The mission:
OK so a few months ago, 844X had flown in primer and no interior
with the standard Lancair avionics package. This package consists
of:
Two Garmin G-900's for primary display and map.
A Moritz display for control of pressurization, air conditioning,
some power switches, and the like.
An L-3 Trilogy backup attitude indicator, to show the pitch, roll,
altitude, and airspeed of the aircraft in the (almost unheard-of)
event that the (really quite good) Garmins stop working.
A bunch of mechanical circuit breakers that customers mount in
various locations that pop out if current ever goes over-limit.
One of the purposes of 844X, however, is to be a test platform for
the Vertical Power VP-400, which is the artificially-intelligent
Runway-Seeker that is designed to choose the best landing site for
you in the event of engine failure or pilot incapacitation, and even
take you down to that runway on autopilot at the push of a single
red button. (I wrote the software for the VP-400, with the exception
of the inter-aircraft communication and control, which is by
Vertical Power). The Vertical Power VP-400, as part of this power,
gives a full synthetic vision display in the upper half of the
touch-screen interface, completely obviating the need for the L-3
Trilogy. As well, the VP-400, in the bottom half of the
touch-screen, lets you control air conditioning, pressurization, and
the like, obviating the need for the Moritz display. As well, the
VP-400 has an electronic circuit breaker panel that can be invoked
in the bottom half of the touch screen, thus obviating the need for
mechanical circuit breakers.
Replacing the Moritz, Trilogy, circuit breakers, and associated
wiring with a single VP-400 unit will result in a huge decrease in
panel-clutter, wiring, and weight, as you will see as you read on.
The primary purpose of the construction of 844X is to be a
development testbed for this new VP-400 technology, and we have to
have it ready for Oshkosh if it is to be shown to all… so we sort of
have a "Monster Garage" situation here, where we have a fascinating
mission to accomplish, and a very tight time-table in which to do
it.
So, with the airplane JUST out of the paint shop (in pieces: the
fuselage, each wing, each control surface, and upper and lower
cowling delivered separately!), the time had arrived to:
1: re-assemble the now-painted pieces of the airplane
2: remove the stock panel with the Moritz and L-3 and circuit
breakers and install an entirely new instrument panel with a
Vertical-Power VP-400 in it.
3: fly the remaining 20 hours that needed to be put on the airplane
to achieve the 40-hour test-flight period… much of this would be
done by me. (gulp)
4: fly this thing to Oshkosh!
We had a pretty tight schedule to do this… 10 days, in fact, since
that is when Oshkosh starts (!)
All of this work had to be done with the following parties:
1: RDD of Redmond, OR (build-shop that oversaw construction of the
airplane)
2: Vertical Power of Albuquerque, NM (company that developed the
VP-400 A-I Runway Seeker)
3: Laminar Research (well… ME, anyway)
All assembly and test-flying would be done in Redmond, OR, at RDD's
facility, before the cross-country to Oshkosh.
My plan was to show up at RDD at T minus 10 days to help with final
assembly and VP-400 installation and programming, and general
oversight of the airplane to be sure that it went together in a way
that I liked.
Play-by-Play to follow:
T minus 10 days: (Wednesday July
11)
So, I arrived by airliner (sigh) in Redmond at midnight, and grabbed
the rental car (a light and springy little Ford Focus… very nice
handling) and went racing up the mountain to the rental house in
Eagle Crest, 10 minutes outside of town. The rental house was in a
gated community (I HATE gated communities!) and of course nobody had
ever given me the gate code, so I was up until 1 am making phone
calls to get that sorted out. Trust me: I was VERY tempted to go
off-road and drive right AROUND the gate and into someone's back
yard and then into the gated community in the little Ford focus, but
there would have been uncomfortable questions to answer when I
returned the car in 10 days with the paint all scratched from
driving through the shrubbery. Finally arriving at the rental house
at 1 am with nothing but airline food in the 12 hour trip from South
Carolina, I discovered that the refrigerator in the rental house was
stocked with the following food: Butter.
T minus 9 days: (Thursday July 12,
switching some writing to present-tense since I am updating daily)
Bright and early to RDD, N844x was sitting in the hangar in, if I
may say so, a rather confused and chaotic state of glory.
To the UNTRAINED eye, one would see a stunning, sleek, swoopy
carbon-fiber body elevated high above the ground, with a complex set
of piping around the long aluminum cylinder that makes up the engine
sticking out the front, and a variety of wires and cabling going
into it's various doors from various tools and plugs in the hangar,
and people scrambling all about it, readying it for flight.
To the TRAINED eye, quite a few other oddities stick out.
=>The plane is sitting on jacks so the landing gear cycling can
be tested, and ladders are needed to access any door into the
airplane.
=>The tail is NOT quite normal. Unlike a stock Evolution, which
has a nearly straight-up vertical stabilizer jutting abruptly out of
the body (much like a P-51 Mustang), the tail on this plane is much
more SWOOPY, following a long, smooth curve from the fuselage up to
the vertical stab. This is much more like a classic Lancair in
appearance.
=>The nose is not quite normal. UNlike a stock Evolution, which
has a fairly flat nose and windshield suddenly jutting up (again,
much like a P-51 Mustang), the nose sort of blends up into the
windshield for a more classic Lancair look.
=>As well, the prop is not the heavy aluminum Hamilton propeller,
but is instead a lightweight composite German MT prop (4-blade,
light grey, with metal leading edges) that can drop into a FEATHERED
position, leaving the blades completely streamlined (turned
perfectly into the wind) as the craft sits on the ramp, giving a
very distinctive look on the ground. As well, this prop sits a few
inches farther forward than on a stock Evo, since it is riding on
the front of a longer engine… a Pratt and Whitney PT6-A-42, which
packs 850 hp, as opposed to the stock 750.
>As well, if the cowl were installed (right now, it isn't) one
would note that the air intake is non-standard, but is instead a
little more curvy, perhaps following local streamlines more closely.
