March 11, 2005
We did a lot of vehicle tests in the past few weeks, but we
ended up making a mess.
Since we were still having problems with the GPS cutting out
due to vibration even after cushioning it as much as we could in the
electronics box, we moved the GPS unit from the electronics box to the
nosecone, and gave it triple isolation: lead acoustic insulation directly
against the board, wrapped in foam, then attached to compact wire rope
We also added new circuits to monitor the regulated voltages
coming from the DC/DC converters, because we had seen some suspicious behavior
in one of the previous tests.
We tried a new warmup procedure to see if we could get the
cloudy engine to run right we left the nitrogen bleed open during pressurization
to hopefully keep any drip leaks from igniting before we wanted them to. When we were ready to go, the engine come up
to temperature (at the thermocouple, at least) fine, but it was still running
very cloudy. We went ahead and hovered
it, even though we could barely see the tip of the nose cone through the
clouds, and the GPS did perform fine.
To try to fix the engine, we cut off the top to replace the
cold catalyst. There was evidence that
the spreading plate had been pushed up quite hard at some point, either during
the rough running before we activated the bleed, or during one of the ignition
pops during pressurization. The top of
the catalyst was still pretty active, but the bottom was dead, and the lower
inch of the two inch catalyst was shriveled quite a bit. When we replaced it, we also tried something
new we caulked around the edges with RTV.
The cold pack shouldnt get abover 250 F or so, so RTV should be fine,
and preventing any leaks around the side should be an improvement. Because the engine was welded to the vehicle
base framework, we werent able to braze the spreading plate back on under the
hydraulic press, so we just put a heavy weight on it for the initial
tacks. We also replaced the throttle
valve with a brand new one to fix the leak.
We were quite disappointed to find that the engine still ran
fully cloudy on the next test. I didnt
bother lifting it off for another hover, we just burned off the propellant load
on the ground.
Process of elimination left the hot pack catalyst, so we
plasma cut a hole in the side of the engine and shook all the catalyst rings
out, then poured new ones in and closed the engine back up. Again, we werent able to put any pressure
on the catalyst.
On the next test, the engine cleared up very rapidly,
conclusively showing that the increased cloudiness of the engines was the
wearing out of the hot pack catalyst.
When it throttled up for liftoff, the engine started
chugging badly, which messes up the hovering auto-throttle, causing it to
continue rising slowly instead of hovering or descending. I had to kill the engine when the nose
started climbing past the top of the lift.
When the vehicle fell down, we were extremely disappointed to see the
recovery strap break, rather than stretching to absorb the fall, so the vehicle
landed hard on its side. We had tied a
knot in the strap to allow us to have a small loop above the GPS antenna, and
the strap broke right at the knot.
Lesson learned never put a knot in a recovery strap.
The impact seems to have killed one of the gyro axis in the
IMU, but the rest of the electronics is unharmed. One of the shocks was wrecked, and the frame is not salvageable,
but most of the gear is fine. The tank
took a hard enough hit that we will retire it to use as a water tank.
Several important lessons:
All engines need to be easily moved from the vehicle to the
test stand. We have been building the
big engines too integrally with the vehicle.
Any change to an engine needs to get a test stand run before
it gets a vehicle run. Replacing the
catalyst without pressure packing it gave us a rough engine, even with the
nitrogen bleed on top.
We will never again use a coated catalyst in the hot
section. They just dont last. This engine had around eight drums of
propellant run through it before it got unusable. While it would be possible to design an engine that was easy to
remove and maintain, that just doesnt fit with my goals for the program.
Both of our last two crashes were due to engine problems,
rather than electronics problems. We
need to get a truly reliable engine before we fly another vehicle.
We have one last direction to try for the mixed monoprop,
using manganese compounds as the catalyst.
Instead of using a thin coating of catalyst, the catalyst bed is made up
of solid mineral crystals, so even if there is erosion, the surface remains
If that doesnt work, we are giving up on our 50% peroxide
mixed monoprop and moving to LOX / methanol exclusively. If we have to do that, it is going to be a
huge change, requiring a completely different vehicle design and control
actuation system. We have laid out a
lot of the changes we want to make for a next generation mixed-monoprop
vehicle, but we probably wont start building it until we have positive engine
Our epoxy shelled LOX engine still leaks even after our
patch attempts, so we are going to give up on that approach. We are looking into gun drilling the cooling
channels, and various rapid prototyping fabrication technologies. So far, selective laser sintering is a bust
for several reasons, and using stereolithography to make a casting pattern is
also unsuitable for our sized cooling channels. We are looking into direct metal deposition next.