_FREQ = 100  ; starting estimate (confirm with INS_VIBE log data)
INS_HNTCH_BW = 40
INS_HNTCH_ATT = 40

The notch filter must be tuned from real INS_VIBE log data taken during the first ground engine runs. Estimate: DLE 35cc at 6,000–8,000 RPM cruise → 100–133 Hz fundamental. Harmonics at 200–266 Hz may also require a second notch.

5.7 Engine Health Monitoring

Sensor	Interface	ArduPilot integration
Cylinder head temperature (CHT)	K-type thermocouple → MAX6675 or MAX31855 SPI amplifier → FC SPI	Log via scripting or companion; no native ArduPilot CHT support — use Lua script to read SPI and forward via MAVLink
Exhaust gas temperature (EGT)	K-type thermocouple (higher range needed: 0–900°C) → MAX31855	Same as CHT
Fuel level	Ultrasonic level sensor or float potentiometer → FC ADC	`BATT2_MONITOR = 7` (fuel flow, if using flow meter) or analog level to ADC for display
RPM	See §5.6	Native ArduPilot RPM library

CHT limits (2-stroke, air-cooled, approximate):

Normal cruise: 150–200°C
High: 200–230°C (acceptable for short time)
Critical: > 250°C — engine damage risk; reduce throttle or enter VTOL mode

EGT is a better real-time mixture indicator than CHT. Rich mixture: lower EGT. Lean mixture: higher EGT and higher CHT. Target EGT range for 2-stroke at cruise: 550–680°C (engine-specific; characterize during break-in).

5.8 Generator Architecture and VESC Starter-Generator

Power Budget

Load	Cruise power draw
Avionics + sensors	~30–50W
Servos (cruise trim, minor inputs)	~5–15W
VTOL battery top-up (trickle)	50–150W (goal)
Total generator target	100–200W

Architecture Decision: VESC Dual-Role Starter-Generator (Option A)

This is the primary architecture to develop. The same brushless motor and VESC controller handles both the electric start function and in-flight power generation, sharing shaft access to the IC engine output.

Concept:

[IC Engine crankshaft / prop shaft]
         │
   [Overrunning clutch or direct coupling]
         │
   [Brushless motor — starter/generator]
         │
   [VESC controller]
    ├── Starting mode: battery → VESC drives motor → spins engine
    └── Generator mode: engine drives motor → VESC rectifies/regulates → charges battery bus

Motor and VESC selection:

At engine cruise RPM of ~6,000–8,000 RPM, the generator motor must produce useful output voltage. For VESC to regulate a target bus voltage (say 50V for 12S charging), the motor KV should be selected so back-EMF at cruise RPM is in the right range.

At 7,000 RPM and target ~50V output (no-load):

KV = RPM / V = 7,000 / 50 = 140 KV (similar to the VTOL motors)
At 7,000 RPM with 140 KV motor: VESC manages field weakening and duty cycle to regulate output

Alternatively, a lower-KV motor allows operation with lower DC bus voltage:

200–400 KV motor at 7,000 RPM: back-EMF ~17–35V; use a boost converter between VESC output and 12S battery

VESC controller options:

Model	Cont. current	Voltage	Weight	CAN	Notes
VESC 6 MkVI	50A cont	60V max	~100g	Yes	Well-supported; open source firmware; widely used in research
Flipsky VESC 6.7	80A cont	60V max	~80g	Yes	Lower cost VESC 6 derivative
VESC 75/300	300A cont	90V max	~300g	Yes	Overkill for this application

For 100–200W generator: VESC 6 at ~50V/3–4A — well within 50A continuous rating.

Shaft coupling and disengagement:

The biggest unresolved mechanical question. Three options:

Option	Description	Issues
Direct coupling, always engaged	Motor shaft permanently connected to engine output. VESC switches between motor and generator mode.	During VTOL hover, engine at idle (~1,500 RPM) drives the motor passively. VESC must be in freewheeling/passive mode to avoid drag. At engine shutdown, motor coasts freely.
Overrunning (sprag) clutch	One-way bearing. Motor drives engine for starting (motor faster than engine). Engine drives motor for generation (engine faster).	Sprag clutch direction must be correct. If motor spins faster than engine during starting, clutch engages. During cruise, engine spins faster, clutch engages other way. Need to verify this orientation works or use bidirectional clutch.
Electromagnetic clutch	Electric clutch that engages on command	Clean control; adds ~150–300g and complexity

Recommendation: start with direct coupling and VESC freewheeling mode for Phase 2. The drag penalty at idle RPM is small (motor back-EMF is low at 1,500 RPM). If idle drag is unacceptable, add electromagnetic clutch in Phase 3.

Battery charging from generator:

The generator output does not use a conventional CC/CV LiPo charger. Instead:

VESC output (regulated DC, ~50V)
        │
  [Schottky diode, 5–10A, 60V]  ← prevents battery backfeed into VESC
        │
  [12S battery bus]

With the VESC regulating its output to slightly above the current battery bus voltage (e.g., bus at 44V → VESC regulates to 45V → ~1–3A flows into battery), the battery's own internal resistance limits the charge current naturally. This is a parallel-source top-up, not a dedicated charger. It is safe for in-flight use because:

Current is self-limiting by voltage differential
Battery is never charged above its full-charge voltage during flight (discharge is happening simultaneously)
Schottky diode prevents reverse current if VESC output drops

VESC parameters for generator mode: set output voltage target via VESC Tool → App Configuration → ADC or via CAN commands from companion computer. Target: ~50V (slightly above nominal 44.4V hover bus voltage).

