August 1, 2004
notes
Flawless Big Hover
We finished the assembly of the 48” diameter vehicle on
Tuesday, and it did a perfect 16 second hover with no hint of any
problems. This is the very first 100% successful
test of a big vehicle. Hopefully all the
new isolation and shielding has banished the electronics gremlins that have
plagued us. We drained out the residual propellant
instead of burning it off, so when we checked the temperature of the various
components, even the outsides of the insulating boxes were barely warm.
http://media.armadilloaerospace.com/2004_08_01/upright.jpg
http://media.armadilloaerospace.com/2004_08_01/base.jpg
The on-vehicle camera view is pretty boring pointing at the
ground, but it is interesting to watch the water on the ground instantly vanish
as the vehicle slides over it. It would
be great for clearing snowy driveways. :-)
http://media.armadilloaerospace.com/2004_08_01/flawlessHover.mpg
Our vehicle loading equipment has been rearranged so that it
just connects directly onto a nitrogen six-pack, instead of being a part of the
test stand loading equipment on the trailer.
We moved back to using the small dual-stage venturi
vacuum pump, because the smaller tank won’t take as long, and it uses a lot
less nitrogen. It still takes four
nitrogen bottles to get it loaded, pressurized, and water washed, but that is a
lot better than the nine it took on the 63” tank.
We found that it is actually easier to unscrew the landing
shock absorbers (with a power drill) for tipping the vehicle up and down than
to compress them down and lock them in place with the little caps we made. We are concerned about damaging the threaded
bodies if we handle them a lot, so we made a custom box to hold them in when
they aren’t on the vehicle.
http://media.armadilloaerospace.com/2004_08_01/shockBox.jpg
We wanted to do a boosted hop on Saturday, but it rained all
day Thursday, so the remote site wasn’t in good enough shape. We were going to do some more hover tests at the shop to try out some code modifications
for the upcoming waivered (long duration) flight, but
we had a surprising incident: we were
just starting to erect the vehicle up by the nose when the entire epoxy bond
holding the cone adapter on popped lose.
http://media.armadilloaerospace.com/2004_08_01/poppedOff.jpg
We were glad that it happened while the vehicle was still in
the cradle instead of hanging off the ground, but it was still a shock, given
how large of a bond area there was. We
had heard a pop during pressurization on Tuesday, which was probably most of
the bond letting go. The primary
difference on this cone was that we had only prepped the aluminum with a
sanding disc instead of a grinding disk, so the epoxy / flox
evidently didn’t get enough bite. We
pulled the cone back off, re-sanded the tank epoxy, ground deep grooves in the
aluminum cone, drilled holes in it for good measure, and bonded it back on.
http://media.armadilloaerospace.com/2004_08_01/gouged.jpg
We had planned on doing hover testing on Sunday after the
epoxy cured, but I managed to get the flight control code pretty broken while I
was working on it, and I didn’t get it fixed in time. A number of issues had been cropping up with
it – I added graphing for the integrated velocity position that is used for the
position hold, and found that it was reversed from the GPS position value, so I
flipped that and negated the gains, but that had me wary enough that I decided
to thoroughly go over it and work out all the known issues. My integrated 6DOF vehicle simulator hadn’t
been working correctly since we switched to jet vanes and ECEF GPS
initialization, and when I started fixing that up, I wound up making several
changes in the flight control code for improved clarity. This turned into an entire day of working out
all the axis transformations and integrations again. It was obviously functioning just fine
before, but there were some important things I wanted to simulate before our
longer boosted hops, so the rework was justified. Some subtle errors in the initialization and
coast-with-GPS-failure modes were uncovered.
Everything should be ready to go now, so we expect to do one
more hover test, then a boosted hop next week.
After that, we are going to do an endurance hover in preparation for our
waivered test flights. If the engine, vanes, or actuators are going
to give us any problems after a 60+ second burn, we would rather have it happen
under the lift instead of a mile up in the sky at the southwest regional
spaceport.
Propellant Disposal
We drained the propellant out of the vehicle after landing
to avoid cooking things in an extended ground burnoff,
but there is still the question of what to do with it. We don’t like to leave the pre-mixed
propellant around, and we don’t want to carry enough water to remote sites to dilute
it (10:1). We added some water to it to
bring the concentration down a bit, then tossed some small chunks of old
catalyst into the drum with it. The
result was surprisingly unexciting, and did not seem to be getting rid of the
peroxide at a very rapid rate. We
eventually dumped in a little bit of potassium permanganate, which did finally
get it decomposing at a pretty good rate, but at the high rate of decomposition
it was also kicking out droplets of liquids, which was making a mess. The permanganate also gets consumed, reacting
a certain mass of peroxide, then turning into an inert
brown gunk. We left the solid catalyst
chunks in the drum, but it was a self-limiting reaction, because as the
peroxide was decomposed the water fraction increased, which slowed the peroxide
decomposition rate. It was still
bubbling away a few days later, down to a concentration of a
couple percent peroxide.
We decided to take one of our old engines and make a
propellant burner that we would just run a hose to the vehicle vent line for. We’ll just set this a ways away from the
vehicle and crack the vent valve until it starts cooking off. Without a spark igniter in it, it won’t run
at high enough temperatures to completely cook everything off, but the foamy
mess will be under 10% peroxide, so it should be fine to just let it run
off. If we aren’t happy with the residue,
we might put together a full fledged engine with an igniter and hot pack, but
no nozzle.
http://media.armadilloaerospace.com/2004_08_01/burner.jpg
7” engine tests
After changing the big vehicle loading equipment, we were
able to move the test stand loading equipment off of the trailer and onto a
dedicated sheld:
http://media.armadilloaerospace.com/2004_08_01/loadingControls.jpg
We replaced the spreading plate in the 7” engine with a more
restricted one (948 x 0.022 holes), but it still ran extremely rough over 550 lbf or so, and was difficult to get started. We believe that the 2 x 1” thick 900 cpsi catalysts in the cold pack is
too much. The current 12” engine is
running very well with a single 1” 900 cpsi catalyst, and the previous engines used 2 x 600 cpsi catalysts.
Since it ran ok at low throttle, we took this opportunity to try out an
experiment I had been meaning to do for a while: replace the methanol mixture ( 7.5 : 1 by mass ) with an isopropal
mixture ( 12 : 1 by mass ), under the theory that the methanol breaking down in
the cold pack before the flameholder is the cause of
the difficulty in getting the initial warmup going on
some engines. We were never able to get
ethanol or isopropal to run steadily before, but now
that we had spark igniters it was worth another try. Didn’t work. It was just as difficult to get the warmup to catch as with methanol, but it did burn ok once
it was going.
We cut the engine apart again, and we found an interesting
piece of evidence: the top of the second
catalyst monolith (between the two monoliths) was severely beaten up. It wasn’t melted, just crushed in places like
it was hit with a small hammer. This was
almost certainly the location of our combustion instability. We are going to put it back together with
only a single 1” 900 cpsi monolith and test again
next week.
http://media.armadilloaerospace.com/2004_08_01/mashed.jpg
News
