November 14, 2004 notes
Sorry about the missed update last week, my time has gotten
a whole lot tighter recently with new engine work at Id going into production
use, but Armadillo work continues at the same pace. No, we arent quitting because the X-Prize has been won. We have a bunch of pictures of the recent
work, but due to a network change I cant get at them right now, and if I dont
do the update today, I would be in danger of missing another week. Ill try and just dump them up here soon.
The custom electronics boards still arent here. Not pursuing the electronics as top priority
immediately after the last vehicle crash is shaping up to be a rather large judgment
error. Without a live flight control
system, the work on the vehicle doesnt get the priority it should, and we are reduced
to tinkering in many cases, which is basically what all the LOX engine work
is at this point.
All of the work with the LOX engines does drive back home a
lot of the advantages of the mixed-monoprop, so we ran a set of experiments to
try to address the warmup issue. We set
up an instrumented hydrogen / air burner to measure the exact flame temperatures
produced. There seemed to be fairly
large variables based on exactly how we positioned and oriented the
thermocouple at the end of the burner tube, but with it clamped in one
position, the numbers were pretty sensible:
(hydrogen flow numbers with an air flowmeter, uncorrected)
CFM air CFH
hydrogen temp C
10 10 370
10 12.5 465
10 15 561
10 20 800
The lower flows couldnt be ignited directly, but if you
ignited it at the richer (still extremely lean) levels, you could dial the
hydrogen flow back down and the flame would be sustained.
We had sworn off the ceramic beads before after they caused
a lot of internal burning of stainless engines, but I wanted to take one more
shot with them in a very high contraction ratio form to see if the reduced
pressure drop could make it viable. We
built an engine with a 7 catalyst ID contracting to a 1.25 throat diameter,
holding 1222 grams of beads in a 2.5 depth.
We choked the maximum propellant flow down to around 5 gpm, which should
prevent an initial inrush flood from quenching the entire engine.
Just flowing the hydrogen / air mixture over the catalyst
would cause the temperature to slowly rise due to catalytic surface burning,
then, at a sufficiently high temperature, it would ignite into a real
flame. At flows of 10 CFM air / 20 CFM
hydrogen, it took about five minutes to get the exhaust flow under the engine
catalyst up to 100 C, but then only five minutes more to get it to 600 C. It doesnt take much hydrogen to get the
engine heated up, but it does take a lot of compressed air. A single high pressure cylinder of air is
only able to heat a small engine like this, it wouldnt be reasonable to heat a
big engine with bottled air. A portable
compressor or blower would be reasonable ground support equipment if it guaranteed
us a smooth, predictable startup.
The engine still ran awful like the earlier bead engines
lots of chugging and shaking and messy exhaust, but it did climb up to full
operating temperature of over 900 C with only a single pack level. Afterwards, you could tell by shaking the
engine that lots of the beads were pulverized inside. The prospect of a simple engine that is just a single poured in
layer of catalyst is very enticing, but we have yet to make it work like our
dual level metal catalyst engines. The
metal ring catalysts are still the only thing we have tried that works reliably
at our full combustion temperatures, but there is probably some hard ceramic berl
saddle random packing that can hold up to the thermal shocks if we were willing
to try more experiments.
We built and tested a couple new LOX engines in the last two
weeks. Our preburners stuck off the
side seem to work pretty well, but the packaging is very awkward. We made two iterations of preburners that
flow radialy in and out between plates, resulting in a preburner the same
diameter as the main engine combustion chamber. The first one we made had the spark plug positioned too far away
from the fuel spray, so it wouldnt light.
We stuck a spark plug in the side, which did get it to light, but it
also served as a local flameholder and let the combustion burn through the
The next one moved the spray nozzle up, and changed a couple
other configurations things, and it ignited and burned fine. There was a slight anomaly in the combustion
temperatures we are using Bete PJ-24 spray nozzles, and the first burner
showed that to be still a bit too high of fuel flow for the amount of lox we
are vaporizing, but the second one had the outlet temperatures much lower,
implying that the movement of some of the passage holes wasnt giving it the
same combustion time before choking it off with all the cold lox.
We welded together our second regeneratively cooled chamber. There are 20 cooling channels of 0.125
width by 0.065 depth running the length of the chamber. Each cooling channel ends with a 0.0325 injection
hole directly into the chamber, so there is no top manifold. The total engine length was 9.8, the throat
was 1.9 diameter, the straight internal chamber was about 7 long by 3.6 ID,
for an L* of about 30. The preburner
exit plate has 20 holes towards the outside, one for each fuel injection point,
so there are 20 separate gas / liquid impingement points.
We have made a new test stand base out of a 3 by 3 piece
of ¾ thick steel, because we have been digging a pit in the concrete with all
the lox engine firings.
The engine made a very good looking plume, but almost
immediately it started melting out the aluminum nozzle, spraying molten aluminum
all over the blast deflector. We didnt
hardcoat the chamber this time, but we are pretty sure the larger issue was
that the new injection arrangement actually got our combustion efficiency out
of the toilet, and the heat was just too much.
We measured a solid 186 Isp at only 125 psi chamber pressure (after the
throat had been eroded quite a bit), which is about 90% of theoretical maximum at
that (rich) O:F ratio. Our previous
runs were grimly low performance under 130 Isp, which was worse than the
mixed monoprop. This new arrangement
should give over 200 Isp with 300 psi tank pressure once we get the oxidizer
flow up a bit more and keep the throat from melting out.
At the low efficiencies we previously had, the heat could be
conducted through the 1 of solid aluminum from the throat to the cooling
channel, but when the combustion got going a lot better, it just chewed the
throat out until it reached the point that it could cool it. We ran it again, and it didnt get noticeably
Our next chamber will have an outside contour to follow the
nozzle, and I will probably cut down the channel depth near the throat. We are going to try a fiberglass wrap closeout
first, but if that doesnt work we can either build a saddle, or try a