Telemetry: VESC supports CAN bus telemetry (voltage, current, RPM, temperature, power). Route to CAN2 on Pixhawk or to companion computer UART.

Alternative: Dedicated Belt-Driven Alternator (Option B)

A small permanent-magnet alternator (PMA) driven by a toothed belt from the engine prop shaft. Research needed:

Candidates: small UAV PMAs from Genasun, Elroy Air, or custom-wound stators
At 100–200W and 6,000–8,000 RPM, a 3-phase PMA produces AC → 3-phase bridge rectifier → voltage regulator → DC bus
Weight target: < 300g for PMA + rectifier + regulator
No engagement mechanism needed — belt slips or is tensioned by design
Simpler than VESC but less flexible (fixed voltage-RPM curve unless regulator compensates)

Decision: develop Option A (VESC) in parallel with researching Option B. Select after bench testing both by early Phase 2.

5.9 Vibration Isolation and Notch Filter

Mechanical isolation:

IC engine mounted on a soft sub-frame isolated from the main fuselage truss using silicone anti-vibration mounts (e.g., Lord Micro Mounts or M3 rubber standoffs)
Do not hard-mount the engine to the carbon tube truss — transmitted vibration at 100–133 Hz will saturate the IMU
Fuel tank should also be soft-mounted or suspended (liquid sloshing dampens vibration naturally)

Electronic isolation:

Pixhawk 6C has triple-IMU redundancy with internal gel isolators. Adequate as a baseline.
If INS_VIBE logs show clipping (values > 30 m/s²) during ground engine runs, add external foam pad isolation below the FC mount.
Ignition CDI and spark plug wire: wrap in braided copper shielding, grounded at one end only. RF from ignition is broadband and will affect GPS if unshielded.
Separate 5V BEC for CDI from avionics BEC. CDI draws pulsed current at spark frequency; this creates noise on shared supplies.

Notch filter setup procedure:

Engine running on bench (props OFF), aircraft secured
Log INS_VIBE and INS_RAW_ACC1/2/3 at 50–100Hz for 60s
FFT of accelerometer data to identify vibration peaks
Set INS_HNTCH_FREQ to the primary peak frequency
Enable INS_HNTCH_MODE = 3 (RPM sensor tracking)
Set INS_HNTCH_HMNCS = 3 (fundamental + first harmonic)
Recheck INS_VIBE with notch enabled — peaks should attenuate by > 20 dB
Log comparison: before/after notch, at multiple throttle positions (idle, 50%, WOT)

5.10 Phase 2 Preliminary BOM

Item	Est. Cost	Notes
IC engine — DLE 35cc (or Stinger 35cc)	~$200	Buy two sizes: 35cc and 55cc for comparison testing
IC engine — DLE/Stinger 55cc	~$280
Carburetor jet needle set (altitude adjustment)	~$20	For altitude mixture correction research
Gasoline (E0, 92+)	~$10/gal	Ongoing cost
2-stroke synthetic oil (Motul 800 or equivalent)	~$25/L	One liter covers ~40L of fuel at 40:1
Ignition CDI + kill wire wiring	~$50	Usually included with engine
Throttle servo (metal gear digital, small)	~$20
Choke servo (micro, < 10g)	~$15
Choke servo mount	~$10	3D printed
High-current relay (30–50A) + flyback diode	~$15	For starter control
Electric starter (Pilot RC DLE35RA)	~$250	If vendor responds; otherwise pull-start for Phase 2
Starter 2S LiPo (1500–2000mAh)	~$20	Isolated starter supply
K-type thermocouple (×2: CHT + EGT)	~$10
MAX31855 SPI thermocouple amplifier (×2)	~$15
RPM hall effect sensor or optical sensor	~$10	If CDI tach output unavailable
Inline fuel filter	~$5
Viton fuel tubing (1m)	~$10	Gasoline-compatible
Fuel tank (6L, aluminum or HDPE)	~$40
Engine vibration isolator mounts (Lord or silicone)	~$30
Separate 5V BEC for CDI	~$15	Noise isolation from avionics
VESC 6 MkVI or Flipsky 6.7	~$120	Starter-generator controller
Brushless motor for starter-generator (~140KV, 200–400W)	~$50	Select after RPM/voltage analysis
Schottky diode (10A, 60V)	~$5	Battery bus protection
Belt/coupling hardware for generator shaft	~$20	TBD by mechanical design
Phase 2 total (est., excl. labor)	~$1,240	Excludes VTOL battery (Phase 1 item)

6. Phase 3: BVLOS (Stubs)

Target: November 2026–January 2027.

