Fuel System

ATA Chapter 28 — N720AK Systems Reference

System Overview

N720AK uses the EFII System32 electronic fuel injection system with an Aeromotive MAP-referenced fuel pressure regulator. This is a modern port EFI system — no venturi, no mechanical fuel injection servo. The fuel system is a pressurized loop: fuel flows from the tanks, through filters and pumps, around a fuel rail past six port injectors, through the regulator, and back to the selected tank via a return line.

The entire fuel tune — injector pulse widths, fuel maps, mixture scheduling — depends on the regulator maintaining a constant pressure differential across the injectors. If this differential changes (due to regulator wear, filter clogging, pump degradation, or plumbing restrictions), the ECU’s fuel calculations become wrong and the tune must be re-evaluated. Every component in this system exists to ensure that differential stays rock-steady.

Physical Plumbing — Tank to Injector and Back

Flow Path (supply side)

LEFT TANK (30 gal)  ──┐                                    ┌── RIGHT TANK (30 gal)
                       │                                    │
              Safety-wired                         Safety-wired
              shutoff valve                        shutoff valve
                       │                                    │
              40μ pre-filter                        40μ pre-filter
              (TS Flight Lines)                    (TS Flight Lines)
                       │                                    │
                       └──── Under seats ────┬──── Under seats ────┘
                                             │
                                     Andair duplex valve
                                     (tank selection)
                                             │
                                ┌────────────┴────────────┐
                                │                         │
                          Walbro 391 pump           Walbro 391 pump
                          (primary)                 (backup)
                                │                         │
                                └────────────┬────────────┘
                                             │
                                    10μ Aeromotive post-filter
                                    (mounted with 8L clamp to
                                     engine mount diagonal)
                                             │
                                        FUEL RAIL
                                    ┌──┬──┬──┬──┬──┬──┐
                                    │  │  │  │  │  │  │
                                   Cyl injectors (×6)
                                    │
                              End of rail
                                    │
                            ┌───────┴───────┐
                            │               │
                     Fuel pressure    MAP reference line
                      regulator ◄──── (from throttle body
                     (Aeromotive)      orifice port)
                            │
                        RETURN LINE

Flow Path (return side)

                        RETURN LINE
                            │
                    Through firewall
                            │
                    Andair duplex valve
                    (routes return to
                     selected tank)
                            │
                  ┌─────────┴─────────┐
                  │                   │
             LEFT TANK           RIGHT TANK

Key Design Points

• Pressurized loop: The pumps pressurize fuel continuously. The regulator bypasses excess fuel back to the tank. The injectors pull from a fuel rail that’s always at pressure.
• Dual pumps: Primary and backup Walbro 391 pumps on a rack. Only one runs at a time. A bus manager switch selects 1/AUTO or 2. In 1/AUTO mode, the bus manager automatically cuts over from pump 1 to pump 2 if fuel pressure drops to 22 PSI or below. The cutover is controlled by the bus manager and performed by a relay mounted under the panel.
• Two-stage filtration: 40μ pre-filter before the pumps (protects pumps), 10μ post-filter after the pumps (protects injectors and regulator).
• MAP reference: The regulator’s vacuum reference comes from an orifice port on the throttle body that provides a stable, damped manifold pressure signal. This is a small restrictive fitting — not a wide-open port — to prevent fuel pressure from chasing rapid MAP transients.
• Return routing: The return line goes back through the firewall to the Andair duplex valve, which routes it to whichever tank is currently selected. This is a direct run — no additional valving on the return side.

Components

Fuel Tanks

Shutoff Valves

Pre-Filters (40 micron)

Andair Duplex Valve

• Spec sheet (GDrive)
• Data sheet (GDrive)

Fuel Pumps — Walbro 391

Pump replacement notes:

Viton crush washers for pump fittings: One Hydraulics SS9500-02V (primary source, where N720AK’s were purchased). Alternate source: Titan Fittings SS-9500V series.

