December 27, 2003 Notes
We received a new 100 gallon tank from Structural last week
to do another hydrotest on. The
previous tank we tested failed at 600 psi, but it had a clear and obvious weak
point in the lightweight and incomplete cast aluminum winding flange on the
bottom of the tank. The new tank has
proper flanges on both sides, and we bolted our own milled aluminum flanges to
Neil borrowed a real hydrotest machine from his work for
this test. We had to change some things
on it, because it was set up with a 20,000 psi gauge. Neil said this was their low pressure one! The machine runs on shop air, and it took a
painfully long time to bring the pressure up.
Neil commented on how different hydrotesting a fiberglass tank was from
the steel vessels they normally work with the steel vessels have so little
stretch that they come up to pressure very rapidly.
The tank made the normal snap-crackle-pop sounds all through
the pressurization, but it passed 600 psi with no problems at all. Our air compressor was having a hard time
supplying the test pump at the higher pressures, running flat out all the
time. We let it sit at 800 psi for 20
minutes while we ate dinner, then went back and continued pressurizing it. Zooming way in with the video camera we could
see a few drops of water leaking past the O-ring, and the ½ thick flange
bowing slightly upwards, but it was still holding fine. We let it continue pressurizing until it
reached 900 psi, but it was going up so slowly that it didnt look like we were
going to burst the tank that night, so we decided to release the pressure and
keep the tank.
Big Vehicle work
I had to order a glow plug harness, because none of the
local auto parts stores stocked them.
We chopped off the connectors and wired everything up so the computer
can control the glow plugs by itself, but it turns out that I accidentally
bought an AC coil power relay, so we were only able to test it by pushing the
We fabricated custom aluminum mount points for the engine
pressure transducers. We were
considering just hose-clamping them to the support struts, but then we would
have had to pull them off and let them dangle whenever we pulled an engine
plate off the vehicle.
We improved the propellant drum loading system so it can
load and pressurize without disconnecting anything. Going through an additional check valve slowed it down a little
bit, taking nine minutes to load a full 55 gallon drum into the vehicle
tank. We also timed gang-pressurizing
an entire six-pack of nitrogen bottles into the vehicle tank. By five minutes, the tank was at 200 psi,
and by nine minutes it was at 265 psi and almost equalized with the tanks. These were transducer readings, which look
to be high compared to the gauge, which read about 218 psi at the end. We were considering bypassing the big regulator
(which was set at 300 psi) completely and just letting the nitrogen bottles
flow out at full pressure, but the times werent so bad getting most of it
through the regulator. The low pressure
times stretch out a lot, so if we decide to cascade load the individual
bottles, we may consider it again. Unfortunately,
even with cascade loading, it doesnt look like we are going to get to 300 psi
tank pressure with a single six-pack, so we are probably going to have to take
two of them with us, which is a pain.
We hung the entire vehicle with the scale to get a current
ready-to-fly weight for the vehicle, which comes out to 1700lb, assuming the
final engines weigh the same as the last couple ones we welded together. This has grown a bit more than we expected
(which should be expected
). We want to
do the initial flight tests with about 500 pounds of propellant, so we really
do need to see 700 pounds thrust minimum from each of four engines to get even
a slow liftoff. We are contemplating
adding a fifth engine in the center to crutch things up if we dont get the
engine efficiency up soon.
Unfortunately, adding the fifth engine now isnt very straightforward
due to the way we have things set up.
We got the 1600 gallon tank cradle fabricated, and we have started
some layout work on the cabin-at-bottom design. It looks like this will actually be roomy and comfortable,
compared to our strap-in-upside-down arrangement on the crush cone
vehicle. The big tank weighed somewhat
less than we expected, it was only 1400 lb, plus about 50 pounds for the flange
Optimization of the engines has continued, still without any
spectacular results. We reworked our
test stand glow plug / valve control electronics to reduce clutter, which
always seems to be a good idea. We keep
hoping we are almost done with this test series, but it looks like it may
stretch on for a while longer.