=>The careful observer, if he looks inside, would also notice
that the panel is completely different.. not even one bit of it is
from a standard Evolution (!) A normal Evo panel has the Garmin
G-900 primary and secondary displays sitting LOW on the panel,
RECESSED BACK a bit, with the various switches for electrical
systems sitting ABOVE them, and slightly CLOSER to the pilot. This
arrangement is great at keeping the Garmins in the shade so you can
always see them, but rides those displays sort of low… sort of lower
than I might find optimal, since I glance down from the windshield
to see generator and battery switches in my immediate line of
vision. In THIS Evolution, though, the panel is totally FLAT..
nothing is recessed or advanced, and the switches re BELOW the
Garmins. This puts the Garmin PFD FRONT AND CENTER, very closely
aligned with your vision, as high as it can be in the panel. The MFD
sits at an equal distance to the right, and the VP-400 sits between
them. The switches are lined up in a row underneath them, flowing
fro left to right across the panel in the order they are access in
flight. This panel is the result of about 3 days of discussion
between myself and RDD, and about 3 revisions on paper. Without a
SINGLE gauge on the airplane (the VP-400 is the backup.. FAR
more reliable than the typical mechanical backups) you see only 3
computer screens on a flat grey panel, with a row of switches
beneath them. It is actually sort of hard to believe that such a
simple setup could do so much! This is an airplane that literally
has no gauges… only 3 computer screens! With only a few computer
screens on flat black metal, the panel looks TOTALLY military.
Really, not civilian at all.
=>An especially trained eye would note that the airplane has a
paint scheme that has very little overlapping stripes, resulting in
less weight in paint. As well, one might notice that the airplane is
NOT ACTUALLY WHITE, but is instead a creamy off-white that WEIGHS
LESS THAN WHITE because it can go on in a thinner layer of paint and
still give total coverage, because it has a light-absorbing color..
perfect light reflection is not needed with any color other than
white, so the layer of paint can be (and is) thinner. In fact, the
flight controls were built with lead in them ahead of the hinge-line
to balance the weight of the control surface that sat behind the
hinge-line… and that lead was measured during the construction of
the flight controls to offset the weight of standard bodywork and
paint. Since I sanded off all the bodywork, and Tom Connors (paint
guy) designed such a thin layer, the lead allocation turned out to
be all wrong and workers had to drill into the flight controls after
the paint was done to REMOVE half the lead from the controls! We
actually did a careful enough job with body and paint to "break" the
design of the airplane.
At first, all I could do was sit there, mouth sort of agape in a
silly grin, and stare at it as RDD workers scrambled about it,
rushing to hook up the myriad of electrical systems needed to make
ready the entire new panel. But, of course, you don't sit there
staring forever, so soon I was asking what I could to help. During
the construction of the airplane, I played a pretty significant
part, because they could give me resin, carbon fiber, and some
plans, and then leave me alone for a while to slave away. The work
was easy to learn, and if I messed up the occasional part,
well, as Jay Leno says about Doritos: "Crunch all you want… they'll
make more". Final assembly is NOT quite like that, though. EACH BIT
of the final assembly is different, and EACH BIT should be performed
by an expert in that area. A guy with FLIGHT CONTROLS EXPERIENCE
should hook up the rudder, elevator, and ailerons to the wings and
control-arms. A guy with DOOR experience should hook the doors back
up. A guy with good ELECTRICAL and PANEL experience should be the
one wiring the panel. My best skill, really, is providing them with
coffee and the occasional fan to blow air through the airplane, or
tool or part, and to answer questions from them about what system I
wanted set up what way. Near the end of the day, it was clear that
my best time-use was to work on software updates to the VP-400, and
updates to X-Plane 10.10 beta, which is going on at the same time.
(yah. busy.)
Below, we see some of the stuff that is NOT in this airplane!
Mechanical circuit breakers? Mechanical standby gauges? Lengthy
wiring harnesses? Lead to balance the weight of the paint on the
control surfaces? All GONE! This, and much much more, does NOT live
in 844X, but does live in other Evos! (As it would turn out, N844X
is about 200 pounds lighter than other Evos because of these careful
design considerations on our part).
T minus 8 days (Friday July 13)
The new instrument panel is now running, and looks BEYOND stunning.
Seeing that dull matte black military-style panel with the dull
matte-black computer displays suddenly light up as the computer
screens and internally-lit switches come to life is stunning. The
layout looks to be perfection, with 3 red buttons on the panel.
Red-button number 1 (center of panel near the bottom): The Garmin
reversionary button. Hit that and BOTH Garmins go to PFD mode,
sacrificing the MAP to have you attitude information if the PFD
fails.
Red-button number 2 (upper-right above): The VP-400 "red button".
Hit this, and the autopilot will engage, the aircraft heading down
to the runway that it deems most likely to result in a successful
power-off landing.
Red-button number 3 (lower-left above): The starter! While most
turbines have a boring light-switch type starter that looks EXACTLY
like the nav-light switch (BORING!!! BORING!!!) I decided that an
850-horsepower jet-prop needed something special. While the
push-button starter from a Ferrari 430 proved problematic (Ferrari
won't sell you the STARTER BUTTON without the STEERING WHEEL!) the
push-button starter from a Honda S-2000 looked JUST the same, and
cost about $29. Internally-lit to glow red like a hot coal, it is
clear to anyone that pushing this button will invoke FIRE!
It was then time to configure the flaps and trims on the airplane,
and this will involve getting into the nitty-gritty of the avionics
design and layout. I hope I do not bore you going into too much
detail here, but you can skim this part if you are not interested in
the fine details of the way the VP-400 works.
Vertical Power designed its VP-xxx line of products to control all
of the various electrical systems in the airplane (flaps, trim
motors, lights, etc). They designed these systems to always be as
fail-safe as possible, and to allow any critical function to be
over-rided if needed. For example, to run a critical system like
elevator trim, a constant series of messages needs to be generated
and sent by the internal workings of the Vertical Power control
unit. If the messages ever stop due to some sort of breakage in the
system, the trim stops as well. This is fail-SAFE (not
fail-dangerous) because if the system breaks, the airplane simply
refuses to do anything exciting. Instead, it just leaves things as
they are now. (In this example, it simply leaves the trim still,
which is a safe occurrence, instead of letting it run without
command, which is a dangerous occurrence).
Brand new for the VP-400, though, is an all-new touch-screen
interface with new software written by me. This software, as
initially written by me, was written in the way that all other
touch-screen interfaces are written: The computer waits for a
message that the user has touched the screen. When the computer gets
that message, it figures that the screen is being touched until it
gets a SECOND message… this message being that the user has lifted
his finger OFF the screen! It seems simple and common-sense, right?
