October 27, 2001 Meeting Notes
In attendance:
John Carmack
Phil Eaton
Russ Blink
Bob Norwood
The conference video we run at the Space Frontier Foundation
conference is now online:
Low bandwidth version (14 megs)
http://media.armadilloaerospace.com/misc/sff_low.mpg
High bandwidth version (50 megs)
http://media.armadilloaerospace.com/misc/sff_high.mpg
We had a lot of successful testing today. A new five foot rod of 2” diameter
polyethylene arrived today, which we will be using for our properly rigorous
hybrid testing in the next few months.
I set everything up so we can log two more channels of data
on the test stand, in preparation for logging chamber pressure simultaneously
with tank pressure, and we tapped a hole in a scrap big engine base we had, but
we held off on testing it to save time today.
The current arrangement is sort of a mess, with our normal dataq DI-195B
conditioning and logging the load cell and one pressure channel (with the valve
control also one of the serial control lines), and a second dataq DI-151 on a
USB serial port adapter. It seems to
work, but it isn’t real pretty.
Sometime, I will just buy a good standalone industrial A/D system that
can take a lot of channels, but I like reading the current ones directly at the
serial port level for accurate latency timing.
We rebuilt the big motor catalyst pack with some changes
that we hope will improve its lifespan.
We have three anti-channel / anti-tunnel rings interspersed in the pack
now. Since there is a blockage directly
in the center, we removed the micro-perf spreading plate completely. We added a great big wave washer at the top
to hopefully keep the pack in compression after multiple thermal cycles. The 1/4" heavy perforated bottom plate
has also been reinforced with an extra bar of steel, because it was bowing
after several firings.
media.armadilloaerospace.com/2001_10_27/controlRing.jpg
We don’t know yet exactly what the fit should be on the
control rings. The current ones are a
little oversize, buckling when initially inserted, and getting crushed
flat. We will see what they look like
next time we open the motor up. We may
need to give it an explicit notch for expansion.
An interesting thing that we noticed when measuring the big
engine is that it has shrunk slightly at the top closure end. We attribute this to the fact that that end
stayed cooler and was clamped to the top closure, so as the bottom end heated
and expanded, it permanently stretched away from the top side a bit.
We test fired the motor, and it made almost 100 pounds more
thrust than the last time we ran it, almost certainly due to the removal of the
micro-perf spreading plate underneath the catalyst disks that fill in the domed
top. Russ is going to make a new top
that is completely flat so we can get rid of the non-uniform sized catalyst
that we are currently using to fill it in.
Short Form Flights
We made three test flights of the big lander in short form
today. The last time we tried the short
for was before the change to asynchronous attitude control, and it didn’t fly
worth a damn, but in theory, with the new control system it should fly smoother
than the wide frame, because the attitude jitters are proportional to the
control authority, and having the engines half the distance out will make a
significant difference.
The theory was right, the flights were very steady in
attitude.
The first two flights were on rather short tethers, but the
third flight was up to a reasonable height, and included some sideways motion.
The trick to landing the rather tippy short frame is to keep
the attitude engines running until it comes to a complete halt on the ground,
so even if you come down at a bit of an angle, the attitude engines will work
to keep you from falling over. I was
pretty sure that if I let it run out of propellant even a foot or so off the
ground, it would tip over after the first bounce, so I was pretty conservative
in how quickly I got it back on the ground.
We only used 250 to 300 psi in the tank, because the vehicle
like this only weighs 165 pounds, basically half the full up and ballasted
weight.
We did have one noteworthy incident that triggered a second
problem, and brought a third one to light.
My laptop hung twice. I have had
the graphics driver lock up on me during development before, but this is the
first time it has happened while just running the application, probably due to
some additional displays that I added recently. This laptop has an S3
SavageMX, which isn’t very well supported anymore, so I am probably going to
migrate all of my work to a new laptop with a GeForce2Go chipset. The second problem is something that has
been mildly annoying for a long time – the flight computer only seems to allow
four logins over telnet or ftp before needing to be restarted. First the telnet session locks up after
entering the password, then future telnet attempts don’t even get to the login prompt. That will probably be an afternoon of wading
through linux code to figure out. The
unexpected problem was that when the laptop crashed, the flight computer didn’t
exit cleanly like it does if the joystick stream is lost under normal
conditions. I believe that this is due
to the laptop having the controlling terminal for the flight computer open on a
telnet session, which eventually blocked on output. I need to experiment to see if I can make the output
non-blocking, or if I am going to have to completely detach from the terminal,
which makes development more of a hassle.
Videos:
http://media.armadilloaerospace.com/2001_10_27/test1.mpg
http://media.armadilloaerospace.com2001_10_27/test2.mpg
http://media.armadilloaerospace.com/2001_10_27/test3.mpg
Telemetry:
media.armadilloaerospace.com/2001_10_27/hop1.gif
media.armadilloaerospace.com/2001_10_27/hop2.gif
media.armadilloaerospace.com/2001_10_27/hop3.gif
The attitude engines were starting to cloud up on the third
run, so we are going to try and make control rings for them on Tuesday.
Based on these positive results today, we are going to build
a new lander base that has the engines inboard of the wide landing pads,
exactly like the little lander did. The
next vehicle will have dual tanks, with the pilot sitting in the middle, right
over the main engine.
News
