September 12, 2004 notes
Matt made a video of last weeks GOX-GH2 rocket engine test:
This is a great example of the abrupt transition from
subsonic to supersonic (choked) flow.
We tried another test with the 7 engine, giving it a very
short 1.5 section between the packs, and the same 1/8 perf plate flameholder
that the smooth running big engine used.
It still ran rough. We welded a
new nozzle on the old 12 motor and ran it at rather low pressure on the test
stand, and it did still run smoothly, so we havent imagined our previous
successes. There are only two remaining
different in that engine versus our new ones: It has a single 2 thick 900 cpsi
catalyst on top, while the new ones have either a single 1 thick, or two 1
thick 900 cpsi noloiths. Instead of
using the square-grid water jet cut support plates, it used heavy perf plates
backed up by a cross of square bars in the hot pack, and a milled pie-section
support plate under the cold pack.
We have been getting so frustrated with our rough running
mixed-monoprop engines that we have decided to do some experimenting with other
We would like to do some 70% peroxide / kerosene biprop
tests, but we are having difficulty getting a few hundred gallons of unstabilized
70% for basic tests. We can definitely get
a tank-car load, but I dont want to commit to that without actually having
some successful engine firings.
I used ISP to run the numbers for 70% peroxide / kerosene at
150 psi chamber pressure and sea level operation:
O:F mass Isp density Chamber
K Exit K
10 : 0.5 156 1.24 1569 1008
10 : 0.6 165 1.23 1732 1133
10 : 0.7 172 1.23 1885 1253
10 : 0.8 179 1.22 2026 1368
10 : 0.9 185 1.21 2152 1479
10 : 1.0 189 1.21 2220 1531
10 : 1.1 187 1.20 2172 1474
10 : 1.2 186 1.20 2109 1420
10 : 1.3 184 1.19 2046 1369
10 : 1.4 183 1.19 1984 1319
10 : 1 by mass is nice and easy to remember...
For comparison, our current 50% peroxide / methanol mix is:
10 : 1.6 (lean) 146 1.11 1304
10 : 1.3 154 1.10 1434
At best, we might see a 25% performance improvement, but we
aren't likely to be able to maintain an exact mixture ratio, especially while
throttling, so the total will be less. The additional tankage and system
complexity would further erode the benefit, but it would still be a somewhat
higher performance setup. The
performance wouldnt justify the change, but more repeatable engines would.
An exit temperature of about 1200 C is at the upper end for
making jet vanes out of superalloys, but we could always go to refractories.
The other combination we are considering is LOX / methanol.
For 150 psi chamber to sea level operation:
O:F mass Isp density Chamber
K Exit K
10 : 5 208 1.0 3042 2486
10 : 5.5 211 0.99 3070 2530
10 : 6 214 0.98 3087 2558
10 : 6.5 217 0.97 3096 2570
10 : 7 219 0.97 3096 2566
10 : 7.5 220 0.96 3086 2540
10 : 8 221 0.96 3067 2480
10 : 8.5 221 0.95 3037 2385
10 : 9 220 0.95 2993 2269
10 : 9.5 219 0.94 2937 2151
Going to higher alcohols, like isopropanol, would slightly
improve Isp and bulk density and shift the O:F ratio higher, but cooling would
get somewhat harder. Going to a
kerosene would improve performance somewhat more, but you need a special grade
to keep it from gunking up your cooling channels, and cooling gets still more
With LOX / methanol you wind up with two roughly equal sized
tanks, which is more of a packaging problem than the big tank / tiny tank
layout of 70% peroxide / kerosene (a cluster of four skinny tanks might be the most
convenient layout). The exit temperature is so high that jet
vanes would have to be made out of a coated refractory metal if we want them to
be reusable (graphite would ablate away during our long, continuous burns). We couldnt use our polyethylene lined
fiberglass tanks for LOX. Alcohol isnt
as good of a coolant as peroxide, there is less of it, and the combustion
temperatures are much higher, so you really have to push it through the cooling
jacket rapidly to pull enough heat out, resulting in the need for a
significantly higher fuel tank pressure, which probably doesnt let us use our fiberglass
tanks for the fuel, either.