The computer gets a message from the touch-screen that the pilot has
touched the screen, and then a little bit later a message that the
pilot has STOPPED touching the screen.
But, what if the message that the pilot has stopped touching the
screen is somehow lost due to some sort of error in the computer or
touch-screen?
Think about it: In that event, the computer will THINK that the
pilot is HOLDING A BUTTON DOWN on the touch-screen down… even though
he is not!
And, what if that button is, for example, an elevator trim button?
In that case, my computer program would be fooled into thinking that
the pilot is holding down the trim button, and would inform the
Vertical Power system accordingly, and the trim would continue to
run… all because a single message from the touch-screen that the
pilot had lifted his finger off the trim button was lost. (And, to
non-pilots: A trim that motors along without command is among the
most dangerous things that can happen in an airplane, because the
airplane could diverge from pilot control, slowly but surely).
Accident reports throughout aviation history are littered with these
odd little cases that only become apparent in retrospect.
So, after actually seeing this happen in our systems-testing on the
ground (!), we decided to build the VP-400 in such a way that it
would be totally immune to this type of failure in flight. But how?
The internal workings of the VP-400 were already fail-safe… now we
just needed to make sure that the touch-screen interface was equally
robust. The answer came to me rapidly: SWIPING. You know how when
your iPhone rings, you have to SWIPE the little button across the
screen to answer the call? This is slow and annoying, but it makes
pretty darn sure that your phone is not answered by mistake as it
moves around in your pocket, right? A very specific action needs to
be taken that only a human would take, and while it may be slow and
annoying to take, it sure does stop your phone call from being
answered by mistake. This is exactly the key to a fail-safe
touch-screen interface, and exactly what we designed into the
VP-400. Here is how it works:
1: The trims (elevator, aileron, and rudder) are normally controlled
by a hat-switch and another switch on the yoke or control stick.
These trims must be held down to close a circuit that runs the trim.
2: If that circuit breaks, then the user may go to the backup trim
control panel in the VP-400. There, the user SWIPES across the
screen for each LITTLE BIT of trim that he wants! So, if he wants a
LOT of trim, he will need to sit there swiping over and over to
drive the trim, much like a cat scratching a scratching post. Each
bit of swiping that he does allows a few more trim message to be
sent from the touch-screen to the Vertical Power hardware. In a
nutshell: The pilot has to perform a specific action to send in each
little bit of trim. If messages are corrupted or lost, then the
following happens: Nothing. If messages from the touch-screen are
lost due to some unforeseen errors, then some swipe actions will not
register as complete, and in that case the trim simply will not move
… fail-SAFE. This makes a trim runaway as close as I can imagine to
impossible. This is a philosophy I am rapidly learning to code into
ALL elements of the VP-400 interface.
Despite system described above that seems to me to be very well
designed, I am still edgily thinking about the somewhat-scary fact
we are running an experimental aircraft on a very tight schedule,
and sometimes the only way to find errors is to experience them, and
that I am supposed to really FLY this thing…
So, with the cool little touch-screen UI above running, I would run
about the airplane, looking at the various flaps and trim tabs,
yelling to whatever poor sod was stuck in the non-air conditioned
cockpit to run the various trims and flaps up and down, and see if
the moving surfaces ON the plane dud what they were supposed to do.
A typical yelled conversation from the aircraft extremities to the
poor sweating boffin in the cockpit would typically run about like
this:
"OK RUN THE ELEVATOR TRIM UP!"
"WHAT?"
"RUN THE ELEVATOR TRIM UP!"
"OK!"
"THAT IS UP? "
"YAH!"
"NO, THAT IS DOWN!"
"IT LOOKS LIKE UP FROM HERE!"
"YOU MEAN LOOKS LIKE UP FROM THE DIRECTION OF THE BUTTON YOU ARE
PUSHING, OR LOOKS LIKE UP ACCORDING TO THE DIRECTION OF THE TRIM
INDICATOR YOU ARE SEEING?"
"WHAT?"
"WAIT, STOP MOVING IT!"
"OK!"
"NOW MOVE IT DOWN!"
"I THOUGHT YOU SAID TO MOVE IT >UP<!"
"I DID!"
"THEN WHY DID YOU JUST SAY TO MOVE IT >DOWN<?!"
"I WANT TO TEST BOTH DIRECTIONS!"
"BUT I ALREADY MOVED IT DOWN!"
And so on and so forth until lunch.
And then until dinner.
T minus 7 days (Saturday July 14):
Below, Marc of Vertical Power confers with builders from RDD on the
proper ways to interface the VP-400 to the Evolutions various
systems.
SO today the app seemed to HANG.
But maybe it didn't.
I don't know.
The VP-400 was installed in the airplane, but the various sensors
and antennae that DRIVE the VP-400 were not yet hooked up. As a
result, the unit (correctly) displayed big red X's for each function
on the screen. But as we hooked up the GPS antenna, the big red X's
remained. Why? Was the GPS not really properly hooked up yet? Or
were we just not getting a signal in the hangar? Or had the program
crashed? Looking only a red-X's, I had no way of knowing. I CANNOT
STAND un-answered questions like that, and this would be especially
frustrating in flight, so the question had to be answered.
So, a few solutions were called for.
First of all, you remember Knight Rider? The black Trans-Am with the
red light cycling back and forth on the nose? That red light always
cycling was how you knew that the car as thinking… so why not put it
on the VP-400 so you would know that it was thinking, even if there
was nothing on the screen that really needed to CHANGE? It only took
a few minutes to code, and now, whenever there are red X's on the
screen because of missing data, there is at least a red light
pulsing left and right (just exactly like on Knight Rider) so you
know that the computer is not crashed, but is instead running, and
simply waiting for valid information to come in the from the GPS and
gyros and other systems to display.
Next was being sure that we never accepted bad data from any sensor.
The VP-400 is designed to go through the following steps,
constantly:
1: listen for data from the GPS, solid-state gyros, and pitot-static
system.
2: flag that data as "received" once it GETS that data.
3: for each bit of info it shows you, make sure that is is only
showing you data that it has "received" within the last one second…
otherwise show a red-X for that display.
For example:
1: get a message about airspeed from the pitot-static pressure
sensor.
2: display airspeed on the display, since it has received that
sensor data.
3: do NOT show a red-X, since step 1 happened just fine.
Now, what if the pressure sensor was broken, and the speed that came
in was 89456379864 knots?
This would be awfully confusing to the VP-400. So, in addition to
seeing if the various packets of data (such as airspeed, for
example) ARRIVE from the sensor, we now ALSO check to see if those
packets are REASONABLE. If a packet for, say, airspeed came in, and
it was over 500 knots indicated, then we now consider it to be
invalid, and discard that incoming data. This packet fails the
'common sense' test and is ignored. If this happens for more than
one second, then a red-X will go up over the airspeed display, and
the system will understand that it does not have airspeed data, and
will act accordingly. (In this case, by not showing airspeed or
planning an emergency descent to the ground, since both of those
things require airspeed).
"Common-sense" tests like this are now put on all of the attitude
and navigation data coming into the VP-400, so hugely erroneous data
should not be able to leak into the system. Red-X's will result if
that happens. Of course, the system is still listening for new data
all the time, so if the data ever gets back into a reasonable range,
the red-X's will disappear and be replaced by appropriate displays.
This will give me a nice warm fuzzy feeling when flying, since I
will know that the VP-400 is constantly listening for all the data
it can, showing me the latest reasonable data it has if some packets
are lost or corrupted, will show a red-X if a sensor fails, and will
resume normal display again if a sensor comes back on line after a
temporary hiccup. As well, the VP-400 will constantly be evaluating
what it needs to draw displays, plan emergency descents, etc, and
doing everything that it can with whatever data it's got.
What I am really describing here is called "degraded mode", or
"unreliable data protocol"… designs that let the system continue to
function (even if in a less-capable state) as other systems in the
airplane fail. It should be very, very, very hard to EVER get the
VP-400 to say "I got nothing".
Another thing we did today is test all the electronic circuit
breakers. (from here on out: "ECBs")
When I first learned that the VP-400 would have ECBs, I was not
really impressed, since I only was excited about seeing the A-I
runway seeker I am developing going into a real airplane. But
now, having designed the nicest ECB interface I can think of, and
using it in the airplane, I see that there is simply no better way
to go. The ECB system is awesome, and the only right way to build a
modern airplane. Here is why:
Imagine you are in a real airplane. Make it at night. Maybe IFR.
Maybe some rain and turbulence going on.
Pop.
A circuit breaker in the airplane pops.
This is a circuit breaker that is on a small panel underneath the
instrument panel.
By your left ankle.
In the dark.
You have no way of KNOWING that it popped, unless something obvious
on the airplane just stops working.
You cannot FIND it even if you know it popped.
You cannot GET TO IT without likely losing control of the airplane,
since you would be fumbling around under the panel trying to find
it.
You cannot tell WHICH breaker popped, since no human on or off the
Earth could ever read the TEENY TEENY TINY LITTLE circuit breaker
label in the dark under the instrument panel.
You would not know WHY the thing popped, so you would not know if
you should reset it.
When you have to fumble around in the dark for a breaker you cannot
see, with a label you cannot read, in a place you cannot reach,
popped for a reason you cannot guess, located in a place you have to
reach down to get so you cannot fly, I am simply going to say: That
interface is TERRIBLE.
Now let's talk about ECBs, as we have here.
On the vertical power display, there is simply a scrolling list of
electrical stuff in the airplane. Flaps. Landing lights. Fuel pumps.
Things like that. This is simply a list of electrical stuff in your
airplane, and you scroll though it by rotating the (single) little
knob on the bottom of the VP-400. Each device lists it's name, and
shows you it's amperage right beside it. You can read it as easy as
you can read the computer you are looking at right now. (easier,
actually, since the fonts are big and more brightly-colored).
Systems running normally are white, and broken ones are in red with
the reason they are red shown. They are on a touch screen, so simply
touch any device on the little scrolling list to play around with
it. (turn it on, off, or reset it if the breaker popped).
That's it!
You can easily see everything going on with the entire airplane,
right down to the amp drawn by every system, in a format that is so
easy to access that it is ridiculous.
As well, if ANY breaker pops, the a little red-alert icon pops up on
the VP-400 ECB screen selector, so you know that there is something
in that screen that needs attention! This way, if a breaker ever
pops, you instantly know it, because the little red-alert icon pops
up to tell you! The, you just go to the ECB screen on the VP-400 (as
easy as launching an app on an iPhone) and scroll through to find
the breaker that is popped or otherwise alerting. It is as easy as
can be, and the only way that makes any sense at all to actually use
in an airplane. You could easily be flying at night, IFR, in the
rain and turbulence, picking up ice, and if the red-alert icon comes
up for a popped breaker, you could easily scroll to it, right on the
front-and-center display, seeing in bright red text exactly what is
going on, without ever getting confused or behind the airplane, or
even taking your focus away from the main instruments on the
instrument panel! Cool!
As well, if a system is NOT HOOKED UP, or simply unplugged or
broken, with a regular circuit breaker, you would NOT have any way
of knowing it! But with this ECB system, where you can see the
amperage going to each system, you can easily see if ZERO amps are
going to any given system. That lets you know that that system is
not working… something that is not possible with mechanical circuit
breakers. A number of people have crashed airplanes with PT-6
engines because the engine quit. The engine quit because a pneumatic
line froze shut with moisture turning to ice. These engine had
heaters on the lie, but those heaters had broken hours, days,
months, or years before the accidents. The pilots never KNEW that,
though, so they flew happily along with their heaters turned ON (and
UNplugged!) until the lines froze over and the engines quit. With an
ECB, we SEE the amperage going to the heater at every moment we feel
like scrolling to the heater item in the ECB list! If we see that
the amperage is zero, we will KNOW we have a problem and fix it when
convenient! Compare this to the alternative if simply never knowing
the thing was unplugged, and thinking everything was fie since the
breaker never popped.
Also ECBs weigh less, and have less wiring. Since wiring tends to
chaff and start fires over the years, this is a good thing.
So, the day was spent inside the hot cockpit on the ground in a hot
hangar, going though every item in the ECB list (every electrical
thingy in the airplane), making sure that I could turn it on or off
(like pushing or pulling the breaker) and that, when turned on, the
system worked and drew amperage.