6.1 Long-Range Telemetry (> 200 km)

E-REQ-07 requires real-time command and telemetry beyond 200 km. This rules out standard RF at any practical power level (100 km practical ceiling for RFD900x with directional antennas).

Candidate systems:

System	Range	Latency	Cost	Weight	Notes
RFD900x (1W, 900 MHz) + high-gain tracker antenna	~80–120 km	< 100 ms	~$300 + tracker	Moderate	Requires tracking ground station; range still short of 200 km
Silvus SC4200 MIMO	~150 km	< 50 ms	~$2,000+	~300g	High cost
Iridium 9603 SBD	Global	2–20 s	~$400 + service	~11g	Low-rate only; not suitable for real-time control
Iridium Edge Solar	Global	300 ms–2 s	~$600 + service	~90g	Better latency; still marginal for control
RockBLOCK 9603 + RC link hybrid	Global telemetry + local RC backup	Mixed	~$450	~20g	Good for BVLOS monitoring with local landing control
Starlink Mini	Global	< 50 ms	$599/mo + $599 hw	~1 kg	Too heavy for 25 kg platform

Decision deferred to Phase 3 design. Likely architecture: satellite link for BVLOS telemetry and command, secondary 900 MHz RF for takeoff/landing proximity ops. Confirm Turkish regulatory requirements for BVLOS satellite operation.

6.2 Obstacle Avoidance

E-REQ-06 requires a front-facing sensor. ArduPilot supports BendyRuler avoidance with PROXIMITY class sensors.

Candidates: RPLidar S2L (360° 2D, 32 m, USB, ~200g), NanoRadar MR82 (80 GHz, 30 m, CAN, ~65g), Intel RealSense D435 (stereo vision, USB, ~72g). Details TBD in Phase 3 design.

7. Phase 4: Payload Bay (Stubs)

Target: February–March 2027.

7.1 Payload Bay Power

Parameter	Requirement
Voltage	12V regulated (E-REQ-09)
Current	TBD based on payload types (surveillance cameras, cargo sensors)
Protection	Per-payload fuse on 12V rail
Source	Dedicated 12V DC-DC converter from HV bus (separate from avionics BECs)
Connector	Quick-release (see §7.3)

7.2 Payload Data Interface

Interface	Use
CAN bus spur	Low-rate sensor data, DroneCAN-compatible sensors
Ethernet	High-bandwidth: video, LiDAR point cloud. Via companion computer or Pixhawk TELEM2 → Ethernet bridge.
Serial (UART)	Fallback for simple payload sensors

7.3 Quick-Release Connector

Arrow payload standard connector (per project requirements §9.7). Candidate: Molex Mini-Fit Jr multi-pin or custom interface combining 12V power and CAN/Ethernet in a single latching connector. Final spec in Phase 4 design, aligned with Arrow platform standard.

8. Work Packages

WP-E01 — Battery and BEC Sizing and Procurement

Field	Value
Owner	Erick
Phase	1
Inputs	MAD Motor spec sheet (motor efficiency vs. throttle curve), MTOW 25 kg, hover time target, CG constraints from Alperen
Outputs	Battery specification (capacity, C-rating, weight, dimensions, brand), BEC selections, recommended battery placement
Dependencies	MAD Motor spec sheet receipt (in transit); Alperen's fuselage CG envelope
Deliverables	Written battery spec with sizing rationale, BEC datasheets, procurement order

WP-E02 — HV Wiring Harness Design

Field	Value
Owner	Erick
Phase	1
Inputs	ESC and motor physical locations (from Alperen's boom and fuselage design), battery placement, peak current from WP-E01, connector selections
Outputs	Harness schematic, routing diagram with cable lengths, AWG selection table, connector BOM, assembly instructions
Dependencies	WP-E01 complete; Alperen's boom and fuselage geometry finalized
Deliverables	Harness diagram (KiCad schematic or detailed hand drawing), wire and connector BOM, assembly guide

WP-E03 — Tail CAN-PWM Adapter and Servo Power Integration

Field	Value
Owner	Erick
Phase	1
Inputs	Matek CAN-L4-PWM adapter, PM12S-3 (Vx servo rail + fixed 5V logic feed), tail bay dimensions from Alperen, servo connector type from WP-E06
Outputs	Mounted and wired tail electronics: HV spur → PM12S-3 → adapter servo rail (Vx) + logic/Here4 feed (5V), CAN daisy-chain, servo harness
Dependencies	WP-E04 (adapter PWM validation); WP-E06 for servo connector type; Alperen's tail bay dimensions
Deliverables	Tail wiring diagram, harness, bench-verified servo response, mounting solution. No custom PCB (off-the-shelf adapter + PM12S-3).
Note	Supersedes the earlier custom tail PCB. Revisit a custom integration board only if mounting two loose modules proves impractical.

WP-E04 — CAN-PWM Adapter Validation

Field	Value
Owner	Erick (executor), Zeynep (ArduPilot config)
Phase	1
Inputs	Matek CAN-L4-PWM, ArduPilot-configured FC (existing Quiver hardware acceptable), oscilloscope or servo response check
Outputs	Pass/fail: adapter enumerates as a DroneCAN node, accepts SERVO_FUNCTION 79/80 mapping, outputs correct PWM timing at 5V logic, servo responds
Dependencies	Adapter in hand; CAN bus config from §4.9
Deliverables	Short test report (DroneCAN node ID, parameter mapping, scope/servo result), posted to Spearhead forum
Note	Replaces the former Here4 breakout PWM test. Confirm ArduPilot 4.x DroneCAN servo output support for the chosen adapter firmware.