Post-Filter (10 micron) — Aeromotive

Fuel Rail

• Configuration: Common rail feeding all 6 cylinders
• Injectors: 6 port fuel injectors, one per cylinder

Fuel Pressure Regulator — Aeromotive

The regulator is the heart of fuel pressure management. See The MAP-Referenced Regulator for full theory.

MAP Reference Line

The vacuum reference line connects from an orifice port on the throttle body to the regulator. The orifice provides a stable, damped manifold pressure signal — it’s a small restrictive fitting, not a wide-open port. This prevents the regulator from chasing rapid MAP transients during throttle changes.

The MAP reference orifice is integral to the EFII-supplied throttle body (stock configuration — no aftermarket modification needed).

Fuel Pressure Sensor — Dynon Kavlico

See Dynon Service Bulletin 120414 regarding blocked baro compensation ports on sensors manufactured July 2013 – June 2014.

Fuel Lines

How It Works: The MAP-Referenced Regulator

How It Maintains Constant Differential

The Aeromotive regulator is a mechanical diaphragm device:

• One side of the diaphragm sees fuel pressure from the fuel rail
• Other side sees manifold pressure (via a vacuum reference line from the throttle body) plus a calibrated spring

The diaphragm balances these forces:

Fuel_absolute = MAP_absolute + Spring_force

When MAP rises (throttle opened), the diaphragm sees more pressure on the reference side, so it allows fuel pressure to rise by the same amount. When MAP drops (throttle closed), fuel pressure drops to match. This is called 1:1 tracking — every PSI change in MAP produces a matching PSI change in fuel pressure.

The result: The spring force sets the differential, and it stays constant:

$$ P_{\text{fuel}} - P_{\text{MAP}} = F_{\text{spring}} \approx 35 \text{ PSI (constant)} $$

What Goes Wrong

The Critical Quantity: Injector Differential Pressure

The pressure drop across each fuel injector determines how much fuel flows during the injector’s open time:

True_differential = Fuel_absolute − MAP_absolute

For N720AK, this should be constant at ~35 PSI regardless of throttle position, altitude, or flight phase.

Why It Matters

The EFII ECU calculates injector pulse width assuming a known, constant pressure differential. If the differential changes:

• Higher differential → more fuel per pulse → richer mixture → wasted fuel, fouled plugs, potentially hydraulic lock
• Lower differential → less fuel per pulse → leaner mixture → hot cylinders, detonation risk, rough running

All fuel tuning depends on this value being stable. If the differential changes — due to regulator wear, filter clogging, pump degradation, or plumbing restrictions — the fuel maps become wrong and must be re-evaluated.

What Changes the Differential

Monitoring the Differential Over Time

Track these metrics at each annual or whenever the fuel system is serviced:

1. Startup fuel pressure (gauge, engine off, pump on) — this is the spring setpoint
2. Run the regulator_diagnostic.py script on a representative flight log
3. Compare against the baseline in the flight log history table below

If any of these metrics change significantly, investigate before flying further. A change in the differential means the fuel tune is no longer valid.

Sensor Reference Frames

This is the most important subtlety in interpreting fuel system data. The two sensors measure in different reference frames:

• MAP sensor measures absolute pressure (relative to vacuum). 29.92 inHg at sea level on a standard day.
• Fuel pressure sensor (Dynon Kavlico) measures gauge pressure (relative to local atmosphere). Reads 0 PSI with no fuel pressure at any altitude.

The Gauge-vs-Absolute Problem

When we naively compute Fuel_gauge − MAP_psi, we get:

Delta_gauge = Fuel_gauge − MAP_psi
            = (Fuel_absolute − Atmosphere) − MAP_absolute
            = (Fuel_absolute − MAP_absolute) − Atmosphere
            = True_differential − Atmosphere

This is not the true injector differential. It’s offset by atmospheric pressure (~14.7 PSI at sea level, ~10.5 PSI at 9,000 ft).

