September 3 and 7, 2002 Meeting notes
Furfuryl Alcohol
We finally got our furfuryl alcohol from Penn Specialty
Chemicals, and we ran some tests.
It is not at all hypergolic with peroxide, so that avenue of
investigation is shut down. It is
supposed to be rapidly and cleanly hypergolic with pure nitric acid, but with
50%, the only grade we had on hand, it just rapidly polymerized and changed
colors.
The density is very high, about 40% more so than kerosene,
which is readily apparent when carrying a five gallon container of it. However, the O/F ratio with peroxide is
about half that of kerosene, so it doesnt really improve total propellant
density, since it is still less dense than peroxide. We might try some dissolved catalyst hypergolic experiments, but
I am still leery of the safety issues with that, and it is more expensive than
kerosene, so this is probably a dead end.
As it seems most have determined in the past, autoignited
kerosene looks to be the best all around choice for fuel with peroxide.
Engine Tests
To test our theory that the bent-pipe injector was
performing worse than the drilled-hole injector due to the shortened distance
inside the venturi, we shortened the exit section up another 1/8, and got
improved performance. We then went back
to the drilled-hole injector, which gave still better performance. We believe that the bent pipe will provide
better performance if we can give it an equal length of venturi underneath it,
but that will require increasing the thickness of our injector plate. The desired venturi height is probably
proportional to throat diameter, so this does mean that sticking with a single
venturi will result in a several inch thick injector plate for the big engine.
We have had a couple problems with accidentally twisting the
tubing inside the injector so that it no longer points straight down. When
tightening an AN fitting into the flare on the end of the tube, it seems we
sometimes rotate the drilled-through swagelok that holds the tube in
place. We may start brazing the tube
either directly in, or brazing the swagelok into an immovable piece.
We tried running a higher performance mixture in the
regeneratively cooled engine. With 250
psi, 0.80 peroxide / 0.50 kerosene jetting and the drilled-tube injector, the
plume got funny after a few seconds of hot firing, so I got off of the
kerosene and let it finish out with monoprop.
The better injector and resulting better combustion was clearly
increasing the heat load into the coolant (we ran a 0.60 fuel jet with the
older injector for 30 seconds without problems previously). We tried a run at 500 psi, which somewhat
decreases the heat load per unit of coolant (at the expense of higher rate of
transfer), but the engine failed abruptly after a few seconds, bursting the
tube that runs from the top of the cooling jacket to the catalyst pack inlet. No gentle warning of imminent failure at the
higher pressure.
We werent sure if it was because we had a vapor detonation in
the line (it wasnt all that violent), or if it was because the aluminum line
had lost too much strength (it was rated for 1400 psi, but at 100F, and it
might have been 300F at the time of rupture.
We replaced the aluminum line with a stainless line, but the engine
seemed to be completely not working when we tried it again, with peroxide
gushing out undecomposed. It turns out
that we had burned a hole through the combustion chamber near the top, which
then fried the line exiting the cooling chamber. It probably wouldnt have just started gushing peroxide out of
the engine instead of bursting the line if we had been using stainless line at
that point.
We had seen on the radiative engine that this injector
caused a hot spot high up in the combustion chamber, and that is likely exactly
where the burn through occurred. We are
hoping that the bent-pipe injector will provide more uniform combustion, which
should help with this. The biggest help
will just be making larger engines, where the square-cube law helps cooling
significantly. The 0.5 throat regen engine
has 25 square inches of cooled surface area to 0.2 square inches of throat,
while the 2 throat regen engine will have 124 square inches of cooled area to 3.2
square inches of throat, so we should only have 1/6th the heat
transfer to the coolant (we cut the volume slightly, otherwise it would be
exactly 1/4th) The question
at that point will be if the thicker section at the throat can transfer the
heat fast enough to keep the throat from melting.
One thing we are still unsure about is the value of trying
to film cool with decomposed peroxide.
The current injector has a ring of 16 1/16 holes around the periphery
to allow some gas to enter the chamber without mixing with fuel in the main
chamber. You can see clean streaks in
the chamber under the holes, but Im not sure if that is good or bad. It seems likely that if stray droplets of
fuel or getting to the wall, then there will be lean burning there, but I can
see how it might reduce total heat transfer in a well-distributed engine. Since total heat load shouldnt be a
problem, we are probably going to skip the holes in the big injector
We have the valves, jets, and lines set up for the 1000 lbf
biprop. I have two separate high flow
tescom regulators on order, so we will perform mixture ratio tweaking by
changing pressures with fixed jets. I
also got enough of the Kzco servo ball valves for us to build a large
four-engine, differentially throttled vehicle, which is looking like our chosen
platform for the high performance vehicles, instead of separate attitude
jets. I broke a 1/16 drill bit inside
one of the balls while venting it for peroxide service, but I think we have
enough remnants from previous valves we have broken parts off to replace it.
Russ is slowly boring out the big chamber. Boring a 14 long chamber from a solid bar
of aluminum is not much fun, so we will probably farm the work out in the
future.
Lander Modifications
I ran the electronics box through a full duty cycle to try
to reproduce the gyro misbehavior that we saw last Saturday. After 80 minutes of operation, my battery
warning light started flashing, but power did not completely fail until 129
minutes, so that is a very conservative indicator.
The inertial units temperature sensor did rise steadily
through the entire time, but it did not reach the levels logged on Saturday
because I was testing inside, and there werent any rockets firing. It did not misbehave in any way, but the
plastic box was noticeably warm underneath the inertial unit, so a thermal
problem is still our best bet. On
opening the box after power finally failed I found the internal temperature to
be somewhat warm, but not too bad. The
inertial unit was fairly warm to the touch, and the switching power supplies
were hot, as expected.
Because the total heat load does not seem excessive, we
decided to try to just improve the cooling of the inertial unit inside the
box. We stood it off about 3/8 from
the box to allow air to circulate underneath it, and we added a fan inside the
box to force air around.
I was also happy this week to come across a store online
that still had some old stock of the Logitech Wingman Warrior joysticks that we
had originally planned to use for the pilot on the lander. It was the last of the RS-232 serial
joysticks with throttle and hat buttons, before everything moved to USB. The PC104 computer we were using for a while
had USB, but we moved back to the larger SBC with more serial ports, and lost
the USB port. I will be getting things
set back up for local serial joystick this weekend.
Since tipping the vehicle over doesnt seem all that
difficult, Russ has started welding in support bars in the front to protect a
pilot in case it rolls over forward.