WP-E05 — Avionics and Sensor Integration

Field	Value
Owner	Erick (integration), Zeynep (ArduPilot configuration)
Phase	1
Inputs	Pixhawk 6C pinout, GPS CAN pinouts (nose module + tail Here4), altimeter selection (Ainstein vs Benewake), pitot tube model from storage, CAN bus topology from §4.9
Outputs	Full avionics wiring diagram, ArduPilot parameter file (.param) covering all sensors, validated bench-tested avionics stack
Dependencies	WP-E04 (CAN-PWM adapter confirmed); altimeter selection is part of this WP
Deliverables	Avionics wiring diagram, parameter file in repo, bench test report with GPS fix confirmation, altimeter readout, and airspeed sensor calibration

WP-E06 — Servo Selection and Control Surface Wiring

Field	Value
Owner	Erick
Phase	1
Inputs	Zeynep's control surface sizing study (final torque requirements), hinge geometry from Alperen, servo bay and linkage dimensions from Alperen, wing panel connector standard
Outputs	Servo model selections for V-tail (×2) and flaperon (×2) with torque analysis, wing panel connector spec, servo wiring diagram
Dependencies	Zeynep's control surface sizing final output; Alperen's tail and wing servo bay dimensions
Deliverables	Servo selection with torque safety factor analysis, wing panel connector spec (connector model, pinout, drawing), servo wiring diagram

WP-E07 — Physical Electrical Layout and Avionics Bay Design

Field	Value
Owner	Erick
Phase	1
Inputs	Component dimensions (Pixhawk 6C, BECs, power module, RC receiver, telemetry radio), avionics shelf dimensions from Alperen (fuselage v5), forward battery bay dimensions, CG target from Alperen
Outputs	Physical layout drawing (top view and side view), cable routing plan, component mass and CG contribution
Dependencies	WP-E01 (battery dimensions locked); Alperen's fuselage truss design finalized
Deliverables	Dimensioned layout drawing (Inventor, Fusion, or dimensioned sketch), component placement table with CG contribution

WP-E08 — RC and Telemetry Link Selection and Integration

Field	Value
Owner	Erick
Phase	1
Inputs	Phase 1 range requirement (VLOS, ≤ 5 km), team's transmitter hardware, GCS preference (Mission Planner or QGroundControl), ArduPilot SERIAL configuration
Outputs	RC system and telemetry radio selections, FC UART configuration, antenna placement plan, range test results from bench
Dependencies	Team transmitter and GCS confirmation from Zeynep/Alperen
Deliverables	Hardware selection with rationale, parameter changes, bench radio-link verification

WP-E09 — Heading LED System

Field	Value
Owner	Erick
Phase	1
Inputs	Wing tip geometry from Alperen, ArduPilot NTF documentation, available GPIO pins on Pixhawk 6C
Outputs	LED strip type and length, wiring plan, ArduPilot NTF parameter configuration
Dependencies	Wing tip dimensions confirmed by Alperen
Deliverables	Wiring diagram, NTF parameter configuration, installed and tested (confirmed LEDs respond to arm/disarm and mode changes)

WP-E10 — IC Engine Electrical Integration (Phase 2)

Field	Value
Owner	Erick
Phase	2
Inputs	Engine selection (DLE vs Stinger, 35cc vs 55cc), ignition module spec, throttle servo travel, starter spec if procured
Outputs	Phase 2 additions to harness diagram, ignition supply regulator design, throttle servo wiring, vibration isolation plan, notch filter tuning procedure
Dependencies	Engine selection by Alperen; Phase 1 electrical installation complete
Deliverables	Updated harness diagram, ignition supply schematic, Phase 2 parameter additions (.param delta from Phase 1 file)

WP-E11 — Generator Architecture Research and Selection (Phase 2)

Field	Value
Owner	Erick
Phase	2
Inputs	Cruise power budget (100–200 W target), weight budget (~1 kg), engine shaft output and RPM at cruise, starter motor specs if Pilot RC unit is procured
Outputs	Generator architecture decision with supporting analysis, preliminary schematic for chosen option
Dependencies	Engine selection; Phase 1 flight data for actual cruise power consumption
Deliverables	Architecture decision document with power budget, efficiency estimate, and weight breakdown; preliminary schematic

WP-E12 — Long-Range Telemetry Research and Selection (Phase 3)

Field	Value
Owner	Erick and Zeynep
Phase	3
Inputs	200 km range requirement (E-REQ-07), latency tolerance for BVLOS command loop, Turkish and international BVLOS regulatory requirements, Phase 2 flight data
Outputs	Telemetry system selection, link budget analysis, integration and antenna plan
Dependencies	Phase 2 flight testing validating platform readiness for BVLOS
Deliverables	Link budget spreadsheet, hardware selection with rationale, wiring and parameter changes, regulatory compliance note