On the upside, LOX is dirt cheap and readily available, and
the combination does give performance more than high enough to compensate for
the increased tank mass.
Based on our need to do fairly deep throttling, we have
started to do some tests on a LOX preburner, which we would use to supply hot
oxygen gas to a primary cooled combustion chamber. We took delivery of two big dewars, one of liquid oxygen and one
of liquid nitrogen for testing. Both
support up to 500 psi pressure.
We built a new concentric gox/gh2 torch, but with an
insulator around the core and various connections so we could avoid a spark
plug altogether, making the spark jump between the two feed tubes. We had a couple unexpected current paths,
and we had to wind up putting a plastic fitting between one of the solenoids
and the torch to keep the spark current from traveling back through all the
plumbing. When we got it all fixed up,
it worked great. Instant push-button
We had been using hydrogen gas running quite rich, with
equal sized orifices and pressure to both the oxygen and hydrogen. When we were doing the gox/gh2 cooled rocket
test we noticed that the regulator needle buzzed around a lot while the
hydrogen was flowing. I thought it
might have been chamber pressure related, but we were also seeing the same
thing on open air burner tests. The
regulator wasnt specifically for hydrogen, so we assumed that the low
molecular weight was causing some issues with it. It didnt seem to be causing a problem, but we had just received
a cylinder of ethane to switch over to.
The liquefied ethane has at least an order of magnitude more mass than
an equal sized hydrogen cylinder, so we can get a lot more tests out of it.
The ethane ran smoothly through the same regulator, but we
had another little surprise from it.
After making a run and closing the solenoid, we saw the line pressure
rise up from the 50 psi we had it set at, sometimes going up to 100 psi or
more. It took us a little while to
figure it out, but apparently the ethane bottle was full enough that our
pulling the gas out from the top fairly rapidly was entraining some liquid
ethane droplets, which would then vaporize in the closed line. It seemed to decrease in intensity as the
tests went on, and I bet that once the bottle gets a quarter or more used up,
it wont happen at all any more.
The original idea was to flow the LOX (liquid nitrogen for
testing) out an annular ring around the torch flame, with the intent of letting
the high velocity torch entrain and vaporize the LOX.
It turns out this doesnt work. Any reasonable flow of liquid nitrogen would promptly extinguish
the torch. In retrospect, this makes
perfect sense. It takes several inches
of space for the torch to actually burn the gox and ethane, and if too much non-participating
gas is added, combustion cant continue.
We added a choke plate so the torch had its own combustion
chamber, then welded another section of tube on the end to be the vaporizer for
the injected cryogen.
Matt caught a nice picture of the first firing, which
ejected all of our machining chips at high temperature:
I burned a thermocouple off trying to get temperature
readings before the LiN flow was started, so we began letting an initial low
flow of LiN go in before lighting the torch off. With a 140 psi oxygen flow through a 0.1 jet and a 50 psi ethane
flow through a 0.1 jet, the full flow from the LiN dewar at 350 psi was raised
to 200 C on exit. As the LiN flow was
increased, the torch flame out the end would shrink down and down, until
nothing but warm gas was coming out the end.
The torch chamber was getting very hot in the last inch or
two before the choke plate, and it looked like we were going to burn
through. We cut out the choke plate,
but we then found out that an initial flow of nitrogen would prevent the torch
from lighting without that amount of backflow prevention.
We realized later that the right way to do this is to have
the vaporizer tube concentric around the burner chamber to keep it cool, then
let the burner just exhaust into the rest of the vaporizer chamber after it has
had enough private combustion space to completely burn. We will get this going on Tuesday, then
attach the preburner to a flange and actually measure some thrust from the gas
flow. After that, we will be ready to
bolt the cooled chamber onto it and try injecting some fuel.