It would basically go like this with me in the cockpit and someone
else scurrying around outside the airplane:
"Next item?"
"Landing lights!"
"Turn 'em on"
"I did!"
"I don't see them!"
"OK they are broken!"
"Put it on the list!"
"OK how about the fuel pump?"
"OK I turned it on!"
"All right I hear it! How many amps?"
"2 amps!"
"Ok turn it off!"
..and so on and so forth, for every system in the airplane. Sitting
in a cockpit that is running about 90 degrees, and has no interior
(not even a pilot's seat yet!!!) this is really not as fun as it may
sound. It is rather surprising how awkward it is to wriggle around
for an hour inside a carbon fiber shell with no proper interior, no
seats, and various (SHARP!!!) metal seat-mounting brackets and other
attach-points sticking at you in every direction as you try to move
around inside the shell.
Back at the house, I connect a (real) VP-400 to a copy of X-Plane in
a little network that we set up on the dining-room table. X-Plane is
actually spoofing the messages that come in from the REAL sensors in
the real airplane, sending those messages to the VP-400 sitting on
the dining room table. Now here is where it gets funny: The VP-400
does NOT know that it is sitting on a dining room table. The VP-400,
since it is getting flight messages from X-Plane, BELIEVES THAT IT
IS IN FLIGHT! HAR! SO here is what happens: X-Plane flies
along for a few moments like a regular pilot would, and then fails
the engine. X-Plane commands that the VP-400 hit the red button,
which the VP-400 effectively does, and the VP-400 then glides
X-PLANE BACK DOWN TO THE BEST RUNWAY TO LAND ON! Once this imaginary
emergency is over, X-Plane resets to some RANDOM location, heading,
speed, and altitude, and does it all over again. And so it goes
throughout the night, with X-Plane imagining engine failure after
engine failure, each failure at a different location and altitude,
each time telling the VP-400 to bring the plane safely down to
earth. For each of these emergencies simulated on the dining room
table, a result is memorized, and the next morning, by me, analyzed.
We start the night of imagined horrors and go to bed, always with
the uneasy feeling that the VP-400 just might show a bad track
record when checked the morning...
T minus 6 days (Sunday July 15):
Well, after 8 hours of simulated nightmares throughout the night,
the VP-400 has scored the following:
IF the plane is high enough to glide to ANY airport at the moment
the engine fails, then:
The VP-400 guided the simulated airplane down to a point just short
of the runway, pointed along the runway heading and at a comfortable
glide-slope, at a comfortable approach speed, so the pilot was
perfectly positioned for a power-off flare and touchdown, 100% of
the time.
The VP-400 then proceeded to attempt to LAND the imaginary airplane,
and intersected the ground within the runway perimeters at a
moderate descent rate, 92% of the time.
The VP-400 then managed to get the airplane stopped on the runway,
simulating a human standing on the brakes, 98% of the time.
As overnight runs go, this is pretty typical. The VP-400 has been
tested through THOUSANDS of emergencies in the simulator like this,
and the scores above are now becoming pretty common: At least in the
simulator, if you have enough altitude to MAKE it to an airport, the
VP-400 can always set up an energy-management path that will put you
looking right at the runway threshold every time, get you bumpily
but without injury on the ground about 90% of the time, and even if
the landing is hard, still get you o the runway in such a location
that if you hit the brakes, you can get stopped on the runway almost
100% of the time (some runways are just too short for an Evolution,
though, so run-offs do sometimes happen).
Now, if the schedule holds, we will fly tomorrow, so this might be a
nice day to reflect on the Lancair Evolution.
This plane has almost no drag because of it's clean shape, retract
gear, and totally-feathering prop… in fact it can glide at a ratio
of almost 20-to-1!!! This is starting to get kind of close to some
gliders in glide ratio. The clean design, long, thin wings, and prop
that can feather so the blades are perfectly aligned with the wind
make this possible.
As well, as I have alluded to earlier, it climbs like a home-sick
angel and goes like stink. (5,500 fpm climb, and can run along at
370 mph).
You already know it has only 3 computer screens for instruments, and
may guess that the power that you can add for take-off is limited
NOT by the engine power, but instead by how much RUDDER AUTHORITY
you have to counter the torque! I have earlier mentioned, I think,
that adding power results in significant ROLL from the power
addition that must be countered by aileron. I have also previously
mentioned that you are limited (by rudder authority) to specifically
550 hp for take-off, but can use 850 hp for climb and cruise
(subject to air density lapse rate at altitude, of course).
Let's talk about stability. The TAIL of an airplane only STABILIZES the airplane if it can dip down into air that is largely un-affected by the rest of the airplane. Think about it: If the plane is flying level and the tail suddenly dips down, the air, which is moving horizontally, simply pushes the tail right back up to where it was before, restoring the nose back down to level again! This is stability. BUT, imagine for a moment: What if, when the tail dipped down, the airflow was NOT horizontal, but instead aligned with the body of the aircraft??? The tail would have ZERO tendency to push back up again, thus restoring the plane to level flight! This is because the airflow over the tail would not change at all in this condition, so there would be no restoring force! The wings on ALL planes (except flying wings) help bring this unfortunate situation about… and a big prop blowing air (along the axis of the airplane!) helps contribute to this effect as well (in ALL single-engine planes with the prop in front.. but a bigger prop blowing more air makes the issue more noticeable). In other words, the huge power and prop of the Evo give incredible performance, but do extract some price in stability: The pilot must always fly this airplane, and not count on it to simply go straight, forever, by itself. (NOTE: The Lancair Evolution is CERTAINLY stable! But, with the huge power and prop and propwah, it is simply not AS docile as, say, a Columbia-400: A Lancair-heritage cousin to the Evolution with one third the power).
Also, airplanes with very TAPERED wings (wings that are narrow at
the tip) have very little DAMPING in roll. Why? Because as the plane
rolls, it is the air 'hitting the wing from above or below' OUT AT
THE TIPS as the airplane rolls that damps out the rolling motion of
the plane. The Evolution has very tapered wings. And the
lower-damping effect is magnified at higher speeds and higher
altitudes where the air pushing up or down on the wingtips as the
plane rolls becomes small in comparison to the forward speed of the
plane, and the air density drops… this results in very little
damping compared to a slower, clunkier, certified airplane, so the
handling is actually quite reminiscent of a helicopter! (which are
among the SLOWEST craft flying!) Like a well-designed helicopter,
the Evo is perfectly responsive and wonderful to fly, but very
intolerant of inattention in flight.