We removed the last two layers of catalyst bale from one of
the test engines and replaced it with 15 very heavy 3-mesh screens that were
used to contain the catalyst bale in commercial use. They were coated with the same catalyst as the bale, but the
total surface area was quite low compared to the other catalysts. On testing, it didnt ever clear up, giving
a densely cloudy (low Isp) exhaust, but it did produce very high thrust:
650 lbf at 278 psi tank and 139 psi chamber (this run
started off high thrust for 1.5 seconds, but then dropped to a lower level)
580 lbf at 272 psi tank and 126 psi chamber
Three x 100 gram layers of bale plus the small amount of
catalyst on the heavy screens was obviously not enough for that amount of flow,
but the bale didnt bunch up and restrict the flow under pressure.
On Saturday we had a new type of catalyst to test that we
have good hopes for. Instead of being a
solid mass of catalyst, it is made up of hundreds of ¼ diameter by ¼ long
metal rings with some extra perforations.
This prevents any straight through flow, and each little ring should be
a great flameholder for combustion in the hot section. We got 3 pounds of the rings, which is about
140 cubic inches of volume. The rings
are made of quite thin material, only 0.008 thick, but they are reasonably
strong at room temperature.
We cannibalized the first welded engine parts to make a 5.5
ID test engine for this. We are
concerned that they will still self-compress under pressure loads at high
temperature, so we built the first test engine with two layers of 200 grams of
rings, each supported by a separate perf-plate. Each layer was about 1 deep.
This engine had two 600 cpsi monoliths for the cold catalyst. This engine was very difficult to get
584 lbf at 271 psi tank and 120 psi chamber
When we tried a re-test, we couldnt get it to fully preheat
at all. It seems like as we have
reduced the flow restrictions in the hot section, the oxygen/methanol vapor
combustion under the cold pack just flows out without transferring much heat to
the hot section catalyst. It is getting
bad enough that the old propane / air forced preheats are starting to look good
We decided to try a different test, switching the 2.2
throat nozzle for our smallest one, a 1.25 throat. This warmed easier, and produced dramatically different pressure
401 lbf at 283 psi tank and 229 psi chamber
That is an extremely efficient feed to chamber pressure for
a catalyst motor, but a 5.5 catalyst diameter to 1.25 throat diameter is a
pretty bulky contraction.
We next switched to a 1.83 nozzle, which is the same 3 : 1 diameter
ratio we have on the big motor ( 12 : 4 ).
This was more difficult to preheat again.
545 lbf at 287 psi tank and 176 psi chamber
The increased flow is starting to lose a lot more by the
time it gets to the chamber. We added
another pressure tap to the engine under the cold catalyst for comparison.
520 lbf at 275 psi tank and 186 psi mid pressure
This shows a very tiny pressure drop across the two ring
layers, but it also shows that the higher flow is starting to choke somewhat
higher, in either the cold catalyst or the spreading plate. The engine with five by 100 gram bale layers
last week produced a higher mid pressure (271 psi tank gave a 221 psi mid
pressure) at a little less thrust. What
we really need to do is have three pressure taps, one under the spreading
plate, one under the cold catalyst, and a third under the hot catalyst. I have ordered more data acquisition
channels so we can start reading all of these simultaneously next week.
When we took the engine apart, we noticed a drawback of the
ring catalysts the supporting perf plates had bowed down enough that the
rings pooled in the dome, exposing the outside edges to straight-through
flow. We hammered it back into place,
but it looks like it will be a good idea to use deeper ring catalyst sections in
the future. We added a third section of
200 grams of rings to the engine to see if engine performance improved with
487 lbf at 270 psi tank and 153 psi chamber
So even the rings are a producing some increased flow restriction
at the full hot flow rates
We tried opening up the top spreading plate with a 0.25 drilled
hole down the center, but that let too much rush through, giving us a central
cloudy stream in the engine.
We went back to the all-welded 2.2 throat engine we made
last week with the 3 x 100 and 2 x 50 bale layers to test it at higher
612 lbf at 410 psi tank, 282 psi mid pressure, 124 psi
The catalyst bale in the hot section gives huge pressure
losses, but the pressure drop to the mid section is still very high. I have a hard time seeing the foil monoliths
being a dominant pressure loss, so I am probably going to have to make some
more spreading plates with different hole sizes to experiment with. It is possible that we just have a critical catalyst
diameter to throat diameter ratio that we are wasting our time fighting, but Im
not quite ready to quit yet. If
necessary, we can reasonably easily use a 7 ID chamber with our existing
nozzles, but it would require another custom order of catalyst, at which point I
would probably just consider building the 12 engines.