April 18, 2005 notes
We have committed to doing demo flights at the X-Prize cup
event in October of this year. While
they would have been fine with us just dusting off our 2 diameter mixed-monoprop
vehicle (the flying crayon), we are going to use this as a firm deadline for
having the new LOX vehicle in robust flying shape. We will be aiming to do several back to back 15 second boosted
hops to demonstrate operational efficiency.
Flying on a biprop, this should be quite exciting.
It looks like we are going to use our last 15 degree crush
cone as the aeroshell for the little vehicle, which will make it look like a
little SSTO model. We would be a lot
more aerodynamic packing it into a 2 diameter cylinder, but I want to give it
a wider base for better landing stability.
LOX Engine Work
We have been doing lots of work in many different areas.
We finished a next-generation lox preburner and engine mount
that is much more compact and easy to inspect.
This burner can attach either directly to the test stand, or to our
gimbal mount. It also uses a larger
methanol injector, http://www.bete.com/products/pages/cw.htm
CW25, because we had to throttle very low to get autoignition temperatures with
the smaller nozzle. This nozzle is
actually a bit too large, so we had to add a restrictor jet in the plumbing to
keep the gox temperatures reasonable even at full throttle.
Our measured Isp was quite low with the previous gox
injector plate. At only around 300
pounds of thrust, the flow was a long way from sonic, and we werent getting
high quality combustion. I made a new
plate that has smaller holes that come down closer to the horizontal fuel
injection points, and also include a splash plate backup, so any fuel that isnt
atomized by the gox should splash back into the stream. This seems to work well, giving us a measured
188 Isp. We dont expect to do too much
better than this until we make a longer chamber, because an L* of only 34 is
almost certainly on the short side, even for gas / liquid injection.
Having the preburner, injector, and chamber all interchangeable
at the flange joint is really convenient, especially since we are still having
trouble with the chambers.
We still dont have a satisfactory solution for our test
stand blast deflector. A half-inch
thick stainless steel plate was fine for 10 seconds or so, but on a 30 second
test run we cut completely through it (and almost cut off the test stand
uprights). We tried making an aluminum
deflector that was cooled by a water hose, but that burned through rather
quickly. We tried extra high
temperature fire bricks, McMaster 9355K1, but they were mostly burned through
at the end of a 30 second run. Flying molten
steel and firebrick also have a tendency to start grass fires nearby. We are going to try the 5000 degree cement
next, but we might have to go to a water cooled deflector supplied with high
The flight computer now runs the test stand, controlling the
burner solenoid, spark coil, purge valve, lox valve, and methanol valve. At the moment we are still logging
pressures, thrust, and flow rate on the separate data acquisition system, but
we intend to move all that to the flight computer (we wont fly with the load
cell and flow meters). At that point,
the test stand will be completely wireless, which will be convenient. One thing we found while investigating low
throttle operation was that there was an over 10% difference in the initial
cracking points on our fuel and oxidizer valves.
We have our gimbal system almost done. We are using Electrak 12v linear actuators,
but we had to swap out the motors, because the factory motors had very
conservative thermal cutouts that shut down after only a few seconds of the
back-and-forth hunting that the gimbal is required to do.
Our biggest issue is that we continue to burn through our
regeneratively cooled chambers.
We made two chambers with normal epoxy for the saddle fill,
but we did get the Cotronics flexible epoxy in, which we used for the next
chamber. Neat stuff. All of the epoxy filled chambers had some apparently
low flowing cooling channels when we blew compressed air through them. We originally attributed this to possibly
not getting all the wax out, but expected that when it got hot it would self
clean. After cutting some of the engines
apart, we found that epoxy had been able to get into some of the channels that apparently
werent completely filled with wax. We
tried using adhesive Teflon tape to cover the channels before epoxy filling,
but it wasnt sticking well enough to give us much confidence. A pretty interesting attempt was made at
using heat shrink tubing to cover everything, which looked good, but we could
only find a polyolefin based tube in the dimensions we needed, and that
probably isnt high enough temperature.
A Teflon tube would probably work well, but I cant find anything at
that diameter with a large enough shrink ratio.
We finally decided to just make a metal outer jacket for the
engine, but it had to be done by hand (many more hours of work), and it wasnt
a completely perfect fit. We eventually
burned through this engine as well, but this is probably the right direction to
We know our coolant velocity is too low, only about 13
ft/sec (not even counting the slop in the outer jacket). We were originally working on engines 2x
3x this size with the same cooling passages, which would have been more
reasonable. Im going to go to cut the
passage width in half for the next chamber.
My mills rotary chuck has a large amount of runout, so I am hesitant to
make the channels much shallower, which would amplify the variability.
Based on comments in Clarks book Ignition!, we tried
mixing silicone oil into the methanol to form internal chamber coatings, but it
came out of solution very rapidly. We
have some ethyl silicate on order, which has been reported to work with
I am getting some larger, accurately CNC machined C145
copper chambers and jackets quoted, which will probably be the real solution,
but we are going to try one more aluminum chamber before Space Access.
Lots of unsorted photos: