June 28, 2003 Notes
Our big parachute arrived on Friday, which is a huge relief
for us. Three previous parachute
companies have declined to work with us because they got the jitters over the
fact that we were building rocket ships (Instead of just jump out of planes.
Sheesh.). Strong Enterprises http://www.strongparachutes.com/
hasnt batted an eye at our application, and the company president has been
working with me on the details. To top
it off, their price and delivery times are quite a bit better than what I got
from Butler Parachutes before they bailed on us. Strong is also the parachute supplier for the Starchaser X-Prize
team, who are using a steerable parachute for their capsule recovery.
Our parachute is a 64 diameter round chute, extracted by a
9 to 15 diameter drogue, depending on application details. It turns out that we use a parachute
basically the same size as the mercury capsule, because our entire vehicle
landing weight is only a bit heavier than just the capsule. For our first test drop, we will have the
drogue on a static line underneath a helicopter, but for free flights the
drogue will be cannon-ejected, as on the subscale vehicle.
The parachute and drogue together only weight about 55
pounds to recover our 2400 lb vehicle, which is lighter than I was originally
budgeting. We expect to have a fully
redundant system when we fly manned, the drogues are going to need to be much
heavier to serve as high speed stabilizers, and the cannons and containers will
add weight, but we should still be around 150lb all up for the recovery
system. The parachute is currently a
loose hand pack in a 17 diameter by 25 long bag, but we will be building a
custom box that will conform to the bottom of the tank, so the parachute wont
protrude below the level of the engines.
We intend to ballast the vehicle to about 2400lb under the
canopy, which will match the straight drop tests we have already done. The tank, with cabin coupler, manway
closure, and all engine mounts, currently weighs 1100lb. We have room on our weight peg for 225 lb of
barbell weights, which is about what our final engines and spare recovery
system should weigh. The cabin and
crush cone weigh about 300 lb, so we need to add about 700lb of additional
ballast in the cabin to simulate the mass of the three passengers and the
various cabin equipment.
We decided to go ahead and build the final passenger
bulkhead in the cabin, so we could put the ballast in the same location the
real passengers would ride. We had
mocked this up with plywood and 2x4s a couple months ago, but we didnt take
build it out with proper materials. The
bulkhead diameter is 51, but because it gets cut away at the hatch to allow
the pilot access to the top seat, we could cut the entire bulkhead as one piece
from a 48 wide section of 2 thick Hexcell composite honeycomb. It works out pretty conveniently one
standard 4 x 8 honeycomb panel provides the material for both bulkheads in
the cabin. Because the cabin is formed
from two welded cone halves, it tends to go somewhat non-circular at the weld
points when free standing. We have two
chain hoists in the shop now, and there is just barely enough height to allow
us to pick the tank up and set it down on the cabin upside down, which rounds
everything out perfectly before bonding.
If the helicopter company is willing to work on the 4th
of July weekend, we should be doing the drop test next weekend.
Catalyst Heating Tests
We wanted to find some way to more consistently heat the
catalyst packs for our 50% tests than just flaming the bottom with a propane
torch until it was glowing, which probably also wasnt heating it very deeply. The platinum catalysts will catalytically
burn propane / air mixtures without a flame, so we plumbed up a T at the top of
the engine with a valve that would allow us to flow a propane / air mixture
into the top of the engine before a test, then close the valve when we are
ready to flow peroxide mixtures into the engine.
Turning on the gas flow through the top of the engine, then
lighting it under the nozzle would just create a flame at the nozzle that would
never get back to the catalyst pack at all.
We eventually found that if we preheated the bottom of the catalyst pack
to red-hot with a separate torch, then turned on the gas flow, that it would
burn at those points inside the catalyst pack.
It takes quite a bit of time (10+ minutes) for the heat to spread from
the initially heated points until the entire base of the pack is glowing red
hot, but it does eventually get completely uniform. The heat then starts slowly propagating up towards the top of the
catalyst block. Even when the entire
catalyst is glowing red hot, the valves at the top of the engine stay cool,
because only cool gas is flowing through that area the burning doesnt happen
until it reaches the platinum.
There are two issues that combine to make the heating take a
long time: We got a really large torch
assembly that connects to a bulk propane tank, but because the engine offers
some flow resistance compared to an open torch tip, there isnt enough air
being pulled into the torch inlet line for very high flows. We will probably try feeding compressed air along
with the propane to let us tailor the mixture more precisely, and enable higher
total heating rates. The other
(probably dominant) problem is that propane / air will only burn on a platinum
catalyst that is already quite hot, so it flows right over the top of the
catalyst, and only lights when it hits the bottom that has been preheated with
the torch. The heat conducts through
the ceramic very slowly, so it takes quite some time before the top of the catalyst
is seeing any burning on its surface.
We are going to try using hydrogen gas instead of propane, because
hydrogen will catalytically burn on platinum at room temperature.
On Tuesday, we did an engine test after heating the pack
with the new method, which we expected to behave far better than anything we
had done before. We were very surprised
to see it run quite a bit worse, with the vast majority coming out completely
undecomposed, and not building enough chamber pressure to make more than a
couple pounds of thrust. We thought
that we might have ruined the catalyst by heating it excessively, but it didnt
seem much worse than the other pieces when tested later. My current theory is that the gridded
ceramic monolith catalysts may cause liquid channeling down the centers of each
pore when they are extremely active.