Altitude Effect

Atmospheric pressure decreases with altitude:

$$ P_{\text{atm}} = 14.696 \times \left(1 - 6.8756 \times 10^{-6} \times h\right)^{5.2559} \quad [h \text{ in feet}, P \text{ in PSI}] $$

As you climb, the gauge delta changes even if the regulator is perfect — because Atmosphere in the equation above is changing.

The Correction

To recover the true injector differential:

$$ \Delta_{\text{true}} = \Delta_{\text{gauge}} + P_{\text{atm}}(h) = P_{\text{fuel,gauge}} - P_{\text{MAP(psi)}} + P_{\text{atm}}(h) $$

Dynon Baro Compensation

Dynon’s Kavlico fuel pressure sensor has baro compensation — an atmospheric vent port that keeps the gauge reading accurate as altitude changes. This does NOT convert the reading to absolute; it just ensures the gauge reading is a faithful measurement of Fuel_absolute − Atmosphere at any altitude.

The baro compensation means the sensor is a good gauge sensor, but it doesn’t eliminate the need for the altitude correction when computing the true injector differential. The correction is required because of the reference frame mismatch between the gauge fuel pressure sensor and the absolute MAP sensor.

Caveat — Dynon Service Bulletin 120414: Some Kavlico sensors manufactured July 2013 – June 2014 had a blocked baro compensation port (environmental seal covering the vent). This causes the sensor reading to drift ~0.5 PSI per 1,000 ft of altitude change. Test by removing the silicone gasket from the connector and checking if altitude-dependent drift decreases. See: https://dynonavionics.com/bulletins/support_bulletin_120414.php

When to Apply Altitude Correction

How to Test Whether Your Sensor Needs Correction

Use the --alt-scan flag on the diagnostic script. It applies a range of correction fractions (0.0 to 1.0) to the altitude term and finds the fraction that minimizes residual sigma:

uv run --with numpy --with matplotlib python3 scripts/regulator_diagnostic.py /path/to/log.csv --alt-scan

• Fraction $\approx 1.0$ → Sensor is gauge, correction needed (e.g., Garmin Kavlico)
• Fraction $\approx 0.0$ → Sensor already behaves as absolute, no correction needed

For Garmin GDU 460 fuel pressure, we confirmed fraction = 1.0 (standard gauge sensor). For Dynon SkyView fuel pressure on N720AK, we found fraction $\approx 0.0$ — this is still under investigation. The regulator noise ($\sigma = 1.4$ PSI) may be swamping the altitude signal (~0.4 PSI/1,000 ft). Fix the regulator first, then re-test.

Diagnostics

Running the Script

cd ~/code/rv10

# Dynon SkyView log
uv run --with numpy --with matplotlib python3 scripts/regulator_diagnostic.py /path/to/dynon_log.csv

# Garmin log (needs altitude correction)
uv run --with numpy --with matplotlib python3 scripts/regulator_diagnostic.py /path/to/garmin_log.csv --alt-correct

# Multiple flights
uv run --with numpy --with matplotlib python3 scripts/regulator_diagnostic.py flight1.csv flight2.csv

# Altitude correction scan
uv run --with numpy --with matplotlib python3 scripts/regulator_diagnostic.py /path/to/log.csv --alt-scan

# Skip plot (console output only)
uv run --with numpy --with matplotlib python3 scripts/regulator_diagnostic.py /path/to/log.csv --no-plot

Step 1: Load and Filter Data

The script auto-detects Dynon vs Garmin format. It extracts:

• Session time, MAP (inHg), fuel pressure (PSI gauge), fuel flow (gal/hr), pressure altitude (ft)
• Engine-on filter: MAP > 15 inHg, fuel pressure > 5 PSI, session time > 5 minutes
• Steady-state filter: |dMAP/dt| < 0.05 inHg/sec (throttle not moving)

Step 2: Compute Delta

$$ P_{\text{MAP(psi)}} = P_{\text{MAP(inHg)}} \times 0.49115 $$

$$ \Delta_{\text{gauge}} = P_{\text{fuel}} - P_{\text{MAP(psi)}} $$

Apply altitude correction if using --alt-correct.

