Subject: Canard EFI Fuel Systems and Experimental Flight Testing
From: Rotary Engine
Date: 5/25/2005, 10:19 AM
To: AA-me


Paul,

I've read some of your comments recently about canard fuel systems, the
Conner and John Denver accidents.

John Denver didn't have EFI or a sump tank. Sump tanks are not standard
on EZs. I built a sump tank as the easiest solution to the EFI problem.
My sump tank only holds one gallon, and holds two Mazda RX-7 fuel pumps.
Fuel return is to the sump. I've never seen any evidence the returned
fuel is heating up the sump, but maybe because my sump is so small it
gets replenished with fresh fuel often enough. I have the stock LEZ fuel
system otherwise, L or R (not both) tank can be selected, and that
drains by gravity to the sump.

My sump was vented from day one. Never any problems. Then I noticed the
engine stumble one morning (this was 3 or 4 years ago), throttle back,
it recovered. It did this a couple more times that winter. Then one time
it did it I noticed fuel pressure low and fluctuating, so I knew the
sump tank was being depleted. I realized the common factors when it
occurred were: very cold mornings, low fuel in selected tank (< 5
gallons), and it happened after I had been doing full throttle speed
runs for an extended period. What was happening? The sump tank vent was
slightly higher and farther out of the stagnant air than the main tank
vents. I believe the sump tank was getting pressurized to a greater
extent than the mains. So I crimped the vent to the sump tank closed
slightly, to let less air in. It's only a 1/4" aluminum tube. That
seemed to clear it up completely. But then a year later one cold morning
it happened again.

I had read that if you didn't vent the sump tank,
that the EFI fuel pumps would actually suck fuel from the main tank. I
thought I would try it, but it made me nervous. I just put a rubber plug
over the sump tank vent tube on top of the fuselage. It worked fine for
several months while I stayed close to home and the plane was hangared.
THEN: I flew up to Puyallup airfair Feb. 2004. Puyallup has a single
4000 foot runway like Paul was flying out of. Flew up in the morning
when it was cool.

The plane was parked outside in the sun nose-down all
day while I was at the airfair. Went out mid-afternoon, raised the nose,
hopped in, started the engine, and taxied out to the runway like I
always do. THANK GOD there were several planes in the pattern, and I had
to sit there at the end of the runway waiting to takeoff. FINALLY I took
the active, punched in full throttle, and the engine died! This had
never happened before. I didn't have a clue at first, just got out and
pulled the plane off the runway. Eventually I noticed I had no fuel
pressure, then I remembered the rubber plug on the sump vent. Popped it
off, heard a "poof" sound and I could hear the fuel flowing into the
sump tank.

Sitting out in the sun, with the fuel lines inside the
cockpit, cockpit warming up like a greenhouse, caused some kind of vapor
bubble to form between the main tank and the sump tank, and fuel would
not flow into the sump tank because it was not vented. So I realized at
that point that the sump tank had to be vented. When I got home, I
lowered the sump tank vent to the same level as the main tank vents, so
the sump tank wouldn't get pressurized higher than the selected main. So
far haven't had the problem reoccur, and I'm going much faster now than
I was then. The area around Puyallup airport is congested city, if my
engine had quit right after takeoff, I probably wouldn't have made it as
there is no crossing runway and there probably would have been nowhere
to land straight ahead. I intend to install a low-fuel level sensor
inside the sump tank for some advance warning it is being depleted.

I think Paul Conner's main error in hindsight was not technical, it was
his choice of airport to do his flight testing.  I picked McMinnville
because it had long runways and crossing runways. When I takeoff on 22
and no one is in the pattern I immediately turn 45 degrees left and
enter a downwind for 35, just in case. When you fly out of a single
runway airport, you don't have that luxury. Also, 4000 feet is very
short for flight testing a canard. I was uncomfortable at first with the
5400' I had. I checked Airnav.com and there are two airports within 10
miles of his home base that he could have flown his 40 hours off from.
They have multiple long runways, the only disadvantage being they have
control towers, which I think are a distraction while flight testing,
but he could have limited his early flight testing to early morning
hours when no one else is flying, that is what I did. I've preached
about this on other lists before.

John Denver's big mistake, besides not putting fuel in the plane, was
failure to fly the airplane. If he had just converted his speed to
altitude instead of fidding with the fuel valve, he would have had time
to make a controlled glide to the beach or the water.

Perry

Putting the high pressures fuel pumps in the sump tank as Perry has
done here obviates vapor lock between pump and the sump tank.
A very good idea. Widely used on cars of course. If, of course, you
use just one sump tank. If you have two sump tanks you should have four
pumps for backup purposes.

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.