April 4, 2004 notes
The thermocouple amplifiers didnt arrive, so we werent
able to fly the big vehicle. Damn
shame, because the weather was perfect, and it is probably going to rain this
coming week. We insulated everything on
the vehicle, mounting a replacement pressure gauge up on the support plate
instead of on the distribution manifold, wrapped the feed lines, rewired the
master cutoff with Tefzel wiring, and built a box around it.
We also had an unexplained failure of our A/D board. The boards digital output still worked, and
the FIFO was operating, but all data read zero. We suspect the on-board DC/DC converter had failed, but it is a
cause for some concern. I replaced it with
a spare that I had fortuitously just ordered, and everything is working fine,
but we will keep a close eye on it.
Since we werent going to be able to fly the vehicle, we
decided to fire the big engine at the 100 acres with the large plumbing and at
a higher pressure. This engine is still
glow-plug based instead of spark based, but we added a thermocouple port so we
could watch the temperatures during warmup.
We did a brief warmup at the shop, and it seemed to catch and warm very
easily. We also built a big blast
deflector plate, in hopes that we could avoid tearing more trim off of our
little shed by the test stand.
On January 11, we did two test firings with this same
2000 lbf at 233 psi tank, 159 psi mid, 123 psi chamber, with
a significant line clearing spike
2440 lbf at ??? (lost the transducer -- about 300?) psi
tank, 199 psi mid, 145 psi chamber, with a line clearing spike of 2550 lbf
Since then, we upsized the valve and flex hose on the engine
from 1 to 2, and hard mounted the engine to the test stand mount so it couldnt
shake on the load cell.
We planned on testing with 48 gallons of propellant at an
initial tank pressure of 450 psi. We
got everything set up at the test site fairly rapidly, but when we tried to
fire the engine, it never started warming up.
The temperature rose to normal cold-pack conditions, but it never
started flameholding, no matter what I did with the throttle. The entire 48 gallons went through without
getting it going.
It turns out that the battery that was running the glow plug
was at fault. It read 12v without any
load, but when the glow plug was on, the voltage just plummeted. At the shop, our control board has a volt
meter on it so we can catch things like this, but at the remote test site we
have to run the glow plug with a relay driven battery, so we couldnt see the
We switched batteries around, but we had to send people all
the way back to the shop to get another drum of propellant.
When we finally got ready to load, we wound up having
problems getting a good enough vacuum in the test tank to draw in all the
propellant. Evidently when we changed
over to the big plumbing, we probably got some 2 pipe thread shavings in the
line, and scuffed the valve seats, because we now had a leak through the
engine. It took us four partial
loadings with our little vacuum pump to get the entire 48 gallons loaded. First thing
we are doing next session is replacing our little electric vacuum pump
with the venturi vacuum pump that we can just drive with nitrogen. Even when we have leaks, it can pull a
really strong vacuum very quickly.
Because of the valve leak, the engine started steaming as
soon as we started pressurizing. It
still took a very long time to get the flameholding to catch in the engine,
wasting over half the propellant before the temperature started climbing. One it lit, the temperature rise was rapid,
and when I opened the valve all the way it ran strong and clear, just not for
very long. The tank pressure had also
bled down quite a bit by the time it was throttled up. The results were:
378 psi tank, 312 psi mid, 226 psi chamber, with no
significant line clearing spike.
The temperature dropped on initial valve opening as it
usually does with the initial inrush, but it came right back up, although there
was an odd drop in the temperature level at propellant depletion that looks
rather like a data acquisition anomaly, because it came back up to a level that
looked like a section of the curve was just chopped off.
Thrust has to be derived from chamber pressure, since the
load cell is no longer in line. The
nozzle has a 4:1 expansion ratio, so this is still overexpanded at sea level,
but closer than the previous runs. The
NASA nozzle simulator at http://www.grc.nasa.gov/WWW/K-12/airplane/ienzl.html
works this out to 3700 lbf.
This was a bit less than we hoped for, but not too bad. The lack of a line clearing spike shows that
the plumbing isnt a significant loss.
We are dropping 66 psi through the spreading plate and cold monolith (a
single 2 thick 900 cpsi monolith), and 86 psi through the hot pack (three
layers of 1kg rings each). This is the
first time we have ever seen a larger drop across the hot pack than the cold
pack. Both catalysts are probably
somewhat excessive we believe that only a single 1 thick 900 cpsi monolith
is required for the cold pack, and we can probably get by with a single layer
of rings of about 2.5 kg. The spreading
plate may also be dropping a fair chunk of the 66 psi.
We need a chamber pressure of 293 psi to make our target
5000 lbf for this size engine, but that would take somewhat over 500 psi feed
pressure with the current engine. We will
probably build another 12 test engine with spark ignition and different
The blast deflector wasnt attached securely enough for this
level of thrust, and smashed down almost immediately when thrust came up. The now non-deflected blast then tore
another piece of trim off the shed
We put together a test engine with one of the catalyst
combinations that we hadnt tried yet: using a 4 tall ceramic monolith as the
hot catalyst. The engines that we had
made with the metal foil monoliths in the hot section were eerily smooth, with
only +/-1.5% thrust roughness, but they ran noticeably cooler (lower Isp), and
the foil monoliths prune up pretty bad in the high heat. The hope was that the taller ceramic
monolith could get the temperatures up and still take the heat.
We built the engine up with an 848 hole spreading plate, two
600 cpsi monoliths, and spark ignition over the ceramic monolith. It warmed up fine, but much to our surprise
it ran with maximum roughness (full scale chug) when the throttle was
opened. With a cracked throttle it
maintained temperature and produced thrust, but opening it much farther lead to
the huge roughness again. This might be
due to having a large area spreading plate with only a 1.7 nozzle, so we may
try swapping the nozzle out.
We did a couple more tests with the 0.25 cavitating venturi
in the plumbing, which held the roughness off, but limited the thrust to a bit
under 500 lbf. Over the three runs we
did, the chamber temperature dropped somewhat each time, so the catalyst is probably
wearing off the ceramic just as we experienced with trying to use it in the
cold pack. We will probably do a few
more tests to make sure this is the case, but it looks like the rings are still
our best hot catalyst.
We have been just pouring the rings in the engines, but
there is probably a good consistency benefit to doing some level of pressure
packing of the rings while building new engines. We took some of our used ring catalyst and measured how much they
544 grams of rings in a 5.5 diameter chamber started out
2.1 tall. At the following press gauge
pressures (1 gauge psi = about 2 lbf) we got the following data:
2000 psi = 1.55
2500 psi = 1.46
3000 psi = 1.40
3500 psi = 1.34
4000 psi = 1.28
When pressure was released, it rebounded to 1.34 height. The rings were formed into a fairly tightly
interlocked mass. This would certainly
trade off some pressure drop, but it would likely make for smoother and more
Differential throttling has been our attitude control
solution since the beginning, but our second choice has been using a single
engine with jet vanes. If warming up
multiple engines gives us serious problems, we might move to a single engine
and save all the extra sensors. We
fabricated a jet vane that we can mount under an engine on the test stand to
start experimenting with this, just in case:
Our first test wound up bending the vane, but we didnt have
things mounted very securely, and we were just controlling the actuator with a
manual switch because I hadnt swapped out the dead A/D board on the computer
yet. We straightened and reinforced the
vane, and now that the computer is back up, we should be able to do some
properly controlled test sequences next week.
Our top priority is flying the big vehicle right now, but we
are considering spending some time converting the old manned-lander to a jet
vane vehicle test bed.