Step 3: Key Metrics

Step 4: Bin Analysis

The script divides steady-state data into MAP bins (15–19, 19–22, 22–26 inHg) and computes $\sigma$ within each bin. High within-bin scatter indicates sticking/hunting independent of the MAP slope.

Step 5: Diagnostic Plot

The 4-panel plot shows:

1. Delta vs MAP — reveals MAP slope and scatter pattern
2. Delta vs Fuel Flow — reveals flow-dependent droop
3. Delta vs Time (colored by |dMAP/dt|) — reveals sticking events and altitude correlation
4. Delta histogram — reveals distribution shape (unimodal = good, bimodal = sticking)

Flight Log Analysis History

Reference Baseline: N88810 (Healthy Regulator)

Same Aeromotive regulator, EFII System32, Garmin GDU 460 (gauge fuel pressure sensor). Flight: X05 → KGAD, 2026-02-09. Altitude range: sea level to 6,161 ft.

After altitude correction (confirmed fraction = 1.0 for Garmin sensor):

• True differential: $35.03 \pm 0.08$ PSI — essentially perfect
• MAP slope: $-0.011$ PSI/inHg — essentially zero
• Fuel flow slope: $+0.046$ PSI/(gal/hr) — flat
• Residual $\sigma$: 0.07 PSI

This proves the Aeromotive regulator CAN perform essentially perfectly with the EFII System32.

N720AK Flight History

Key findings:

• MAP slope is consistent at $-0.30$ PSI/inHg across all flights — this is a structural characteristic of the regulator, not flight-dependent
• Sticking/hunting varies dramatically (residual $\sigma$: 0.18 to 1.25) — worse on longer flights with more throttle changes, possibly temperature-dependent
• Startup fuel pressure varies (32.0 to 35.8) — may indicate the regulator settles at different stick points on startup

Diagnosis: N720AK Regulator

Problem 1: MAP Under-Tracking (Slope = −0.30 PSI/inHg)

The regulator does not follow MAP changes on a 1:1 basis. For every 1 inHg increase in MAP, the injector differential drops by 0.30 PSI. Over the 10 inHg operating range, that’s ~3 PSI of sag — about 9% of the 33 PSI setpoint.

Likely causes:

• Vacuum reference line partially restricted or leaking
• Diaphragm stiffness or age
• Spring rate mismatch

Effect: At high power, injectors see less differential than intended → leaner than the ECU expects. At idle, richer than expected. The ECU’s fuel map is calibrated assuming constant differential — this slope means the actual air-fuel ratio shifts with power setting.

Problem 2: Mechanical Sticking / Hunting (Variable, up to σ = 1.25 PSI)

Random, non-repeatable variation in the differential, worst at mid and high power settings, worse on longer flights.

Likely causes:

• Diaphragm friction or contamination
• Valve seat wear or debris
• Spring fatigue causing hysteresis

Effect: Random variation in fuel delivery per injector pulse → uneven cylinder mixture, potentially rough running.

Recommendations

1. Inspect vacuum reference line — check for kinks, cracks, loose fittings, contamination
2. Replace or rebuild the regulator — the sticking won’t be fixed by a vacuum line repair
3. After replacement, re-fly and re-analyze — target σ < 0.2 PSI and MAP slope near zero
4. After regulator fix, re-test the baro port question — with low regulator noise, the altitude signal will be detectable

Open Questions

• Dynon baro compensation behavior: Our analysis showed the altitude correction fraction = 0.0 for N720AK, suggesting the Dynon is removing altitude effects internally. But the physics says a baro-compensated gauge sensor should still need the correction. The regulator noise (σ = 1.4 PSI) likely swamps the altitude signal (~0.4 PSI/1,000 ft). Fix the regulator first, re-test with clean data.
• Startup fuel pressure variation: 32.0 to 35.8 PSI across flights. Is this temperature-dependent? Different regulator stick points? Pump output variation?
• Baro port status: Cannot conclusively test from current flight data. Fix regulator first.
• Dynon differential display: Dynon offers a built-in Fuel_gauge − MAP differential display, but this shows True_differential − Atmosphere, not the true injector differential. It shifts with altitude. Not sufficient for regulator health monitoring without manual altitude correction.

