Subject: ACRE AD on Andair fuel selector valves.
From: Rotary Engine
Date: 5/24/2005, 8:12 PM
To: AA-me


If you think there is a better way of doing it send me a sketch.
I am open to suggestions.

Paul,
You're a much better artist than I, so I'll describe in words and if you
think my idea has merit, please feel free to draw it up and post it.

Assumption #1: The most failure prone component of the fuel system is
that piece of meat holding on to the yoke/stick.  Therefore, any design
should minimize the chance that failure of that component will cause
loss of fuel supply to the engine.

Assumption #2: Fuel is a highly flammable solvent.  Keeping it contained
is a major pain.  We should avoid making additional tanks of fuel where
possible.

Design:
Fuel controller (microcontroller)
 Inputs:
   I1. fuel tank enable (from panel switch); one per fuel tank
   I2. tank outflow encoder; one per tank
   I3. tank inflow encoder; one per tank
   I4. throttle position switch (over/under 50% switch)

 Outputs:
   O1. fuel pump; one per tank (one pump per tank)
   O2. fuel return enable; one per tank (connects to NC fuel valve)
   O3. fuel tank/pump status LED (red/green bi-color); one per tank
   O4. LCD panel for fuel flow rates & totalizers

Other electrical components
   E1. SPST tank on/off switch (one per tank)
   E2. red/green bi-color LED (one per tank)
   E3. LCD panel
   E4. Emergency switch (single pole, multi throw)

Fuel components:
   F1. Immersable EFI-pressure fuel pumps (one per tank)
   F2. Fuel pump & screens (one per tank)
   F3. Fuel flow encoders (two per tank; one inflow, one outflow)
   F4. Solenoid controlled normally closed (NC) fuel valve (one per
tank)
   F5. High pressure fuel line (from tanks to fuel rail)
   F6. Low pressure fuel line (from fuel rail to tanks)
   F7. Two N-to-1 fuel tees (if you have two tanks, this is a 2-to-1)
   F8. Fuel filter
   F9. Fuel rail, pressure regulator, and injectors

How it all goes together:
Put the fuel pumps in the low point of the fuel tanks (immediately above
a sump cock) with input screens.  This immerses the pump inlet in fuel,
practically eliminating the possibility of heat-aggravated inlet vapor
lock.  Using high pressure fuel line, plumb the pump outlet to a tank
outflow encoder, and from there to one of the N-to-1 tees (the "N"
side).  Repeat for each of N tanks.  Using high pressure fuel line,
plumb the N-to-1 manifold to the fuel filter, and the fuel filter to the
fuel rail assembly (fuel rail, injectors, and pressure regulator).

Using low pressure fuel line, plumb the output of the pressure regulator
to the other N-to-1 tee (the "one" side).  Using low pressure fuel line,
plumb the N-to-1 tee to a solenoid fuel valve, the solenoid fuel valve
to a tank inflow encoder, and the tank inflow encoder to the fuel tank.
Repeat for each of N tanks.

Connect the throttle position switch to the throttle plate so that it
goes on/off as the throttle passes the 50% position (or 40%; it isn't
critical).  Switch should close with the throttle above the transition
point and be open below that point.

Wire the fuel pumps, tank outflow encoders, tank inflow encoders,
solenoid valves, tank switches, tank status LEDs, throttle position
switch, and LCD panel to the fuel controller.  Wire the emergency switch
to each pump and power.  The emergency switch should normally be open.

Program the microcontroller as follows:

   P1. Count tank inflow-outflow and record as tank change (for each
tank)
   P2. Sum all tank's individual changes for engine consumption rate
   P3. If tank enable switch is ON and throttle sensor open circuit
         (i.e. below 50%) :
         Turn on pump and open fuel return valve for the tank
   P4. If throttle sensor transitions to closed (i.e. above 50%):
         Start "climb out timer" (10 minutes or so)
   P5. If tank enable switch is on and "climb out timer" still
counting:
         Turn on pump and open fuel return valve for the tank
   P6. If throttle sensor closed and "climb out timer" is zero:
         Operate ENABLED pump/return valves in round robin fashion
         (i.e. 10 minutes pump #1, then 10 minutes pump #2, then #3,
...)
   P7. If tank outflow encoder is turning, matching LED is green
   P8. If tank outflow encoder is NOT turning and pump is commanded OFF
         (ENABLED, but off in round-robin mode), matching LED is off.
   P9. If tank outflow encoder is NOT turning and pump is commanded ON
         (ENABLED and ON), matching LED is RED (maybe BLINKING RED).
         When this happens, switch to another enabled tank (i.e. act
         as if the round-robin timer reached zero).  LED remains red
         until that tank comes up again in the round-robin rotation AND
         the tank outflow counter indicates flow (resets from
         temporary port uncovering).
   P10. If tank is DISABLED (i.e. switch OFF), matching LED is amber
         (idiot warning)

Pilot's checklist:
At all times before takeoff (ignition on), fuel tank LEDs should be all
green unless tank is DISABLED, in which case LED is amber.

During climb out, the LEDs should stay all green (except if tank
DISABLED) until the "climb out timer" counts down.  If an LED goes RED,
it may indicate that tank's port became uncovered or that tank is out of
fuel.

During cruise, the LEDs should have one green and the remaining off.
The green LED should "march" across the tanks as the round-robin
advances.

During descent (throttle less than 50%), LEDs should be all green unless
tank is DISABLED, in which case LED is amber, or unless the tank is
empty/unported and enabled, in which case the LED is RED.


What I like:
============
Immediate trouble indication (don't wait for the engine to cough & die).

Pilot is relieved of fuel management tedium.

Plane always runs on tank with gas.  If plane runs out of gas, it has
*NO* gas (i.e. none of the stupid "wrong tank" errors).

No fuel lines near people.

No fuel pumps without flooded inlets (vapor lock eliminated).

Unconsumed fuel is returned to a tank rather than a header (minimizes
heat build-up in fuel rail fuel).

Cruise "all" mode uses fuel in a balanced manner.

No bunker of ready fuel (header tank) in engine compartment to supply to
the engine fire (Heaven forbid!)

Emergency switch runs all pumps & opens all valves if fuel controller
fails.


What I don't like:
==================
Cost is unknown (considering the cost of the Andair valves, this *might*
be cheaper).

Potential for tank overflow on the ground, during climb out, and descent
caused by unbalanced returns while in "all" mode.  This could be
reduced/eliminated if one wants to get clever with the controller
(periodically turn off the return valve for the tank that would
overflow).

Potential for unintended tank-to-tank transfer if a non-running pump can
be driven backwards by the pressure in the high pressure side.  I don't
think that this can happen, but it should be checked.  If so, add some
check valves on the high pressure side between the pumps and the tee.

People believe mechanical fuel mechanisms (manual valves, mechanical
pumps) are more reliable than the electrical equivalents, and they may
be right.  But I believe that the reliability of those components are at
least an order of magnitude greater than the reliability of the meat
component (pilot).

Russell Kent

That sounds interesting Russell. Can we postpone this for a few days until
I get this P-port engine running on the dyno?

Paul Lamar

-- 
The Rotary Engine NewsLetter.  Powered by Linux.
ACRE NL web site. http://www.rotaryeng.net
Copyright 1998-2005 All world wide rights reserved.