9. Open Questions

ID	Question	Owner	Priority	Gates
OQ-01	Battery closed June 17 (single 12S 22Ah 15C semi-solid pack, motorobit TR, 3,709g, 190×78×126mm, QS8-S anti-spark). Only remaining: the forward-bay CG placement, pending the CAD session with Alperen (WP-E07), plus price confirmation	Erick / Alperen	Medium	WP-E07
OQ-02	CAN-PWM adapter (Matek CAN-L4-PWM): enumerates as a DroneCAN servo node and outputs correct PWM under ArduPilot 4.x?	Erick / Zeynep	High	WP-E04
OQ-03	~~Altimeter: Ainstein US-D1 vs Benewake TF03-180?~~ Closed June 11: US-D1 selected (radar robustness in prop wash dust). Quote pending, longest lead item	Erick	—	—
OQ-04	RC side closed June 18 (Radiolink AT9S Pro + R9DS, SBUS). Telemetry side: first flight borrows STORK's Holybro SiK V3 on TELEM1; the long-term Spearhead telemetry radio (RFD900x-class) is still open	Alperen / Erick / Zeynep	Medium	WP-E08
OQ-05	~~HV kill: must be RC-switchable (June 11 criterion)~~ Closed June 17: no hardware HV kill for PT1. Software E-stop + QS8-S unplug for ground safing, recorded §9.3 deviation (§4.3). LV kill retained. Full HV kill returns for the final product	Erick / Alperen	—	—
OQ-06	Pitot tube (from storage): model and interface confirmed?	Alperen	Medium	WP-E05
OQ-07	Shared or separate battery for VTOL vs IC ignition in Phase 2?	Erick	Low	Phase 2 design
OQ-08	Electric starter generator mode: what is the mechanical limitation?	Erick	Low	WP-E11
OQ-09	Wing panel connector: Molex Mini-Fit Jr confirmed, or alternative?	Erick	Medium	WP-E06
OQ-10	Payload quick-release connector: align with Arrow platform standard?	Erick / Alperen	Low	Phase 4 design
OQ-11	~~Here4 breakout board: does Thomas's kit include one?~~ Closed, superseded by the standalone adapter. Here4 no longer used as converter.	Thomas	—	—
OQ-12	~~Primary GPS: second Here4 (matched RTK pair, GPS yaw) or F9P-class/salvage (compass yaw)?~~ Closed June 17: u-blox F9P on a serial UART (GPS1), compass yaw, on-hand antennas (parts list D1)	Team	—	—