References

• EFII System32 Installation Manual (Rev 9-13)
• EFII System32 Installation Manual (Rev 6-19)
• EFII System32 Operating Procedures (12-20)
• EFII System32 Fuel Flow & RPM Config (Rev 10-19)
• EFII System32 Initial Tuning — CSP (Rev 6-20)
• EFII System32 Upgrade Installation Manual
• Regulator Diagnostic Script — auto-detects Dynon/Garmin CSV, computes diagnostic metrics, generates 4-panel plot
• Regulator Diagnostic Plan — workflow for analyzing regulator health
• Andair FS2020-D2 Duplex Valve Spec Sheet
• Andair FS20 Fuel Selector Data Sheet
• Aeromotive 13129 EFI Bypass Regulator Manual
• Dynon Service Bulletin 120414 — blocked baro compensation ports on Kavlico sensors
• JDAir Fuel Drain & Vent Fairings Combo — silver drain fairing, blue vent fairing
• Flight log data: maintenance/fuel-system/flight-logs/

TODO: Information Needed

Component Data Sheets and Part Numbers

• Aeromotive regulator — Aeromotive Compact EFI Regulator with 0.020“ bypass orifice (ProTek Performance / Robert Paisley)
• Walbro 391 pumps — pump curves filed: GDrive, also images/walbro-gsl391-pump-curves.png
• ProTek Performance pump rack — dual electric fuel pump, supplied by EFII
• Aeromotive post-filter — Aeromotive 12347, 10-micron, serviceable
• TS Flight Lines pre-filter — private label, no part number available; 40 micron, serviceable
• Andair duplex valve — FS20-20-D2-6M, with 5x EF20 elbow fittings
• EFII fuel injectors — part number, flow rating
• Fuel rail — manufacturer, part number
• Kavlico fuel pressure sensor — exact Dynon part number, data sheet
• Shutoff valves — Peterson Fluid Systems 09-0910, -6 AN panel mount ball valve

Fuel Lines

• Supplier name and contact info
• Line type (AN size, material — stainless braided PTFE?)
• Lengths for each run (tank to valve, valve to filter, filter to duplex, duplex to pump rack, pump to post-filter, post-filter to rail, regulator return through firewall, firewall to duplex return, duplex to tank return)
• AN fitting sizes at each connection
• Any adapters or reducers in the system
• Torque specs for AN fittings
• Photos of routing under seats and through firewall

Pump Replacement

• Viton crush washers — Titan Fittings SS-9500V
• Step-by-step replacement procedure
• What does it take? Time, tools, access?
• Photos of pump internals (healthy pump)
• How to tell when pumps need replacement

Filter Servicing

• Pre-filter cleaning procedure (solvent? compressed air?)
• Post-filter cleaning procedure
• Photos of clean filter element vs contaminated
• What does contamination look like? What does it mean?
• Replacement schedule or inspection criteria

Photos Needed

• Overall fuel system routing (overview)
• Pre-filter location and mounting
• Post-filter mounted on engine mount diagonal
• Pump rack
• Fuel rail and injectors
• Regulator and MAP reference line connection
• Throttle body orifice fitting
• Andair duplex valve
• Firewall passthrough
• Clean vs dirty filter elements

System Questions

• Fuel type — 100LL or premium unleaded 91 octane mogas minimum
• Total usable fuel per tank — 29.5 gallons per tank
• Pump selection — bus manager switch (1/AUTO or 2), auto cutover at 22 PSI via relay under panel
• MAP reference orifice — integral to EFII throttle body, stock configuration
• Fuel pressure sensor tap location on the rail