844X has synthetic vision and artificial intelligence to find and
display the way down after an engine failure… the Boeing 787 has
NEITHER of these things in it's avionics suite.
So the Evolution becomes a fascinating contradiction in (now
out-dated) assumptions.
The airplane is the most COMPLEX single-engine prop that I know of…
but it's exterior shape is the CLEANEST and SIMPLEST that I have
EVER seen.
Being a carbon fiber shell with a turbine engine, it is so light
that the HEAVIEST single thing in the airplane is the FUEL it
carries.
The plane goes like a bullet… but in the event of power loss, it
glides like a glider.
It is the fastest single-engine prop you can buy… but it handles
like a helicopter.
It has avionics that exceed a Boeing 787 in multiple types of
sophistication… but it has not a single gauge on the panel.
It is certainly serious… but is flown with a joystick.
It contains a big jet engine… but is pulled by a prop.
The design elements above result in the following (in order of the
above, line for line):
speed
speed
speed
speed
safety
speed and safety
speed and efficiency
This is a pretty fascinating airplane…
T minus 6 days (Monday July 16):
3:34 pm now and no flying yet. We just got a weight and balance:
2,501 pounds, behind 850 hp... yikes.
Google the weight and power of some fast cars to see where that
range falls... that is 2.9 pounds per horsepower.
Think about this: The imagine something with the POWER of a HORSE,
but weighs 2.9 pounds! Wow.
Below, we go on the scales to find our weight and center of gravity
location.
Below, on jacks, we retract the landing gear to make sure that all
the doors are pulled up tight.
One note about swinging the landing gear: The system will NOT raise
the gear if the airplane is on the ground!
The way we tell if we are on the ground is the AIRSPEED indicator.
If the airspeed is too low, the gear will not retract.
But, how do we TEST the gear on the ground?
There is only one way: We have to pressurize the pitot tube so the
plane thinks it is in the air and will raise the gear.
So, how do you pressurize a pitot tube, exactly?
Well, let's just say, if anyone wants to know what I will do for my airplane, this picture may denote it a bit too graphically:
OK, enough pictures of me blowing my pitot tube. Now let's talk
about how the day was spent.
Basically all of the day had "hurry up and wait" written all over
it, since there was nothing myself or Vertical Power could do but
wait for the electrical guys to finish wiring up circuits. So, I
decided to use that time to design our next-gen landing gear system
instead.
This airplane has a STANDARD LANCAIR landing gear system. It does
NOT currently interact with the VP-400.
This being the case, myself and Vertical Power were able to spend
the day designing the ultimate landing gear system to go into the
VP-400 without a recklessly-tight time-table, since 844X will fly to
Oshkosh with the standard Lancair gear system that is already well
built and tested.
So, how DO you design a landing gear system for a VP-400 to go into
a Lancair Evolution?
Well, if you are like me and Vertical Power, you sit in a tiny, hot,
sweaty office while techs scramble around the airplane outside and
have the following conversation:
"All right so what is the gear indicator supposed to do? I just need
3 lights that are red or green on the display, right?" "Well, you
need green for down, and I don't know what red means." "Broken?"
"But how will the system KNOW that it is broken?" "If the gear is
not where it is supposed to be?" "That is called 'in transit'" "OK"
"That should be indicated by some sort of in transit indication,
like maybe yellow... or maybe a yellow and black barber pole." "Now
how do we know if we are in transit?" "The handle should say one
thing, and the contact-switches in the gear system should say
something else" "But what if the gear hydraulic PUMP is running?
Shouldn't THAT indicate that we are in transit?" "Maybe, or maybe
not.. .what if there is a hydraulic leak? Then, the pump could run
even if the gear is not in transit" "But that only happens a little
bit, right, to re-build pressure?" "It depends on how fast the leak
is" "But what if the solenoid fails, so that the pump is working
fine, but the gear won't move? What is the indication there?" "Can
we even TELL if the solenoid has failed?" (pause) "And what if the
pilot has commanded gear down, but the gear has NOT GONE DOWN
BECAUSE HE IS ABOVE V-LE?" "Wait, you mean V-le or V-o?" "V-e is is
landing gear extension... how fast you can lower your gear..." "NO!
V-le is landing gear EXTENDED... how fast you can go with the gear
EXTENDED, even if that is FASTER than V-lo, which is how fast you
can OPERATE the gear." "Ok so we do not let them OPERATE the gear
above V-lo, but there has to be an over-ride in case the airspeed
sensor fails... we don't want that to keep the pilot from lowering
the gear!" "Yah there needs to be an over-ride, but if the handle is
set above V-lo and the over-ride is not engaged and the plane cannot
lower the gear, how do we color the indicators?" "Wait, don't we
need to to INDICATE to the pilot somehow that he is going too fast
to lower the gear? How do we SHOW that?" (pause) "How about each
indicator looks like a little up and down arrow if he can raise or
lower the gear, and a square if he cannot?" "No! Too busy! You will
have SIX arrows for THREE indicators! We always had a little circle
around the gear lights that was white to indicate that the gear
could not move, and green to say that it can." "HUH? NO WAY! That is
too many colors! It won't look good and the green circle will be
confused with a green gear light!" "We never had any complaints..."