Liquid contact with the sides of each pore may cause an immediate flash
to gas that keeps additional liquid away from the walls, like blow-apart with
hypergolic propellants, or liquid drops skating on a steam cushion on a hot
surface. This would explain why several
of our runs started out good, then got worse the pack was heating up (as
seems reasonable, given the flaming exhaust plume), and eventually reached a
level of activity where it was causing liquid channeling down the pores. When we preheated the entire pack to very
high temperatures, it immediately went to the channeling behavior.
If this is the case, we cant expect any better behavior
from either the 4 thick ceramic catalyst, or the rolled metal foil catalyst we
have on order, because they will also offer straight shot channels. What we will need to do is segment the
catalyst to break up the straight shots all the way to the nozzle. We tried to cut one of the 2 thick catalyst
blocks into thirds on the bandsaw, but a small amount of side pressure caused
the ceramic grids to crack off. We took
the second one and just cut it in half, which worked ok. We fabricated a 2 long chamber extension
and packed all four catalyst sections (we refit the cracked ones together by
hand) into the engine, separated by three stainless screens each.
We werent sure if the propane preheating front would
propagate across the screens between the catalyst blocks, but it did eventually
reach all the way to the top of the engine.
The first test firing started off a little cloudy, but cleared up for a
nice clean run, with a broad plateau at 100lbf. We had adjusted our mixture ratios to account for the fact that
our food grade peroxide is only 40% concentration, not the 50% we assumed, and
the plume color was notably different than our previous runs, being a pale blue
instead of the more luminescent orange we had when running richer. There is a big rise in thrust to 200 lbf at
the end of the run, due to the long, restrictive plumbing we had on the test
stand at the time. Since we had been
working with very low flow rates for a while, we had a 6 hose from the tank to
the ¼ servo valve, then another short 6 hose to a 1/8 restriction right before
the engine. As the 6 hoses clear out,
the total pressure drop in the system goes down, letting the thrust go up. We have rediscovered this many times, and
always seem to forget about it when looking at test runs with a nicely repeating
thrust spike at the end.
For the second run, we exchanged the tank to valve hose for
a 10, and removed the 1/8 restriction.
The run was good again, with steady thrust increasing from 100lb to
150lb, but it still peaked another 50lb at the end, due to the second short
length of 6 hose. The run was a little
rougher, probably because the pressure drop from the 1/8 restriction was
providing some flow damping. The
measured Isp across these two runs was about 100s, which isnt too bad for the
low chamber pressure and notable transient behavior at the start. The thrust peaked (after clearing all the
restrictive plumbing) at 200 lbf at 173 psi tank pressure with a 1.25 diameter
For the third run, we put the 1/8 restriction back in to
smooth it out, but moved the servo valve directly to the side of the engine,
eliminating the second length of hose. We
loaded 4x the propellant, which would give a long enough run to get everything
fully stabilized and measured. On
firing, the thrust ramped up, but instead of lighting, it went cloudy, and
thrust dropped off rapidly.
My only current theory, assuming the catalyst isnt
deteriorating with each run, is that we only have enough active catalyst area
to make about 200 lbf of thrust, and the initial inrush of liquid before there
is chamber pressure is contributing to a cooling of the catalyst, which must
race against the building of chamber pressure.
We can test this by trying to crack the throttle to build some chamber
pressure before opening it all the way up.
A larger cavitating venturi would be the proper solution. If this does turn out to be the case, it would
mean that we would need a huge amount of this catalyst to make big 5000 lbf
engines. If we can get the ceramic monoliths
made in thinner slices, say ½, that would improve things a lot if the
channeling theory is correct.
We tested our new TDS (total dissolved solids) meter this
week. As expected, distilled water read 0 ppm. Our local tap water
read 146 ppm. The 40% food grade peroxide (Solvay brand) was 18
ppm. The 45% technical grade peroxide (FMC brand) read 232 ppm.
We also repeated our catalyst poisoning tests for the
ceramic catalyst with additional controls.
One chip went in 100 ml of food grade peroxide, one chip went in 100 ml
of technical grade peroxide, one chip went in methanol, one chip went in water,
and one chip was left as a control. The
technical grade sample clearly showed some reduced activity due to poisoning,
but what was interesting was that even the water soaked chip (after drying
under a heat gun) was less active than the control chip. We believe that the process of catalyst
impregnation in the ceramic carrier leaves some loosely bonded catalyst on the
surface, which gives high activity, but is easily washed away by liquids, leaving
the strongly bonded catalyst with less surface area below.
We have another style of catalyst on order that consists of
a roll of thin stainless foil that is platinum plated and corrugated to form
vertical passages. Because this is a
metal-metal bond, it should not show any wash-away behavior, but the vertical
channels may still have the flow-through problem. We may also have some problems at high temperatures, because one
end of the foil roll is held together with a braze.
If flow-through is the primary problem, we may wind up being
best off with just platinum plated stainless screens, instead of these more
advanced catalysts. We can certainly
put that together, but the light weight and lower pressure drop of the advanced
catalysts are strong motivators to try and get them working.