December 13, 2003 notes
If you havent seen our commemorative 100th anniversary
of flight video yet, go look on the main page: www.armadilloaerospace.com
We went through 70 gallons of peroxide this week. We have worked out a nice system for extended
length testing we pump four gallons of peroxide at a time into five gallon
carboys, and fill one gallon jugs with the right amount of methanol (about 3
liters). To make a run, all we have to
do is dump a methanol jug into a carboy and shake it up. Because quenching is one of our major
concerns, we arent bothering with any shorter runs. Our peroxide supply seems to be all straightened out, so we might
as well make the runs long.
For larger runs, we moved the 80 gallon tank onto the
trailer and plumbed it up so we could use the same filling equipment for either
the small tank or the big tank. The Structural
composite tanks are only rated to take about -4 psi of vacuum, but we didnt
have any problem vacuum loading 20 gallons of propellant into the tank.
The first test we did was an experiment to see if a much
lower open area flameholder plate between the cold catalyst and the hot
catalyst section would make things work better. The plate had eight ½ diameter holes, and this was over the same
chamber used last week, with 80 grams of catalyst bale on top of two sections
of 1 thick x 400 cpsi monolith. This
made a good deal less thrust, basically the same as the restrictive bale
chamber, showing that the peroxide decomposition alone was producing enough gas
to choke at these hole sizes.
The next runs were retests of the bale engine that we built
on November 1, which had started out with three sections of 330 grams catalyst compressed
to 1.5 height each, then had the top section removed. It worked just like it did before.
All of these were right around 350 lbf at 280 psi tank
pressure, 400 lbf at 280 psi line clear
We cut the chamber open again, and removed another section of
catalyst, leaving only a single 330 gram section. This still worked fine, but produced little more thrust. The 330 gram blocks were pretty solidly
compressed, and obviously fairly restrictive.
We built a new chamber that had three sections of 110 grams
of catalyst each, individually supported by welded perf-plate.
440 lbf at 264 psi tank pressure, 520 lbf at 264 psi line
This is a tiny bit less than we made with the high flowing
monolith chamber, but the engine was obviously running a lot hotter, so Isp was
back up where we wanted it. We repeated
this, and then ran 10 gallons of it, which it handled with no sign of
On Sunday, we converted our tank-to-test-stand hose to 16 from
10, which has nearly eliminated the plumbing loss for this size of motor. We also replacing a leaky valve, changed a
tank o-ring, and relocated our data acquisition tank pressure transducer so it
would read from both the small tank and the big tank.
The next thing we tried was swapping the 2 throat nozzle
for a bored out 2.2 throat nozzle. It
didnt make any more thrust:
540 lbf at 280 psi (very little line clear difference now)
We saw several interesting things: The thrust actually rose slightly as the run went on. The nozzle was glowing red hot, but the
catalyst wasnt red at all after the run.
We conclude that significant combustion was going on underneath the final
perforated plate, inside the nozzle converging section, which isnt
optimal. Even though the exhaust was
clear, there was a smell of peroxide in the air, and the run was completed in
less time than with the smaller nozzle, implying that Isp was down, even though
thrust stayed constant.
This appeared to have crossed the line of how much flow 330
grams of bale catalyst could handle. If
it was actually quenching, then adding more catalyst layers wouldnt help
because it would just progressively quench the additional layers out. We made a 10 gallon run with the same
configuration, and were pleased to see that it showed no sign of quenching,
and, in fact, showed signs of combustion getting somewhat better towards the
end, with the side of the chamber by the initial flameholder at the top
beginning to glow red. This seems to be
a fundamental difference between the bale catalyst and the straight-through
monoliths if the monoliths dont stay completely hot, they will eventually
quench, while the bale seems to be able to partially handle larger flows
without becoming useless.
We are going to add a fourth layer of 110 grams of catalyst
to the chamber, which will probably bring the combustion quality back up, and
should give us increased thrust with the bored out nozzle. When we cut the chamber open, we could see
that the 110 gram patties of catalyst bale had compressed significantly under
their own flow resistance, such that they are rattling loose between the
perforated plates. That probably doesnt
hurt anything, giving an open area for turbulent flameholding above each catalyst
patty, but it means we can probably precompress it into less space without
hurting the flow characteristics. 1
between perf plates is probably what we are going to do on the next fresh
The next thing that may be limiting our thrust is the spreading
plate hole size, which may be a little tight at 204 holes of 0.032 diameter
for 0.164 square inches of area. We
might crudely try to drill an extra large hole in the middle of the existing
plate, but we will probably have to wait for good experiments on this. I have a big batch of modular flanged chamber
sections being machined, which will allow us to mix-and-match lots of different
pieces. It will be interesting to
answer questions like: Is one cold catalyst enough? Does three improve things?
How much better are the 900 cpsi monoliths than the less dense
ones? Does adding space between the
bale patties help? Would a monolith
added underneath three bale sections stay hot? Etc.
If tests on Tuesday go very well with the four patty chamber,
we may go ahead and start assembling the four engines for flying the big
vehicle, but it would be nice to optimize a little more first.
Preparing for flight
The big vehicle is back together again after insulating.
We installed the electronics, and started doing prep tests. The valve potentiometer feedback points were
calibrated, and everything was leak checked again. We used the plasma cutter to make big cutouts on the mounting
plate so we can get at the flange bolts easier. This electronics mounting arrangement is just temporary, designed
so that we can swap the electronics between the big vehicle and the small
vehicle, but now that it looks like we may not fly the small vehicle, we might
do something more permanent.
With the electronics board inside the aluminum cabin,
communication was almost impossible, even when there was a pretty straight shot
through the open hatch, probably due to huge multipath interference inside. We have an external antenna to mount, we
just havent installed it yet. It looks
like it will be necessary even for our hover tests.
We built a 50 16 braided hose to use to fill the vehicle
tank instead of the 2 PE hose we were using.
We may need to upsize when we are filling 850 gallons of propellant, but
at this point liquid loading is not the dominant time, and using a high
pressure hose allows us to pressurize through the same hose that the liquid
loading is done through, saving a manual swapping step. We are building a custom drum pressurization
loading system to push propellant directly from drums into the tank.