10. Revision History

Rev	Date	Author	Changes
0.19	June 26, 2026	errrks	Reconciled with the June 18–19 call notes. RC link closed (D3 RC side): Radiolink AT9S Pro transmitter + bundled R9DS receiver on SBUS into the 6C RC-IN port, superseding the ExpressLRS/Crossfire CRSF preference for PT1. Telemetry (D3 telemetry side): first flight borrows STORK's Holybro SiK V3 (915 MHz, 100 mW) on TELEM1; the long-term unit (RFD900x-class) stays open, to be researched by Zeynep + Alperen. Updated §2.2 comms topology, §4.6 H17/H18, §4.12 (RC Link + Telemetry Radio), §4.15, §4.16 BOM, OQ-04. RC and ground telemetry stay on separate links since SBUS is unidirectional.
0.18	June 17, 2026	errrks	Decisions and reconciliation pass against the June 1–15 call notes. HV kill: none for PT1 — software E-stop (ArduPilot motor emergency stop on RC, AMPX 2 s CAN watchdog backstop) + QS8-S anti-spark unplug for ground safing. Recorded §9.3 deviation, E-REQ-02 annotated (§3), OQ-05 and D2 closed; LV kill retained; full HV kill returns for the final product. F9P confirmed salvaged (on hand, $0). Battery reclassed as semi-solid (~290–300 Wh/kg); clarified that "motorobit" is the TR retailer, not the pack brand. Avionics compartment dimensioned ~180×250mm × 90mm (June 12) with a WP-E07 fit check (§2.3, §4.15). Altimeter noted optional for first flight with a Feather LiDAR fallback (§4.8). OQ-01 narrowed to forward-bay CG placement pending the CAD session. Bench FC confirmed as the full 6C (same unit as the build, no Mini mismatch). Cross-doc consistency pass: reconciled `docs/open-questions.md` (battery/connector/FC/GPS/servos/POWER2/HV-kill), info note (SPH-E-002) updated, parts list (SPH-E-003) dated June 17. Updated §2.1, §3, §4.1, §4.3, §4.6 H1, §4.8, §4.15, §4.16, OQ-01, OQ-05.
0.17	June 17, 2026	errrks	FC changed from the Pixhawk 6X to the full-size Pixhawk 6C + analog PM02 V3 (rx-dynamic TR, no GPS in the bundle). The 6C's IO PWM still drives the flaperons (wing CAN node stays dropped) and its two analog POWER ports enable the FCHUB vehicle V/I to feed POWER2 directly (PM02 V3 on POWER1 = supply + BATT2 voltage; FCHUB on POWER2 = BATT1 current + voltage), which the digital PM02D could not do. UART map reworked for the 6C's 5 ports. Primary GPS moved from a TBD DroneCAN node to a u-blox F9P on a serial UART (GPS1, compass yaw), closing OQ-12/D1 and removing CAN node 10. Battery changed from the ProFuse 16Ah 60C to the motorobit 12S 22Ah 15C solid-state pack (3,709g, 190×78×126mm, QS8-S anti-spark socket; ~113g lighter than the ProFuse so the weight trade is moot; 15C = 330A continuous, covers the 320A peak but watch sag), reopening/reclosing WP-E01 and superseding the June 11 D5. QS8-S being anti-spark removes the connector argument from the HV-kill decision (§4.3). Servos: PT1 now runs the final-class Kingmax CLS3015S HV servos directly (35 kg·cm @ 8.4V, 80g), dropping the 5V test-servo stage; note the +100g aft tail mass for CG. Servo rails: forward PM12S-3 dropped, replaced by a robocombo DC-DC buck @ 8.4V for the flaperons; the tail PM12S-3 stays at Vx 8V for the ruddervators plus its fixed 5V for the adapter and Here4. Both servo rails are HV-derived, so §4.3 LV kill no longer lists any servos. Battery monitoring documented to the parameter level (§4.13, Setup B): FCHUB on POWER2 = BATT1 (`BATT_VOLT_PIN 5`, `BATT_CURR_PIN 14`, MULT 21.0, PERVLT 133.3), PM02 V3 on POWER1 = BATT2 (`BATT2_VOLT_PIN 8`, `BATT2_CURR_PIN 4`, MULT 18.18, PERVLT 36.36), instances deliberately swapped so the real current sensor is primary. 6C ADC pin numbers taken from the ArduPilot Pixhawk6C hwdef. Updated §2.1/§2.2/§2.3, §4.1–§4.4, §4.6 (H1/H2/H6/H8/H9/H15/H20 + wing/tail connector specs), §4.7, §4.8, §4.9, §4.10, §4.11, §4.12, §4.13, §4.15, §4.16 BOM, OQ-12.
0.16	June 11, 2026	errrks	Reconciled with the June 5 and June 11 calls. FC: Pixhawk 6C or 6X by sourcing (Zeynep approved either). Power module follows the FC and the pairing error is fixed: PM02D is I2C, 5X/6X only, so the 6C path uses the analog PM02 V3 and the 6X set includes the PM02D HV. Battery Y-splitter dropped: the FC power module taps the FCHUB battery input pads (Y stays as fallback). Component zones reworked for fuselage v5: forward battery bay, top-access avionics shelf, slimmed nose, boom-routed tail harness. HV kill: added the RC-switchable criterion (June 11) and a new candidate table (EV200 + relay driver, MOSFET anti-spark switch, software E-stop + recorded deviation), superseding the PP75 option (undersized at 75A class vs 98–124A hover). Altimeter selected: US-D1, specs corrected to 110g / 100 Hz / UART+CAN per Ainstein. Regulators standardized on Matek BEC12S-PRO (5.2/8/12V selectable). Added the PT1 harness connection table (§4.6, H1–H21), renamed §4.6, and split the procurement view into `docs/pt1-parts-list.md` (SPH-E-003). §5: DLE55-class layout model, starter-generator moved to PT3 research. Late June 11 additions: battery selected (ProFuse Super Nano 12S 16Ah 60C, XT150 socket, 3,822g, 186×76×160mm, robotsepeti TR), battery interface moved from AS150U to XT150 with the anti-spark burden shifted to the inline kill (§4.3 connector interaction note), FCHUB-12S confirmed as V2 (built-in 1K:20K VBat divider closes the §4.13 voltage sense question, 12VSW PINIO pad can drive the EV200 coil directly), tail CAN boards sourced in Turkey at Aykut Havacılık with the CAN-L431 documented as alternate, US-D1 believed on hand in Turkey pending confirmation. Regulators changed to 2× Matek PM12S-3 (Vx 15A selectable 5.25/6/8V + fixed 5V + fixed 12V per unit, Aykut Havacılık), superseding the BEC12S-PRO plan: tail unit feeds the ruddervator Vx rail plus adapter/Here4 5V, forward unit feeds the flaperon Vx rail, and the final 8V servo change becomes a pad setting. A wing CAN-PWM node (13) was added and dropped the same evening with the FC decision. Final state: Pixhawk 6X selected (D4 closed), chosen because its IO PWM outputs 1–2 drive the wing flaperons directly, saving the second adapter and shelf space. CAN-L4-PWM order stays at 2 (tail node 12 + spare), `CAN_D1_UC_SRV_BM = 0x000C`, and the §4.6 EMI rules re-apply to the ~1–1.5 m wing PWM runs. 6X sourcing: Aykut lists the set at 38,220 TL (~3× Holybro direct) and is out of stock, so Holybro EU at $321 is the order path. The set's M9N GPS is kept as a bench/backup unit. PM12S-3 TR markup vs Matek direct flagged in the parts list. Battery harness corrected from 4 AWG to 8 AWG: 4 AWG does not fit XT150 solder cups and the connector is the thermal limit on the short run.