"What about some lines above or below the gear lights to indicate
that the lites cannot move?" "But they DON'T move. They change
color." "Maybe if we put a hashed background of diagonal warning
stripes behind the gear, then that will make it clear that the gear
cannot be moved." "What does RDD say about all this?" "We have
noticed that sometimes the pump runs when there is a leak or
something... you can NOT hear it in flight, an indication that the
pump is running would be good..." "How do we SHOW that?" "An
annunciator?" "No way... too busy.. too many annunciators.. it will
look like something is WRONG" "Well, maybe something is... he said
it can be caused by a LEAK" "How about the gear lights PULSE to show
that the pump is running, so you see some motion there?" (pause)
"Well, you could TRY it..." "Now what about the gear over-ride? How
do we ACTIVATE that?" "I say it goes in the SYSTEM screen! Yah! Put
a button in the SYSTEM screen to over-ride the gear. "Why not put it
in the main PFD screen where the gear lights are shown?" "That is
too close to the airport-info page... people will be hitting it by
mistake all the time." "OK if in the system screen, I can add a
button there I guess... " "Do we have ROOM there?" "Oh that is no
problem at all... there is plenty of room there. So they just hit a
button to over-ride?" "No way! Too easy to hit by mistake! They need
to do your swipe-thingy you just coded... they have to SWIPE to
confirm." "Wait... if they over-ride on the ground, then they will
retract the gear on the ground?" "Yup." "Oh crap. OK the swip thing
is needed and let's take a page from the test-firing of the pyro
charges on my model rockets: Once you swipe, the thing will count
down from 10 before it activates the gear, and you can cancel at any
time." "OK, but I want more than one message sent from the Seeker to
the main controller before I activate the gear." "How about I send
one message for each second in the count-down, and you only fire off
the gear if you get every message in the count-down?" "That should
work..." "OK now when do we fire off the audio warnings to RAISE the
gear?"
And on and on and on it goes, for hour and hours, talking through
every possible failure and interface and fallback and backup and
over-ride... for HOURS, just for the landing gear activation and
indication!
So, we spent the day talking through our next-gen systems while the
techs wired up the plane....
Going back home, we set up the overnight simulator for another night
of continuous engine failures...
T minus 5 days (Tuesday July 17):
We get panel power even if no batteries are on if we just engage the
CROSS-TIE... that seems wrong...
There is a freon leak in the air conditioning...
Not all of the switches are yet secure enough in their mounting
brackets to be safe for flight...
My little swipe-trim code will sometimes run the trim in the wrong
direction (but only a little bit, due to the fail-safe nature of the
code)...
The terrain databases in my Runway-Seeker are not yet verified on
start-up of the VP-400... this is a self-test we need to do but it
need not hold up VFR flight.
I am reading the POH for the airplane, but not sure what to believe
since this is such a non-standard Evo... we are updating the POH
accordingly...
And, with all the hardware stuff above done: (the lighting in this
picture most accurately shows the colors in cloudy lighting, I
think)
Final paperwork given to Dave McRae of RDD: the punch-list of to-do
items for the first flight:
And:
Flight!
Dave checks out the plane and finds a half-dozen little things to
tweak, and I will be flying it soon!
T minus 4 days (Wednesday July 18):
OK yesterdays test flight resulted in no fatalities ;-P, but of
course there are always things to tweak. The plane was a stunning
beauty to look at and fly, of course, the PT-6 whined on the ground
and growled in the air to perfection, as expected, but we DID have a
leaky pitot-static system. This meant that the airplane was losing
air through the line from the cockpit to the pitot tube, and thereby
NOT getting enough pressure to drive the airspeed up to a proper
value. As a result, the airplane SAID that it was going 150 knots,
when it was actually going faster. MUCH faster. So, back on the
ground today, we hooked up some pressurized hoses to the static port
and pitot tube that go to a box that pressurizes air to exactly
simulates various speeds, and confirm that those speeds are
displayed on the instruments in flight. This test caused on the
techs to ask why even when the system was pressurized (or
DE-pressurized!) to simulate high altitude, the VP-400 did NOT
indicate a higher elevation... and the answer is that the altitude
of the VP-400 is determined by a GPS signal, not an air pressure.
(See my adventure in the Austin's adventures "Your Altitude is
Wrong" to get a thorough understanding of why this is).
Anyhoo, after pressuring and testing the system, we found THIS
little thingy in the system:
(as you can see, the fitting to the pitot tube was cracked, causing
a loss of air pressure)
We are fixing this as I write, and our test-pilot Dave McRae is
preparing to fly again to make sure that this, and all other,
systems are running perfectly.
Once Dave has flown solo enough to be sure that the plane is safe, I
will come on-board to monitor VP-400 operations, and then take
control of flying.
Now, after flying with the pitot static system fixed, the plane
would NOT pressurize... there seems to be a leak in the pressure
vessel somewhere!
As well, as I am running the VP-400 through stress-tests, I discover
that when X-Plane suddenly repositions the airplane to a random
location instantly, the Runway Seeker code in the VP-400 CAN CRASH.
This should never happen in the real airplane, unless you find a way
to reposition the real airplane a few hundred miles in a fraction of
a second... but nevertheless, if you shut down the VP-400 in flight,
and then turned it back on in a different location 30 minutes later,
then the system just might get equally confused... so I am working
to fix that software crash now, while RDD tries to find where the
air is leaking out of 844X... Luckily, the VP-400 can be tested on
the ground by being hooked to X-Plane, which allows a huge amount of
testing at no cost and in very little time...
Coming into 6 pm and we are still on the ground so that is annoying.
Slow as things may be, Dave is test-flying, which is what he is best
at, his techs are fixing and finishing, which they are best at, and
I am coding, which I am best at, so everyone is doing what they are
best at to get us complete.
T minus 3 days (Thursday July 19):
Still no flying for me.
Dave, the test pilot, has found the following issues at a 6:30
flight this morning:
The plane wants to ROLL left ever so slightly, due to an
imperceptibly-small difference in the left and right flap rigging.
The effect is doubtless much smaller than the gigantic roll moment
that comes from engine power application, but is still a few percent
away from perfect. The left brake is just barely touching the
landing gear wheel well when retracted, and the vibration from the
airplane is actually causing bubbles or drainage of the brake fluid
from that area, resulting in very little left brake authority on
touch-down until the brakes are pumped a few times to work the fluid
though. There were small leaks where the wings were re-attached
after painting that let little bits of air out of the airplane,
making it impossible to pressurize (those are now fixed). The rudder
cables were a bit too loose, resulting in slop in the rudder cables,
and were tightened yesterday afternoon, but now they are a bit too
tight, resulting in a rudder that will stay off-center until the
pilot actively PUSHES it to center to get a true-flying airplane.
The angle of attack sensor is not fully installed yet (I actually
got to do a little bit of work starting to install that system
yesterday! I LOVE working with resin and carbon-fiber and flox... so
fun and easy to do... like arts and crafts in school, really... the
only difference is that once the 'glue' dries, it is much stronger
than metal! But the idea of putting the glue on tubes and fabrics
and stuff like that to glue them together really is the same. For
the angle of attack indicator, some little tubes are glued inside
the wing that sense the air pressure through tiny holes drilled in
the wing that lead to those tubes, and from that pressure they can
estimate the angle of attack). The Garmins are both indicating a
little bit nose-low in pitch, so they have to be re-calibrated, and
the VP-400 needs to have it's magnetometer re-calibrated since it
does not seem to know what direction it is facing.
So, we are tweaking things like that, and Dave only wants me to go
up with me monitoring the VP-400 (and then flying) once the plane is
behaving like an airplane,
not a huge assembly or mechanical
parts that are not all quite tuned to fit together
perfectly quite yet. Now this brings up a really interesting point
that may be the difference between a pilot and a test pilot: As a
pilot, I fully expect both brakes in the airplane to respond equally
on touch-down. If I hit the brakes, and one works, and one doesn't,
then for me that is a surprising scenario and a potential accident.
For a test pilot though (particularly one whose shop BUILT the
airplane, this is not such a confusing or surprising event. You see,
the true home-builder knows THAT THERE IS NO SUCH THING AS BRAKES ON
AN AIRPLANE! Instead, there are simply these levels under your feet
that are connected to hydraulic lines, and those hydraulic lines go
to some calipers on the left and right gear, and those calipers are
squeezed onto some discs when the the levers on the rudder pedals
are pushed, and those discs are welded to the landing gear wheels.
Such a pilot can push on the brake pedals and know all of the
mechanical reactions that will result, and not be so surprised if
something is out of adjustment or imperfect along the way. When
seeing systems in this kind of detail, it suddenly becomes obvious
how little we pilots think about what really goes into the airplane:
We just push the brake pedals and expect a given response. Pilots
fly airplanes. Home-builders fly complex assemblies of systems and
parts that re all carefully designed to work together to give the
desired result of hurtling through the air. There is a big
difference, trust me!
Today is mostly me waiting for Dave to flight-test more... so I am
working on X-Plane 10.10 Beta a lot...
OK one of the things we need to do is install the static wicks... I
basically had to force the wicks away from one of the techs so I
could do the job myself, since I am tired of just doing the CODING
for this airplane on this trip. So here is how it is done:
The static wick has screw on the forward end... you widen the hole
the static wick goes into with a drill to drill the paint out... you
insert the static wick into the hole and screw it in securely from
the other side... and you have your static wick. Repeat 8 times for
the plane: Twice for each wing or horizontal stab tip.
But still, I cannot fly it. The flap adjustment is not QUITE
right... the plane rolls to the left just a little in cruise,
requiring about ONE DEGREE of right aileron. Also, when you go over
bumps, the nosewheel "rattles". So, we need the tiniest tweak to the
flap adjustment to run with NO aileron input, and a check of the
nosegear system and bearing.
T minus 2 days (Friday July 20):
OK I am flying! Full write-up to follow later. (bottom line: pretty
sweaty afterward)
OK I am maybe a bit too tired to write all the details right now,
but after a few really fast-paced, hot, 'dynamic' flights in the
heat and thermals and turbulence and wind in an airplane that needs
a different trim or control input for every flap, gear, and throttle
setting, I can say that my Sunday July 15 notes were pretty much
right on... but incomplete. Here is what I did not know to mention:
The DRAG and NOISE from the LANDING GEAR is HUGE! When you are
flying along at low power with the gear up, the turbine is just sort
of humming along with a whine, and the prop is sort of
choppa-choppa-chopping along as well... all is smooth and nice. But,
the minute you lower the gear: WHAM! There is a bug THUNK as it
releases, and the drag is HUGE! The plane suddenly begins to slow
down, and there is a huge buffeting loud wind noise as the air ROARS
around the gear. I guess because the PLANE is so sleek, the GEAR
(comparatively speaking) is really loud and draggy. Lowering it just
puts big draggy brakes on the airplane and down you come unless you
just DRAG the airplane through the air with power. The flaps also
add drag, but seemingly not so much! At least they do not make loud
buffeting wind-noise when lowered.
The take-offs are also a bit odd. You do NOT firewall the throttle
at all (that would over-boost the engine and run the plane off to
the left). Instead, it is like shooting an arrow: How much do you
pull back on the bow? As much as you think you can control. Same
thing with the throttle. How much do you advance it? As much as you
think you can control. The thrust is huge, you are pushed back in
the seat, and you may be hitting a good bit of right rudder... and
this would be a take-off at about
half power. So, engine management winds up having an almost organic
feel to it: You just add the amount of throttle that pushes you back
in the seat about a much as you want. Then, once established in
climb or cruise, you can set the throttle to the performance you
want.
Anyway, I am still scared to death of it, but will get used to it
with more practice.
T minus 1 days (Saturday July 21):
Having had a night to sleep on it, I am now a lot LESS scared of the
airplane, and ready to fly it some more! So into the plane with Dave
and Mark from RDD and off to Oshkosh! Flight was at Mach 0.50 (not
kidding) at 27,000 ft. About 90 degrees in the cockpit since the
cooling of the pressurization is not quite perfected yet. Ice still
formed on some bits of the carbon-fiber walls since it was -15 F
outside. So we were sitting in 90-degree heat with bits of ice
forming on the carbon fiber cockpit walls around us. Wow. Once the
interior is installed and the enviro system is dialed in, it will be
less 'dramatic' but for now it is test-flying, baby!!! The plane
hums along nicely and you really feel the turbine spinning up there
in front of you... kind of a like whining spinning sound... which I
guess is exactly what you would expect. Anyway here are pictures
from the flight:
Before departure, a nice profile:
In the cockpit:
The VP-400 showing the best place to land if an engine failed.
Actual altitude, GPS-measured, is actually 29,000 ft.
Green airports have much longer runways than the yellow or orange
airports, so are better to go to if the engine fails, even though
they are a bit farther.
The view from the cockpit, with weather below and sky dark up above
as it darkens towards.... space.
And here we are at Oshkosh!
Come see the plane, and fly a simulation of it at the X-Plane booth
at the hangars!