0.15	June 2026	errrks	Reconciled with the June 1 call (PT1 decisions). Servos: PT1 uses 5V test servos for wing and tail, dropping both 7.8V BECs. Flaperons run on the peripheral 5V rail, the tail runs on a local HV→5V BEC that also powers the adapter and Here4, which resolves the tail-5V run and puts the whole tail control path on the HV kill only. Final design keeps 6–8.4V HV servos. Battery splitter reverted from XT90 to AS150U/XT150 (full current, 4AWG). Added a §4.6 routing and EMI subsection. Kill switch: PT1 implementation left TBD (contactor/SSR vs manual disconnect), §9.3 requirement noted. Updated §2.1/§2.3, §4.1, §4.3, §4.4, §4.6, §4.10, §4.11, and the BOM.
0.14	June 2026	errrks	Battery changed from a DroneCAN smart pack to a dumb 12S LiPo at 16Ah (Alperen's May mission-energy rerun, saves ~1kg). Monitoring moves to the FCHUB-12S 440A analog current sensor plus a voltage sense (BATT_MONITOR = 4), with the PM02D I2C as a redundant voltage reading. No per-cell or CAN battery data. Reworked §4.1, §4.2, §4.13, §2.3, and the BOM. Reconciled against the May 2026 meeting notes.
0.13	June 2026	errrks	Flight controller powered by a Holybro PM02D (XT60 input, 5.2V/3A + I2C) on its own leg of an XT90 splitter at the battery, isolated from the PDB. Single FC supply: POWER1 = PM02D, POWER2 unused (no hot-standby for V1). Restructured §4.4 5V rails into FC power, peripheral 5V, and LED 5V. XT90 accepted for the high-current path (short hover duty, monitor connector temps), high-current leads drop to ~10AWG to fit XT90, the PM02D leg is a smaller-gauge XT60 branch. PM02D I2C is an optional BATT2 voltage monitor and its current reads only the FC branch. Updated §2.1/§2.3, §4.3, §4.4, §4.6, §4.7, §4.13, and the BOM.
0.12	June 2026	errrks	Power distribution: selected the off-the-shelf Matek FCHUB-12S PDB for V1 (no custom PDB). Corrected its specs (4×70A continuous and 4×110A burst per ESC pad, 8–60V input, on-board 5V/5A + 12V/4A + 3.3V BECs, 440A current sensor left unused, solder pads not screw terminals). Only the power pads are used since the ESCs run DroneCAN. CAN1 to the ESCs now fans out through a passive CAN splitter/hub instead of a custom board. Updated §2.1/§2.2/§2.3, §4.2, §4.4, §4.9, and the BOM.
0.11	June 2026	errrks	Tail architecture change: the tail Here4 is now GPS only and a standalone DroneCAN-to-PWM adapter (Matek CAN-L4-PWM, node 12) drives the ruddervators. Custom tail PCB dropped. Ruddervator 7.8V now comes from an off-the-shelf HV→7.8V BEC on the HV spur. Confirmed two redundant GPS (AIP-006 §9.4): primary nose module (DroneCAN, TBD) plus secondary tail Here4. Rewrote §4.10 and updated §2.1/§2.2/§2.3, §4.3/§4.4/§4.6/§4.7/§4.8/§4.9, the BOM, WP-E03/E04/E05, and OQ-02/OQ-11. Added OQ-12. Note: a mixed GPS pair does not support moving baseline yaw.
0.10	May 2026	errrks	Added STATUS callout at the top of §5 (Phase 2): the entire section is preliminary, not a design baseline. Listed the specific items requiring further research before commitment (engine selection, starter, VESC dual-role viability, generator topology, ignition battery, CHT/EGT, vibration spectrum, throttle servo). Phase 1 (§4) remains the working baseline.
0.9	May 2026	errrks	Corrected §3 requirements traceability source: replaced "Charter §X.Y" citations with "AIP-006 §X.Y" (DAO forum proposal, March 2026). The formal project charter referenced in proposal §5 has not yet been published, so AIP-006 is the authoritative requirements source. Added clarifying note above the table.
0.1	May 2026	errrks	Initial draft — Phase 1 specified, Phases 2–4 stubbed
0.2	May 2026	errrks	Updated §4.1 and §4.5 with MAD V8013 PRO datasheet data: revised hover current (70A → 98–124A), battery floor (16Ah → 22Ah), added full performance table and operating zone limits
0.3	May 2026	errrks	Updated §4.5 ESC with full AMPX 80A V2 datasheet: BEC 5V/200mA (avionics-inadequate), IP67, thermal shutdown thresholds (125°C/140°C), 2s CAN watchdog, wire specs (12AWG/800mm power, 14AWG/150mm phase). Corrected §4.6 harness AWG. Updated §4.13 with ESC thermal/CAN watchdog notes. Corrected BOM ESC unit price ($95 → $60).
0.8	May 2026	errrks	Confirmed smart battery: removed balance tap, PM07, and current sensor from design. Updated §4.1 (smart battery note), §4.2 (distribution), §4.3 (kill switch scope), §4.6 (harness), §4.7 (FC power), §4.13 (battery monitoring via DroneCAN BatteryInfo). BOM updated — nose GPS and RC/telemetry TBD pending team decisions.
0.7	May 2026	errrks	Corrected §4.1 hover analysis: 12S voltage range closely matches datasheet 48V conditions; motor operates in continuous zone for most of flight discharge curve, short-term zone only near low-battery threshold. Corrected capacity sizing to average draw. Separated §4.13 ESC thermal (junction temp, 125/140°C) from motor thermal (winding current zones, 24A/78A, motor datasheet source). Added motor temp monitoring proxy approach.
0.6	May 2026	errrks	Removed motor tilt (decision reversed). Renamed V-tail surfaces to ruddervators throughout. Restructured servo power: 7.8V regulation split — flaperon BEC in CG bay, ruddervator regulation on tail PCB from HV spur (no LV servo run through fuselage). Updated tail PCB spec to HV input with on-board switching regulators. Added smart battery decision point in §4.1.
0.5	May 2026	errrks	Updated §4.5 propeller with full FLUXER PRO 26×7.8 MATT datasheet: 69g/blade, Ø10mm bore, Ø20mm M3×4 hub, 1900–4000 RPM optimum, 18 kgf thrust limit, prop test data cross-validated against motor datasheet. Corrected BOM prop price ($90 → $193.90/pair). Phase 1 grand total updated ($4,668 → $5,081).
0.4	May 2026	errrks	Expanded §5 Phase 2 from stubs to detailed architecture: fuel mix and storage (§5.1), throttle and ArduPilot ICE parameters (§5.2), choke actuation and cold-start sequence (§5.3), ignition and kill wiring (§5.4), electric starter circuit (§5.5), RPM sensor options and ArduPilot notch filter setup (§5.6), engine health monitoring CHT/EGT (§5.7), VESC dual-role starter-generator architecture with battery top-up design (§5.8), vibration isolation procedure (§5.9), Phase 2 BOM (§5.10).

Table of Contents​

1. Purpose and Scope​

2. System Architecture​

2.1 Power Topology​

2.2 Communications Topology​

2.3 Physical Location Overview​

3. Requirements Traceability​

4. Phase 1: Electric VTOL (Detailed)​

4.1 Battery System​

Sizing​

Specification​

4.2 Power Distribution​

4.3 Kill Switches​

4.4 Voltage Regulation​

4.5 VTOL Motors and ESCs​

Motors​

ESCs​

Propellers​

4.6 Wiring Harness and Connection Plan​

Wire Sizing​

CAN Bus Wiring​

Servo Wiring (Tail Harness)​

Wing Panel Connector​

Routing and EMI​

PT1 Harness Connection Table​

4.7 Flight Controller​

4.8 GPS and Navigation Sensors​

GPS (×2 redundant, AIP-006 §9.4)​

Airspeed Sensor​

Radar Altimeter — Ainstein US-D1 (selected June 11, OQ-03 closed)​

4.9 CAN Bus Architecture​

4.10 Tail Electronics​

CAN-to-PWM Adapter — Matek CAN-L4-PWM​

Tail Servo Power — Matek PM12S-3​

Wing Flaperon Drive — 6C IO PWM (wing CAN node considered and dropped)​

4.11 Servo System​

Ruddervator Servos (×2)​

Flaperon Servos (×2)​

4.12 Telemetry and Communications​

RC Link​

Telemetry Radio​

4.13 Health Monitoring​

Battery Monitoring​

ESC Telemetry and Thermal​

Motor Thermal​

Pre-flight Diagnostics​

Phase 2 Additions (stub)​

4.14 Heading LEDs​

4.15 Component Locations​

4.16 Phase 1 BOM and Cost Estimate​

5. Phase 2: Hybrid Integration​

5.1 Fuel System and Mix​

2-Stroke Fuel Mix​

Fuel Storage and Delivery​

5.2 Throttle Control​

5.3 Choke and Cold-Start Sequence​

Choke Actuation Options​

Cold-Start Sequence (proposed, to be refined with engine testing)​

5.4 Ignition System​

5.5 Electric Starter​

Pilot RC Auto-Starter (primary candidate)​

Fallback: Manual Pull-Start​

5.6 RPM Sensor​

Sensor Options​

5.7 Engine Health Monitoring​

5.8 Generator Architecture and VESC Starter-Generator​

Power Budget​

Architecture Decision: VESC Dual-Role Starter-Generator (Option A)​

Alternative: Dedicated Belt-Driven Alternator (Option B)​

5.9 Vibration Isolation and Notch Filter​

5.10 Phase 2 Preliminary BOM​

6. Phase 3: BVLOS (Stubs)​

6.1 Long-Range Telemetry (> 200 km)​

6.2 Obstacle Avoidance​

7. Phase 4: Payload Bay (Stubs)​

7.1 Payload Bay Power​

7.2 Payload Data Interface​

7.3 Quick-Release Connector​

8. Work Packages​

WP-E01 — Battery and BEC Sizing and Procurement​

Table of Contents